the modern firm, corporate governancedocshare01.docshare.tips/files/6637/66375734.pdf · 2016. 6....

The Modern Firm, Corporate Governance and Investment

NEW PERSPECTIVES ON THE MODERN CORPORATION

Series Editor: Jonathan Michie, Director, Department for Continuing Education and President, Kellogg College, University of Oxford, UK

The modern corporation has a far-reaching infl uence on our lives in an increasingly globalized economy. This series will provide an invaluable forum for the publication of high-quality works of scholarship covering the areas of:

● corporate governance and corporate responsibility, including environmental sustainability

● human resource management and other management practices, and the relationship of these to organizational outcomes and corporate performance

● industrial economics, organizational behaviour, innovation and competitiveness

● outsourcing, offshoring, joint ventures and strategic alliances ● different ownership forms, including social enterprise and employee

ownership● intellectual property and the learning economy, including knowledge

transfer and information exchange.

Titles in the series include:

Corporate Governance, Organization and the FirmCo-operation and Outsourcing in the Global EconomyEdited by Mario Morroni

The Modern Firm, Corporate Governance and InvestmentEdited by Per-Olof Bjuggren and Dennis C. Mueller

The Modern Firm, Corporate Governance and Investment

Edited by

Per-Olof Bjuggren

Jönköping International Business School, Sweden

Dennis C. Mueller

University of Vienna, Austria

NEW PERSPECTIVES ON THE MODERN CORPORATION

Edward ElgarCheltenham, UK • Northampton, MA, USA

© Per-Olof Bjuggren and Dennis C. Mueller 2009

All rights reserved. No part of this publication may be reproduced, stored ina retrieval system or transmitted in any form or by any means, electronic,mechanical or photocopying, recording, or otherwise without the priorpermission of the publisher.

Published byEdward Elgar Publishing LimitedThe Lypiatts15 Lansdown RoadCheltenhamGlos GL50 2JAUK

Edward Elgar Publishing, Inc.William Pratt House9 Dewey CourtNorthamptonMassachusetts 01060USA

A catalogue record for this bookis available from the British Library

Library of Congress Control Number: 2009922767

ISBN 978 1 84844 225 2

Printed and bound by MPG Books Group, UK

v

Contents

List of contributors viiPreface viii

1 Introduction: the modern fi rm, corporate governance and investment 1

Per-Olof Bjuggren and Dennis C. Mueller

PART I KEY ISSUES

2 Opening the black box of fi rm and market organization: antitrust 11 Oliver E. Williamson 3 The corporation: an economic enigma 43 Dennis C. Mueller

PART II THE THEORY OF THE FIRM FROM AN ORGANIZATIONAL PERSPECTIVE

4 A contractual perspective of the fi rm with an application to the maritime industry 63

Per-Olof Bjuggren and Johanna Palmberg 5 The use of managerial authority in the knowledge economy 82 Kirsten Foss 6 Competence and learning in the experimentally organized

economy 104 Gunnar Eliasson and Åsa Eliasson

PART III INVESTMENTS AND THE LEGAL ENVIRONMENT

7 Corporate governance and investments in Scandinavia – ownership concentration and dual-class equity structure 139

Johan E. Eklund 8 The cost of legal uncertainty: the impact of insecure property

rights on cost of capital 167 Per-Olof Bjuggren and Johan E. Eklund

vi Contents

9 The stock market, the market for corporate control and the theory of the fi rm: legal and economic perspectives and implications for public policy 185

Simon Deakin and Ajit Singh

PART IV THE BOARD, MANAGEMENT RELATIONS AND OWNERSHIP STRUCTURE

10 Institutional ownership and dividends 225 Daniel Wiberg11 Contracting around ownership: shareholder agreements in

France 253 Camille Madelon and Steen Thomsen12 Board governance of family fi rms and business groups with a

unique regional dataset 292 Lluís Bru and Rafel Crespí13 Better fi rm performance with employees on the board? 323 R. Øystein Strøm14 The determinants of German corporate governance ratings 361 Wolfgang Drobetz, Klaus Gugler and Simone Hirschvogl15 Top management, education and networking 382 Mogens Dilling-Hansen, Erik Strøjer Madsen and

Valdemar Smith

Index 401

vii

Contributors

Per-Olof Bjuggren, Jönköping International Business School, Sweden

Lluís Bru, Universitat de les Iles Balears, SpainRafel Crespí, Universitat de les Iles Balears, SpainSimon Deakin, The Faculty of Law, Cambridge University, UKMogens Dilling-Hansen, School of Economics and Management, University of Aarhus, DenmarkWolfgang Drobetz, Department of Corporate Finance, University of Basel, SwitzerlandJohan E. Eklund, Jönköping International Business School, SwedenÅsa Eliasson, IBMP CNRS Strasbourg and Vitigen GmbH, Siebeldingen, GermanyGunnar Eliasson, KTH, StockholmKirsten Foss, Copenhagen Business School, DenmarkKlaus Gugler, Department of Economics, University of Vienna, AustriaSimone Hirschvogl, Department of Economics, University of Vienna, AustriaCamille Madelon, HEC School of Management, Paris, FranceErik Strøjer Madsen, Department of Economic, Aarhus School of Business, DenmarkDennis C. Mueller, University of Vienna, AustriaJohanna Palmberg, Jönköping International Business School, SwedenAjit Singh, Queen’s College, Cambridge University, UKValdemar Smith, Centre for Industrial Economics, University of Copenhagen, DenmarkR. Øystein Strøm, Østfold University College, NorwaySteen Thomsen, Copenhagen Business School, DenmarkDaniel Wiberg, Jönköping International Business School, SwedenOliver E. Williamson, University of California, Berkeley, USA

viii

Preface

This book is a collection of papers from two workshops held in Jönköping, Sweden, in 2006 and 2007. The theme of the workshop in 2006 was ‘Corporate Governance and Investment’ in a wide sense. Topics of papers could be: to describe and analyse the ownership and corporate governance structure of a given country; to make a comparative analysis of govern-ance structures in diff erent countries; to study corporate governance and performance in diff erent types of fi rms (for example, family and non-family owned fi rms); to explain the levels of investment of companies; and to draw policy implications about how capital markets might be altered to improve the allocation of capital and the overall performance of companies. The workshop arranged in Jönköping was one in a series of annual meetings of a European Corporate Governance Network. The network’s fi rst meeting was at Cambridge University (UK) in 1998 by initiative of Professor Dennis C. Mueller and Professor Alan Hughes. Since then several meetings have been organized. The 2006 workshop in Jönköping was the seventh.

The second workshop, held in Jönköping in September 2007, was the fi rst of its kind inspired by the emerging literature on the economics of the fi rm. The background to the workshop was the revolutionary development of the theory of the fi rm that has taken place during the last 35 years. In spite of all the progress in the fi eld, traces of the new developments in micr-oeconomic and industrial organization textbooks are scant. The comments made by Ronald Coase in 1971 at an NBER meeting about a non-existent treatment of organization of economic activities within and between fi rms in industrial organization textbooks are still valid.

But in other ways the situation today is quite diff erent from 35 years ago. At the same NBER meeting Coase also commented upon his celebrated article from 1937 (‘The Nature of the Firm’) with the words ‘much cited and little used’. This comment turned out to be a truthful description of the situation in 1971, but not true for the rest of the 1970s. In the same year as the NBER meeting (1971) Oliver E. Williamson published a seminal article in the American Economic Review, which was the start of a large number of books and articles that, like Coase, centred on the importance of transac-tion costs in analyses of economic organizations. A new fi eld of transaction cost economics emerged.

Some other articles from which new fi elds of research have emanated

Preface ix

were also published in the 1970s. The team production and the property rights perspective introduced by Alchian and Demsetz (1972) and the corporate governance perspective in Jensen and Meckling (1976) have been especially infl uential. From the 1980s the evolutionary theory of the fi rm presented by Nelson and Winter (1982) and the new property rights approach by Grossman and Hart (1986) have reshaped research in a similar fashion.

These and other branches of the growing tree of the theory of the fi rm were the sources of inspiration for the workshop on ‘The Economics of the Modern Firm’.

The output from the two workshops is merged in this book under the title The Modern Firm, Corporate Governance and Investment. To merge contributions from the two workshops makes sense given the close con-nection between the topics and papers presented at the workshops. For example, several papers at the second workshop were on corporate govern-ance. In all, 14 papers from the two workshops have been selected, nine from the second workshop and fi ve from the fi rst. The keynote addresses of Oliver E. Williamson and Dennis C. Mueller at the second workshop are the fi rst two chapters after the introduction.

The participants at the workshops and the referees of the diff erent articles in this book have helped to improve the contents. We thank them for their questions and comments. The workshops and the preparation of this book were fi nanced by Sparbankstiftelsen Alfa, Torsten and Ragnar Söderberg´s foundation and CESIS (Center of Excellence for Science and Innovation Studies). We are grateful for their support that allowed us to engage in this research. We are also indebted to Ibteesam Hossain and Maria Eriksson for excellent research assistance with this book.

1

1. Introduction: the modern fi rm, corporate governance and investmentPer-Olof Bjuggren and Dennis C. Mueller

The book is organized into four parts. Part I contains overviews of the theory of the fi rm. Part II is devoted to fi rms and organization of eco-nomic activities. Part III deals with how the institutional framework of an economy aff ects investments made by fi rms. Part IV looks at the impact of ownership structure and board composition on fi rm performance.

I. OVERVIEWS

Part I contains two overviews of the theory of the fi rm from diff erent per-spectives. ‘Opening the black box of fi rm and market organization: anti-trust’ by Oliver E. Williamson presents an overview of the characteristics of the transaction cost approach to the study of economic organization. The antitrust implications of this new view of economic organization are also considered. Thus, this chapter reviews both the positive and norma-tive aspects of Williamson’s theory of the fi rm, and off ers a contractual view of economic organization. The black box of the fi rm is opened in the sense that the governance attributes that distinguish the fi rm from the market are outlined. The market of the ‘pure vanilla’ type (spot contract character) found in most textbooks is complemented by the contractual deviations that can be characterized as hybrids of market and fi rm. The new explanations of antitrust phenomena provided by transaction cost analysis are discussed. Instead of solely focusing on market power aspects of vertical market relations, pricing practices and horizontal and conglom-erate mergers, a transaction cost analysis provides a broader picture by also including cost-reducing explanations. Williamson shows how these alternative explanations gradually have been recognized by US antitrust authorities.

‘The corporation: an economic enigma’ by Dennis C. Mueller looks

2 The modern fi rm, corporate governance and investment

at how the view on the corporate form of business has changed amongst economists since Adam Smith. The chapter addresses the key issue in cor-porate governance about effi ciency implications of ownership and control in corporations. An overview is provided on how economists’ views of the corporation and its performance had changed over time. The early economists such as Adam Smith, John Stuart Mill and Alfred Marshall off ered descriptions of corporate behaviour based on their observations of how companies function. Berle and Mean’s book The Modern Corporation and Private Property from 1932 is also based on observations that are sup-ported by an impressive amount of descriptive data. The neoclassical view emerging during the 1930s and 1940s represents a diff erent way of doing research on the fi rm and corporate form. For pedagogical and simplifying reasons the fi rm is looked upon as a profi t maximizing entity. The mana-gerial challenge of this neoclassical view and the ongoing debate between these two schools of thoughts are then discussed. One way to resolve the confl ict between these two views is to look at the return on invest-ments. Such studies have been done recently and show that investment effi ciency has actually improved since the 1990s in some countries like the United States. Possible explanations are disciplining takeovers, increased product competition due to globalization and the growth of institutional shareholding.

II. ORGANIZATION OF ECONOMIC ACTIVITIES

In Part II, chapters studying the fi rm from an economic organization per-spective are found. The fi rst chapter, by Per-Olof Bjuggren and Johanna Palmberg, (entitled ‘A contractual perspective on the fi rm with application to the maritime industry’) introduces a contractual model of the fi rm and applies it to explain how the maritime sector is organized. The capacity of the fi rm as a legal person to enter into contracts with suppliers of goods and services, customers and creditors is highlighted. It is argued that mutual dependency is what determines the character of contractual relations. The employment contract and ownership of assets in adjacent vertical stages enables the fi rm to supplant price as coordination mechanism in the pro-duction of goods and services. The maritime industry off ers a rich fl ora of contractual relations due to diff ering degrees of mutual dependence between shipper and carrier. Both the fi rm and the freight contract are analysed from a contractual perspective. A contractual explanation is also off ered for the phenomenon of third-party management.

A second chapter, by Kirsten Foss, (entitled ‘Authority in the knowledge economy’) takes a closer look at authority relations between employer and

Introduction 3

employees. In transaction cost economics, it is claimed that the possibility to use authority as a mode of coordination is what primarily characterizes fi rms. Authority is a key concept in the theory of the fi rm, and Foss throws light upon it. She claims that the emerging knowledge economy makes it necessary to take a closer look at the diff erent dimensions of the authority concept in order to understand ongoing changes in economic organization. From a review of authority in the economic literature, the conditions under which it is effi cient to use authority for coordination, contract enforce-ment, and dispute resolution are identifi ed. Finally, how these conditions have to be adapted to an emerging knowledge economy is discussed.

The third chapter, by Gunnar and Åsa Eliasson, (entitled ‘Competence and learning in the experimentally organized economy’) off ers a new evo-lutionary perspective on how fi rms emerge and disappear. The authors picture an economy with boundedly rational and myopic actors who try to take advantage of perceived business opportunities. Success is to a large extent dependent on the interaction of actors in so called competence blocs. In a successful competence bloc, customers, innovators, entrepreneurs, fi nanciers, exit markets and industrialists interact effi ciently in the sense of minimizing the cost of keeping losers on for too long and losing the winners. The customer has a key function in a bloc by being the ultimate arbiter of value. In the experimentally organized economy that Gunnar and Åsa Eliasson envision, economic organization and ‘the fi rm’ are a result of how the competence bloc is structured. Effi cient ways to organize the relations between the actors in a bloc will be rewarded by increasing returns, which make the organization viable.

III. IMPORTANCE OF THE INSTITUTIONAL ENVIRONMENT

Part III consists of chapters dealing with investment and legal environ-ment. Johan Eklund’s contribution (entitled ‘Corporate governance and investment in Scandinavia – ownership concentration and dual-class equity structure’) looks at how ownership structure aff ects investment perform-ance in the diff erent Scandinavian countries. As a measure of investment performance, the marginal q developed by Dennis Mueller and Elisabeth Reardon is used. From legal and political perspectives the Scandinavian countries are rather similar. But there are still distinctive diff erences in separation of ownership and control through the use of dual-class shares. Sweden has the highest fraction of listed fi rms that use dual-class shares, while Norway has the lowest fraction. The implications of these diff erences on investment performance are investigated. It turns out that in Norway


the marginal q estimate indicates overall effi cient investment performance among the listed fi rms, while the estimate for Swedish and Danish fi rms indicates over-investment – marginal returns on investment are lower than costs of capital. A non-linear eff ect of ownership concentration in Scandinavian fi rms is also found, implying a positive but marginally decreasing eff ect of ownership concentration on investment returns.

In a second chapter (‘The cost of legal uncertainty: the impact of insecure property rights on cost of capital) Per-Olof Bjuggren and Johan Eklund study how institutional risk infl uences the required return on international investments. Institutional risk due to weak property rights and investor right protection represents a non-diversifi able risk to international inves-tors, as these rights are fairly stable over time. Hence investors are likely to demand a risk premium in those countries where these rights are weak. The required return on investment in such countries will accordingly be higher. The capital asset pricing model (CAPM) is used to test for the importance of taking institutional risk into consideration, and to fi nd out the risk premium associated with institutional risk. It turns out that the explanatory power of the CAPM is considerably increased if such a risk is taken into account. Furthermore, the risk premium due to institutional risk is found to be signifi cantly higher for developing than for developed countries.

A third chapter, by Simon Deakin and Ajit Singh, (‘The stock market, the market for corporate control and the theory of the fi rm: legal and economic perspectives and implications for public policy’) takes up the question of how important an active market for corporate control is for economic effi ciency. The authors have severe doubts about whether hostile takeovers have positive eff ects on effi ciency and growth in developed countries. Their discussion of the pros and cons of a market for corpo-rate control starts with the observation that shareholders do not ‘own a company’ in the sense of being entitled to ‘a particular segment or portion of the company’s assets, at least while it is a going concern’. Furthermore, directors’ fi duciary interests of loyalty and care are owed to the company. Even though in practice it is foremost the interests of the shareholders that are catered to, other stakeholders’ interests are to a diff ering degree also recognized. This is especially the case in civil law systems. It is argued that it cannot be taken for granted that company law and articles of associa-tion that serve as obstacles to takeovers are at the expense of economic effi ciency.

Two strands of thought in the economic literature are referred to in Deakin and Singh’s economic analysis; On one hand, there is the principal–agent view of the market for corporate control that is used in fi nancial eco-nomics. On the other hand, there is a more antitrust oriented analysis used

Introduction 5

in industrial organization. The views of these schools diff er. While fi nancial economists have focused on takeovers and mergers as a mechanism to dis-cipline managers, industrial economists also stress their negative eff ects on the overall economy. Balancing diff erent views of effi ciency implications, Deakin and Singh come to the conclusion that hostile takeovers are likely to harm the prospects for growth in developing and transition economies.

IV. OWNERSHIP STRUCTURE, BOARD COMPOSITION AND FIRM PERFORMANCE

Part IV contains chapters that examine how the composition of the board, management relations and ownership structure aff ect fi rm per-formance. A fi rst chapter by Daniel Wiberg (‘Institutional ownership and dividends’) studies the relationship between institutional ownership and dividends. Wiberg wants to see both whether there is a positive relation between institutional ownership and dividends, and whether there are more rational reasons than ‘short-termism’ for explaining such a relation. Swedish data are used to test the hypotheses. The Swedish institutional framework is interesting as there is widespread use of dual-class shares and the tax rules make dividends more attractive to institutional than to other ownership categories. Through the use of dual-class shares, ownership can be separated from control, leading to pronounced agency problems. One way to overcome these agency problems is to insist on dividends. If such a relationship exists it implies that dividends are higher in fi rms with greater separation between ownership and control due to dual-class shares. Wiberg fi nds this to be the case, and that institutional ownership has a posi-tive impact on dividend growth.

Camille Madelon and Steen Thomsen (‘Contracting around ownership: shareholder agreements in France’) use data from large, French listed fi rms to examine the eff ects and determinants of shareholder agreements. These agreements represent a way to contract around the offi cial ownership structure. While the relationship between observable formal ownership and behaviour/performance has been extensively studied, there has been no study of the relationship between real ownership structure (consider-ing contracts between shareholders as well) and behaviour/performance. Madelon and Thomsens’s study is one of the fi rst steps towards ‘fi lling this void’. It is an explorative study that analyses agreements from a trans-action cost approach view. The costs and benefi ts of acquiring control through contracting amongst shareholders are compared with the alterna-tive of outright ownership. Several theoretical propositions are derived that consider ownership and industry characteristics and network ties as


explanations as to why contractual agreements are chosen. Propositions about the impact of shareholders’ agreements on economic performance are also derived.

‘Board governance of family fi rms and business groups with a unique regional dataset’, by Lluís Bru and Rafel Crespi, is both a methodological paper about how to empirically study family business and a description of what family businesses look like in the Balearic region of Spain. They have managed to trace family ownership and management by use of the Spanish ‘two-surnames’ system. This system has two features. Married women usually do not change their name and newborns have both the father’s and the mother’s surname. This surname system has made it possible to trace both ownership and involvement in boards and management by family members. Besides family companies, it has been possible to trace business groups under the control of associated families. The importance of family fi rms and groups in the Balearic economy and the characteristics of the diversifi cation patterns of family business groups are described.

Reidar Øystein Strøm (‘Better fi rm performance with employees on the board?’) uses data from Norwegian listed fi rms to analyse how co-determination aff ects performance. He distinguishes between direct and indirect eff ects of employee directors. In the theoretical literature both positive and negative eff ects of employee directors on performance are envisioned. Employee directors might contribute to positive performance by bringing more information about how the fi rm functions and enhanc-ing the incentives to invest in fi rm-specifi c human capital. On the other side, owners’ and employees’ interests might not be aligned, producing a negative eff ect on performance. Most empirical studies fi nd a negative impact on performance. Strøm takes the analysis one step further by taking account of the reactions of the shareholders to anticipated negative eff ects of employee directors. By adjusting the composition of the board and the fi nancial leverage of the fi rm, these negative eff ects can be counteracted. This indirect eff ect is taken into consideration in a simultaneous equations framework. A three-stage least squares methodology with fi xed eff ect is used to estimate both the direct and indirect eff ects. Even though indirect endogenous eff ects are taken into account, the results of the empirical analysis show a negative impact of employees on the board.

Wolfgang Drobetz, Klaus Gugler and Simone Hirschvogl (‘The determi-nants of German corporate governance ratings’) analyse corporate govern-ance rating in Germany. As in many other European countries, Germany has since 2002 had a Corporate Governance Code of a ‘comply or explain’ kind. Instead of assessing the impact of code fulfi lment on performance of fi rms, Drobetz, Gugler and Hirschvogl choose to analyse the determinants of corporate governance rating. One advantage with this approach is that

Introduction 7

an endogeneity problem due to self-selection is avoided. The analysis is to a large extent based on the assumption that there is a positive relation between fi rm performance and a high corporate governance rating. The determinants of performance studied are ownership concentration, the size of the supervisory board, choice of strict accounting rules, and the use of an option-based remuneration plan. Ownership concentration is hypothesized to be non-linearly related to ratings. The rationale is that it is only at high levels of ownership concentration that the entrenchment eff ect of ownership is balanced by the rewards associated with better per-formance. For the other determinants, a negative eff ect is found for board size, stricter accounting rules are positively related to performance, and stock-option schemes have a positive relation to rating. All the hypotheses are corroborated in the empirical analysis based on survey data from 91 German fi rms.

Mogens Dilling-Hansen, Erik Strøjer Madsen and Valdemar Smith (‘Top management, education and networking’) use Danish data to study how management can benefi t from networking. Networking takes place through board participation by top management. They look at network ties between fi rms linked by ownership and between independent fi rms. The former ties are labelled internal ties while the latter are called external ties. Networking through external ties can increase the top management’s knowledge of the competitive and technological environment of the fi rm. It can also facilitate collusion. Networking can then be expected to improve performance. Networking of an internal character can improve the control of subsidiaries. The extent to which networking will have a positive infl u-ence on performance can be expected to be dependent on education. An empirical analysis of a large sample of Danish fi rms fi nds a signifi cant positive eff ect of internal network activities on fi rm performance, and that education implies a positive attitude towards networking. No other signifi cant relationships can be traced.

PART I

Key issues

11

2. Opening the black box of fi rm and market organization: antitrust*Oliver E. Williamson

The task of linking concepts with observations demands a great deal of detailed knowledge of the realities of economic life.

– Tjalling Koopmans

Opening the black box of fi rm and market organization and examining the mechanisms inside is a defi ning characteristic of the transaction cost approach to the study of economic organization (Arrow, 1987, 1999; Dixit, 1996; Kreps, 1990). But questions remain. Do the details matter for a wide range of phenomena or only a few? Which, among the endless number of details that could be recorded, have conceptual and operational signifi -cance? What, if any, are the public policy ramifi cations?

My responses to these queries are that the details matter for a wide range of phenomena, that many relevant details are uncovered by examining economic organization through the focused lens of contract/governance,1 and that public policy toward business has been a benefi ciary. Antitrust applications are developed here. Regulatory applications are examined elsewhere (Williamson, 2007a).

I begin with a statement of the crisis in antitrust as of 1970. A synopsis of the microanalytic setup is then sketched in Section 2. The paradigm problem for transaction cost economics is the intermediate product market transaction, as described in Section 3. Antitrust applications are developed in Sections 4–8. Concluding remarks follow and there is an Appendix on the antecedents on which transaction cost economics builds.

1. THE CRISIS IN ANTITRUST

Victor Fuchs opens his foreword to the National Bureau of Economic Research 50th anniversary volume, Policy Issues and Research Opportunities in Industrial Organization, with the query ‘Whither industrial organiza-tion?’, to which he opines that ‘all is not well with this once fl ourishing fi eld’

12 Key issues

(1972, p. xv). Of the various answers that could be advanced to explain this decline, the ones to which I attach the greatest weight are the all-purpose reliance by industrial organization economists (and others) on a black box theory of the fi rm and a plain vanilla theory of markets. Because the fi rm was described as a production function that transformed inputs into outputs according to the laws of technology, non-technological or non-price theoretic explanations for reshaping the boundary of the fi rm were thought to be deeply problematic. Contractual deviations from simple market exchange were likewise regarded as suspect.

Since economists were dismissive of the possibility that the internal organization of transactions had important economizing consequences,2 vertical integration and other organizational practices that lacked a ‘physi-cal or technical aspect’ were presumed to have the purpose of increasing the ‘market power of the fi rms involved rather than reduction in cost’ (Bain, 1968, p. 381). Vertical market restrictions (and other deviations from simple market exchange) were also regarded as deeply problematic. As the then head of the Antitrust Division of the US Department of Justice put it, ‘I approach customer and territorial restrictions not hospitably in the common law tradition, but inhospitably in the tradition of antitrust.’3 Indeed, some protectionist antitrust enforcement offi cials regarded pro-spective effi ciency gains from a merger to be anticompetitive because less effi cient rivals would be disadvantaged.4 Such upside-down reasoning encouraged respondents to merger litigation to disclaim that any effi ciency benefi ts would accrue.5 Ronald Coase summarized the prevailing state of disarray as follows: ‘If an economist fi nds something – a business practice of one sort or other – that he does not understand, he looks for a monopoly explanation. As in this fi eld we are very ignorant, the number of ununder-standable practices tends to be rather large, and the reliance on monopoly explanation, frequent’ (1972, p. 67).

An altogether diff erent lens through which to examine complex con-tracting and economic organization would be needed to break the grip of such convoluted thinking. As described in Section 2, the lens of contract/governance describes fi rms and markets as governance structures, the dif-ferent mechanisms of which matter in effi ciency respects. An economics of organization takes shape in the process as monopoly is reduced to an important but special case.

Opening the black box of fi rm and market organization 13

2. THE MICROANALYTICS: A SYNOPSIS

2.1 General

As Herbert Simon observes (1984, p. 40):

In the physical sciences, when errors of measurement and other noise are found to be of the same order of magnitude as the phenomena under study the response is not to try to squeeze more information out of the data by statistical means; it is instead to fi nd techniques for observing the phenomena at a higher level of resolution. The corresponding strategy for economics is obvious: to secure new kinds of data at the micro level.

Inasmuch, however, as the social sciences are ‘hypercomplex’ (Wilson, 1998, p. 183; Simon, 1957, p. 89), the details proliferate. Where precisely do the relevant microanalytics reside? That will vary with the phenomena to be investigated and the lens through which the phenomena are viewed.

2.2 The Rudiments6

Rather than operate out of the neoclassical lens of choice (with emphasis on prices and output, supply and demand, in relation to which organiza-tion is held to be unimportant), transaction cost economics works out of the lens of contract/governance. The building blocks are transactions and governance structures and the effi cient alignment thereof, whereupon organization is not only important but is susceptible to analysis.7 In addi-tion to simple market exchange (contract as legal rules), provision is made for hybrid contracting (contract as framework, for which continuity of the exchange relationship is important) and hierarchy, each of which is described as an alternative mode of governance. Note that the decision to use one mode of governance rather than another depends on the transac-tions for which governance support is required. Hitherto neglected trans-action costs take their place in the analytical fi rmament.

If both transactions and governance structures diff er, then the relevant microanalytics for describing both of these will need to be worked out. Herbert Simon’s advice, to little discernible eff ect, that ‘Nothing is more fundamental in setting our research agenda and informing our research methods than our view of the nature of the human beings whose behavior we are studying’ (1985, p. 303) is pertinent in this connection. Cognitive competence is especially relevant, but so too is the manner in which self-interest is described.

If, for example, human actors possess the cognitive ability to imple-ment comprehensive contingent claims contracting, then we are in the

14 Key issues

world of Arrow–Debreu and, in such a contractual world, organization is unimportant. If instead the list of six assumptions that are made by Drew Fudenberg, Bengt Holmstrom and Paul Milgrom (1990) apply, then we are in a world where a sequence of short-term contracts can implement an optimal long-term contract.8 More generally, the point is this: diff er-ent assumptions about cognition lead into diff erent theories of contract and organization (and the same holds for descriptions of self-interest (Williamson, 1985, pp. 64–7)).

Transaction cost economics describes both cognition and self-interest in a two-part way. Specifi cally, cognition combines bounded rationality with feasible foresight while self-interest joins benign behavior with opportun-ism. Thus all complex contracts are unavoidably incomplete (by reason of bounded rationality) yet human actors are assumed to have the capacity to look ahead, recognize hazards, work out the mechanisms, and, albeit imperfectly, factor the ramifi cations back into the ex ante contractual design (which is a manifestation of feasible foresight). Also, most human actors will do what they agree to and some will do more most of the time (benign behavior), but outliers for which the stakes are great will elicit defection and/or posturing (which are manifestations of opportunism) with the purpose of inducing renegotiation.

Whether contractual incompleteness (bounded rationality) and defec-tion hazards (opportunism) pose serious governance issues depends on the attributes of transactions. Transactions for which continuity of the exchange relationship is important and for which coordinated adaptations are needed to restore effi ciency are those for which the effi cacy of simple market exchange breaks down. Behavioral attributes in combination with transactional attributes thus underpin the need for added governance supports (or not) – where governance is defi ned as the means by which to infuse order, thereby to mitigate confl ict and realize mutual gain.

The three attributes of transactions that are especially important for preserving continuity by implementing coordinated adaptations are asset specifi city (which is a measure of bilateral dependency, to which maladap-tation hazards accrue), uncertainty (the disturbances, small and great, to which transactions are subject), and the frequency with which transactions recur, which has a bearing on both reputation eff ects (in the market) and private ordering mechanisms (within fi rms).

Governance structures are described as discrete structural syndromes of attributes that diff er in their adaptive capacities, of which two types are distinguished: autonomous adaptation to changes in relative prices as described by Friedrich Hayek (1945) and coordinated adaptations of a con-scious, deliberate, purposeful kind as described by Chester Barnard (1938). Incentive intensity, decision and administrative control instruments, and


contract law regimes are the defi ning attributes with respect to which gov-ernance structures are described.

As discussed in Section 3, the three main modes of governance for organ-izing intermediate product market transactions are market, hybrid, and hierarchy. Interestingly, but not surprisingly, spot markets and hierarchies are polar opposites – in that spot markets are characterized by high-pow-ered incentives, negligible administrative control, and a legal rules contract law regime, thereby to support autonomous adaptation, whereas hierar-chies use low-powered incentives and hands-on administrative control, and settle internal disputes administratively under a forbearance law regime9 in support of coordinated adaptation. Hybrid contracting is located between market and hierarchy in all three attributes and in both adaptation respects and thus can be thought of as a compromise mode.

The discriminating alignment hypothesis provides the predictive link between transactions and governance structures – to wit, transactions, which diff er in their attributes, are aligned with governance structures, which diff er in their costs and competences, so as to eff ect a transaction cost economizing match.

3. TRANSACTIONS IN THE INTERMEDIATE PRODUCT MARKET: THE PARADIGM TRANSACTION

The intermediate product market transaction (or in more mundane terms, the make-or-buy or outsourcing decision) is the obvious paradigmatic transaction for transaction cost economics for two reasons. First, this is the transaction to which Coase referred in pointing up a lapse in economic theory in 1937: ‘The purpose of this paper is to bridge what appears to be a gap in economic theory between the assumption (made for some purposes) that resources are allocated by means of the price mechanism and the assumption (made for other purposes) that this allocation is dependent on . . . [hierarchical mechanisms]. We have to explain the basis on which, in practice, this choice between alternatives is eff ected’ (Coase, 1937, p. 389). Secondly, intermediate product market transactions are simpler than are labor market, capital market, and fi nal product market transactions because they are less beset with asymmetries of information, budget, legal talent, risk aversion, and the like. The simple contractual schema, as described herein, focuses on the intermediate product market transaction but applies (with variation) to the study of transactions more generally.

Thus assume that a fi rm can make or buy a component and assume

16 Key issues

further that the component can be supplied by either a general purpose technology or a special purpose technology. Letting k be a measure of asset specifi city, the transactions in Figure 2.1 that use the general purpose technology are ones for which k 5 0. In this case, no specifi c assets are involved and the parties are essentially faceless. Transactions that use the special purpose technology are those for which k . 0. Such transactions give rise to bilateral dependencies, in that the assets cannot be redeployed to alternative uses and users without loss of productive value. The parties therefore have incentives to promote continuity, thereby to safeguard specifi c investments. Let s denote the magnitude of any such safeguards, which include penalties, information disclosure and verifi cation proce-dures, specialized dispute resolution (such as arbitration) and, ultimately, integration of the two stages under unifi ed ownership. An s 5 0 condition is one for which no safeguards are provided; a decision to provide safeguards is refl ected by an s . 0 result.

Node A in Figure 2.1 corresponds to the ideal transaction in law and economics: there being an absence of dependency, governance is accomplished through competition and, in the event of disputes, by court awarded damages. Node B poses unrelieved contractual hazards, in that specialized investments are exposed (k . 0) for which no safeguards (s 5 0) have been provided. Such hazards will be recognized by farsighted players, who will price out the implied risks.10 Confronted with the added costs of these hazards, buyers have the incentive to mitigate the hazards (in

B (unrelieved hazard)

A (unassisted market)

h = 0

h > 0

s = 0

s > 0

D (integration)

administrative

market safeguard

C (crediblecommitment)

Figure 2.1 Simple contractual schema


cost-eff ective degree), which is to say that node B is an ineffi cient mode of governance for ongoing (as against episodic) supply purposes.

Added contractual supports (s . 0) are provided at Nodes C and D. Node C governance corresponds to what Karl Llewellyn referred to as con-tract as framework, as distinguished from contract as legal rules, where the former better preserves continuity of the transaction through ‘a framework highly adjustable, a framework which almost never accurately describes real working relations, but which aff ords a rough indication around which such relations vary, an occasional guide in cases of doubt, and a norm of ultimate appeal when the relations cease in fact to work’ (1931, pp. 736–7). This is the aforementioned hybrid transaction where credible contracting mechanisms are introduced in support of cooperative adaptation. Such hybrids are not, however, ‘indefi nitely elastic. As disturbances become highly consequential, . . . an incentive to defect [arises]. The general propo-sition here is that when the “lawful” gains to be had by insistence upon literal enforcement exceed the discounted value of continuing the exchange relationship, defection from the [cooperative] spirit of the contract can be anticipated’ (Williamson, 1991a, p. 273). Benjamin Klein subsequently describes ‘the self-enforcing range’ similarly: if and as ‘changes in market conditions move outside the self-enforcing range, . . . the one-time gain from breach [will] exceed the private sanction’ (1996, p. 449).

But this is not the end of the governance story. As the expected mal-adaptation costs of hybrid contracting progressively mount, best eff orts to craft cost-eff ective credible commitments notwithstanding, transaction cost economics predicts that transactions will be removed from the hybrid and organized under unifi ed ownership (vertical integration). Inasmuch as added bureaucratic costs accrue upon taking a transaction out of the market and organizing it internally, hierarchy is usefully thought of as the organization form of last resort: try markets, try hybrids, and have recourse to the fi rm (Node D) only when all else fails.

Node D governance (hierarchy) involves (1) unifi ed ownership of succes-sive stages, (2) coordinated adaptation at the interfaces by the application of routines (to manage disturbances in degree) and by the use of hierarchy (to manage disturbances in kind), (3) internal dispute resolution for settling disputes that cannot be resolved by the parties by appealing these to a common ‘boss’, and (4) the aforementioned bureaucratic cost burdens.

Transaction cost economics thus predicts that generic (k 5 0) transac-tions will be assigned to Node A (the market mode, where continuity is of no importance and disputes are settled in court), more complex transac-tions (k . 0) to Node C (the hybrid mode, where continuity matters and adaptations are accomplished under the more elastic concept of contract as framework), that very complex transactions (k . . 0) will be taken

18 Key issues

out of the market and organized within hierarchy at Node D, and that few transactions (mistakes or adventitious transactions) will be located at ineffi cient Node B.

What is furthermore noteworthy is that empirical tests of the predictions of the theory have ensued and have been broadly corroborative. Indeed, ‘despite what almost 30 years ago may have appeared to be insurmount-able obstacles to acquiring the relevant data [which are often primary data of a microanalytic kind], today transaction cost economics stands on a remarkably broad empirical foundation’ (Geyskens, Steenkamp and Kumar, 2006, p. 531). This applies, moreover, not merely to the tests of the paradigm problem of vertical integration but to a vast variety of other phenomena that are interpreted as variations on a theme (Macher and Richman, 2006). There is no gainsaying that transaction cost economics has been much more infl uential because of the empirical work that it has engendered (Whinston, 2001).

4. APPLICATIONS TO ANTITRUST: GENERAL

The over-reaching excesses of monopoly reasoning during the 1960s contained the seeds of their own destruction. Confronted with escalat-ing implausibility, Supreme Court Justice Potter Stewart, in a dissenting opinion, observed that the ‘sole consistency that I can fi nd is that in [merger] litigation under Section 7, the Government always wins.’ 11 Alarmist excesses of monopoly reasoning eventually elicited a series of challenges – to include both allocative effi ciency and transaction cost reasoning,12 where the latter made express provision for economies of organization, the myopic quality of entry barrier reasoning was confronted with remediableness considerations, nonstandard and unfamiliar contracting practices that had been declared to be anticompetitive under the inhospitality tradition were re-examined and found, often, to serve credible contracting purposes, and selective appeal to zero transaction costs was supplanted by an insistence that positive transac-tion costs be recognized wheresoever they may reside.13

Antitrust scholars from the ‘Chicago School’ (Stigler, 1968; Demsetz, 1974; Posner, 1976; Bork, 1978) receive and deserve much of the credit, but transaction cost economics was also a contributing factor. Thus whereas ‘the Chicago School focused on explaining why vertical integration and nonstandard vertical contracts did not create or enhance market power . . . transaction cost economics focused on why these vertical arrangements emerged as cost-reducing responses to certain transactional characteris-tics’ (Joskow, 1991, p. 56; emphasis added). Without such an affi rmative rationale,


it is hard to believe that the Chicago critique of antitrust policies regarding vertical arrangements would have had as much infl uence, especially among professional economists and antitrust scholars, . . . for [whom] the theoretical and empirical work in transaction cost economics . . . demonstrated that previ-ously suspect vertical arrangements often could be explained as contractual and organizational responses motivated by a desire to reduce the cost of transacting. (Joskow, 1991, p. 57)

As between critiques of wrong-headed reasoning and explanations for the practices in question for which the mechanisms have been expressly worked out, the latter is the more demanding.

Interestingly, Timothy Muris, during his term as chair of the Federal Trade Commission, held that much of the New Institutional Economics ‘literature has signifi cant potential to improve antitrust analysis and policy. In particular, . . . [the transaction cost branch has] focused on demystifying the “black box” fi rm and on clarifying important determi-nants of vertical relationships’ (2003, p. 15). Opening the black box and acquiring an understanding of the mechanisms inside has had an impact, moreover, on practice:

The most impressive recent competition policy work I have seen refl ects the NIE’s teachings about the appropriate approach to antitrust analysis. Much of the FTC’s best work follows the tenets of the NIE and refl ects careful, fact-based analyses that properly account for institutions and all relevant theories, not just market structures and [monopoly] power theories. (Muris, 2003, p. 11; emphasis in original)14

A comprehensive examination of the applications of transaction cost economics to antitrust is beyond the scope of this chapter. My purpose is merely to illustrate the ways in which examining the microanalytics of complex contract and economic organization through the lens of contract/governance has served to alter and deepen our understanding of many antitrust related phenomena. I successively examine applications to verti-cal market relations, price theoretic issues, credible contracting, and the modern corporation.

5. VERTICAL MARKET RELATIONS

Lateral integration into components, backward into raw materials, and forward into distribution are successively examined. The analysis through-out tracks the logic of the simple contractual schema – in that the move from market to hybrid to hierarchy is predicted as asset specifi city and outlier disturbances increase. Asset specifi city refi nements − as among

20 Key issues

physical, human, site, dedicated, and brand name capital – are also con-sequential. With respect, for example, to mobile physical assets (such as specialized dies), it may be possible for the specialized investments to be made by the buyer, who relieves bilateral dependency by assigning the spe-cialized dies to the winning bidder for the duration of the supply contract and repossessing and reassigning these to a successor if the original bidder does not win the renewal contract.15 The need for unifi ed ownership is also relieved by the use of credible commitments to support hybrid contract-ing – as with exchange agreements, or for organizing distribution through a large number of geographically dispersed outlets by franchising rather than by forward integration (although there is also merit in dual distri-bution). As, however, asset specifi city and disturbances increase, unifi ed ownership is predicted.

5.1 Lateral Integration

Economies are commonly ascribed to the integration of successive stages in the ‘technological core’, an example of which is the unifi ed ownership of iron- and steel-making stages by reason of thermal economies (Bain, 1968, p. 381). By contrast, lateral integration into components that lack such a ‘physical or technical aspect’ is (under technological reasoning) believed to be deeply problematic. As discussed above, monopoly purpose and eff ect were commonly ascribed to these.

Transaction cost economics disputes such reasoning. All that is implied by thermal economies (or, more generally, by the physical or technical aspects to which Bain refers) is that the two stages be located adjacent to each other. The governance issue is whether the exchange of product across these co-located stages should be mediated by market or by hierarchy. Unless contractual problems are projected, there is no reason why each stage could not be independently owned and the two stages joined by an interfi rm contract. If, therefore, co-located stages are integrated, that is because transaction cost economies are thereby realized: unifi ed owner-ship relieves the contractual hazards that would otherwise arise between independent, site-specifi c trading entities.

But there is more: transaction cost economics also selectively off ers an economizing interpretation for transactions that lack the ‘physical or technical aspects’ to which Bain refers. As discussed in Section 3 above, the outsourcing of separable components of a non-site-specifi c kind is the paradigm problem on which transaction cost economics is based and to which empirical tests were fi rst applied (Monteverde and Teece, 1982; Masten, 1984).16 The upshot is that the same comparative contractual logic applies to the organization of asset-specifi c transactions of all kinds,


site-specifi c or not. The contrast with earlier antitrust predilections is stark.17

5.2 Raw Materials Procurement

Except perhaps for very atypical cases, an effi ciency case for vertical integration backward into raw materials is believed to be rare if not non-existent. Surely the lesson of the Ford Motor Company’s ‘fully integrated behemoth at River Rouge, supplied by an empire that included ore lands, coal mines, 700,000 acres of timberland, sawmills, blast furnaces, a glass works and coal boats, and a railroad’ (Livesay, 1979, p. 175) is that this was vertical integration run amok.

Exactly right: maybe comprehensive vertical integration has the appear-ance of being an engineer’s dream, but it is not an economic ideal. As John Stuckey’s examination of backward integration from the refi ning into the raw materials stage in the Australian aluminum industry reveals, the transactional details matter. Bauxite ore, it turns out, is not a uniform mineral but, instead, is ‘a heterogeneous commodity, . . . [where] the ore in any deposit has unique chemical and physical properties’ (Stuckey, 1983, p. 290). That is consequential: the cost diff erence of processing a mixed-hydrate bauxite, which is effi ciently processed with a high-temperature technology, in a low-temperature refi nery instead, comes to almost 100 percent (Stuckey, 1983, pp. 53–4). Other details also matter. Bauxite storage covers are needed for some ores and not for others (p. 49); residue processing costs vary greatly (p. 53); and air pollution equipment is tailored to the attributes of the bauxite (p. 60). Moreover, although smelting is less idiosyncratic, there is, nevertheless, an ‘art part of smelting’, which is upset if the aluminum supply is varied (p. 63).

Not every refi nery, however, is dependent on a specifi c bauxite deposit. Thus, whereas most of the above described economies are realized by spe-cializing the characteristics of a local refi nery to a local bauxite deposit (as in Australia), the same cannot be said for remotely located refi neries, as in Japan, where a general purpose refi nery that can process bauxite ores procured on the world market has countervailing advantages.

Interestingly, regulatory concerns sometimes get in the way of back-ward integration – an example of which is the bilateral dependency that sometimes arises between fuel source and operating stages in electricity generation by coal-burning generators (Joskow, 1987). Lest utilities ‘inte-grate backward into coal production to shift profi ts from a regulated to an unregulated activity, the regulatory process has discouraged this’ (Joskow, 1987, p. 284, n. 17).

As with bauxite, ‘The type of coal that a generating unit is designed to

22 Key issues

burn aff ects its construction and its design thermal effi ciency’ (Joskow, 1987, p. 284). In some regions, as in the Eastern United States, coal of relatively uniform quality is available from a large number of small nearby mines; in other regions, as in the West, deposits are large and coal quality variation among mines and the distances for shipment are great (1987, p. 284). ‘Mine mouth’ generating plants of specifi c design are often observed for the latter. More generally, comparative contractual reasoning predicts that longer-term and more nuanced contracts will be observed for the West than in the East, which is borne out by the data: ‘as relationship-specifi c investments become more important, the parties . . . fi nd it advantageous to rely on longer-term contracts that specify the terms and conditions of repeated transactions ex ante, rather than relying on repeated bargaining’ (Joskow, 1987, p. 296).

5.3 Forward into Distribution

A huge franchising literature in economics and marketing examines the decision of whether producers should own some or much of their distribu-tion system or contract with others to manage the distribution of goods and services instead. In the event of the latter, vertical market restrictions often apply, a common purpose being to protect the network against brand name devaluation (Klein and Leffl er, 1981).

Many economizing issues are posed by forward integration into market-ing and the uses of vertical market restrictions, of which asset specifi city (especially in the form of brand name capital) is only one. Transaction cost reasoning nevertheless plays a central role in the marketing decision (Coughlan et al., 2005) as to which contractual mode to choose and, if the market, whether contractual restrictions should be imposed. Contrary to the ‘inhospitality tradition’ in antitrust,18 vertical market restrictions will yield social benefi ts if the requisite transaction cost pre-conditions are satisfi ed.

6. PRICE THEORETIC ISSUES

6.1 Price Discrimination

The price theoretic argument in favor of price discrimination (especially perfect price discrimination) is that discrimination permits parties whose valuation is below a uniform monopoly price but above marginal costs to buy the good or service in question, as a result of which allocative effi ciency benefi ts accrue. A problem with the argument is that perfect


price discrimination assumes that the transaction costs of discovering true customer valuations and of policing against arbitrage are zero, which are heroic assumptions. Upon taking the costs of discovering price valuations and enforcing arbitrage restrictions into account, it can be shown that costly price discrimination can lead to both private benefi ts (monopoly profi ts increase) and social losses (Williamson, 1975, pp. 11–13).19

6.2 Robinson-Patman

Transaction cost economics also has a bearing on the Robinson-Patman Act, which has been interpreted as an eff ort ‘to deprive a large buyer of [dis-counts] except to the extent that a lower price could be justifi ed by reason of a seller’s diminished costs due to quantity manufacture, delivery, or sale, or by reason of the seller’s good faith eff ort to meet a competitor’s equally low price.’ 20 The concern, plainly, is that large buyers will use their muscle to extract better deals from suppliers, as a result of which smaller buyers will be disadvantaged. To this, however, should be added the possibility that diff erent buyers are prepared to off er diff erent contractual supports for the same good or service. With reference to Figure 2.1, suppose that a supplier uses specialized assets to produce the same good or service for two buyers. Assume that one of the buyers refuses to off er contractual safeguards while the other does off er safeguards. These two correspond to Node B and Node C contracting, respectively. Plainly, the supplier will sell on better terms to the Node C buyer than to the Node B buyer.

The upshot is that quantity and meeting competition considerations do not exhaust the legitimate reasons for off ering lower prices to some buyers than to others. Application of the lens of contract/governance, as against all-purpose reliance on textbook micro theory, serves to uncover these additional purposes.

6.3 Predatory Pricing

Transaction cost economics disputes the merits of the marginal cost pricing test for predatory pricing, as advanced by Philip Areeda and Donald Turner (1975), in two respects. First, although marginal cost pricing can be thought of as a hypothetical ideal (second best considerations aside), such an ideal is a deceptive standard if the measurement of marginal costs invites accounting manipulation and deceit in the courtroom. Additionally, Areeda and Turner apply the same marginal cost pricing test to price reductions of both continuing and temporary kinds – which is to say that they make no provision for strategic price reductions: now it’s there, now

24 Key issues

it isn’t, depending on whether a new entrant has appeared or been van-quished. That is unwarranted, since the welfare benefi ts of temporary price cuts are at best small and could easily be net negative.

Here as elsewhere, however, objections to a proposed criterion do not, alone, carry the day. There is an obligation to advance a superior feasi-ble alternative. The output test proposed in Williamson (1977) has three advantages over the marginal cost pricing test: (1) repositioning, (2) meas-urement, and (3) contingent versus continuing responses. Repositioning makes allowance for the possibility that parties to which predatory pricing rules apply will adapt (reposition) in relationship to them. Areeda and Turner ignore this incentive, yet it is noteworthy that their test has infe-rior repositioning properties in comparison with the output test. Output, moreover, is much easier to measure than is marginal cost. And the output test expressly favors continuing over contingent supply – now it’s here, now it isn’t, depending on whether an entrant has appeared or perished – by the established fi rm. The upshot is that transaction cost considerations are very relevant for uncovering the effi ciency ramifi cations of two price theoretic tests for predation.

6.4 Over-searching

The market for gem-quality uncut diamonds employs two nonstandard contracting practices that are puzzling at best and are easily interpreted as eff orts by de Beers to exercise muscle in its dealings with the buyers of uncut diamonds. The two trading restrictions in question are the ‘all-or-none’ and ‘in-or-out’ trading rules. Inasmuch as de Beers had market power in the supply of uncut diamonds, these trading rules were believed to have the muscular purpose of extracting profi t from diamond cutters.

Although the web of cooperative practices among diamond cutters in New York (Richman, 2006) might be interpreted as collusive, the de Beers trading rules applied to a global market. Consider therefore the possibility raised by Roy Kenney and Benjamin Klein (1983) that these rules have effi ciency purposes.

Whereas uncut diamonds are classifi ed into more than two thousand cat-egories, signifi cant quality variation in the stones evidently remains. How can such a market be organized so as to reduce the oversearching costs that would be incurred if buyers were to evaluate every stone, or at least every grouping of stones, off ered by de Beers? The combination of all-or-none with in-or-out trading rules arguably serves to reduce over-searching.21

The all-or-none trading rule requires that a buyer accept the entire grouping of diamonds assembled by de Beers (a ‘sight’) or none at all. Buyers are thereby denied the opportunity to pick and choose among


individual diamonds, yet nonetheless have the incentive to inspect each sight very carefully. Refusal to accept a sight would signal that the sight was over-priced – but no more.

Suppose now that an in-or-out trading rule is added. The decision to refuse a sight now has much more serious ramifi cations. To be sure, a refusal could indicate that a particular sight is egregiously over-priced. More likely, however, it refl ects a succession of bad experiences. It is a public declaration that de Beers is not to be trusted. In eff ect, a disaff ected buyer announces that the expected net profi t of dealing with de Beers under these constrained trading rules is negative.

Such an announcement has a chilling eff ect on the market. Buyers who were earlier prepared to make casual sight inspections are now advised that there are added trading hazards. Everyone is put on notice that a confi -dence has been violated and is warned to inspect more carefully.

On this interpretation, the in-or-out trading rule is a way of encouraging buyers to regard the procurement of diamonds not as independent trading events but as a related series of trades. If, overall, things can be expected to ‘average out’, then it is not essential that the payment made for value received corresponds exactly on each sight. In the face of systematic under-realiza-tions of value, however, buyers will be induced to quit. If, as a consequence, the system is moved from a high to a low trust trading culture, then the costs of marketing diamonds increase. That is an adverse outcome to the system which de Beers has strong incentives to avoid. Accordingly, in a regime where both all-or-none and in-or-out trading rules apply, de Beers will take greater care to present sights such that the legitimate expectations of buyers will be achieved. The combined rules thus infuse greater integrity of trade.

7. CREDIBLE COMMITMENTS

Although credible contracting is the core purpose served by hybrid modes of governance, such a purpose was slow to register in antitrust enforcement – mainly because of the monopoly predisposition with which nonstandard and unfamiliar contracting practices were viewed. But whereas ‘traditional market power theories [were so predisposed], . . . TCE can [frequently] . . . illuminate the meaning of facts – particularly in the context of complex contractual relations – that cannot otherwise be explained, or worse, are explained incorrectly’ (Muris, 2003, p. 18). Credible commitment reason-ing (of a Node C versus Node B kind) has been applied to a wide range of contractual practices – including franchise restrictions, exchange agree-ments, take-or-pay agreements, and a host of other nonstandard con-tracting practices (Masten, 1996). Exchange agreements are an especially

26 Key issues

interesting illustration of opening the black box and interpreting the pur-poses served by the mechanisms inside.22

Petroleum exchanges have puzzled economists for a very long time and have been routinely challenged in antitrust cases and investigations of the petroleum industry. The 1973 case brought by the United States Federal Trade Commission against the largest petroleum fi rms maintained that exchanges were instrumental in maintaining a web of interdependencies among major fi rms, thereby helping to eff ect an oligopolistic outcome in an industry that was relatively unconcentrated on normal market struc-ture criteria.23 A later study, The State of Competition in the Canadian Petroleum Industry, likewise held that exchanges were objectionable.24 The Canadian Study, moreover, produced documents – contracts, internal company memoranda, letters, and the like – as well as deposition testimony to support its views that exchanges are devices for extending and perfect-ing monopoly among the leading petroleum fi rms.25 Such evidence on the details and purposes of contracting is usually confi dential and hence una-vailable. But detailed knowledge is clearly germane – and often essential – to a correct assessment of the transaction cost features of a contract.

Engineers, managers, and lawyers in the major petroleum companies all had a benign interpretation of exchanges. If X has a surplus of product in region A and a defi cit in region B while Y has a surplus of product in region B and a defi cit in region A, and if both wish to market their product in both areas, then the exchange of product will save on cross-hauling. That, however, omits another possibility: why not create a central market into which each fi rm can report its surpluses and defi cits and procure in an anon-ymous rather than bilateral way? Petroleum industry engineers, managers, and lawyers found this query unsettling, yet the critical issue that needs to be faced is why bilateral exchange rather than simple market exchange?

The Canadian Study lists four objections to exchanges, the fi rst two of which I will pass over here (but see Williamson (1985, p. 148)). The other two are more intriguing: competition is impaired by conditioning supply on the payment of an ‘entry fee’ (pp. 53–4) and by exchange agreements that impose limits on growth and supplementary supply (pp. 51–2).

The antitrust concerns posed by the entry fee are supported by the fol-lowing documentation and interpretation (pp. 52–3; emphasis added):

Evidence of an understanding that a fee relating to investment was required for acceptance into the industry can be found in the following quotation from Gulf:

‘We do believe that the oil industry generally, although grudgingly, will allow a participant who has paid his ante, to play the game; the ante in this game being the capital for refi ning, distributing and selling products.’ (Document #71248, undated, Gulf)


The signifi cance of the quotation lies equally in the notion that an ‘entry fee’ was required and in the notion that the industry set the rules of the ‘game.’ The meaning of the ‘entry fee’ as well as the rules of the ‘game’ as understood by the industry can be found in the actual dealings between companies where the explicit mention of an ‘entry fee’ arises. These cases demonstrate the rules that were being applied – the rules to which Gulf was referring. Companies which had not paid an ‘entry fee,’ that is, companies which had not made a suffi cient investment in refi ning capacity or in marketing distribution facilities would either not be supplied or would be penalized in the terms of the supply agreement.

Once a comparative contractual perspective is adopted, a diff erent inter-pretation of these practices presents itself. So as to keep the comparison simple, suppose that there are two would-be buyers and that each places an order for a signifi cant and identical amount of product for delivery over the same time interval with the same supplier. The buyers diff er, however, in that one of the buyers is prepared to create a safeguard to deter premature termination while the other is not. It is elementary that the seller will charge a higher (Node B) price to the latter.

But wherein do exchange agreements relate to such trades? Given that the amount of product to be supplied is signifi cant, and assuming that the supply interval is long and that the surplus/defi cit geographic relations described above apply, then buyers and sellers so situated will fi nd that an exchange agreement between them not only saves on cross-hauling costs but, additionally, provides a reciprocal credible commitment – in that ter-mination by one party is deterred by the expectation that it will be answered in kind. Especially if both parties to the exchange agreement experience correlated disturbances, in which event both will want to adapt similarly, exchange agreements have good adaptation and security properties.

Assuming that each party to such a supply agreement constructs and main-tains a larger plant than it otherwise would, the specifi c investments made by these fi rms take the form of ‘dedicated assets’ – large incremental additions to plant, the output from which is earmarked for a specifi c buyer – as secured by an exchange agreement. Little wonder that petroleum fi rms will contract on better terms with other petroleum fi rms that have ‘paid the ante’ to ‘play the game’ than they will with buyers whose purchases are unsecured.

Consider therefore the use of growth and supplementary supply restraints, an example of which is the Imperial–Shell exchange agreement, under which Imperial supplied product to Shell in the Maritimes and received product from Shell in Montreal (p. 51):

The agreement between Imperial and Shell, originally signed in 1963, was rene-gotiated in 1967. In July 1972, Imperial did this because Shell had been growing too rapidly in the Maritimes. In 1971–72, Imperial had expressed its dissatisfac-tion with the agreement because of Shell’s marketing policies. Shell noted:

28 Key issues

‘[Imperial’s] present attitude is that we have built a market with their facilities, we are aggressive and threatening them all the time, and they are not going to help and in fact get as tough as possible with us.’ (Document #23633, undated, Shell)

Specifi cally, Imperial renewed the agreement with Shell only after impos-ing a price penalty if expansion were to exceed ‘normal growth rates’ and furthermore stipulated that ‘Shell would not generally be allowed to obtain product from third party sources’ to service the Maritimes (p. 52; emphasis added).

The Canadian Study notes that Gulf Oil also took the position that rivals receiving product under exchange agreements should be restrained to normal growth: ‘Processing agreements (and exchange agreements) should be entered into only after considering the overall economics of the Corporation and should be geared to providing competitors with volumes required for the normal growth only.’ 26 It furthermore sought and secured assurances that product supplied by Gulf would be used only by the recipi-ent and would not be diverted to other regions or made available to other parties (p. 59).

Limits on ‘normal growth’ and prohibitions on ‘third parties’ could well have anticompetitive purpose and were so regarded by the Canadian Study. Examined, however, through the lens of contract/governance, it is also possible that these same restrictions had the purpose and eff ect of preserving symmetrical incentives between the parties to exchange agree-ments, thereby allowing them to reach Node C credible commitments. Without use restrictions, bilateral dependence could become unbalanced. Also, symmetry could be placed under strain if one party was to grow ‘in excess of normal’ – in which event it might be prepared to construct its own plant and scuttle the exchange agreement. Marketing restraints that help to forestall such outcomes encourage parties to participate in exchanges that might otherwise be unacceptable.

To be sure, credibility benefi ts that are valued by the parties may not be equally valued by society. Such restraints may in some cases have both market power and secure transaction purposes. My purpose is merely to emphasize that, whereas the Canadian Study viewed these entirely in a one-sided (monopoly) way, the perspective of credible contracting adds another. To repeat, transaction cost economics can sometimes ‘illuminate the meaning of facts [and words] – particularly in the context of complex contractual relations – that otherwise cannot be explained, or worse, are explained incorrectly’ (Muris, 2003, p. 14).27


8. THE MODERN CORPORATION

The lens of contract/governance applies to the modern corporation in numerous ways: limits to fi rm size, scaling up, divisionalization, horizontal merger, conglomerate mergers, corporate governance, Japanese outsourc-ing practices, disequilibrium forms of organization, and the list goes on. My discussion here is restricted to limits to fi rm size, scaling up (to include corporate governance), and horizontal and conglomerate mergers.28

8.1 Limits to Firm Size

The puzzle of fi rm size was posed by Frank Knight in 1921 when he observed that the ‘diminishing returns to management is a subject often referred to in economic literature, but in regard to which there is a dearth of scientifi c discussion’ (Knight, 1965, p. 286, n. 1). He elaborated in 1933 as follows (1965, p. xxxi; emphasis added):

The relation between effi ciency and size of fi rm is one of the most serious prob-lems of theory, being, in contrast with the relation for a plant, largely a matter of personality and historical accident rather than of intelligible general principles. But the question is peculiarly vital, because the possibility of monopoly gain off ers a powerful incentive to continuous and unlimited expansion of the fi rm, which force must be off set by some equally powerful one making for decreased effi ciency.

Tracy Lewis’s later remarks that large established fi rms will always realize greater value from inputs than small potential entrants are apposite (1983, p. 1092; emphasis added):

The reason is that the leader can at least use the input exactly as the entrant would have used it, and earn the same profi ts as the entrant. But typically, the leader can improve on this by coordinating production from his new and existing inputs. Hence the new input will be valued more by the dominant fi rm.

If the dominant fi rm can use the input in exactly the same way as the entrant, then the larger fi rm can do everything the smaller fi rm could. If it can improve on the input usage, it can do more. Applied to vertical integra-tion, the parallel argument is that the acquisition of an independent com-ponent supplier is always preferred to outsourcing because the combined fi rm will never do worse (by reason of replication) and will sometimes do more if the acquiring stage always but only intervenes when expected net gains can be projected (by reason of selective intervention).

The puzzle of fi rm size thus reduces to this: what are the obstacles to the implementation of replication and selective intervention? As I discuss

30 Key issues

elsewhere (Williamson, 1985, Chapter 6), promises to replicate and selec-tively intervene are not costlessly enforceable. An acquired supplier can neither trust the acquirer to do the accounting (on which the supplier’s net receipts are calculated) in an unbiased way nor trust the acquirer to intervene always but only for good cause; and the acquirer cannot trust the supply stage to operate the plant and equipment (now owned by the acquirer) with unchanged due care and to adapt appropriately to autono-mous disturbances. The upshot is that neither replication nor selective intervention can be implemented without cost, as a result of which the gov-ernance mechanisms of markets and hierarchies diff er in kind (Williamson, 1991a). The recurrent point to which I call special attention, however, is this: the bureaucratic burdens of integration are discerned only upon opening the black box and examining the microanalytics.

8.2 Scaling Up

Solow observes that ‘The very complexity of real life . . . [is what] makes simple models so necessary’ (2001, p. 111). The object of a simple model is to capture the essence, thereby to explain hitherto puzzling practices and make predictions that are subjected to empirical testing. But simple models can also be ‘tested’ with respect to scaling up. Does repeated application of the basic mechanism out of which the simple model works yield a result that recognizably describes the phenomenon in question?

The test of scaling up is often ignored (possibly out of awareness that scaling up cannot be done) or is sometimes scanted (possibly in the belief that scaling up can be accomplished easily). The infl uential paper by Michael Jensen and William Meckling, ‘Theory of the Firm: Managerial Behavior, Agency Costs, and Capital Structure’ (1976), is an exception. The authors work out of a simplifi ed setup where an entrepreneur (100 percent owner-manager) sells off a fraction of the equity of the fi rm, as a result of which his incentive intensity is reduced and effi cacious monitoring arises as a response. What the authors are really interested in, however, is not entrepreneurial fi rms but the ‘modern corporation whose manag-ers own little or no equity’ (1976, p. 356). Although the latter project was beyond the scope of their paper, they expressed belief that ‘our approach can be applied to this case . . . [These issues] remain to be worked out in detail and will be included in a future paper’ (1976, p. 356).29

Alas, Jensen and Meckling never produced the follow-up paper, but many others have since examined the effi cacy of the board of directors as monitor in the large corporation where the ownership is diff use. The jury is still out, but I ascertain that serious obstacles stand in the way of acquiring the relevant information to support vigilant monitoring and, furthermore,


contend that the advisability of assigning the role of vigilant monitor to the board of directors is extremely problematic (Williamson, 2007b). In that event, corporate governance does not scale up from the entrepreneurial fi rm to the diff usely owned modern corporation.

Scaling-up issues relevant to the modern corporation are also posed by the theory of the fi rm as team production (Alchian and Demsetz, 1972) and the theory of the fi rm as governance structure. The theory of team pro-duction works through technological nonseparability, which Alchian and Demsetz illustrate with the example of manual freight loading: ‘Two men jointly lift heavy cargo into trucks. Solely by observing the total weight loaded per day, it is impossible to determine each person’s marginal pro-ductivity’ (1972, p. 779). Accordingly, rather than each person being paid his (unmeasurable) marginal product, such activities are organized cooper-atively, with a team whose members are paid as a team and are monitored by a boss lest they engage in shirking. This is instructive, but does techno-logical nonseparability scale up to explain the modern corporation?

One possibility is that the large corporation is a vast, indecomposable whole, in which event everything is connected with everything else and the model of technological nonseparability goes through. Another possibility is that, as Simon describes in ‘The Architecture of Complexity’ (1962), large hierarchical systems evolve from nearly decomposable subsystems – within which subsystems interactions are extensive and between which they are attenuated.30

Simon’s examination of social, biological, physical, and symbolic systems as well as the logic of complexity supports the proposition that decomposability ‘is one of the central structural schemes that the architect of complexity uses’ (1962, p. 468). Inasmuch as such decomposability relieves the condition of technological nonseparability on which Alchian and Demsetz rely, scaling up from small groups to which nonseparability applies (such as manual freight loading and, possibly, groups as large as the symphony orchestra) does not extend to the decision to join a series of tech-nologically separable stages, thereby to form the modern corporation.

So how does that transaction cost economics setup fare in scaling-up respects? Does successive application of the make-or-buy decision, as it is applied to individual transactions, scale up to describe something that approximates a multi-stage fi rm? Note in this connection that transaction cost economics assumes that the transactions of interest are those that take place between technologically separable stages. This is the ‘boundary of the fi rm’ issue as described elsewhere (Williamson, 1985, pp. 96–8). Upon taking the technological ‘core’ as given (possibly as derived from site specifi c investments, of which thermal economies are an example (see Section 5.1 above)), attention is focused on a series of separable make-or-buy decisions

32 Key issues

− backward, forward, and lateral – to ascertain which should be outsourced and which should be incorporated within the ownership boundary of the fi rm. So described, the fi rm is the inclusive set of transactions for which the decision is to make rather than buy – which does appear to implement scaling up, or at least is a promising start (Williamson, 1985, pp. 96–8).31

8.3 Horizontal Mergers

My initial inclination was to regard oligopoly to be outside the scope of transaction cost reasoning, mainly because I had become accustomed to thinking about oligopoly in terms of the prevailing structure–conduct–performance paradigm, where concentration ratios and barriers to entry were the coin of the realm. Upon viewing oligopoly as a cartel problem, however, its contractual nature is immediately evident.

Consider in this connection the claim that monopoly and oligopoly are nearly indistinguishable in competitive respects.32 Such a claim fails to make allowance for (1) the advantages of hierarchy (within a monopoly) as com-pared with interfi rm contracting (among oligopolists) for dispute settlement and coordinating purposes and (2) the diff erential incentives and the related propensity to cheat that distinguish internal from interfi rm organization. Examining the cartel as a fi ve-stage contracting process – contract specifi ca-tion, joint gain agreement, implementation under uncertainty, monitoring contract execution, and penalizing contract violations – is instructive.

As discussed elsewhere (Williamson, 1975, pp. 238–44), oligopolies diff er in their ‘complexity’ in all fi ve of these contractual respects. Simple oligopolies – where numbers are few and products are homogeneous, shares are easy to agree upon, disturbances are small, price and output are public knowledge, and penalties for violations are assuredly meted out – will surely recognize their interdependence and behave accordingly. As, however, deviations from these simple conditions arise, cartel contracts become progressively more complex and undergo slippage and fracture during contract execution. Interestingly, even in the 1870s and 1880s, when express collusion was not unlawful, repeated eff orts by the railroads to curb competitive pricing – fi rst by informal alliances, then by managed federations – were undone by cheating.33 When the railroads ‘found to their sorrow that they could not rely on the intelligence and good faith of railroad executives’ to manage the cartels (Chandler, 1977, p. 141), they gave up on interfi rm agreements and turned to merger.

Contractual reasoning is thus instructive in making oligopoly– monopoly comparisons. Because, however, the predictions of the contractual approach to oligopoly are very similar to many other oligopoly theories, empirical research on oligopoly has been little aff ected.


8.4 Conglomerates

The conglomerate form of organization was a matter of grave concern in the 1960s, especially among those with populist predilections, of which H.M. Blake (1973) was one. According to Blake, the anticompetitive hazards of conglomerate mergers, in potential competition and other respects, were ‘so widespread that [these] might appropriately be described as having an eff ect upon the economic system as a whole – in every line of commerce in every section of the country’ (1973, p. 567). So regarded, the conglomerate was a menace.

Can a contractual approach to conglomerate diversifi cation help to inform the issues? The basic proposition is this: whereas vertical inte-gration is viewed as taking transactions out of the intermediate product market and organizing them internally, the conglomerate can be inter-preted as taking transactions out of the capital market and organizing them internally. So described, the conglomerate experiences a breadth for depth tradeoff in managing capital market transactions.

The argument relies on part in the distinction between centralized (unitary or U-form) and decentralized (multidivisional or M-form) corporations, as developed by Alfred Chandler (1962) and interpreted in effi ciency terms in Williamson (1970). Specifi cally, the conglomerate can be understood as a logical outgrowth of the divisionalized strategy for organizing complex economic aff airs. Thus, once the merits of the M-form structure for manag-ing separable, albeit related, lines of business (such as diff erent automobile brands or diff erent chemical divisions) were recognized and digested, its extension to manage less closely related activities was natural, although that is not to say that the management of diversifi cation is without prob-lems of its own. The basic M-form logic, whereby strategic resource alloca-tion and oversight are assigned to the general offi ce and operating decisions are the responsibility of the operating divisions, nevertheless carries over.

To the degree to which conglomerates employ the M-form logic (as against the go-go conglomerates of the 1960s) and possess deep knowl-edge (as compared with the capital market) within a large but delimited set of diversifi ed investment opportunities, then the indicia for an effi -cient resource allocation interpretation of the conglomerate take shape.34 Plainly, however, not all conglomerates qualify to be so described.

9. CONCLUSIONS

Opening the black box of fi rm and market organization is accomplished by taking transaction cost economizing to be the main case, in relation to

34 Key issues

which adaptations (of autonomous and coordinated kinds) are especially important, and examining economic organization through the lens of con-tract. The transaction is made the basic unit of analysis and governance is the means by which to infuse order. In conformity with Simon’s advice that our research agenda and research methods are shaped by our descrip-tion of human actors, the cognitive and self-interestedness attributes of human actors that bear on contracting are expressly identifi ed, after which the ramifi cations of human actors as these relate to the key attributes of both transactions and governance structures are worked out. Aligning transactions, which diff er in their attributes, with governance structures, which diff er in their cost and competence, so as to eff ect a transaction cost economizing outcome is where the predictive content resides.

The key features of the foregoing are these:

(1) The lens of contract/governance is an instructive way by which to open the black box of fi rm and market organization and examine the mechanisms inside.

(2) This project subscribes to Jon Elster’s dictum that ‘explanations in the social sciences should be organized around (partial) mechanisms rather than (general) theories’ (1994, p. 75; emphasis omitted).

(3) Especially relevant to public policy analysis is that nonstandard con-tractual practices and organizational structures that were believed to be anticompetitive when examined through the lens of price theory are often revealed to serve effi ciency purposes as well or instead.

(4) Subsequent uses of the lens of contract/governance to examine complex contract and economic organization reveal that many ‘superfi cially disconnected and diverse phenomena . . . [are] mani-festations of a more fundamental and relatively simple structure’ (Friedman, 1953, p. 33).35

Such applications notwithstanding, many conceptual, empirical, and public policy challenges await.

NOTES

* This is the fi rst of two papers dealing with ‘opening up the black box’. This paper deals with applications to antitrust. The other paper describes applications to regula-tion (Williamson, 2007a). The introduction and Section 2 of this chapter overlap with Section 1 of Williamson (2007a).

1. The importance of a focused lens is crucial. The distinction here is between promising but sprawling concepts that invite ex post rationalizations for any outcome whatsoever (which is a chronic problem with vaguely defi ned concepts – of which ‘power’ is one (March, 1966)). A focused lens both delimits the set of factors that can be invoked to explain


complex phenomena and reveals the mechanisms through which these factors work. The promising but vague concept of transaction costs which Coase introduced in his 1937 article, ‘The Nature of the Firm’, remained in a state of disuse 35 years later (Coase, 1972, p. 63) precisely because the key ideas had not been operationalized (Coase, 1992, p. 718).

2. As Harold Demsetz observes, it is ‘a mistake to confuse the fi rm of [neoclassical] eco-nomic theory with its real-world namesake. The chief mission of neoclassical economics is to understand how the price system coordinates the use of resources, not the inner workings of real fi rms’ (1983, p. 377).

3. The quotation is attributed to Donald Turner by Stanley Robinson (1968), p. 29. 4. The Federal Trade Commission’s opinion in Foremost Dairies states that the necessary

proof of violation of Section 7 ‘consists of types of evidence showing that the acquiring fi rm possesses signifi cant market power in some markets or that its overall organization gives it a decisive advantage in effi ciency over its smaller rivals’ (In re Foremost Dairies, Inc., 60 FTC, 944, 1084 (1962), emphasis added).

5. See Williamson (1985, pp. 366–7) for an elaboration upon the convoluted status of antitrust enforcement during the 1960s.

6. This terse summary is elaborated elsewhere (Williamson, 1985, 1991a, 2002, 2005). The intellectual antecedents are set out in the Appendix.

7. R.C.O. Matthews describes the New Institutional Economics (with emphasis on trans-action cost economics) in precisely these terms in his Presidential Address to the Royal Economic Society (1986, p. 903).

8. Because I judge several of the listed six assumptions to be implausible (Williamson, 1991b, pp.172–6), I take the lesson of Fudenberg et al. (1990) (which is an intellectual tour de force) to be that a sequentially optimal contract is infeasible. Especially prob-lematic are their assumptions of three-way costless knowledge of public outcomes (by principal, agent, and arbiter) and common knowledge of both technology and preferences over action-payment streams. If and as the attainment of logical consistency (in theory) yields infeasibility (in practice), applied economists will understandably be chary of the operational signifi cance of the theory.

9. For a discussion of contract law regimes as these relate to governance, see Williamson (1991a).

10. Note that the price that a supplier will bid to supply under Node C conditions will be less than the price that will be bid at Node B. That is because the added security features at Node C serve to reduce the contractual hazard, as compared with Node B, so the con-tractual hazard premium will be lowered. One implication is that suppliers do not need to petition buyers to provide safeguards. Because buyers will receive goods and services on better terms (lower price) when added security is provided, buyers have the incentive to off er credible commitments.

11. United States v. Von’s Grocery Co., 384 U.S. 270, 301 (1966) (Stewart, J., dissenting).12. As, for example, in ‘Economies as an antitrust defense: the welfare tradeoff s’ (Williamson,

1968).13. This is an insistent theme in Coase (1960, 1964, 1972).14. Stephen Stockum’s summary of Muris’s position is as follows (2002, p. 60):

Muris describes his economic approach as neither Chicago School nor Post-Chicago, but rather ‘New Institutional Economics’, which combines theory with a study of real world institutions, . . . [is] heavily empirical, . . . [and provides relief from economic ideology in favor of] more practical discussions of how economic analysts can contribute to rational enforcement of the antitrust laws.

15. This relieves problems of valuing such dies if they are owned by the supplier, although user-cost abuses of dies become a concern if the buyer owns them.

16. For discussions, see Williamson (1979, 1987).17. Thus whereas industrial organization specialists and the Antitrust Merger Guidelines

once advised that an antitrust issue is posed should a fi rm with a 20 percent market share acquire a 5 or 10 percent share in any industry from which it buys or to which it

36 Key issues

sells (Stigler, 1955, p. 183), transaction cost economics counsels that the attributes of the transaction tell us a lot more about the purposes of integration, especially for such small market shares.

To be sure, vertical integration sometimes serves strategic purposes. As Alfred Marshall observed, if, in a small country, spinning and weaving were joined, ‘the monopoly so established will be much harder to shake than would either half of it separately’ (1920, p. 495). Bain warned that vertical integration can be used as a means by which to ‘disadvantage, weaken, eliminate, or exclude non-integrated competitors’ (1968, pp. 360–62). And Stigler advised that integration ‘becomes a possible weapon for the exclusion of new rivals by increasing the capital requirements for entry into the combined integrated production processes’ (1955, p. 224). I do not disagree. Because, however, strategic entry deterrence is so easy to invoke, those who would make such claims should describe the details of the underlying mechanisms and explain when we should expect alleged adverse eff ects to rise to the level of public policy signifi cance.

18. For a discussion of both the Jurisdictional Statement and the Brief for the United States in United States v. Arnold, Schwinn & Co., 388 U.S. 365 (1967), see Williamson (1979; 1985, pp. 183–9).

19. Applied welfare economics apparatus is used to display these two eff ects. The tradeoff s had gone unnoticed, however, until positive transaction costs were expressly introduced into the calculus.

20. FTC v. Morton Salt Co., 334 U.S. 37 (1948); emphasis added.21. The remainder of this subsection is based on Williamson (1996, pp. 77–8).22. The remainder of this subsection is based on Williamson (1985, pp. 197–201).23. FTC v. Exxon et al. Docket No. 8934 (1973).24. Robert J. Bertrand, Q.C., Director of Investigation and Research, Combines

Investigation Act, coordinated the eight-volume study, The State of Competition in the Canadian Petroleum Industry (Quebec, 1981). All references in this chapter are to Vol. V, The Refi ning Sector. That study will hereinafter be referred to as the Canadian Study.

25. Page numbers here and below that do not name the source all refer to Vol. V of the Canadian Study (see note 24, above).

26. The Canadian Study (p. 59) identifi es the source as Document #73814, January 1972.27. Muris (2003, pp. 15–23) discusses a variety of other applications of transaction cost

economics to complex contracting. Also see Joskow (2002).28. For discussions of divisionalization, see Williamson (1970, 1985, Chapter 11); for

Japanese economic organization, Williamson (1985, pp. 120–123); for corporate govern-ance, Williamson (1988, 2007b); for disequilibrium contracting, Williamson (1991a).

29. Other examples where scaling-up tensions are posed include Thomas Schelling’s treat-ment of the evolution of segregation in the ‘self-forming neighborhood’ (Schelling, 1978, pp. 147–55), the expansive uses sometimes made of the so-called paradox of voting (Williamson and Sargent, 1967), and the move from project fi nancing to composite fi nancing in the modern corporation (Williamson, 1988).

30. ‘The loose . . . coupling of subsystems . . . [means that] each subsystem [is] independent of the exact timing of the operation of the others. If subsystem B depends upon subsystem A only for a certain substance, then B can be made independent of fl uctuations on A’s production by maintaining a buff er inventory’ (Simon, 1977, p. 255).

31. In consideration of the diffi culties and importance of scaling up, it is judicious to hold theories of the modern corporation for which scaling up has not been demonstrated in public policy abeyance.

32. John Kenneth Galbraith took the position that ‘the fi rm, in tacit collaboration with other fi rms in the industry, has wholly suffi cient power to set and maintain prices’ (1967, p. 200).

33. Note that while the courts tolerated collusion, they refused to enforce price setting agreements.

34. There is an additional contractual wrinkle, in that conglomerates once posed an acquisi-tion threat to underperforming fi rms, including fi rms that had allowed their debt–equity


ratio to fall below the optimal level (Williamson, 1988). Leveraged buyout specialists such as Kohlberg-Kravis-Roberts are precisely attuned to such opportunities and have since taken over many of these takeover functions.

35. In the spirit of pluralism, we will benefi t from any theory that deepens our under-standing of complex phenomena and satisfi es the precepts of pragmatic methodology (Williamson, 2007c).

REFERENCES

Alchian, A. and H. Demsetz (1972), ‘Production, Information Costs, and Economic Organization’, American Economic Review, 62 (December), 777–95.

Areeda, P. and D.F. Turner (1975), ‘Predatory Pricing and Related Practices Under Section 2 of the Sherman Act’, Harvard Law Review, 88 (February), 697–733.

Arrow, K. (1987), ‘Refl ections on the Essays’, in George Feiwel (ed.), Arrow and the Foundations of the Theory of Economic Policy, New York: NYU Press, pp. 727–34.

Arrow, K. (1999). ‘Forward’, in Glenn Carroll and David Teece (eds), Firms, Markets, and Hierarchies, New York: Oxford University Press, pp. vii–viii.

Bain, J. (1968), Industrial Organization, 2nd edn, New York: John Wiley and Sons.Barnard, C. (1938), The Functions of the Executive, Cambridge, MA: Harvard

University Press (15th printing, 1962).Blake, H.M. (1973), ‘Conglomerate Mergers and the Antitrust Laws’, Columbia

Law Review, 73 (March), 555–92.Bork, R.H. (1978), The Antitrust Paradox: A Policy at War With Itself, New York:

Basic Books.Chandler, A.D. (1962), Strategy and Structure, Cambridge, MA: MIT Press.Chandler, A.D. (1977), The Visible Hand: The Managerial Revolution in American

Business, Cambridge, MA: Harvard University Press.Coase, R. (1937), ‘The Nature of the Firm’, Economica, N.S., 4, 386–405.Coase, R. (1960), ‘The Problem of Social Cost’, Journal of Law and Economics, 3

(October), 1–44.Coase, R. (1964), ‘The Regulated Industries: Discussion’, American Economic

Review, 54 (May), 194–7.Coase, R. (1972). ‘Industrial Organization: A Proposal for Research’, in V.R.

Fuchs (ed.), Policy Issues and Research Opportunities in Industrial Organization, New York: National Bureau of Economic Research, pp. 59–73.

Coase, R. (1992), ‘The Institutional Structure of Production’, American Economic Review, 82 (September), 713–19.

Coughlan, A., E. Anderson, L. Stern, and A.El-Ansary (2005), Marketing Channels, 7th edn, Upper Saddle River, NJ: Pearson/Prentice Hall.

Demsetz, H. (1974), ‘Two Systems of Belief About Monopoly’, in V. Fuchs (ed.), Policy Issues and Research Opportunities in Industrial Organization, New York: National Bureau of Economic Research.

Demsetz, H. (1983), ‘The Structure of Ownership and the Theory of the Firm’, Journal of Law and Economics, 26 (June): 375–90.

Dixit, A. (1996), The Making of Economic Policy: A Transaction Cost Politics Perspective, Cambridge, MA: MIT Press.

Elster, J. (1994), ‘Arguing and Bargaining in Two Constituent Assemblies’,

38 Key issues

unpublished manuscript, remarks given at the University of California, Berkeley.

Friedman, M. (1953), Essays in Positive Economics, Chicago: University of Chicago Press.

Fuchs, V. (ed.) (1972), Policy Issues and Research Opportunities in Industrial Organization, New York: Columbia University Press.

Fudenberg, D., B. Holmstrom and P. Milgrom (1990), ‘Short-Term Contracts and Long-Term Agency Relationships’, Journal of Economic Theory, 51 (June), 1–31.

Galbraith, J.K. (1967), The New Industrial State, Boston: Houghton-Miffl in Company.

Geyskens, I., J.B.E.M. Steenkamp and N. Kumar (2006). ‘Make, Buy, or Ally: A Meta-analysis of Transaction Cost Theory’, Academy of Management Journal, 49 (3), 519–43.

Hayek, F. (1945), ‘The Use of Knowledge in Society’, American Economic Review, 35 (September), 519–30.

Jensen, M. and W. Meckling. (1976), ‘Theory of the Firm: Managerial Behavior, Agency Costs, and Capital Structure’, Journal of Financial Economics, 3 (October), 305–60.

Joskow, P. (1987), ‘Contract Duration and Relationship-Specifi c Investments’, American Economic Review, 77 (1), 168–85.

Joskow, P. (1991), ‘The Role of Transaction Cost Economics in Antitrust and Public Utility Regulatory Policies’, Journal of Law, Economics, and Organization, 7 (March), 53–83.

Joskow, P. (2002), ‘Transaction Cost Economics, Antitrust Rules and Remedies’, Journal of Law, Economics and Organization, 18 (1), 95–116.

Kenney, R. and B. Klein (1983), ‘The Economics of Block Booking’, Journal of Law and Economics, 26 (October), 497–540.

Klein, B. (1996), ‘Why Hold-Ups Occur: The Self Enforcing Range of Contractual Relationships’, Economic Inquiry, 34 (3), 444–63.

Klein, B. and K. Leffl er (1981), ‘The Role of Market Forces in Assuring Contractual Performance’, Journal of Political Economy, 89 (August), 615–41.

Knight, F. (1965), Risk, Uncertainty, and Profi t, New York: Harper & Row, Publishers, Inc.

Kreps, D. (1990), A Course in Microeconomic Theory, Princeton, NJ: Princeton University Press.

Lewis, T. (1983), ‘Preemption, Divestiture, and Forward Contracting in a Market Dominated by a Single Firm’, American Economic Review, 73 (December), 1092–1101.

Livesay, H.C. (1979), American Made: Men Who Shaped the American Economy, Boston: Little, Brown.

Llewellyn, K.N. (1931), ‘What Price Contract? An Essay in Perspective’, Yale Law Journal, 40, 704–51.

Macher, J.T. and B. Richman (2006), ‘Transaction Cost Economics: A Review and Assessment of the Empirical Literature’, unpublished manuscript.

March, J.G. (1966), ‘The Power of Power’, in David Easton (ed.), Varieties of Political Theory, Englewood Cliff s, NJ: Prentice-Hall, pp. 39–70.

Marshall, A. (1920), Principles of Economics, 8th edn, New York: Macmillan and Co., Ltd.

Masten, S. (1984). ‘The Organization of Production: Evidence from the Aerospace Industry’, Journal of Law and Economics, 27 (October), 403–18.


Masten, S. (1996), Case Studies in Contracting and Organization, New York: Oxford University Press.

Matthews, R.C.O. (1986), ‘The Economics of Institutions and the Sources of Economic Growth’, Economic Journal, 96 (December), 903–18.

Monteverde, K. and D. Teece (1982), ‘Supplier Switching Costs and Vertical Integration in the Automobile Industry’, Bell Journal of Economics, 13 (Spring), 206–13.

Muris, T. (2003), ‘Improving the Economic Foundations of Competition Policy’, George Mason Law Review, 12 (1), 1–30.

Posner, R. (1976), Antitrust Law, Chicago: University of Chicago Press.Richman, B. (2006), ‘How Communities Create Economic Advantage: Jewish

Diamond Merchants in New York’, Law and Social Inquiry, 31 (2), 383–420.Robinson, S. (1968), ‘Comment’, New York State Bar Association, Antitrust

Symposium.Schelling, T. (1978), Micromotives and Macrobehavior, New York: Norton.Simon, H. (1957), Models of Man, New York: John Wiley & Sons.Simon, H. (1962), ‘The Architecture of Complexity’, Proceedings of the American

Philosophical Society, 106 (December), 467–82.Simon, H. (1977), Models of Discovery, Boston: D. Reidel Publishing Co.Simon, H. (1984), ‘On the Behavioral and Rational Foundations of Economic

Dynamics’, Journal of Economic Behavior and Organization, 5 (March), 35–56.Simon, H. (1985), ‘Human Nature in Politics: The Dialogue of Psychology with

Political Science’, American Political Science Review, 79 (2), 293–304.Solow, R. (2001), ‘A Native Informant Speaks’, Journal of Economic Methodology,

8 (March), 111–12.Stigler, G. (1955), ‘Mergers and Preventive Antitrust Policy’, University of

Pennsylvania Law Review, 104, 176–85.Stigler, G. (1968), The Organization of Industry, Homewood, IL: Richard D.

Irwin.Stockum, S. (2002), ‘An Economist’s Margin Notes: The Antitrust Writings of

Timothy Muris’, Antitrust (Spring), 60.Stuckey, J. (1983), Vertical Integration and Joint Ventures in the Aluminum Industry,

Cambridge, MA: Harvard University Press.Whinston, M. (2001), ‘Assessing Property Rights and Transaction-Cost Theories

of the Firm’, American Economic Review, 91 (2), 184–99.Williamson, O.E. (1968), ‘Economies as an Antitrust Defense: The Welfare

Tradeoff s’, American Economic Review, 58 (March), 18–35.Williamson, O.E. (1970), Corporate Control and Business Behavior, Englewood

Cliff s, NJ: Prentice-Hall.Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust

Implications, New York: Free Press.Williamson, O.E. (1977), ‘Predatory Pricing: A Strategic and Welfare Analysis’,

Yale Law Journal, 87 (December), 284–340.Williamson, O.E. (1979), ‘Assessing Vertical Market Restrictions’, University of

Pennsylvania Law Review, 127 (April), 953–93.Williamson, O.E. (1985), The Economic Institutions of Capitalism, New York: Free

Press.Williamson, O.E. (1987), ‘Vertical Integration’, in J. Eatwell et al. (eds), The

New Palgrave Dictionary of Economics, Vol. IV, London: Macmillan, pp. 807–12.

40 Key issues

Williamson, O.E. (1988), ‘Corporate Finance and Corporate Governance’, Journal of Finance, 43 (July), 567–91.

Williamson, O.E. (1991a), ‘Comparative Economic Organization: The Analysis of Discrete Structural Alternatives’, Administrative Science Quarterly, 36 (June), 269–96.

Williamson, O.E. (1991b), ‘Economic Institutions: Spontaneous and Intentional Governance’, Journal of Law, Economics, and Organization, 7 (Special Issue): 159–87.

Williamson, O.E. (1996), The Mechanisms of Governance, New York: Oxford University Press.

Williamson, O.E. (2002), ‘The Theory of the Firm as Governance Structure: From Choice to Contract’, Journal of Economic Perspectives, 16 (Summer), 171–95.

Williamson, O.E. (2005), ‘The Economics of Governance’, American Economic Review, 95 (2), 1–18.

Williamson, O.E. (2007a), ‘Opening the Black Box of Firm and Market Organiza-tion: Regulation’, working paper, University of California, Berkeley.

Williamson, O.E. (2007b), ‘Corporate Boards of Directors: In Principle and In Practice’, Journal of Law, Economics, and Organization, 24 (October), 247–72.

Williamson, O.E. (2007c), ‘Pragmatic Methodology’, working paper, University of California, Berkeley.

Williamson, O.E. and T. Sargent (1967), ‘Social Choice: A Probabilistic Approach’, The Economic Journal, 77 (308), 797–813.

Wilson, E.O. (1998), Consilience, New York: Alfred Knopf.


APPENDIX 2.1 INTELLECTUAL ANTECEDENTS TO THE LENS OF CONTRACT/GOVERNANCE

The comparative contractual approach to economic organization is inspired by a series of key ideas, many of which fi rst surfaced in the 1930s (or thereabouts). Of special importance are these:

(1) The organization of economic activity as among fi rms, markets, and other modes of governance should be derived rather than taken as given (Coase, 1937).

(2) Such a derivation should make explicit allowance for positive trans-action costs (Coase, 1937, 1960).

(3) Unstated assumptions about the nature of the human beings whose behavior we are studying should be revealed (Simon, 1957, 1985).

(4) The unit of analysis should be named, of which the transaction is a candidate (Commons, 1932), and dimensionalized.

(5) Moving beyond the economics of simple market exchange, ongoing contractual relations should also be brought under scrutiny – with emphasis on the triple of confl ict, mutuality, and order (Commons, 1932).

(6) Provision also needs to be made for the contract law diff erences between modes. Simple market exchange (to which the concept of contract as legal rules applies) gives way to long-term contracting (to which the more elastic concept of contract as framework (Llewellyn, 1931) applies) and to hierarchy (whereby internal organization becomes its own court of ultimate appeal).

(7) The central problem of economic organization to which transaction cost consequences accrue is that of adaptation, of which two kinds are distinguished: autonomous adaptations (Hayek, 1945) and coor-dinated adaptations (Barnard, 1938).

(8) Going beyond the neoclassical lens of choice, complex economic organization is also usefully examined through the lens of contract/governance, where the latter implements the proposition that ‘mutu-ality of advantage from voluntary exchange . . . is the most fundamen-tal of all understandings in economics’ (Buchanan, 2001, p. 29).

References

Barnard, C. (1938), The Functions of the Executive, Cambridge, MA: Harvard University Press.

Buchanan, J. (2001), ‘Game Theory, Mathematics and Economics’, Journal of Economic Methodology, 8 (March), 27–32.

42 Key issues

Coase, R. (1937), ‘The Nature of the Firm’, Economica, N.S., 4, 386–405.Coase, R. (1960), ‘The Problem of Social Cost’, Journal of Law and Economics, 3,

(October), 1–44.Commons, J. (1932), ‘The Problem of Correlating Law, Economics, and Ethics’,

Wisconsin Law Review, 8, 3–26.Hayek, F. (1945), ‘The Use of Knowledge in Society’, American Economic Review,

35 (September), 519–30.Llewellyn, K.N. (1931), ‘What Price Contract? An Essay in Perspective’, Yale Law

Journal, 40, 704–51.Simon, H. (1957), Models of Man, New York: John Wiley & Sons.Simon, H. (1985), ‘Human Nature in Politics: The Dialogue of Psychology with

Political Science’, American Political Science Review, 79 (2), 293–304.

43

3. The corporation: an economic enigmaDennis C. Mueller

The Anglo-Saxon version of corporate organization – widely dispersed ownership, and professional managers with small ownership stakes – has been somewhat of an enigma throughout its 200 or so year history. Some economists have thought it to be an ineffi cient organizational structure; others have proclaimed its superiority over all other ways to organize busi-ness activity. The performance of US corporations during the 1970s and 1980s seemed to confi rm the judgments of the corporate form’s critics. One market after another was lost to companies from Japan or Europe. Articles and books appeared proclaiming the ‘German model’ or the ‘Japanese model’ superior to the Anglo-Saxon model.1 One student of American capi-talism even predicted ‘the eclipse of the public corporation’ (Jensen, 1989).

The impressive performance of the United States’ economy during the 1990s, alongside the stumbling performances of the Japanese and the German and some other European economies, has led some to now claim that it is the Anglo-Saxon model which is superior and toward which all others must converge.2

In this chapter I review some of the arguments and evidence regard-ing the effi ciency of the corporate form. I shall argue that one reason for the diff erent views of economists about corporations is that they tend to see what they want to see. Although most of the evidence cited seems to support the position of the critics of the corporate form, I shall close the chapter with the suggestion that developments over the last decade or two have altered the environments in which corporations operate in such a way as to justify the effi ciency claims of their defenders. I begin, appropriately enough, with Adam Smith.

1. THE EARLY ECONOMISTS

Corporations, or joint stock companies as they were commonly called at the end of the 18th century, were relatively rare when Adam Smith wrote

44 Key issues

The Wealth of Nations. But Smith had seen enough of them to off er the following observations:

The directors of such companies . . . being the managers rather of other people’s money than of their own, it cannot well be expected, that they should watch over it with the same anxious vigilance with which the partners in private copartnery frequently watch over their own. Like the stewards of a rich man, they are apt to consider attention to small matters as not for their master’s honour, and very easily give themselves a dispensation from having it. Negligence and profusion, therefore, must always prevail, more or less, in the management of the aff airs of such a company . . . It is upon this account that joint stock companies for foreign trade have . . . very seldom succeeded without an exclusive privilege; and frequently have not succeeded with one. Without an exclusive privilege they have commonly mismanaged the trade. With an exclusive privilege they have both mismanaged and confi ned it. (1776, p. 700)

John Stuart Mill regarded this conclusion as ‘one of those overstatements of a true principle, often met with in Adam Smith’ (1885, p. 140). Nevertheless, he also thought the principle to be true. After discussing the advantages of joint stock companies, Mill took up ‘the other side of the question’.

[I]ndividual management has also very great advantage over joint stock. The chief of these is the much keener interest of the managers in the success of the undertaking.

The administration of a joint stock association is, in the main, administration by hired servants. Even . . . the board of directors, who are supposed to super-intend the management . . . have no pecuniary interest in the good working of the concern beyond the shares they individually hold, which are always a very small part of the capital of the association, and in general but a small part of the fortunes of the directors themselves; and the part they take in the management usually divides their time with many other occupations, of as great or greater importance to their own interest; the business being the principal concern of no one except those who are hired to carry it on. But experience shows, and prov-erbs, the expression of popular experience, attest, how inferior is the quality of hired servants, compared with the ministration of those personally interested in the work, and how indispensable, when hired service must be employed, is ‘the master’s eye’ to watch over it. (Mill, 1885, pp. 138–9)

Smith and Mill were, of course, giants in the development of econom-ics. That each had a brilliant mind goes without saying. But their genius stemmed not only from their capacity to reason. Each was also a keen observer of people and institutions. Each could draw generalizations from what he saw that others would recognize to be true upon hearing. Another good example of this is Smith’s statement that ‘People of the same trade seldom meet together, even for merriment and diversion, but the conversation ends in a conspiracy against the public, or in some

The corporation 45

contrivance to raise prices’ (1776, p. 128). As with his statements about corporations, what is particularly interesting about this famous passage is that it is not a proposition about what businessmen under certain assumptions will do. It is a statement about what they do do. Smith’s classic treatise is not a series of hypotheses about how individuals and markets will behave, but rather a treasure chest of observations of how they do behave.

Alfred Marshall was fi rst and foremost a keen observer of the world of business. He, too, had some doubts about the effi ciency of joint stock companies. In a section entitled ‘Temptations of joint stock companies to excessive enlargement of scope’ in Industry and Trade (1923), Marshall notes that ‘unfortunately many [outside directors] are unable to give the large time and energy needed for obtaining a thorough mastery of the aff airs of the companies for which they are responsible’ (p. 321). The slack thereby created can lead to ‘excessive enlargement of scope’, because company managers ‘cannot always approach a proposal for enlarging an existing department, or starting a new one, without some bias’ (p. 322). Nevertheless, he was much less concerned about the potential ineffi cien-cies of the corporate form than Smith and Mill, for he judged there to be a countervailing development that protected the ‘powerless’ position of the shareholders.

It is a strong proof of the marvellous growth in recent times of a spirit of honesty and uprightness in commercial matters, that the leading offi cers of great public companies yield as little as they do to the vast temptations to fraud which lie in their way . . . There is every reason to hope that the progress of trade morality will continue . . . (Marshall, 1920, p. 253)

The founders of neoclassical economics on the western side of the Atlantic were also sanguine in their views about the newly emerging cor-porations and trusts formed through merger. Anticipating the reasoning underlying transaction costs economics and the ‘new learning’ in industrial organization by 80 years, they deduced that Darwinian competition could be relied upon to select only the most effi cient combinations of assets.3

The next landmark in our intellectual history is The Modern Corporation and Private Property. As with many of the arguments contained in The Wealth of Nations, much that Berle and Means (1932) wrote about the corporation was known at the time they wrote. But they combined their thesis with an exhaustive history of the evolution of the corporate legal form, and amassed data demonstrating the extent of the separation of ownership from control, and the rise in aggregate concentration that had occurred in the fi rst third of the 20th century. If the dangers of dispersed

46 Key issues

shareownership forewarned by Smith and Mill were real, Berle and Means’s book suggested that the United States had much to fear.

The bulk of The Modern Corporation and Private Property must have been written in the late 1920s, and thus the book cannot be construed as an account of the collapse of 1929. But the timing of the book’s publication could not have been better. The arguments put forward by Berle and Means about the potential for managerial abuse of discretion created by the sepa-ration of ownership and control resonated with the thunder of falling stock prices and profi ts. The handful of examples of abuse of managerial power in the book were duplicated and dwarfed by the accounts appearing daily in the business press.4 The modern corporation appeared to have fulfi lled the worst fears of Smith and Mill, and dashed the hopes of Marshall. The Victorian noblesse oblige that Marshall saw protecting shareholders as late as 1920 had by 1930 vanished, at least in the United States.

2. THE ‘MARGINALIST’ CONTROVERSIES

With the publication of The Modern Corporation, the Great Crash and its aftermath of revelations of misuse of position by managers, the issue of corporate effi ciency could not be ignored, or so it would seem. But most economists did ignore it.5 For by the 1930s the neoclassical revolution, in which Alfred Marshall had played such an important part, had triumphed. When a new issue arose, the economist would no longer turn to his fi rst-hand knowledge of the relevant facts and institutions for addressing this issue, or lacking fi rst-hand knowledge proceed to gather it. The economist’s fi rst reaction would now be to turn to one of the models he had used to analyze similar problems in the past. The neoclassical models had proved themselves to be insightful analytic tools for laying bare the basic elements of certain economic problems. To achieve their pedagogic potential they needed to be kept simple, however, and so it was often the case that indi-viduals were assumed to choose a single instrument (for example, price) to achieve a single goal (profi t). Thus, although the managerial corporation was by the 1930s the dominant economic institution of Western capitalism, the fi rm (entrepreneur) remained the main business actor in the economics literature, and it (he) maximized profi t.

The 1930s were diffi cult for mainstream economics to digest. Much seemed to be happening that the newly developed neoclassical models could not explain. Keynes’s response is the most famous reaction, of course. But attacks on the micro front were also afoot. A number of economists were troubled by the failure of prices to fall to eliminate signifi cant amounts of excess supply. Gardiner Means (1935) attempted to account for this with

The corporation 47

the thesis that large corporations in oligopolistic markets enjoyed consider-able freedom to administer prices independently of market forces. Hall and Hitch were struck by interview evidence that revealed a gross inconsistency between ‘the way in which business men decide what price to charge for their products and what output to produce’ and the behavior assumed in neoclassical models (1939, p. 12). They concluded that the evidence ‘casts doubt on the general applicability of the conventional analysis of price and output policy in terms of marginal cost and marginal revenue, and suggests a mode of entrepreneurial behavior which current economic doctrine tends to ignore’ (p. 12). They off ered as an alternative to marginal analysis a ‘full cost’ or mark-up model of pricing. Richard Lester was left with ‘grave doubts as to the validity of conventional marginal theory and the assumptions on which it rests’ from answers given by 58 entrepreneurs from the South to a questionnaire circulated in the mid-1940s (1946, p. 81). Kaplan et al. (1958) conducted interviews of chief executives in the late 1940s and mid-1950s and uncovered a variety of objectives and rules of thumb for setting prices that did not resemble marginal cost equals marginal revenue. Thus, evidence gathered over two decades and two countries on how managers actually do set prices directly contradicted the assumptions upon which most economic modeling of pricing was at that time, and is today, based.

Not surprisingly these challenges to the mainstream view were vigor-ously repelled (Machlup, 1946, 1947; Kahn, 1959).6 What is interesting is that the defenders of the neoclassical model off ered neither contradic-tory interview and questionnaire evidence to support their positions nor empirical evidence that would allow one to reject one hypothesis and not the other. Rather the argument was made that it was not important that individuals consciously maximize as posited in economic models, but that they act as if they did. Examples from the interview/questionnaire evidence or from everyday life were then used to suggest that the data, indeed, would sustain, if systematically garnered, the neoclassical model.

Although the direct rejoinders to the attacks on marginalist pricing models did not present data to support their positions, others did. Among these, the most famous perhaps is George Stigler’s (1947) demolition of the Hall–Hitch–Sweezy kinked-demand schedule explanation of price rigidity in the 1930s. Stigler argued quite correctly that a kink should only exist for oligopolies, and thus that the relationship between price changes and concentration or number of sellers should be U-shaped. Only oligopo-lies should change price less frequently than profi t maximization would imply. Stigler presented data on numbers of price changes in markets with diff erent numbers of fi rms that dramatically rejected the U-shape predic-tion. The number of price changes in a market increased directly with the

48 Key issues

number of sellers. In the two markets with one seller, aluminum and nickel, there were respectively two and zero price changes between June of 1929 and May of 1937.7

It is diffi cult to believe that, in a decade as unusual as the 1930s, the demand for a basic industrial product like nickel did not shift suffi ciently to induce at least one change in the profi t-maximizing price for this monopo-list, especially since the coeffi cient of variation of output for this industry was the sixth largest of the 21 industries Stigler examined. The Stigler results, while destroying the kinked-demand schedule hypothesis, raised the puzzling question of why price rigidity increases with concentration, and Stigler admitted ‘that the neoclassical theory does not provide a satisfactory explanation for this extraordinary rigidity of monopoly prices’ (1947, p. 428). The lesson drawn by the profession from Stigler’s paper was, however, only that the data had rejected the challenges to neoclassical theory off ered by Hall and Hitch (1939) and Sweezy (1939). That the data were equally inconsistent with what neoclassical theory predicted was ignored.

In the mid-1940s one would not have had to cast about far to fi nd a hypothesis that fi t these results, however. Gardiner Means’s administered price hypothesis argued that large corporations held prices constant for long periods in markets dominated by a ‘relatively small number of con-cerns’ (National Resources Committee, 1939, p. 143, as quoted in Scherer, 1980, p. 350). That George Stigler would not immediately seize upon the administered price hypothesis to explain his results is not surprising. Indeed, a generation later he and James Kindahl (1970) were to publish a major empirical study that claimed to refute the administered price thesis. In fact, the price rigidities that were observed could be reconciled with ‘traditional theory’ if the latter was appropriately modifi ed by additional assumptions regarding long-term contracts and transaction costs (Stigler and Kindahl, 1973, p. 719). Both Means (1972) and Leonard Weiss (1977) followed with empirical studies that they claimed were consistent with the administered price thesis. Many additional studies examined the fl exibility of prices. I shall not dwell on this literature,8 but merely assert that the work of that period did not produce a resounding victory for the marginalist model in any standard form. But the profession proceeded ahead as if it did.

3. THE MANAGERIALIST CHALLENGE

The attacks on economic orthodoxy just discussed all questioned the implications of profi ts maximization with regards to pricing decisions. In the 1950s and 1960s studies appeared that directly questioned the profi ts maximization assumption and neoclassical predictions regarding decisions

The corporation 49

other than price. William Baumol (1959, 1966) hypothesized that managers maximized sales; Oliver Williamson (1963) added staff and emoluments to the managers’ objective function; Robin Marris (1963, 1964), the growth of the fi rm. Cyert and March (1963) posited four objectives in addition to profi ts pursued by the fi rm. Most fundamentally, Herbert Simon (1957, 1959) argued that managers did not maximize any objective function at all; they satisfi ced.

What needs to be stressed about these examples is that they all stemmed from observations about how managers and corporations actually behave. Simon’s satisfi cing hypothesis originated from his work in psychology and his study of organizational behavior. His colleagues, Cyert and March, built on Simon’s behavioralist approach and set out to describe the decision-making processes in actual, large corporations rather than to model an ideal, representative fi rm. To do so they constructed program-ming models of actual corporate decision structures. Williamson, a student of Simon, was also seeking a more realistic description of the ‘managerial preference function’ than existed at that time. Baumol’s hypothesis arose from observations about the importance of sales to managers as an index of the health of their fi rm, and as a source of status (1966, pp. 44–8). Marris launches his study with a lengthy review of the literature on organizational behavior, which dwells on motivation, compensation formulae and the like.9 Thus each was seeking to model in a more accurate way the behavior of managers as they had actually observed it, or as they had come to under-stand it from reading a literature that came from outside of economics.

These challenges to economic orthodoxy were dismissed with arguments similar to those used to repel the attacks against marginalist price theory (Baldwin, 1964; Peterson, 1965; Machlup, 1967).

In 1970 William Baumol and colleagues published estimates of rates of return on reinvested cash fl ows during the 1950s and 1960s ranging from 2 to 6 percent. These returns were signifi cantly below both the returns shareholders were earning over this period and the returns Baumol et al. estimated on new debt and equity issues. They corroborated the hypoth-esis that managers not subject to the discipline of external capital markets would, relative to what was optimal for their shareholders, overinvest their internal cash fl ows. As with every study that seemingly contradicts the con-ventional wisdom in economics, the Baumol et al. results were immediately challenged, and several additional studies followed.10 In one of these Henry Grabowski and I brought in a life-cycle hypothesis (1975). Mature fi rms in industries with mature technologies earned signifi cantly lower returns than young fi rms in industries with newer technologies. Our results also cast light on another paradoxical fi nding in the literature – the seemingly ‘irrational’ preference of shareholders for dividends over retained earnings.

50 Key issues

For our sample, it was only the shareholders of mature companies earning relatively low returns on investment who preferred dividends to retentions. The preference for dividends tended to disappear for fi rms earning high returns on investment. Shareholders were not so irrational after all.

In the late 1960s Dale Jorgenson and Calvin Siebert (1968) developed and tested a neoclassical theory of investment. The key determinant of invest-ment for the shareholder-wealth-maximizing fi rm was the Modigliani and Miller cost of capital. Cash fl ow had no place in a neoclassical investment equation.11 Of course Jorgenson was right. But cash fl ow did belong in the investment equation for the managerial fi rm, since it was this source of capital over which managers could exercise the most discretion. Although the neoclassical cost of capital invariably outperformed cash fl ow in Jorgenson’s articles, other studies (Elliot, 1973; Grabowski and Mueller, 1972) continued to fi nd that cash fl ow was superior to measures of the neoclassical cost of capital.

Thus, a pattern of empirical results was visible in the 1970s that was fully consistent with a managerial discretion/size-growth maximization hypothesis about the corporation. The greater a corporation’s cash fl ow, the more it spent on capital equipment and R&D; reinvested cash fl ows earned relatively low rates of return; mature corporations earned lower returns than young companies or those with new technologies; the market priced the shares of mature fi rms in a way that implied a preference for greater dividends and less reinvested cash fl ows. At the same time evidence was accumulating to suggest that the conglomerate merger wave of the 1960s had reduced corporate effi ciency. The wealth of the shareholders of acquiring fi rms steadily declined relative to other shareholders as the market learned more and more about the conglomerate mergers.12

But this pattern of evidence either went unnoticed or, if it was discerned, failed to dislodge the view that managers maximized profi ts or shareholder wealth. The managerial theories joined the mark-up pricing models that had preceded them as valiant but futile attempts to replace the simplistic view of managerial decision making that characterized the neoclassical model of the fi rm. It should be noted that this outcome is not peculiar to the fi eld of industrial organization. Robert Frank (1985) focuses on the inadequacy of neoclassical theory in explaining wage patterns within fi rms, but notes also, citing Mayer (1972), that ‘the evidence for [a] relationship [between income and savings] is so strong and so consistent that it would appear diffi cult for proponents of the permanent income and life-cycle [saving] theories to continue to insist that savings rates are unrelated to income. Yet these claims persist in all major undergraduate and graduate texts in macroeconomics’ (Frank, 1985, p. 160). Thus, despite a broad consensus among economists that hypotheses should be formulated in such a way

The corporation 51

that they can be ‘rejected by the data’, no empirical evidence is ever deemed strong enough to reject a hypothesis that assumes that agents maximize one of the standard behavioral objectives, that is, in the case of the fi rm, profi ts or shareholder wealth, or at least so it seemed up into the 1970s.

4. CONSTRAINTS ON MANAGERIAL DISCRETION

In trying to explain why managers maximize profi ts or shareholder wealth, economists have off ered essentially fi ve diff erent arguments. Four rely on the existence of a particular market.

4.1 Product Market Competition

The most obvious way to curb managerial discretion and force managers to maximize profi ts is to ensure that product markets are perfectly competi-tive. If managers have to maximize profi ts simply so that their company survives, they will maximize profi ts.

4.2 An Effi cient Capital Market

It should be obvious to potential shareholders that the incentives managers have to maximize shareholder wealth are attenuated if the managers own only a fraction of a company’s shares. If the capital market is effi cient, it will recognize when an owner-manager fi rst announces a sale of shares that the owner-manager will engage in more on-the-job consumption following the sale – purchase more staff and emoluments, pursue growth to a greater degree, and so on. The assumption of capital market effi ciency implies that the share price drops immediately upon the sale’s announcement to refl ect the manager’s additional on-the-job consumption. All of the agency costs from a separation of ownership and control are thus borne by the original owner-manager and s/he therefore has an incentive to minimize these costs.13

4.3 The Market for Managers

Eugene Fama (1980) has claimed that agency problems are eliminated through the workings of the market for managers. Managers will wish to develop a reputation for maximizing profi ts (not engaging in on-the-job consumption) to improve their chances for promotion within the fi rm that they currently work for, and to generate attractive job off ers from other companies.

52 Key issues

4.4 The Market for Corporate Control

Robin Marris (1963, 1964) hypothesized that the constraint upon growth-maximizing managers which prevented them from ignoring shareholders’ interests entirely was the threat of takeover by outsiders if the share price fell too low, and subsequent loss of job. Henry Manne (1965) coined the term ‘market-for-corporate-control’ and claimed that it tended to solve the agency problems created by the separation of ownership and control.

4.5 Principal–Agent Contracts

The fi fth constraint on managerial discretion emphasized in the literature comes through the incentives built into the manager’s compensation con-tract. The principal, that is, the shareholder, is assumed to be concerned only with his wealth or the utility of his wealth. The agent gets utility from his wealth and disutility from the eff ort expended on behalf of the princi-pal. One or both may be risk averse. The principal cannot fully monitor the agent and thus must try to induce the agent to maximize the principal’s wealth or utility by incorporating the proper incentives into the employ-ment contract. The optimal contract typically does not maximize share-holder wealth, because it needs to insure the agent from some of the risks of the company.

5. HOW STRONG ARE THE CONSTRAINTS?

In this section, I briefl y question each of the hypothesized constraints on managerial discretion put forward in the previous section.

5.1 Product Market Competition

Few industrial organization economists believe that all or even most markets are perfectly competitive. Neither Microsoft’s nor Coca-Cola’s managers need lie awake at night worrying about whether their fi rm can survive the intense competition it faces.

5.2 An Effi cient Capital Market

The initial share off erings of most companies occur when they are young and small. The main concern of potential buyers of these shares at that time is not over managers’ on-the-job consumption, but whether the young fi rm will survive. If it does, and if it grows big, a day will come when its

The corporation 53

managers can engage in on-the-job consumption at their shareholders’ expense. But by this time the company will most likely have ceased issuing shares. It presses the assumption of rational expectations on the part of the capital market very hard to assume that a possible share price drop upon the initial sale of a company’s shares in anticipation of managerial on-the-job consumption in the distant future can curb this activity.

5.3 The Market for Managers

The market for managers seems more likely to be an eff ective disciplinary force for middle managers than for senior managers, at least for large com-panies. Once a manager has become CEO or president of a BP or a General Electric their next job is likely to be hitting golf balls. Should they choose to engage in a little on-the-job consumption or empire-building they are unlikely to worry about the eff ects on their future employment activities. Potential agency problems with respect to the top managers of large corpo-rations are unlikely to disappear because of the market for managers.

5.4 The Market for Corporate Control

If a corporation has a potential market value of $100 billion and a current market value of $80 billion, then a potential gain exists from taking over the company and replacing the current managers to realize the company’s potential value. If the managers of the undervalued fi rm are unwilling to sell out through a friendly merger, a tender off er must be made. This would require raising $40 billion at the current share price, and even more consid-ering that a premium will have to be paid to acquire at least 50 percent of the shares. Few potential bidders have that amount of money, and it may be diffi cult to raise from banks if the assets of the potential target cannot be quickly turned into cash once control is gained.

5.5 Principal–Agent Contracts

The basic problem with the principal–agent incentive contract story is that we do not observe managerial compensation contracts with the characteristics that this story implies. If the shareholders were to design an incentive contract to mitigate the principal–agent problem, they would tie managerial compensation to some combination of reported profi ts and share price, as perhaps with stock bonuses. The measure of profi ts or share price in the contract would be some estimate of the increase in profi ts or share price caused by the managers. Compensation contracts typically reward managers, however, with stock options and bonuses that

54 Key issues

go up with general market swings. A wise shareholder would also retain the authority to fi re the manager for poor performance, for example, by giving managers only nonvoting shares. However, if nonvoting shares are issued, they typically go to the shareholders. Increasingly popular among managers in recent years have been multiple-vote, super shares. Perhaps, the clearest evidence of the failure of managerial compensation contracts to provide proper incentives to managers has been the practice following the drop in stock prices at the end of the late 1990s bull market of revaluing mangerial stock options downward. Why the discrepancy between theory and fact? It arises because the fundamental premise of the principal–agent model, in the case of the shareholder and manager, is false. Shareholders do not write the contracts that defi ne managerial com-pensation, and do not hire and fi re managers. To a considerable degree managers select themselves and design their own contracts (Vancil, 1987; Bebchuk and Fried, 2004).

6. DEVELOPMENTS OVER THE LAST TWENTY YEARS

In 1993 Elizabeth Reardon and I published a study in which we estimated marginal qs (ratios of returns on investment to costs of capital) for 699 companies over the period end of 1969 to end of 1988. A fi rm which maximizes its shareholders’ wealth should have a marginal q equal to or slightly greater than 1. Our estimates were less than 1 for eight out of ten companies. The median estimated marginal q was 0.71. Cumulated over the 19-year period, the 699 companies have collectively destroyed roughly $1 trillion by investing in projects with returns less than their costs of capital. General Motors alone, with a marginal q of 0.48 destroyed around $150 billion. These fi gures vividly reveal the signifi cant agency problems in US corporations that existed during the 1970s and part of the 1980s with respect to investment policies. Shareholders in the 699 large companies in our sample would have been $1 trillion richer if the managers of these companies had invested in the way that economics textbooks say they do.

Others observed the poor performance of US corporations and com-mented upon it at the time. As late as 1983, in a survey of the merger lit-erature with Richard Ruback, Michael Jensen requested more ‘knowledge of this enormously productive social invention: the corporation’ (Jensen and Ruback, 1983, p. 47). By 1989, Jensen had acquired the requested knowledge and predicted that the ineffi ciencies of the corporation with ownership and control separated were so signifi cant that it was destined

The corporation 55

to disappear. A more rapid conversion had not taken place since Paul journeyed to Damascus. Even as Jensen was predicting the corporation’s demise, however, developments were taking place that would enhance its effi ciency.

Many economists refer to a merger wave of the 1980s in the United States. Compared with merger activity in the 1990s, that of the 1980s was barely a ripple. It is legitimate, however, to speak of a wave of hostile takeo-vers in the mid to late 1980s. Tender off ers, of which hostile takeovers are an important part, rose to 25 percent of all mergers during this period, a fi gure never before or since seen (Gugler et al., 2007). Moreover, many of the hostile takeovers were headline-grabbing takeovers of major US com-panies. The business and popular press was fi lled with stories about them, and they even became the subject for a popular movie, Wall Street. For the fi rst time in US history, the takeover constraint that Marris and Manne had postulated existed began to have some real teeth.

Managers reacted. Unprofi table and unrelated divisions were sold off . Managers began buying back their companies’ shares rather than under-taking bad investments or mergers. Terms like ‘back to core competences’, ‘downsizing’ and ‘shareholder value’ began to fall from managers’ lips. Managers’ concerns about shareholder wealth changed radically following the hostile merger activity of the 1980s.

The constraint on managers from product market competition can also be said to have increased in recent years due to globalization. Forty years ago a US corporation needed to worry only about the response of other US companies to an innovation or price change. Today, Schumpeter’s gale of creative destruction storms over the innovator from around the world.

A third development that has increased constraints on managers is the growth of institutional shareholdings. In 1950, only one in ten shares was held by a pension fund, mutual fund, or some other institutional share-holder. By the mid-1990s the fi gure was one in two (Friedman, 1996). The managers of these institutional portfolios are full-time stockholders who appear to be increasingly willing to intervene to block a merger or other management decision that the portfolio managers believe will lower share price.

These developments have had a noticeable impact on the investment performance of US companies. Estimates of marginal q for the United States over the period 1985–2000 yield a mean of 1.02, a dramatic improve-ment over the 1970–88 period (Gugler et al., 2004). Thus, at a point in time when many economists are fi nally beginning to acknowledge the existence of agency problems and to take them seriously, institutional changes may be making them less important.

56 Key issues

7. AN END TO MANAGERIAL DISCRETION?

Despite the institutional developments discussed in the previous section, and the much better investment performance of US companies, it may still be too early to declare all managerial discretion issues to be totally resolved. The study that reported the 1.02 estimate for the 1985–2000 period in the United States reported much lower fi gures for Continental European and Latin American countries. Thus, even if agency problems seem to have become less important in the United States and some other Anglo-Saxon countries, they appear to be alive and well in many other countries.

The fi rst phrase that comes to mind when trying to describe the merger wave of the late 1990s is not ‘back to core competences’ or ‘downsizing’. The merger wave looked to be fueled by soaring stock prices much like all other waves. And, as in other merger waves, the shareholders of the acquiring companies appear to have suff ered substantial losses (Gugler et al., 2007; Moeller et al., 2005). This wave also illustrated that institutional shareholders are not as powerful a check on managers as some think. They appear to have been just as swept up by the euphoria of the bull market, and just as willing to believe the various ‘theories’ about why this or that merger will generate synergies – theories which, as with other merger waves, have not received a lot of empirical support.

Finally, it must be noted that managerial salaries continue to climb to unprecedented heights despite all the attention that they have received.

NOTES

1. See Gilson and Roe (1993), and Charkham (1994). 2. See Hansmann and Kraakman (2000). 3. See in particular the quote of John Bates Clark in Letwin (1965, p. 74) and Stigler (1950,

p. 76). 4. On these see Galbraith (1972). 5. When not ignoring it, the economist would often ridicule it. As late as 1982, at a con-

ference held ostensibly to ‘celebrate’ the fi ftieth anniversary of the publication of The Modern Corporation and Private Property, the tenor and tone of the papers reveals that many came not to praise the book but to bury it (see special issue of Journal of Law and Economics, June 1983). Douglass North’s (1983) comment is a nice exception.

6. Friedman’s (1953) and Becker’s (1962) famous essays could also be cited here. 7. Stigler’s initial fi ndings have been substantiated in several other studies; for example,

Simon (1969); Primeaux and Bomball (1974); and Primeaux and Smith (1976). 8. For a survey, see Scherer (1980, pp. 350–62), and a more recent re-examination (Carlton,

1986). 9. I also know from personal conversations with him that his thinking on these matters has

been importantly infl uenced by personal experiences with a fi rm in his family.10. This literature is reviewed in Mueller (2003a, pp. 145–8).11. The pioneering study of the role of cash fl ow in an investment equation is of course

The corporation 57

that of Kuh and Meyer (1957). Although they interpret the strong performance of cash fl ow in their investment equations as consistent with the profi ts maximization hypoth-esis, cash fl ow enters their list of possible explanatory variables not as a result of the application of the marginal analysis, but because ‘by far the most outstanding aspect of . . . direct inquiries [about the determinants of investment] is their virtual unanimity in fi nding that internal liquidity considerations and a strong preference for internal fi nanc-ing are prime factors in determining the volume of investment’ (p. 17). The third of the three reasons they give for the strong preference for internal funds is ‘the hierarchical structure and motivations of corporate management which make outside fi nancing asymmetrically risky for the established or in-group’ (pp. 17–18).

12. See the extensive references in my surveys (Mueller, 1977, 2003b, pp. 163–70).13. See Jensen and Meckling (1976).

REFERENCES

Baldwin, W.L. (1964), ‘The Motives of Managers, Environmental Restraints, and the Theory of Managerial Enterprise’, Quarterly Journal of Economics, 78 (May), 236–58.

Baumol, W.J. (1959), Business Behavior, Value and Growth, New York: Macmillan; 2nd edn, 1966.

Baumol, W.J., P. Heim, B.G. Malkiel and R.E. Quandt (1970), ‘Earnings Retention, New Capital and the Growth of the Firm’, Review of Economics and Statistics, 52, (Nov.), 345–55.

Bebchuk, L. and J. Fried (2004), Pay Without Performance, Cambridge, MA: Harvard University Press.

Becker, G.S. (1962), ‘Irrational Behavior and Economic Theory’, Journal of Political Economy, 70, (Feb.), 1–13.

Berle, A.A. and G.C. Means (1932), The Modern Corporation and Private Property, New York: Commerce Clearing House; rev. edn, New York: Harcourt, Brace, Jovanovich, 1968.

Carlton, D.W. (1986), ‘The Rigidity of Prices’, American Economic Review, 76, (Sept.), 637–58.

Charkham, J.P. (1994), Keeping Good Company, Oxford: Oxford University Press.Cyert, R.M. and J.G. March (1963), A Behavioral Theory of the Firm, Englewood

Cliff s, NJ: Prentice-Hall.Elliot, J.W. (1973), ‘Theories of Corporate Investment Behavior: Revisited’,

American Economic Review, 63 (March), 195–207.Fama, E.F. (1980), ‘Agency Problems and the Theory of the Firm’, Journal of

Political Economy, 88 (April), 288–307.Frank, R.H. (1985), Choosing the Right Pond, Oxford: Oxford University Press.Friedman, B.M. (1996), ‘Economic Implications of Changing Share Ownership’

Journal of Portfolio Management, 2 (3), 59–70.Friedman, M. (1953), ‘The Methodology of Positive Economics’, in Essays in

Positive Economics, Chicago, IL: University of Chicago Press, 3–43.Galbraith, J.K. (1961), The Great Crash, 1929, Boston, MA: Houghton Miffl in,

3rd edn, 1972.Gilson, R.J. and M.J. Roe (1993), ‘Understanding the Japanese Keiretsu: Overlaps

Between Corporate Governance and Industrial Organization’, The Yale Law Journal, 102 (4), 871–906.

58 Key issues

Grabowski, H.G. and D.C. Mueller (1972), ‘Managerial and Stockholder Welfare Models of Firm Expenditures’, Review of Economics and Statistics, 54, Feb, 9–24.

Grabowski, H.G. and D.C. Mueller (1975), ‘Life-Cycle Eff ects on Corporate Returns on Retentions’, Review of Economics and Statistics, 57, Nov, 400–409.

Gugler, K., D.C. Mueller and B.B. Yurtoglu (2004), ‘Corporate Governance and the Returns on Investment’, Journal of Law and Economics, 47 (Oct.), 598–633.

Gugler, K., D.C. Mueller and B.B. Yurtoglu (2007), ‘The Determinants of Merger Waves’, mimeo, University of Vienna.

Hall, R.L. and C.J. Hitch (1939), ‘Price Theory of Business Behaviour’, Oxford Economic Papers, os-2 (1), 12–45.

Hansmann H. and R. Kraakman (2000), ‘The Essential Role of Organizational Law’, Yale Law Journal, 110 (3), Dec, 387–440.

Jensen, M.C. (1989), ‘The Eclipse of the Public Corporation’, Harvard Business Review, (Sept./Oct.), 61–74.

Jensen, M.C. and W.H. Meckling (1976), ‘Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure’, Journal of Financial Economics, 3 (4), 305–60.

Jensen, M.C. and R.S. Ruback (1983), ‘The Market for Corporate Control’, Journal of Financial Economics, 11 (April), 5–50.

Jorgenson, D.W. and C.D. Siebert (1968), ‘A Comparison of Alternative Theories of Corporate Investment Behavior’, American Economic Review, 58 (4), 681–712.

Journal of Law and Economics (1983), Special Issue, June.Kahn, A.E. (1959), ‘Pricing Objectives in Large Corporations: Comment’, American

Economic Review, 49 (4), 670–80.Kaplan, A.D.H., J.D. Dirlam and R.F. Lanzillotti (1958), Pricing in Big Business.

A Case Study Approach, Washington, DC: Brookings Institution, 128–9.Kuh, E. and J.R. Meyer (1957), The Investment Decision, Cambridge, MA:

Harvard University Press.Lester, R.A. (1946), ‘Shortcomings of Marginal Analysis for Wage–Employment

Problems’, American Economic Review, 36 (1), 63–82.Letwin, W. (1965), Law and Economic Policy in America, New York: Random

House.Machlup, F. (1946), ‘Marginal Analysis and Empirical Research’, American

Economic Review, 36 (Sept.), 519–54.Machlup, F. (1947), ‘Rejoinder to an Antimarginalist’, American Economic Review,

37 (1), 148–54.Machlup, F. (1967), ‘Theories of the Firm: Marginalist, Behavioral, Managerial’,

American Economic Review, 57 (March), 1–33.Manne, H.G. (1965), ‘Mergers and the Market for Corporate Control’, Journal of

Political Economy, 73 (April), 110–20.Marris, R. (1963), ‘A Model of the “Managerial” Enterprise’, Quarterly Journal of

Economics, 77 (May), 185–209.Marris, R. (1964), The Economic Theory of Managerial Capitalism, Glencoe: Free

Press.Marshall, A. (1920), Principles of Economics, 8th edn, London: Macmillan.Marshall, A. (1923), Industry and Trade, 4th edn, London: Macmillan.Mayer, T. (1972), Permanent Income, Wealth and Consumption, Berkeley, CA:

University of California Press.Means, G.C. (1935), Industrial Prices and Their Relative Infl exibility, Washington,

DC: U.S. Senate Document 13, 74th Congress, 1st Session.

The corporation 59

Means, G.C. (1972), ‘The Administered Price Thesis Reconfi rmed’, American Economic Review, 62 (June), 292–306.

Mill, J.S. (1885), Principles of Political Economy, 9th edn, London: Longmans, Green and Co.; fi rst published 1848.

Moeller, S.B., F.P. Schlingemann and R.M. Stulz (2005), ‘Wealth Destruction on a Massive Scale? A Study of Acquiring-Firm Returns in the Recent Merger Wave’, Journal of Finance, 60 (2), 757–82.

Mueller, D.C. (1977), ‘The Eff ects of Conglomerate Mergers: A Survey of the Empirical Evidence’, Journal of Banking and Finance, 1, December, 315–47.

Mueller, D.C. (2003a), The Corporation – Investment, Mergers, and Growth, London: Routledge, 2003.

Mueller, D.C. (2003b), ‘The Finance Literature on Mergers: A Critical Survey’, in M. Waterson (ed.), Competition, Monopoly and Corporate Governance, Essays in Honour of Keith Cowling, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 161–205.

Mueller, D.C. and E. Reardon (1993), ‘Rates of Return on Corporate Investment’, Southern Economic Journal, 60 (Oct.), 430–53.

National Resources Committee (1939), The Structure of the American Economy, Part I, Washington, DC: National Resources Committee.

North, D.C. (1983), ‘Comment on Stigler and Friedland: The Literature of Economics: The Case of Berle and Means’, Journal of Law and Economics, 26 (2), 269–71.

Peterson, S. (1965), ‘Corporate Control and Capitalism’, Quarterly Journal of Economics, 79 (Feb.), 1–24.

Primeaux, W.J. Jr. and M.R. Bomball (1974), ‘A Reexamination of the Kinky Oligopoly Demand Curve’, Journal of Political Economy, 82 (4), 851–62.

Primeaux, W.J. Jr. and M.C. Smith (1976), ‘Pricing Patterns and the Kinky Demand Curve’, Journal of Law and Economics, 19 (1), 189–99.

Scherer, F.M. (1980), Industrial Market Structure and Economic Performance, 2nd edn, Chicago, IL: Rand McNally.

Simon, H.A. (1957), ‘The Compensation of Executives’, Sociometry, 20 (March), 32–5.

Simon, H.A. (1959), ‘Theories of Decision Making in Economics and Behavioral Science’, American Economic Review, 49 (June), 253–83.

Simon, J.L. (1969), ‘A Further Test of the Kinky Oligopoly Demand Curve’, American Economic Review, 59 (5), 971–5.

Smith, A. (1776), The Wealth of Nations, reprinted New York: Random House, 1937.

Stigler, G.J. (1947), ‘The Kinky Oligopoly Demand Curve and Rigid Prices’, Journal of Political Economy, 55 (5), 432–49; reprinted in G.J. Stigler and K.E. Boulding (eds), 1952, Readings in Price Theory, Chicago, IL: Irwin, 410–39.

Stigler, G.J. (1950), ‘Monopoly and Oligopoly by Merger’, American Economic Review, 40 (2) (May), 23–34; reprinted in R.B. Hefl ebower and G.W. Stocking (eds), 1958, Readings in Industrial Organization and Public Policy, Homewood, IL: Irwin, 69–80.

Stigler, G.J. and J.K. Kindahl (1970), The Behavior of Industrial Prices, New York: Columbia University Press.

Stigler, G.J. and J.K. Kindahl (1973), ‘Industrial Prices, as Administered by Dr. Means’, American Economic Review, 63 (4), 717–21.

60 Key issues

Sweezy, P.M. (1939), ‘Demand Under Conditions of Oligopoly’, Journal of Political Economy, 47 (4), 568–73.

Vancil, R.F. (1987), Passing the Baton, Boston, MA: Harvard Business School.Weiss, L.W. (1977), ‘Stigler, Kindahl, and Means on “Administered Prices”’,

American Economic Review, 67 (Sept.), 610–19.Williamson, O.E. (1963), ‘Management Discretion and Business Behavior’,

American Economic Review, 53 (Dec.), 1032–57.

PART II

The theory of the fi rm from an organizational perspective

63

4. A contractual perspective of the fi rm with an application to the maritime industryPer-Olof Bjuggren and Johanna Palmberg*

1. INTRODUCTION

In 1776 Adam Smith developed a theory for markets as the coordinating device in an economy. However, economic activities are not only coordi-nated through the price mechanism of the market but are also guided by fi rms within which the production of goods and services takes place. In contrast to markets the fi rm is not so well developed in economic theory. This is refl ected, for example, in basic economic textbooks that usually assume the fi rm as something exogenously given that need not be explained or analyzed. After acknowledging the existence of fi rms, textbooks quickly turn to the market and analyze its importance for a well-functioning economy.1

However, since the early 1970s there has been rapidly expanding research on the theory of the fi rm, largely inspired by an article dating back to 1937 about the nature of the fi rm written by the Nobel laureate, Ronald H. Coase. However, it took more than thirty years for researchers to draw inspiration from the ideas put forward by Coase. In the 1970s (primarily) Oliver E. Williamson continued along the line of research that Coase had outlined. Since then theories of the fi rm have been a new expanding area of research in economics. In this chapter, a contractual perspective on the fi rm is used, with the concept of institution being an important corner-stone, and a synthesis of diff erent contractual perspectives on the fi rm as the coordinating institution within the maritime industry.

This chapter recognizes the fact that the fi rm itself can enter into con-tracts with other fi rms and physical persons. In a competitive environment there is a strong tendency for the most cost-effi cient composition of con-tracts within a given institutional environment to survive (see, for example, Fama and Jensen, 1983). We will apply such a contractual view. In addi-tion to the production costs (remuneration to suppliers of diff erent kinds

64 The theory of the fi rm from an organizational perspective

of inputs such as labor and capital services) these costs include information costs and the costs for negotiating and monitoring contracts.

In line with transaction cost economics, we consider all transactions to be costly. These transaction costs can then be used to explain market structures as well as fi rms and other institutions (see, for example, Chapter 2 in this volume). The focus in this chapter is on the relationship between contracts and transaction characteristics. The maritime industry off ers an interesting area of research for economists. Diff erent theories developed within the fi elds of fi nancial economics, institutional economics, corporate governance and industrial organization can be applied with good analyti-cal results. A quick glance through the volumes of Maritime Policy and Management illustrates this. Furthermore, the maritime industry is inter-esting for contract theory since all diff erent types of contracts ranging from spot contracts to vertical integration are used in the industry. Also, the organization of the shipping company is aff ected by third-party manage-ment which in its extreme form separates ownership from control.

In the early 1990s an article by Stephen Pirrong demonstrated how transaction cost analysis could be applied in the analysis of contracting practices in maritime transport (Pirrong, 1993). We extend Pirrong’s anal-ysis by applying transaction cost analysis both to the freight contract and to the fi rm. The idea behind this chapter originates from a larger research project on the Swedish shipping industry (see Johansson et al., 2006).2 Therefore the Swedish maritime industry is used for illustration. The mari-time industry is a truly global industry; the Swedish ship-owners and their counterparts all act on international markets. Hence, characteristics valid for Sweden also apply to larger shipping nations.

Section 2 of the chapter starts with a presentation of our contractual view of the fi rm, with maritime illustrations, and Section 3 off ers a brief presenta-tion of maritime transport and the diff erent types of fi rms that operate in the market. A closer look at diff erent contractual compositions that character-ize maritime transport fi rms with transaction cost and institutional explana-tions is presented in Section 4, and Section 5 concludes the chapter.

2. A CONTRACTUAL PERSPECTIVE ON THE FIRM

In neoclassical economics the fi rm is often treated as a ‘black box’, but viewing the fi rm as a ‘nexus of contracts’ (Jensen and Meckling, 1976; Ståhl, 1976) results in a much richer analysis of market processes and the allocation of resources in an economy. In this section the fi rm is dis-cussed from a contractual perspective. First, an overview of the relation between specialization and productivity and the need for coordination

A contractual perspective of the fi rm 65

and institutions (Section 2.1) is presented. Section 2.2 discusses the fi rm as a contractual entity in depth. Section 2.3 concludes with an analysis of the relation between contracts, markets and fi rms.

2.1 Specialization and Institutions

In an analysis of the organization of an industry, specialization is a crucial concept since it fosters increased productivity. This is an important message of the fi rst chapter of Adam Smith’s Wealth of Nations (1776). With the help of the famous pin factory example, Smith demonstrates in detail how productivity is increased by specialization. However, specialization implies increased mutual dependence and hence production must be organized in a fashion that solves this mutual dependency. In Figure 4.1 this is expressed as demands on proper institutional solutions created by the coordination problem caused by mutual dependence due to specialization. (Institutions are thereby defi ned as rules for individual interactions/cooperation.)

Firms and markets are alternative institutional solutions to the coor-dination problem. We will start by looking at the fi rm as an institu-tional arrangement to solve the coordination problem caused by factor specialization.

2.2 The Firm as a Nexus of Contracts

An important feature of a fi rm is that it is a legal entity and can, just like a physical person, enter into binding agreements (contracts) with other physical and legal persons (see Ståhl, 1976). From this perspective the fi rm can be seen as a ‘nexus of contracts’ that coordinates fi nancial investors, suppliers of intermediate goods, services and labor, and customers in the production of goods and services. Figure 4.2 shows the fi rm from such a contractual perspective.

The production factors human capital (H) and physical capital (K) can serve as a starting point in a description of this contractual view of the fi rm. These are the independent variables commonly used in production functions such as Cobb–Douglas and CES. The fi rm can choose either to own or to rent its physical capital. Ownership implies property rights

Increasedproductivity bydivision of labour(specialization)

Mutualdependence

Coordinationrequired

Institutionalsolutions

Figure 4.1 Specialization and institutions


with an exclusive right to use physical capital and the return from its use. Furthermore, private property right is associated with an exclusive right to transfer the property right through an agreement (contract) to another person. A rental agreement implies a much more limited scope for deci-sions about the use of the physical asset. In the maritime sector the most important physical capital is the vessel. Both ownership and rental agree-ments and combinations of these two alternatives are common amongst ship-owners.

For labor there is a choice between an employment contract and hiring of a consultant. An employment contract is much more open than a contract with a consultant with regard to use of labor. The employment contract makes it possible to use a hierarchical type of decision making regarding the use of human capital, and thereby provides an opportunity to replace the price mechanism of the market with administrative decisions about resources allocation (see, for example, Coase, 1937; Masten, 1988). One could say that the invisible hand of the market is replaced by the visible hand of an organization.

In the shipping industry a wide range of contracts with labor can be found. A phenomenon of special interest that will be discussed in more detail below is the so-called third-party ship management, where the owners of physical capital, the vessel, hire parts of or all labor and management services from another company. (See Section 4.1 for further discussion.)

A fi rm’s fi nancial contractual relations have governance implications.

= Guaranteed contracts = Residual contract

The firm as aproduction unitand a nexus of

contracts

Shareholders

Creditors

Users/customers

Suppliers of rawmaterial and goods

Suppliers of humancapital services

Suppliers of physicalcapital services

Figure 4.2 The fi rm as a nexus of contracts


The shareholders are considered the owners of the fi rm. Their contrac-tual relation with the fi rm is characterized by a claim on the residual that remains when all other contractual obligations of the fi rm have been met. (They are residual claimants.) The return on their investment is therefore directly related to how well the fi rm is managed. This dependency makes it important to have a mechanism through which shareholders can control how the corporation acts as a legal person. In most cases it is the board of directors and the CEO who, on behalf of the fi rm, enter into binding contracts with, for example, suppliers, employees and customers. It is thus logical (understandable) that the shareholders directly or indirectly choose who will have these positions.3

On the fi nancial side of the fi rm there are also lenders (investors) with fi xed claims contracts (banks and bondholders). In contrast to the share-holders they have specifi ed claims on the fi rm in terms of mortgage plans, maturity and interest claims. If the fi rm cannot meet these fi xed claims it can be forced into liquidation/bankruptcy. The remuneration that lenders and also suppliers can get is then dependent on the value of assets to enti-ties other than the bankrupted (liquidating) corporation. Fungible assets with a well-functioning second-hand market are valuable to others and can therefore serve as collaterals for loans. Consequently fi rms that have such assets can to a larger extent than other fi rms use loans as a source of fi nance (see Williamson, 1988).

In the maritime sector the vessel is in most cases a fungible asset. There are well-functioning second-hand markets for vessels. The number of alternative carriers and customers for carrier services is for some types of vessels large enough to make the second-hand market competitive (Stopford, 1997).

Finally, we turn our attention to the fi rm’s contractual relation with suppliers and customers (contracts to be found on the input and output sides of the fi rm in Figure 4.2). Value added chains, vertical integration and supplier-specifi c/customer-specifi c specialization are important concepts here.4 A value added chain shows the diff erent stages in the processing of a raw material to fi nal consumer product; for example, from axe to loaf, from stone to house, from iron ore to car. In a value added chain there are several technologically separate stages. Between all these stages it is pos-sible to envisage transport by ship that brings the output of one stage to a succeeding stage. Just as a number of diff erent contractual relations can be found by the supplier and user in a value added chain, an equally rich fl ora of contracts is found for the shipping of raw materials, intermediate goods and consumer goods.

If the automotive industry is taken as an example, shipping can enter into the value added chain in diff erent stages of the processing of raw material


such as iron ore and steel in the manufacturing of a car for sale to a fi nal customer. Raw material such as iron ore might have to be transported by a dry bulk carrier to a steel mill. Diff erent automotive parts made of steel might have to be transported over the sea in containers to the car manu-facturer. The ready-made cars in turn have (if exported) to be shipped in specially designed vessels to other countries.

2.3 Contracts, Markets and Firms

What is especially important from a contractual perspective is whether there are assets whose productive value is dependent on uninterrupted transac-tions between a specifi c supplier and customer.5 As indicated in Figure 4.1 it is usually specialization of assets that is the source of the mutual depend-ence. Here, it is important to note the problems of sunk costs and quasi-rents associated with investment in transaction-specifi c assets. A bilateral dependence evolves immediately after the investment has been made and there is no alternative equally attractive transaction partner within a spe-cifi c price span. Klein et al. (1978) call this span the ‘appropriable quasi-rent’ as it amounts to the part of the value of a specialized investment that can be taken away without a change of employment of the factor.6

A rational supplier or customer is not prepared to make an investment in an asset of a transaction-specifi c character without at least some guarantee in the form of a long-term contract with a price that makes the investment profi table. But, as pointed out by Williamson (1985), it is sometimes diffi cult to construct such a long-term contract suffi ciently watertight in terms of no loopholes and at the same time suffi ciently fl exible to allow for changed circumstances. This is especially the case if uncertainty and complexity char-acterize the business in which the supplier and the customer are engaged. Both of the requirements ‘water tightness’ and fl exibility have to be met if the long-term contract is to serve as a perfect institutional solution to the problem of mutual dependence. But the rationality of human beings sets, as noted by Williamson (1975, 1985 and 1996), a limit to what can be achieved in terms of ‘water tightness’ and fl exibility (the assumption of bounded rationality). In a complex world, contracts are therefore bound to be more or less incomplete with loopholes that invite opportunistic behavior.7

In transactions between fi rms with diff erent owners there are confl ict-ing profi t incentives in contract negotiations. While the supplier has a profi t incentive to obtain as high a price as possible, the customer’s profi t incentive is to get as low a price as possible. The confl icting profi t incen-tives make transactions costly if an appropriable quasi-rent due to asset specifi city exists. Hence, a long-term contract could be preferred to a spot contract and, if the transactional problems are severe, vertical integration


(joint ownership) could be the most cost-effi cient solution. At the same time the strong (high-powered) incentives of independent fi rms to mini-mize cost have to be factored in when the decision to vertically integrate is contemplated. The cost minimization incentive is likely to be less powered for a salaried manager than for an owner-manager (see Williamson, 1985, and Chapter 2 in this volume).

With vertical integration, transaction costs may be avoided, as the incen-tives are diff erent in a fi rm and in the market. An employee-manager of a division X, which supplies inputs to a division Y within the same fi rm, will not be rewarded if she/he engages in costly negotiations that increase the profi t of division X at the expense of division Y and at the expense of the overall profi t of the fi rm. Instead the manager runs the risk of being fi red. Within the fi rm it is behavior of co-operation and not costly rival behavior that will be rewarded.

The structure of vertical integration can take diff erent paths, as can be seen in the structure of shipping services in relation to cars and car manu-facturers. For instance in cases such as that of container traffi c there might be a need of vertical integration forward with terminals in ports and train carriers in order to secure smooth transport of parts to the car manufac-turer. If, for example, parts of a car produced in Asia are transported on a larger container vessel for assembly on the American East Coast there might be the problem that the vessel cannot pass through the Panama Canal. Consequently, the vessel has to go through a Californian port and land transport might have to be used over the continent from west to east. In order to speed up the handling in the port, special equipment might be needed and it may thus be more effi cient to use special railway wagons. These are so-called dedicated assets that have to be available when the vessel arrives.8 Vertical integration with terminals and wagons is therefore an attractive solution (see Midore et al., 2005). In other cases, such as oil tankers, there has instead been a trend of vertical disintegration (see Veenstra and De la Fosse, 2006).

To summarize, a contractual relation depends on the nature of the asset invested in. If assets are designed in one way or another to complement each other, it is unlikely that the coordination of the use of assets in suc-cessive stages of a supply chain will be left to an atomistic market. Some safeguard is needed in order to make transactions effi cient in such cases. Vertical integration is one solution where the assets are controlled under the same ownership. In other cases it will be suffi cient to have some hybrid contractual solutions aligning the interests of supplier and customer.9

Without any safeguard the confl icting profi t incentives of supplier and customer are a source of costly transactional problems. Mutual depend-ence means that the interest of the supplier to increase revenue through


a higher price and the reverse interest of the customer to decrease cost through a lower price will lead to clashes. If there are no alternative trans-action partners to turn to, confl icts will have to be solved within a con-tinuing transactional relation. Otherwise costly investments will lose value. The vertical integration of container carriers can be explained in this way. Terminals and wagons are dedicated assets that are necessary for keeping down the costs of the vessels at ports (see Midore et al., 2005).

3. CHARACTERISTICS OF MARITIME TRANSPORT – MARKET AND FIRMS

The maritime sector off ers a wide spectrum of institutional arrangements, which makes it an interesting fi eld for contractual research. Economists with a background in fi nancial economics, institutional economics and corporate governance will fi nd contracts and organizational solutions a prospec-tive area for empirical research. This section gives an overview of the rich contractual and organizational pattern that can be found in the maritime sector.10 Moreover, it attempts to give a transaction economics explanation for the wide array of institutional arrangements in that sector.

3.1 The Freight Market

The maritime sector can roughly be divided into four types of markets (see Figure 4.3). First there is tramp shipping. The tramp market is traffi cked primarily by tankers and bulk carriers. In this market, spot contracts and forward contracts are the most commonly used contracts. Long-time char-ters are also used but they are less common. The bulk market is diversifi ed with respect to various types of cargo, such as coal, ore, grain, and forest products. Some of these products demand specially constructed (designed) vessels; this implies that the ship-owner becomes more specialized and more vulnerable to long-term changes in the market structure. Magirou et al. (1992) give an excellent review of the freight market.

A second market is liner shipping. The goods shipped in this market are higher up in the value added chain and are often transported in contain-ers and various types of ro-ro vessels. In this market the shipper–client relationship is more long-term than in tramp shipping. Moreover, vertical integration forward in the logistic chain is found in the form of ownership of terminals in ports and special train wagons. Some of the goods have to be delivered just in time, which increases the mutual dependence of the assets in these later stages of the logistic chain. Another contractual dif-ference between the tramp and liner markets is that the freight contract in


the tramp market is focused on the vessel whereas in the liner market the contract is focused on the transport.

There are diff erences also in terms of the market structure between the tramp and the liner market. The tramp market is characterized by (almost) perfect competition. Sea transport is supplied by ship-owners and bought by charterers. In contrast to the tramp market the liner market is to a large extent cartelized. Groups of shippers come together in so-called liner con-ferences in order to negotiate prices and to supply sea transport to diff erent trades. The ship-owners therefore have to compete on factors other than price such as service and times of transport. Due to containerization the liner conferences have declined in importance to the pricing strategy.

A third shipping market is the ferry/cruise market. A strong parallel can be drawn between this market and the aircraft industry. A certain route is used. Terminals and other dedicated assets have to be invested in. However, the assets (the vessels) are not transaction specifi c to a specifi c route. They can rather easily be transferred to another ferry route.

Finally there is special cargo shipping, to which car carriers and forest sea transport belong. Special shipping is characterized by long-term cus-tomer adaptation; usually the COA (contract of aff reightment) is applied. Swedish ship-owners mainly operate within the tramp market or special cargo shipping. These vessels are designed for goods that are processed so far that they are downstream near the end of the value added chain. The receiver of the good is a retailer selling a good to the fi nal consumer. Here, we fi nd long-term contracts and vertical integration.11

Spot contracts or time charters(forward contracts)

Ship-owners

Tramp marketPerfect competition

Liner shippingMarkets are often

cartelized

Tank Bulk

Ferry traffic Special cargo

shippingLong-term contracts

Standard cargo(Container-, ro-ro vessels)

General cargo

Source: Johansson et al. (2006) based on Magirou et al. (1992).

Figure 4.3 Structure of the freight market


3.2 The Freight Contract

The shipping industry is characterized by at least four types of market structures, ranging from almost perfect competition on the tramp market to cartelization on the liner market. The diff erent types of markets also use diff erent types of freight contracts. This section focuses on the bulk market. (The other types of markets are to a larger extent characterized by long-term freight contracts.) In principle there are three types of freight contracts used in the bulk market: (i) the voyager charter, (ii) the time charter and (iii) the contract of aff reightment (COA). The time span and the intensity in the contract between the shipper and the carrier depend partly on the type of freight contract and partly on the extent to which the carrier uses the same shipper (Pirrong, 1993).

The voyager charter implies that the shipper transports a single cargo or a series of shipments during a given time period. This freight contract has the shortest time span. The charter is directed at the vessel, which means that the ship-owner transports one cargo from one port to another. The contract is traded on the spot market and ceases, in normal cases, directly after the cargo has been discharged at the port. The time charter, on the other hand, is used by a carrier who wants to carry out the transport process on their own. The time charter implies that the carrier exercises command over the vessel for a specifi ed time period. The charterer is responsible not only for the commercial operation of the vessel but also for the variable costs of the vessel such as port charges, canal dues, costs for loading and discharging, stowing, and so on. The ship-owner is, on the other hand, responsible for the maintenance and the nautical operation of the vessel and for the fi xed costs of the vessel, interest on equity, deprecia-tions, and so on (Gorton et al., 1989; Johansson et al., 2006).

The third type of freight contract is the COA. This is a long-term con-tract spanning between 3–4 years and 15 years. The COA implies that it is only the vessel that is leased. The charterer is responsible for the crew, insurance, maintenance, inspections, and all variable costs. The contract is often detailed and specifi es the type of cargo, minimum and maximum volumes carried over specifi c time periods, destination and origin ports, and so on (Pirrong, 1993).

3.3 The Shipping Company

The organization of the shipping company is decided by factors such as type of shipping activities, type of market on which the company operates, and the size of the company. The structure of a company operating on the spot market diff ers from the structure of a liner shipping company. A


shipping company operating on the tramp market has, in general, only two departments: administrative and technical. Shipping companies that time-charter their vessels for shorter time periods also have an aff reightment department. The liner shipping company has the largest administrative department, with many employees and agents across the globe. Smaller shipping companies usually have a small administrative department and it is common that the onboard crew takes many of the decisions regarding the fi rm (Nya Sjöfartens Bok, 2006).

Figure 4.4 shows how a shipping company can be organized. It is only the larger shipping companies that have all the diff erent departments dis-played in the fi gure. In general, the shipping company operates on either the liner market or the spot market. Smaller shipping companies purchase administrative and technical services instead of providing them internally in the company, as shown in the fi gure.

The liner department is responsible for all the regular traffi c that the ship-ping company controls. The business is carried out both from the head offi ce and from subsidiaries. In addition, the liner shipping company purchases services from shipping agencies (brokers) in ports where it is not located. The tramp department is responsible for all the vessels that are leased on the open spot market. That is, the department is responsible both for the charter-ing and for the commercial operation of the vessels. It is also responsible for the purchasing and the selling of vessels. The administrative department is, in the ideal case, dynamic and well able to follow the fast development that characterizes the shipping industry. In general, one can say that the larger the shipping company the more of the administrative services are handled

Technological department– Construction– Operating– Inspection– Maintenance– Commissariat service

Administrative department– Financial and accounting– Personnel department– Law– IT–‘Ship management’– External and internal information

Tramp department– Affreightment– Calculation– Operational department

The shippingcompany

Liner department– Operational department– Freight booking– Freight office and accounting– Reloading– Marketing/sales

Source: Based on Nya Sjöfartens Bok, 2006 (2005).

Figure 4.4 Business structure of a large shipping fi rm


internally in the company; smaller shipping companies purchase various administrative services to a larger extent (Nya Sjöfartens Bok, 2006).

4. THE ECONOMIC ORGANIZATION OF MARITIME TRANSPORT AND MARITIME FIRMS FROM A CONTRACTUAL PERSPECTIVE

A general fi nding is thus that the contractual relation seems to depend on the good’s place in the value added chain. Raw material is transported according to spot contracts. In these markets perfect competition can be recognized. When we go further up the value chain and come closer to the fi nal consumer good, the contractual relations exhibit more of a long-term character and asset specifi city explanations can be found.

The distinction between fi rm and market also varies in an interesting way. Third-party management is a phenomenon that illustrates the fact that labor (the crew) is not tied to the asset (the vessel) to the same extent as in other industries. Instead, it is possible to have an arm’s-length relation between labor and capital to an extent not found in other industries.

4.1 Third-party Ship Management

Third-party management is a denomination for a complete separation of ownership and control in maritime transport. That is, a professional man-agement company manages ships owned by another company. In our con-tractual model (see Figure 4.2) it is at its extreme the same as breaking up the fi rm into two halves. Ownership and fi nancing of physical capital are put in one company that contracts with another company for the supply of human capital services and other inputs necessary for the production of maritime transport services. There is no employment relation connecting capital ownership and labor. The ship-owner cannot infl uence the use of the ship through fi at. A contract between two separate fi rms decides how the use of capital and labor shall be coordinated. Contract (market) has replaced the fi rm as a coordination mechanism.

In the extreme case, a ship-management company takes over all the activities outlined in Section 3.3 (Figure 4.4). Mitroussi (2003) describes it thus:

once a ship owner assigns activities, like crewing, technical and freight manage-ment, insurance, accounting, chartering, provisions, bunkering operations and sale and purchase of a vessel, in essence he or she gives up, together with the full management, control of his assets to third parties with no ownership rights in them. (p. 81)


However, in practice ship managers seldom go to this extreme. In another study Mitroussi (2004) shows that in the majority of cases the fi rms assign crewing, technical management, service, accounting and operations to independent managers. The extreme division of ownership and control indicates that the asset specifi city in the relationship between capital and labor is not large. The crew of a tanker, a bulk carrier or a container ship can serve equally effi ciently on vessels owned by diff erent fi rms. In line with Williamson (1996, Chapter 3), the high-powered incentives in an arm’s-length (market) relation can be used and the price for transport services is in most cases given. With a given price, cost minimization becomes important. A ship-management company has higher-powered incentives to minimize cost than an employee. Furthermore, as also noted by Williamson (1996, Chapter 3, p. 66), ‘markets can sometimes aggregate demands to advantage, and thereby realize economies of scale and scope’. The possibilities to reap economies of scale and scope are mentioned by Panayides and Cullinance (2002) as the main rationale for use of third-party management.

The ship-management companies are usually subsidiaries of larger ship-owners. The knowledge and the experience from operating vessels thereby remain within the company. Important customers of the ship-management fi rms are large multinational oil and manufacturing companies that have built up their own fl eet. The rationale behind owning a fl eet is lower costs of transport as well as the possibility to control the supply of sea transport. Furthermore, the cost of letting a ship-management company operate the vessel is often lower than the cost of developing a department within the existing company. It is also common that the ship-owner purchases man-agement services in order to learn how to operate a vessel (Nya Sjöfartens Bok, 2006; Branch, 1988).

4.2 A Wide Array of Contracts in Maritime Transport

The prevalence of the diversity of contracts in maritime transport cannot fully be explained by the types of asset specifi city enumerated in Williamson (1985), for example, site, physical asset and human asset specifi city. Vessels, in general, are not built to serve a specifi c shipper (customer), with the exception of vessels used in special shipping. Most of the vessels used in bulk shipping are, as pointed out by Pirrong (1993), not customer-specifi c in the same sense as site specifi city and asset specifi city. Furthermore, the use of third-party management indicates that human asset specifi city is not as important as in other industries. Instead time and space consid-erations give rise to what Masten et al. (1991) call ‘temporal specifi cities’ and Pirrong (1993) claims that this type of specifi city is common in bulk


shipping. Temporal specifi cities are, according to Pirrong, the explanation as to why some contracts are more commonly used at some freight markets than at others. Appropriable quasi-rents exist because ‘delays in shipment cause great harm on the shipper and the carrier’s next best cargo shipping opportunity is suffi ciently distant’ (Pirrong, 1993, p. 942). This temporal specifi city can be costly to handle in spot market transactions. Some type of safeguard represented by a long-term contract or other type of binding between shipper and carrier has to be imposed.

In markets where forward and time contracts are common a second type of specifi city can arise due to the time span of the contract, namely contractual specifi cities. These arise in the vacuum after the expiration of the contract when all other vessels and cargos are engaged in other freight contracts. During this time period, when all other shippers and carriers are tied up in other contracts, the previously contractually related carrier and shipper are in a bilateral monopoly situation. The transaction costs of entering into a new agreement then increase considerably and there are incentives for strategic behavior. The cost of contractual specifi cities can, however, be mitigated by an increasing number of contracts that expire repetitively. Both the shipper and the carrier then have a number of poten-tial new partners. Costs associated with forward contracting, such as cost of lower fl exibility and the cost of opportunistic behavior, increase with the time span of the contract, whereas contractual specifi cities lead to longer freight contracts and vertical integration (Pirrong, 1993).

In freight markets where specialized vessels are common, for example, for car transports and wooden products, the probability of long-term gaps between expiration dates of diff erent short- and medium-term contracts is high. From a transaction costs perspective the shipper can then gain by buying its own vessel, that is, be independent of ship-owners. Long-term contracts are yet another way of reducing the transaction cost.

Temporal and contractual specifi cities do determine the design of the freight contract. Spot contracts are used at markets where temporal spe-cifi cities lack importance, whereas long-term freight contracts and vertical integration exist on markets associated with signifi cant temporal specifi ci-ties caused by ‘i) market thinness, ii) the unavailability of supplies of the shipped commodity from non-fi rm specifi c resources, or iii) effi ciencies arising from the use of specialized tonnage’ (Pirrong, 1993, p. 951). In Table 4.1 the diff erent types of bulk markets are analyzed from a contrac-tual perspective.

Spot contracts should be used when there is no temporal specifi city apparent, that is, when there is a no in each of the three boxes following each commodity respectively. Accordingly spot contracts are used for freights for commodities such as grain, oil (post-1973), fertilizers and


scrap, transported with general vessels, in thick markets and without fi rm-specifi c supply. Various kinds of time charters are used when the market is characterized by fi rm-specifi c supplies, specialized vessels or thin market. In these markets both time specifi cities and contractual specifi cities are signifi cant.

Freight market characteristics can change over time; for example, the characteristics of the market for oil shipping changed considerably during the late 1970s, from forward contracts and time charter (MTC or VI) into spot contracts.12 This change displays the major role that temporal specifi cities play in contracting practices. Also, the change in contracting practices took place when fi rm specifi city became less important. Before 1973 most of the contracting with crude oil producers consisted of forward

Table 4.1 Contracting practices in bulk shipping

Commodity Firm-specifi c supplies

Specialized vessels

Thin market

Typical contract

Grain No No No SpotOil post-1973 No No No Spot Oil pre-1973 Yes No No MTC or VIThermal coal pre-1980

No No No Spot

Thermal coal post-1980

Yes Yes/No No LTCOA for large ships, MTC for others

Coking coal Yes Yes/No – LTCOA for large ships, MTC for others

Fertilizer No No No SpotScrap No No No SpotIron ore Yes Yes Yes LTCOA or VIGreat Lakes ore Yes Yes Yes LTCOA or VIWood chips Yes Yes Yes LTCOA or VILumber Yes Yes Yes LTCOA or VICement Yes Yes Yes LTCOA or VIBauxite Yes Yes Yes LTCOA or VILiquifi ed natural gas

Yes Yes Yes LTCOA or VI

Autos Yes Yes Yes LTCOA or VI

Note: LTCOA 5 long-term contracting, VI 5 vertical integration, Spot 5 spot chartering, and MTC 5 medium term chartering.

Source: Adopted from Pirrong (1993, p. 974).


and time charters. At this time the crude oil producers in the Middle East, West Africa, and Indonesia supplied the major oil refi ners – British Petroleum, Exxon, Gulf, Mobil, Royal Dutch Shell, Socal and Texaco – with crude oil on equity contracts or preference agreements. These market conditions resulted in temporal specifi cities. At the same time there was an excess supply of tankers which mitigated contractual specifi cities. During the 1970s political as well as economical changes took place in countries belonging to the Organization of Petroleum Exporting Countries (OPEC). These changes resulted in a drastic reduction in the use of forward con-tracts and time charters. The process, towards spot contracts, was sup-ported by the development of spot oil markets in the Netherlands, the US and the UK.

To summarize, the above reasoning demonstrates that fi rm specifi cities can create temporal specifi cities which in turn give rise to the use of forward contracts, time charters and vertical integration. The increased use of spot contracts increased the fl exibility, which lowered the transaction costs of spot contracting.

5. CONCLUDING REMARKS AND DISCUSSION

This chapter presents a contractual perspective of the fi rm that highlights the function of the fi rm as a common contracting partner to suppliers, customers, labor, capital and fi nanciers. The nature of contracts concluded is in our model dependent on the degree of mutual dependency. A high degree of mutual dependency requires safeguards that can be provided in the form of long-term contracts, an employment relation or joint owner-ship of assets in diff erent stages of a value added chain.

The maritime industry is analyzed from a contractual perspective with special focus on the link between the carrier and the shipper. We present a synthesis of diff erent contractual perspectives on the fi rm as a coordinating institution. The maritime industry is interesting due to the wide variation of contracts used, ranging from spot contracts, forward contracts and time charters to vertical integration. It is also interesting that in comparing dif-ferent freight markets there is a large variation going from one extreme to another. The tramp market is characterized by (almost) perfect competi-tion whereas the liner market is characterized by cartels.

In general there are three types of freight contracts, the spot contract, the time-voyager and the contract of aff reightment (COA). The choice of contract depends not only on the position of the commodities in the value added chain but also on the existence of temporal specifi cities. The spot contract is used on tramp markets mainly for raw material such as oil and


grain where no temporal specifi cities prevail. Commodities higher up in the value added chain are most often transported with forward contracts or time charters. In addition, these markets are characterized by temporal specifi cities due to political and economic changes in the contract practices over time, such as in the shipping of oil. This example highlights the fact that fi rm specifi cities can create temporal specifi cities which induce the use of forward contracts and time charters in an industry that due to its char-acteristics normally should apply spot contracts.

The existence of third-party ship management in the maritime industry is also an interesting phenomenon. It sheds light on complete separation of ownership and control. In this case the relation between physical and human capital is regulated by a contract. This in turn indicates that there is a relatively low degree of asset specifi city regarding capital and labor in the maritime sector. When third-party management is applied, the allocation of resources is replaced by a contract.

There is also a strong relation between commodity specialization and fi rm structure. For example, fi rms operating mainly on the spot market are usually smaller whereas fi rms on the liner and ferry markets usually are relatively large. The existence of temporal specifi cities therefore aff ects both the fi rm structure and the design of the freight contract.

NOTES

* Corresponding author. 1. However, it is a plain vanilla theory of markets. No attention is paid to diff ering con-

tractual aspects of markets (see Williamson, Chapter 2 in this volume). 2. The report has benefi ted from excellent comments and branch-specifi c knowledge sup-

plied by P.A. Sjöberger, Swedish Shipowners’ Association. 3. In many cases CEO, chairman and largest shareholder can be one and the same person. 4. Value added chains can be used to show the diff erent stages in shipping. For example the

Swedish shipping companies Broström AB and Wallenius AB use comprehensive freight contracts that include several steps in the supply chain. Instead of buying these serv-ices from other logistic fi rms the ship-owners develop them internally in the company (Johansson et al., 2006).

5. Williamson (1985, 1996 and Chapter 2 in this volume) uses the term ‘asset specifi city’ in his description of such a dependency.

6. Klein et al. (1978) make a distinction between use and user when defi ning an appropri-able quasi-rent. While the concept quasi-rent according to them refers to how much the value of an asset in current use exceeds the value of an asset in its best alternative use, appropriable quasi-rent denotes the diff erence in value in relation to what the next highest valuing user is prepared to pay for the service of the asset.

7. According to Williamson (1975, 1985, 1996) bounded rationality refers to the limited capacity of the human mind to conceive and evaluate all alternatives pertinent to a decision. Opportunistic behavior in turn means to give false or self-disbelieved prom-ises about the future or self-interest seeking with guile; to include calculated eff orts to mislead, deceive, obfuscate, and otherwise confuse.


8. According to Williamson (1996, p. 105) dedicated assets are ‘discrete investments in general purpose plant that are made at the behest of a particular customer’.

9. By hybrid one means ‘Long-term contractual relations that preserve autonomy but provide added transaction-specifi c safeguards, compared with the market’ (Williamson, 1996, p. 378).

10. Section 3 is based on Johansson et al. (2006). 11. See for example the Swedish companies Wallenius Wilhelmsen Line and SCA.12. The following discussion is based on Pirrong (1993).

REFERENCES

Branch, A.E. (1988), Economics of Shipping Practice and Management, 2nd edn, Chapman & Hall, London.

Coase, R.H. (1937), ‘The Nature of the Firm’, Economica N.S. 4:386–405, reprinted in O.E. Williamson and S. Winter (eds) (1991), The Nature of the Firm: Origins, Evolution, and Development, New York: Oxford University Press, pp. 18–33.

Fama, E.F. and Jensen, M.C. (1983), ‘Separation of Ownership and Control’, Journal of Law and Economics, 26:301–26.

Gorton, L, Ihre, R. and Sandevärn, A. (1989), Befraktning, edition 3–1, Liber Hermods, Läsprodukter AB, Halmstad.

Jensen, M. and Meckling, W. (1976), ‘Theory of the Firm: Managerial Behavior, Agency Costs, and Capital Structure’, Journal of Financial Economics, 3:305–60.

Johansson, B., Karlsson C., and Palmberg, J. (2006), Den svenska sjöfartsnäringens ekonomiska och geografi ska nätverk och kluster, Institutet för Näringslivsanalys, http://www.ihh.hj.se/doc/3707.

Klein, B., Crawford, R.A. and Alchian, A.A. (1978), ‘Vertical Integration, Appropriable Rents, and the Competitive Contracting Process’, Journal of Law and Economics, 21:297–326.

Magirou, E., Psaraftis, H.N. and Christodoulakis, M.N. (1992), ‘Quantitative Methods in Shipping: A Survey of Current Use and Future Trends’, Centre for Economic Research, Athens University of Economics and Business, Report no. E115.

Masten, S. (1988), ‘A Legal Basis for the Firm’, Journal of Law, Economics, and Organization, 4:181–98.

Masten, S., Meehan Jr., J. and Snyder, E. (1991), ‘The Cost of Organization’, Journal of Law, Economics, and Organization, 7:1–22.

Midore, R., Musso E. and Parola, F. (2005), ‘Maritime Liner Shipping and the Stevedoring Industry: Market Structure and Competition Strategies’, Maritime Policy and Management, 32:89–106.

Mitroussi, K. (2003), ‘Third Party Ship Management: The Case of Separation of Ownership and Management in the Shipping Context’, Maritime Policy and Management, 30:77–90.

Mitroussi, K. (2004), ‘The Ship Owners’ Stance on Third Party Ship Management: An Empirical Study’, Maritime Policy and Management, 31:31–45.

Nya Sjöfartens Bok, 2006 (2005), Svensk sjöfartstidningsförlag, Nr 23, 16 December.

Panayides, P.M. and Cullinane, K.P.B. (2002), ‘The Vertical Disintegration of Ship Management: Choice Criteria for Third Party Selection and Evaluation’, Maritime Policy and Management, 29:45–64.


Pirrong, S.C. (1993), ‘Contracting Practices in Bulk Shipping Markets: A Transaction Cost Explanation’, Journal of Law and Economics, 36(2):937–76.

Smith, A. (1776), The Wealth of Nations, reprinted with a foreword by A. Skinner (1976), Harmondsworth: Penguin Books.

Stopford, M. (1997), Maritime Economics, 2nd edn, Routledge Taylor and Francis Group, London.

Ståhl, I. (1976), ‘Ägande och makt i företagen – ett debattinlägg’, Ekonomisk Debatt, nr 1.

Veenstra, A.W. and De la Fosse, S. (2006), ‘Contributions to Maritime Economics – Zenon S. Zannetos, the Theory of Oil Tankship Rates’, Maritime Policy and Management, 33:61–73.

Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications, New York: Free Press.

Williamson, O.E. (1985), The Economic Institutions of Capitalism, New York: Free Press.

Williamson, O.E. (1988), ‘Corporate Finance and Corporate Governance’, Journal of Finance, 43:567–91.

Williamson, O.E. (1996), The Mechanisms of Governance, New York: Oxford University Press.

82

5. The use of managerial authority in the knowledge economyKirsten Foss

1. INTRODUCTION

This chapter contributes to the ongoing debate in diverse literatures, including management, sociology and economics, on the issue of the use of authority. The debate has re-surfaced with the focus on the emerging knowledge economy. In that debate a number of management academics and sociologists have argued that authority relations will strongly dimin-ish in importance or at least change signifi cantly in character.1 More spe-cifi cally, these writers have argued that the exercise of authority is, if not outright impossible, at least not effi cient and will be substituted by diff erent means of organizing the production of knowledge intensive goods and services. More discretion will be granted to knowledge workers and this will result in an eruption of the hierarchical structures of organizations.

Such predictions are a serious concern for those theories, notably trans-action cost theories (Coase, 1937; Williamson, 1985) and property rights theory (Hart and Moore, 1990; Hart, 1995, 1996), that place emphasis on authority relations as the mode of coordination that primarily character-izes fi rms. Thus, explanations of the existence and boundaries of fi rms that rely on authority are challenged. Part of the reason behind the prediction of the decline of the importance of authority relations is the assertion that fi rms will experience a tension between the exercise of authority by supe-riors based on their position in the hierarchy and the exercise of authority by knowledge workers based on their great expertise in certain areas. As superiors come to lack information about the tasks that are carried out by knowledge workers, and as knowledge workers gain bargaining power because of their control of crucial information, the economic importance of the authority of the position in a hierarchy will diminish.

However, this claim is based on a much too narrow conception of what it means to exercise the authority of the position. Orders are merely one way of exercising authority; superiors may, for example, also defi ne goals, set restrictions for employees regarding the use of fi rm resources, or implement

The use of managerial authority in the knowledge economy 83

various restrictions (N. Foss, 2001). If all such uses of authority decline in importance as means of coordinating activities in fi rms, we should be able to fi nd an economic rationale for this. However, in order to do so we must fi rst understand the economic rationale for the use of authority as a means of coordinating economic activities in order to discover whether the condi-tions for the use of such authority have changed. In this chapter I provide effi ciency explanations for the use of diff erent forms of authority in fi rms and argue that we may expect the use of diff erent types of authority of the offi ce in conjunction with experts who may possess authority based on their expertise. The questions to be addressed in this chapter are the following:

Can the use of authority co-exist with the delegation of discretion ●

within a contract?If so, what factors can change the use of authority relative to the use ●

of market contracts as a means of organizing productive activities?Is it likely that a move from the capital intensive to the knowledge ●

intensive economy will change the relative use of authority?

2. WHAT IS AUTHORITY?

The concept of authority is closely linked to the sociological literature on bureaucracy (for example, Weber, 1946, 1947) and organization and behavioral theories, usually drawing on sociology and psychology, present a number of interpretations of authority (Simon, 1951; Blau, 1956; Thompson, 1956; Grandori, 2001). It would be a hopeless task to present a full review and critical evaluation of the multitude of defi nitions and ideas regarding the concept of authority. For the purpose of this chapter the concept of authority as defi ned by Barnard (1938) and Simon (1951, 1991) serves as a starting point for the development of the terminology adopted here. Both Simon and Barnard employ the concept of ‘willingness to obey’ (Simon, 1951, p. 126).

Simon (1951) defi nes authority as obtaining when a ‘boss’ is permitted by a ‘worker’ to select actions, A0 , A, where A is the set of the worker’s possible behaviors. More or less authority is then defi ned as making the set A0 larger or smaller. Simon develops a model (specifi cally, a multi-stage game in the context of an incomplete contract with ex post governance), where, in the fi rst period, the prospective worker decides whether to accept employment or not. In this period, none of the parties know which actions will be optimal, given circumstances. In the next period, the relevant cir-cumstances as well as the costs and benefi ts associated with the various possible tasks are revealed to the boss. The boss then directs the worker to


a task, which – for the latter to accept it – must lie within his or her ‘zone of acceptance’.

An important feature of authority thus is that the authority of a superior is constrained by the subordinate’s acceptance of it. ‘A subordinate may be said to accept authority’, Simon (1951, p. 22) explains, ‘. . . whenever he permits his behavior to be guided by a decision reached by another, irrespective of his own judgment as to the merits of that decision.’ Simon’s use of the notion of authority is akin to that of Coase (1937), who defi nes an authority relation as ‘one whereby the factor, for a certain remunera-tion (which may be fi xed or fl uctuating), agrees to obey the directions of an entrepreneur within certain limits. The essence of the contract is that it should only state the limits to the powers of the entrepreneur. Within these limits, he can therefore direct the other factors of production’ (idem, p. 242). Coase (1937), Bernard (1938), and Simon (1951) all linked the notion of authority to entering into an employment contract. Others, however, have argued that an employment relation is not a necessary condition for the use of authority (Alchian and Demsetz, 1972; Cheung, 1983).

3. AUTHORITY IN FIRMS AND MARKETS

Many contributions to the understanding of authority share the – some-times implicit – assumption that if complete contingent markets existed, price coordination would suffi ce. This was indeed Coase’s point of depar-ture in ‘The Nature of the Firm’ where he posed the question, ‘Why do fi rms exist?’ The reason for the existence of fi rms is that there are costs of using the price mechanism and that ‘[t]he most obvious cost of “organiz-ing” production through the price mechanism is that of discovering what the relevant prices are’ (Coase, 1937, p. 21). When there are high transac-tion costs fi rms pose an alternative to markets as a means of achieving coordination, because the fi rm allows for the use of authority as a means of allocating resources comparable to bargaining and contracting in markets.

In parts of the economic literature, authority has been linked to the employment contract (Coase, 1937; Simon, 1951; Williamson, 1985) and fi rm coordination is seen as diff erent from market coordination due to the use of employment contracts. According to Coase an employment contract is preferred: ‘owing to the diffi culty of forecasting, the longer the period of the contract is for the supply of the commodity or service, the less possible, and indeed, the less desirable it is for the person purchasing to specify what the other contracting party is expected to do’ (Coase, 1937, p. 21). Thus, the employment contract is a long-term contract for an unspecifi ed labor


service. Managed direction of resources substitute for price direction of resources when parties to transactions realize that contingencies of dif-ferent sorts may in an unpredictable manner disrupt the choice of action or the timing and sequencing of interdependent activities (see Wernerfelt, 1997).2

However, there are also costs of using authority (that is, managed direction) which defi ne the effi cient limits for the use of authority. In an economy characterized by uncertainty with respect to the actions that need to be taken, the main costs of using authority stem from the ‘increas-ing opportunity costs due to the failure of entrepreneurs to make the best use of the factor of production’ (Coase, 1937, p. 23). According to Coase (1937),

the costs of losses through mistakes will increase with an increase in the spatial distribution of transactions organized, in the dissimilarity of the transactions, and in the probability of changes in the relevant prices. As more transactions are organized by an entrepreneur, it would appear that the transactions would tend to be either diff erent in kind or in diff erent places. (p. 25)3

Managers, in other words, have limited capacity to ‘discover the relevant prices’ and this increases mistakes as more dissimilar transactions are organized in a fi rm.

Simon (1951) also links fi rms and authority to the use of the employ-ment contract. Like Coase, he perceives of the employment contract as a contract for unspecifi ed labor services. Uncertainty also constitutes the essence in the explanation provided by Simon (1951) of the use of employ-ment contracts and authority as a means of coordination. The employment contract grants the right to the employer to postpone the decision about what services to demand from the employee until he obtains the relevant information on which to base the decision. However, limits to the use of authority are not due to managerial mistakes, as in Coase, but to the diff er-ences in costs of fulfi lling the participation constraints of an employee and an independent contractor. Therefore the use of authority is effi cient only when contractual adaptation is critical to one party while the other party is nearly indiff erent as to the actions he carries out.

The Coase–Simon view compares adaptations to uncertainty by means of authority and by means of re-contracting among independent agents. Authority diff ers from bargaining power in that authority is voluntarily granted by one party to another on the basis of effi ciency considerations. This clear-cut distinction between bargaining power and authority is not present in later developments of the concept of authority within property rights theory as defi ned by, for example, Hart (1991, 1995). This theory is based on the assumption that the main problem of coordination is that of


providing the right incentives for investing in situations where there are high costs of describing the relevant investments to be made and the output to be delivered. This kind of transaction cost makes it impossible for the contracting parties to use a third party (courts) as a means of enforcing original promises, thus creating a situation where the parties expect re-contracting over the created surplus. Firms represent an effi cient choice of enforcement of contracts because of the bargaining power they posses in the re-contracting situation. For example, Hart defi nes the fi rm as the physical assets over which a legitimate owner has formal residual user rights. Having residual user rights over assets provides the fi rm with a bar-gaining power that is diff erent from that which characterizes transactions that take place between individual owners of assets. The limits to the use of authority depend on the parties’ incentives to invest in non-contractible assets when ownership of assets is centralized as in fi rms and when it is dispersed as in markets.

The property rights theory shares its strong focus on incentive alignment as the central coordination issue with agency theory. Both assume that the best uses of resources are already known and that the problem of coor-dination arises due to asymmetric information on some relevant aspects relating to the contractual execution. In principal–agency literature asym-metric information introduces the need for diff erent contractual designs, of which some may include ‘authority’. In the work of Alchian and Demsetz (1972) teamwork creates one of the situations in which the use of authority serves the purpose of improving the effi cient use of labor inputs for given ends. With teamwork it is costly to separate the contribution of each par-ticipant, creating incentives for moral hazard. The solution to such a team problem is to set up an organization (but not necessary a fi rm) which will economize on metering costs so as to better allocate rewards in accord with the eff ort delivered. A monitor specialized in metering eff ort is granted by the members of the team the authority (or right) to alter membership of the team in order to improve incentives for work eff ort.4 Thus, specializa-tion advantages in monitoring and more advanced incentive schemes than those that can be devised for re-contracting between independent indi-viduals (with defi nite time horizons) help overcome incentive problems. Authority in this conception is based on a granted right (as in Coase and Simon) but it is not necessarily related to the use of employment contracts, nor is it granted because of the need to adapt to unforeseen changes.

These ideas are similar to those of Cheung (1983), who emphasizes that there is a wide spectrum of contracts, ranging from pure spot market contracts to order contracts, from piece rate contracts to employment con-tracts. All contracts, with the sole exception of pure spot market contracts, include some instructions or restrictions or are accompanied by orders.


As Coase (1937) did, Cheung (1983) stresses the cost of discovering the relevant prices as the reason for this wide spectrum of contracts. However, the reasons he provides as to why prices are substituted by other means of coordination diff er somewhat from those of Coase (1937), since he stresses the cost of measuring the relevant attributes of goods and services as the main reason for lack of complete contingent contracts. According to Cheung, such measurement costs are likely to be high ‘[i]f the activities performed by an input owner change frequently [and] if these activities vary greatly’ (Cheung, 1970, p. 7). Moreover, it may simply be too costly to separate the contribution of each party to the production of a consumer good. In other words, measurement costs may cause team problems. In contractual relations characterized by high measurement costs, ‘it tends to be more economical to forgo any direct measurement of these activities and substitute another measurement to serve as a proxy’ (ibid.). For example, instead of contracting for the use of a secretary to type a certain letter at a certain place at a certain time, one may contract for the unspecifi ed labor service of the secretary for a given period of time and a given pay per time period. However, when a price for each job to be done by the secretary cannot be specifi ed, the contract needs to be supplemented by directions/instructions (or orders). The relation between remuneration of the input and the valued attributes of the output may be more or less imprecise and therefore require more or less instructions.5 We should, according to Cheung, expect an extensive use of orders where measurement costs makes it effi cient to use fl at wages rather than payments that more fully refl ect the contribution of the subordinate to the quality of the fi nal consumer good.

The principal–agency literature on fi rms (Cheung, 1970; Alchian and Demsetz, 1972) emphasizes that orders and restrictions are not unique to the employment relation and the exercise of authority not confi ned to fi rms. Nor does it view diff erences in bargaining power as discriminating fi rms (or employment contracts) from markets. This point of view is most clearly expressed in Alchian and Demsetz (1972), who claim that ‘[i]t is common to see the fi rm characterized by the power to settle issues by fi at, by authority, or by disciplinary actions superior to that available in the conventional market. This is a delusion’ (p. 72). An employer has, in their opinion, no diff erent means at his disposal for punishing disobedience than individual contractors have. According to this view every single instance of the exercise of authority is based either on implicit agreements or on the bargaining power of the parties in the relation. The conclusion one may derive from Cheung (1970) and Alchian and Demsetz (1972) is that author-ity exists in the same form and to the same degree in markets and in fi rms.

This view is contrasted by those who emphasize the diff erences in the legal conditions that support the private exercise of authority (as


enforcement of contracts) within fi rms compared with the use of market contracts (Williamson, 1985; Masten, 1991; Vandenberghe and Siegers, 2000). First, employment law requires that orders or instructions are carried out, restrictions implemented by an employer are respected and disputes about their merits and legitimacy are settled after the completion of the order. Williamson (1985), for example, describes fi rms as their own ultimate court of appeal. Second, the rights of the employer to monitor and sanction actions diff er in employment and market contracts (Masten 1991). The extended right to monitor employees complements the functioning of authority as a means of solving disputes. For example, an employer may have better access to knowledge of relation-specifi c non-contractible invest-ments than courts. Thus, although restrictions, directions and orders can be found in market contracts, they diff er from employment contracts in the legal frame that supports the contracts. Market contracts do not provide one of the parties with the formal (and legally supported) right to make decisions without the consent of the other party. Adaptations can be made if there is no confl ict of interest between the parties or a verifi able mecha-nism (such as elevator clauses) of contractual adjustment is agreed on in the contract.6 That is, measurement costs explain why contracts contain more stipulations than just the price and the item. However, only uncertainty or partly unpredictable volatility makes it necessary to use managerial discre-tion to make ex post adaptation of these instructions and stipulations.

3.1 Formal and Real Authority

The notion of authority as supported by fi rm governance structure does not imply that fi rms (or more precisely managers) are always able to exer-cise the authority they are legally entitled to exercise. That is, employees may exercise discretion and avoid the restrictions, instructions and orders issued by managers. This introduces a distinction between formal and real authority (Aghion and Tirole, 1997, p. 1). Whenever there is a legally rec-ognized employment relation we fi nd formal authority in the sense of an attribute that is attached to the role of an employer (for example, Weber uses the notion of bureaucratic authority or authority of the offi ce). The employment relation supports the use of authority of the position (formal authority) but formal authority does not necessarily convey real authority which is ‘an eff ective control over decisions, on its holder’ (Aghion and Tirole, 1997, p. 1).7 That is, employees may exercise discretion within the authority relationship.

Discretion can broadly be defi ned as the ability of an agent to control or consume resources over which he/she does not have formal ownership. Discretion includes instances where an agent behaves morally hazardously


or exerts infl uence indirectly or where peers in a group voluntarily agree to have their points of view represented by a group member or to have confl icts resolved by one member of the group.8 The exercise of discretion by employ-ees does not fully preclude their recognition of the formal authority granted to an employer through employment law. For example, an employee may recognize the legitimacy of the formal authority relation with a superior and be willing to accept the limits within which the superior can direct the actions of the subordinate, while not fulfi lling the agreement when given specifi c directions in areas where he is not subject to eff ective control. Also, the employee may recognize the legitimacy of the formal authority relation and the specifi c ways in which the authority is exercised, while spending resources on altering the guidance to which he is subject to by infl uencing superiors to change the way in which the authority is carried out.

Employees’ ability to exercise discretion depends on the employer’s costs of monitoring and on their bargaining power. For example, a knowl-edge worker’s technical competences and expertise may be an important constraint on the ability of the employer to exercise his formal authority – this is the assertion that seems to underlie the prediction that knowledge workers are less likely to be subject to the authority of offi ce and more likely to possess discretion or authority themselves due to their specifi c qualifi cations.

Authority and discretion may be delegated to one employee or a group of employees. Delegated (or formal) discretion can be defi ned as instances where an agent is delegated the rights by a superior to allocate resources (including his own labor services) to those ends he desires without the formal approval of an owner or employer.9 Delegated authority is to be distinguished from delegated discretion by the fact that the discretion does not include formal rights to direct the actions of other members of the fi rm. Both delegated authority and delegated discretion are limited by the formal rights of a superior to overrule the decisions made by a subordinate (Baker et al., 1999) as well as by the real discretion exercised by lower-level employees.

Delegation of discretion and authority is a means of overcoming the ineffi ciencies that stem from centralized decision making (Jensen and Meckling, 1992). In the Coase–Simon–Cheung view it is implicitly assumed that the superior receives the relevant information and has the relevant knowledge to make use of the information by ordering the sub-ordinate to carry out a specifi ed action. With such a narrow conception of authority there is no room for the use of dispersed knowledge or informa-tion that rests with employees within the employment relation. Thus, the use of centralized decision making by an authority becomes ineffi cient in a knowledge economy where employees or subordinates (at least at certain


points in time) have the relevant information or the relevant knowledge that makes them superior in using the information.

However, Simon (1991, p. 31) himself pointed out four decades after his paper on authority, ‘[a]uthority in organizations is not used exclusively, or even mainly, to command specifi c actions.’ Instead, he explains, it is a command that takes the form of a result to be produced, a principle to be applied, or goal constraints, so that ‘[o]nly the end goal has been sup-plied by the command, and not the method of reaching it.’10 The concept of authority introduced by Simon (1991) sets focus on the many ways in which contractual relations can be altered ex post contracting. Moreover, the defi nition allows for the delegation of discretion to subordinates with respect to their choice of actions. Authority on the part of the superior may then be interpreted as the right to unilaterally change the degree of delegation ex post contract agreement and to veto decisions made by the subordinate (see also Aghion and Tirole, 1997; Baker et al., 1999).

It does not alter the economic rationale for formal authority if it is broadened from the narrow focus on orders to encompass situations where the superior ex post contracting vetoes decisions or implements restrictions that prevent the subordinate from picking the actions most preferred by him. In accordance with Coase and Simon, formal authority still serves the overall purpose of achieving coordination of productive activities in a setting where it is economically effi cient for the parties to adapt the con-tractual relation to unpredictable changes in contractual circumstances. Adaptation must be costly to obtain through re-negotiations or by means of contingent plans (Coase, 1937).

Given the above, I put forward the following defi nition of authority. Authority is a formal right granted to a superior to use managerial judg-ment to unilaterally decide on changes or maintenance of aspects of the term of contract ex post contracting irrespective of the contracting party’s judgment as to the merits of that decision. In employment relations the formal right to exercise authority is supported by labor law.

4. CAUSES OF CHANGE IN THE RELATIVE USE OF AUTHORITY

4.1 The Use of Authority from an Incentive Perspective

The tension between real and formal authority is center stage in some of the writings on the diminishing importance of authority compared with, for example, market contracts (or hybrids such as collaborative arrangements supported by, for example, norms (Grandori, 2001)). Two diff erent types of


arguments support this idea. First, a move toward the knowledge economy may be expected to increase the relative importance of investments in knowledge assets compared with investments in capital. With this change, employers’ costs of monitoring and bargaining with knowledge workers over the actions to be chosen may increase. In turn, this leads to increased costs from moral hazard (Jensen and Meckling, 1992) and from bargaining in all transactions dealing with knowledge workers. Also, investments in knowledge assets and allocation of bargaining power change endogenously (Hart, 1995). That is, in order to create effi cient incentives for knowledge workers to invest in knowledge assets we should expect a reallocation of the real authority that follows from ownership of co-specialized physical assets from managers to knowledge workers.

If we hold everything but the discretion and costs of bargaining with employees constant we should expect the benefi t function from the use of formal authority to move downward, causing a relative decrease in the use of formal authority. However, there are also factors that can off set this move. For example, increased investments in knowledge assets may create more assets specifi city, which in turn raises costs of market contracting and introduces a need for private courts that can handle the specifi c types of confl icts that arise in incomplete contract situations (Williamson, 1985). Thus increasing knowledge workers’ discretion (or authority) does not nec-essarily imply more market transactions. However, the formal authority of the employer may to a lesser extent be accompanied by real authority.

A second type of argument focuses on how information is diff erently dispersed in the knowledge economy compared to the capital intensive one. Grandori (2001), for example, has forcefully stated that changes in the distribution of information may cause ‘authority [as a centralized decision-making system] to fail in all its forms’ (p. 257). Along with the diff erently dispersed knowledge, the choice set of actions available to knowledge workers may become much greater than that available to workers in the capital intensive economy. In such a situation it may be too costly for a central manager to be informed about the entire choice sets of an employee. Knowledge workers then become better able to select actions that create joint value than the employer. However, the employer may interpret the employees’ choices as morally hazardous when they do not benefi t the employer and he may not allow these choices although they maximize the joint satisfaction function. If as Simon (1951) argues the employee cannot make the employer commit to choose actions that maximize their joint satisfaction function, the use of authority becomes ineffi cient. For a given level of uncertainty (variance) employees will demand a higher compensa-tion (compared with the compensation they get from market transactions) in order to meet their participation constraints.11 The increased costs of


meeting employee participation constraints raise the costs of using author-ity relative to market contracting. Thus, the diminished use of authority in a knowledge economy should be attributed to an increased confl ict of interests leaving little scope for the use of authority that satisfi es the par-ticipation constraints of both the employer and the employee. A limited scope within which the participation constraints of both employee and employer are fulfi lled may also arise if changes in work content produce a wider gap between the preferences of the knowledge workers and those of their employers (with no changes in the size of the choice set). For example, it may well be the case that knowledge workers to a greater extent prefer to work on projects that add to their general (non-relation-specifi c) stock of human capital, whereas an employer to an increasing extent wants employ-ees to work on projects that mainly build relation-specifi c experience.

Two mechanisms can ease the problem. One is to introduce diff erenti-ated pay for activities in accordance with the merits to the employer and the costs to the employee. It has in fact been argued that in the knowledge economy we see an increasing use of high-powered incentives within fi rms. Of course this re-introduces the issue raised by Coase (1937), Cheung (1983) and others of the costs of discovering the relevant prices. Another mechanism is for the fi rm to delegate some discretion to knowledge workers and build a reputation for allowing the employee to select actions that maximize the joint satisfaction function (Aghion and Tirole, 1997).12 The reputation has to be robust to evolving changes in the pay-off to the employer and employee of emerging actions (see, for example, Kreps, 1990). The recognition of fi rms as legal entities and the legal protection of corporate identity are factors that ease the use of reputation mechanisms as a means of private enforcement of implicit promises within fi rms as com-pared with across spot markets. Thus, increased distribution of knowledge and the accompanying need for delegation need not result in a relative diminished use of formal authority.

However, as delegation of discretion introduces costs in the form of moral hazard, there will have to be limits to the delegation that takes place within fi rms. The optimal delegation of discretion is, according to Jensen and Meckling (1992), one that balances ‘the costs of bad decisions owing to poor information and those owing to inconsistent objectives’ (p. 264). In order to constrain moral hazard, restrictions are often used in employment relations where assets have many diff erent uses and where only a subset of these uses optimize the joint satisfaction function. This conclusion is in line with the work of Holmstrom and Milgrom (1994), Barzel (1989) and Holmstrom (1999), who argue that employers use restrictions in order to avoid costs of morally hazardous behavior when incentive payments are too costly to implement. Thus, the employer may replace direct monitoring


of work eff ort by supervision of a set of constraints that the employer has imposed on the employee with respect to the use of the labor services and capital assets (Barzel, 1989). However, since contracting takes place in a dynamic setting in which new opportunities for value creation and for moral hazard arise, the constraints and restrictions may need to be changed over time. The employment contract ensures that the employer has the formal rights to make these changes unilaterally, thereby saving contracting costs. Thus, an increased need for delegation need not imply a diminished role for authority.

To sum up: writers on the knowledge economy have emphasized the importance of increasing misalignment of preferences between employer and employee or the relative increase in employees’ ability to discover value-creating actions which in turn should diminish the relative importance of the use of authority. However, there is no reason to assume that there is a diminished need for the role of formal authority as supported by a fi rm governance structure. It may still be effi cient to use authority supported by fi rm governance to solve confl icts of interest in situations of incomplete contracting (as argued by Williamson, 1985) and to support the building of a credible reputation for delegating discretion. Moreover, delegation of discretion within an employment contract allows for diff erent types of monitoring compared with market contracts, making it easier to counter moral hazard within the fi rm governance structure. Finally, the fi rm gov-ernance structure allows for a fl exible adaptation of constraints to delega-tion as employers have the right to make these decisions unilaterally.

The above discussion has centered on how incentive issues infl uence the relative use of authority in the knowledge economy holding the nature of the coordination problem constant. However, it is also possible that the nature of the coordination problem changes with a move from a physical capital to a knowledge intensive economy. In the following, I make use of an example of a coordination problem that arises between a marketing and a product development function in order to illustrate how changes in the nature of the coordination problem infl uence the relative costs and benefi ts of the use of authority under diff erent conditions of dispersion of information and knowledge between an employer and two employees. For the moment I leave aside the discussion of the incentive issues, thus taking common goals as the standard assumption throughout the example.

4.2 The Use of Authority from a Production Coordination Perspective

The coordination problem consists in carrying out a product develop-ment project. Two employees (in marketing and product development respectively) and a manager are engaged in the project. The choice set for


the marketing employee contains two diff erent product concepts and the choice set for the product developer contains two diff erent technical solu-tions. The coordination problem is that of selecting the concept and the technological solution that under the prevailing contingencies generate the greatest revenue to the fi rm (which in this example is the choice that maximizes joint value). The contingencies facing the fi rm consist of the values (which enter as parameters in the revenue function) for the state of customer preferences and for the state of technological knowledge. Both customer preferences and technological knowledge can change over time.

The way in which the coordination problem can be solved depends on the nature of interdependencies and on the distribution of information and knowledge between the employer and the two employees. Information refers to information about a realized state and the solutions available, while knowledge refers to the ability to specify the revenue function and solve for the optimal solution. Either the superior or the two employees in the fi rm may posses information regarding the choice sets and the kind of states that have emerged. Moreover, either the superior or the subordinate or both may have the knowledge needed to select the optimal combination of product concept and technical solution, given the information available on customer preferences and technological knowledge.

When the informational interdependencies between the choice of product concept and technical solution are complex, coordination requires that the decision maker has information about the contingencies and the solutions facing both design and product development. The coordination problem is characterized by decisiveness when decisions on product concept and tech-nical solution can be made sequentially by diff erent decision makers. For example, customer preferences may be decisive for the choice of product concept and for the choice of technique. This implies that the marketing employee (who selects product concepts) can make a decision without information about the state of technological knowledge or technical solu-tions. Moreover, he only needs to communicate his choice of concept to the employee in product development, who selects the technical solution on the basis of his investigation of the state of technological knowledge. Finally, the coordination problem can be characterized by complete independ-ence, in which case the marketing employee only needs information about the state of consumer preferences and product concepts and the product development employee only needs information on the state of technologi-cal knowledge and technical solutions.

Centralized authority in a setting of complex interdependenciesThe employer is the only one who can make the relevant decision if he is the one who possesses all relevant information of states and solutions


available and the knowledge needed to use that information. Some ex ante communication of the information possessed by the employer will be necessary though, as employees will be unwilling to accept an employ-ment contract unless it specifi es the nature and limits of the choices that the employer can make (Simon, 1951). However, compared with the use of market contracting, which requires a more detailed specifi cation of actions and instructions, the use of order in employment contracts may save some communication costs (Demsetz, 1995). If contingencies or the choice set of actions change, the benefi ts from the use of formal authority supported by fi rm governance structure to make adaptations increase, as more transac-tion costs are saved compared with carrying out the adaptation by means of spot market transactions (Coase, 1937).13

The arguments presented above indicate at least three important vari-ables that may infl uence the relative use of centralized decision making (assuming there are no confl icts of interest). First, the costs of complet-ing contracts over markets in the knowledge economy could have been reduced. For example, the introduction of IT technology in the knowledge economy and the embodiment of information needed for coordination, as well as the development of interface and measurement standards, are factors that reduce costs of re-contracting. However, these factors may infl uence fi rm internal costs of using centralized decision making to the same extent. Second, uncertainty may have been reduced. In the example above, this would be the case if the states of consumer preferences or tech-nological knowledge did not change or if no new product concepts and solutions could be identifi ed. However, a reduction in the level of uncer-tainty is contrary to what most writers on the knowledge economy assume. Finally, the costs to the central manager of obtaining relevant information and of using this may have increased. The latter falls in line with the argu-ments presented in much of the literature on the use of authority in the knowledge economy, where many writers argue that an increased disper-sion of knowledge makes it increasingly diffi cult for the holder of formal (and centralized) authority to reach effi cient decisions (see, for example, Minkler, 1993; Cowen and Parker, 1997; Hodgson, 1998; Radner, 2000). For example, an increase in specialization in production and knowledge in the knowledge economy results in more diverse types of transaction (in terms of choice sets, contingencies and sources of interdependencies) and this increases the costs to managers of ‘discovering the right prices’, result-ing in a reduction of the effi cient size of fi rms and an increase in the relative use of market transactions.

Part of the confusion about the status of authority in a knowledge economy seems to arise because the use of authority is confused with highly centralized decision making. Indeed, the costs that arise due to the limited


mental capacity of managers can be reduced if some discretion and author-ity is delegated to lower-level employees. Delegation has been mentioned in the organization literature as a means of improving decision making under uncertainty (Miller, 1992), economizing on principals’ opportunity costs (Salanié, 1997) and avoiding decision delays under circumstances of volatility and uncertainty (Thompson, 1956; Burns and Stalker, 1961; Mintzberg, 1983). The underlying idea is that delegation of discretion provides an effi cient use of distributed knowledge in fi rms (Jensen and Meckling, 1992) that is costly to communicate to a central decision maker (Casson, 1994).14 Uncertainty, unpredictable volatility and some level of distributed knowledge and information create the conditions under which there is an economic rationale for the coexistence of authority and del-egation of discretion within hierarchies. This conclusion is based on the assumption that the coordination problem is at least partly decomposable (Simon, 1962). One such situation arises if the coordination problem is characterized by decisiveness.

Delegation in a setting of decisiveness, dispersed information and centrally or dispersed knowledgeWhen the coordination problem is characterized by decisiveness it is possible to achieve the optimal solution with delegation of discretion to employees. The employer can delegate decisions to the employee who has the relevant information about states and solutions and the knowledge needed to make the choice. When marketing information is decisive for the coordination problem the marketing employee can make the deci-sion and communicate it to the employer, having them select the optimal technique, or they can communicate their decision directly to the product development employee who has obtained information about the state of technology. The choice between centralized decision making and decen-tralized corporation (through markets or within the hierarchy) depends on the trade-off between benefi ts from specialization in decision making and costs of communicating to the employer the states and solutions that have been observed by marketing and product development.

In cases where the design problem is not characterized by natural decisiveness, it may sometimes be effi cient to have the employer impose decisiveness on problems by dispensing with the communication of the decision premises. As an example, the employer can choose to take cus-tomer preferences or technological knowledge as given and make that the ‘normal state’. Decisions will only be made in a consultative manner when the employer or the employee in a marketing department discovers an unusual state. In all other situations they will be made in a sequen-tial manner. An even stricter way of imposing decisiveness is to restrict


employees from all examination of states. In the latter case decisions are taken in a routine manner and the only role for a central authority is to monitor the adherence to and feasibility of this restriction.

Restrictions on the delegation of discretion may be needed as long as there is some level of interdependence. Costs from delegation of discre-tion arise when knowledge workers do not possess all relevant knowledge. Employees’ exercise of discretion can produce spillover eff ects (that is, ‘externalities’) due to unintended consequences of the actions taken. These harmful spillover eff ects include coordination failures, such as schedul-ing problems, duplicative eff orts (for example, of information gathering, R&D), cannibalization of product markets and other instances of decen-tralized actions being inconsistent with the fi rm’s overall aims, and so on.

The use of authority to restrict harmful consequences is effi cient if the employer is better able to form a judgment regarding what types of actions are appropriate. The employer defi nes constraints that only allow the knowledge workers to choose among actions he deems appropriate (Armstrong, 1994). In this way the employer prevents the knowledge worker choosing actions that he knows or believes to be infeasible. If the employer does not know where to set the restrictions ex ante to contracting, he may instead overrule or veto decisions made by the employee (as in, for example, Aghion and Tirole, 1997). That is, even with perfect alignment of incentives between employer and knowledge there is a role for employers as monitors and enforcers of restrictions. The use of restrictions and veto brings attention to the function of authority as a means of constraining ‘the method[s] of reaching’ an end goal, in Simon’s (1991) terminology. Under conditions of uncertainty where the choice set of the knowledge worker and interdependencies in decision change over time, constraints will have to be adjusted. Thus, the role of authority in a setting of distributed knowl-edge and uncertainty may well be that of unilaterally altering constraints on decisions made by lower-level knowledge workers and adjusting criteria for accepting or rejecting decisions made by such knowledge workers.15

The way in which delegation of discretion can ease the constraints on the use of authority in settings characterized by disperse information and knowledge move the focus from the discussion of authority versus market contracts to a discussion of centralized versus decentralized decisions within the employment relation. From the above it is clear that the role of the employer as one who solves coordination problems decreases as one moves from coordination problems characterized by interdependency to those characterized by (imposition of) decisiveness. Changes in production techniques can cause a shift in the nature of interdependencies. Moreover, it may have become more attractive to impose decisiveness on coordina-tion problems with the move to the knowledge economy. This is the case


if the costs of investigating states increase drastically compared with the benefi t of reaching the optimal solution. However, a move away from the use of centralized authority toward more decentralized decisions may also be a consequence of a relative change in the importance of knowledge workers’ information about actions in the choice set and contingencies compared with employers’ knowledge of interdependencies. That is, some interdependencies are ignored or suppressed at the expense of achiev-ing the optimal fi t between subsets of the solution. The movement in the knowledge economy toward the design of modular products is an example of the latter. Firms imposed decisiveness on problem solving, dispensing with some interdependencies in the product design, and the interfaces that are specifi ed ex post detailed product development constitute the natural state of the environment which is to be taken for granted in the choice of the specifi c design solutions.

5. CONCLUSION: IS THE USE OF AUTHORITY DIMINISHING IN THE KNOWLEDGE ECONOMY?

The notion of authority is an important one in economics. In particular, it underlies important contemporary theories of economic organization (Williamson, 1985; Hart, 1995). Given this, it is surprising that so few fun-damental discussions of the notion are to be found in the literature. This is problematic, because the notion of authority has recently been subject to much discussion in neighboring fi elds, such as sociology and business administration. Economists have had diffi culties entering this discussion because of their relatively crude conception of authority, in which authority is seen as synonymous with either bargaining power or the use of ordered direction by a highly informed employer (Simon, 1951). However, the theory of economic organization is in no way necessarily limited to those notions of authority. Thus, the theory suggests other possible, yet comple-mentary, understandings of authority, such as the unilateral right to set and change constraints on the activities of subordinates, overrule decisions made by subordinates, and implement and change reward systems. Seen in the light of this latter understanding of authority, it is not so apparent that authority will strongly decline in importance in the emerging knowledge economy, as asserted by a number of writers. Although knowledge workers may have more bargaining power and be more knowledgeable about the activities they carry out than ‘traditional’ industrial workers, they too will be subject to authority, as long as productive activities are characterized by uncertainty and measurement costs which make complete contracting


prohibitively costly. It may be conjectured that under these circumstances delegation will be more widespread. However, because of diff ering prefer-ences, asymmetric information and externalities in decision making, such delegation is likely to be circumscribed.

The literature on authority points to several factors that can diminish the relative use of authority in the knowledge economy:

changes in the distribution or importance of information and knowl- ●

edge held by employees and employers respectivelyless need for adaptation – more interface standards, more modular ●

products, and so onincreased complexity and diversity in transactions ●

increased confl ict in preferences over actions between employer and ●

employeemore bargaining power to employees ●

less asset specifi city ●

greater importance of knowledge workers’ investments in knowledge ●

compared with employers.

In order to fully answer the question of whether or not authority is dimin-ishing in importance, one must empirically investigate how changing from the capital intensive to the knowledge intensive economy aff ects all of these variables.

NOTES

1. Some representative sources are Boisot (1998), Foss (2001), Ghoshal et al. (1995), Grandori (2001), Harrison and Leitch (2000), Hodgson (1998), Liebeskind et al. (1995), Matusik and Hill (1998), Minkler (1993), Vandenberghe and Siegers (2000), and Zucker (1991).

2. For example, the employer can be ignorant about the best use of diff erent labor services as in the case of innovative activities (N. Foss, 2001). Also, in cases of rather complex production or product innovations, it may be desirable to conduct a number of control-led experiments before one decides on what labor service is required for a certain task (K. Foss, 2001)

3. Richardson (1972) has a very similar argument for the boundaries of fi rms 4. Agency theory, a body of literature to which Alchian and Demsetz (1972) belongs,

ascribes all contracting costs to the costs of observing variables. Monitoring denotes ‘measur[ing] output performance, apportioning rewards, observing the input behavior of inputs as means of detecting or estimating their marginal productivity and giving assignments or instruction in what to do and how to do it’ (Alchian and Demsetz, 1972, p. 782). However, the reason for the last kind of activity is left unexplained.

5. Barzel (1989) argues that orders are used in conjunction with fl at wages, not because employees lack information, but because they have no incentives to devise ways of exer-cising use rights over assets that would produce utility.

6. In market contracts one party may allow the other party to make some specifi c type of


decision ex post contracting. For example, a painter who enters a market contract with a customer may allow the customer to decide on the color and type of paint ex post contract-ing. This can be interpreted as an instance where the painter is indiff erent between colors and types of paint to be used and therefore implicitly negotiates and accepts the order.

7. Thompson (1956) distinguishes between authority and power as the basic means of obtaining obedience. He defi nes power as ‘the ability to determine the behavior of others, regardless of the bases of that ability’, and authority as ‘that type of power which goes with a position and is legitimated by the offi cial norms’ (p. 290).

8. For example, an employee or a group of employees may formally be delegated rights to decide, among themselves, on the use of resources that infl uence their decisions about how to carry out certain activities. I reserve the notion of delegation of authority to those instances where a superior has formal rights to infl uence the decisions made by subordinates, whereas I consider it an instance of delegation of discretion when a group of employees are to decide among themselves on the use of resources.

9. The authorization to give well-defi ned orders under well-defi ned circumstances does not fall under the defi nition of discretion.

10. In fairness to Simon, it should be noted that the more expansive notion of authority in the 1991 paper can be found already in Simon (1947). Thus, Simon’s views of authority did not change between 1951 and 1991. What arguably happened was that Simon in the 1951 paper developed a formal model of authority and that tractability of the formal analysis required that a relatively simple concept of authority be employed.

11. In a reinterpretation of Simon (1951), one may consider restrictions on the subordinate as a means of reducing the set of actions that the employee has the discretion to choose at a given wage. The authority relation should then be preferred when it is effi cient to allow the superior to postpone the decision about the preferred type of work activity and/or the preferred set of restrictions on the work activity.

12. For example, Aghion and Tirole (1997) have investigated the use of authority in a setting in which the agent has asymmetric information about his expected private benefi ts from various projects. The principal may veto projects to protect him from the agent’s adverse selection of projects that reveals high private benefi t to the agent and little or no benefi t to the principal. However, in such situations the superior must be able to credibly commit to choose projects that do not generate negative expected benefi ts to the agent (see also Baker et al., 1999).

13. The employment contract could be interpreted as providing a stock of labor services that within limits could be allocated to diff erent uses by the direction of a manager in response to unforeseen contingencies (Coase, 1937; 1991). However, managers only need to bear the cost of carrying such a stock if they cannot appropriate the benefi ts of their knowledge as new contingencies emerge. Three factors may explain why they cannot sell their knowledge in markets. First, there is the well-known problem of infor-mation as a public good that, if revealed before the transaction, cannot be protected from capture (Arrow, 1962); second, negotiations may take longer time than direction by orders and, because of this, the opportunity for profi table action may be gone; and third, managers’ knowledge may be specifi c to particular transactions.

14. Employers also grant discretion to employees for a number of other reasons, including improving motivation through ‘empowerment’ (Conger and Canungo, 1988), fostering learning by providing more room for local explorative eff orts, and improving collec-tive decision-making by letting more employees have an infl uence on decisions (Miller, 1992). However these reasons arise also if there is no uncertainty.

15. The rather considerable literature on delegation in organizations (for example, Galbraith, 1974; Fama and Jensen, 1983; Jensen and Meckling, 1992) does not explain why delegation should be associated with the exercise of authority. Part of the reason may lie in the static nature of the analysis: all costs and benefi ts associated with delega-tion are given (hence, optimum delegation is known immediately to decision-makers), and there is no role for authority, except perhaps monitoring the use of delegated deci-sion rights.


REFERENCES

Aghion, Philippe and Jean Tirole (1997), ‘Formal and Real Authority in Organizations’, Journal of Political Economy, 105: 1–29.

Alchian, Armen and Harold Demsetz (1972), ‘Production, Information Costs, and Economic Organization’, American Economic Review, 62: 777–95.

Armstrong, Mark (1994), ‘Delegation and Discretion’, Discussion Paper in Econo-mics and Econometrics, Department of Economics, University of Southampton.

Baker, George, Robert Gibbons and Kevin J. Murphy (1999), ‘Informal Authority in Organizations’, Journal of Law, Economics and Organization, 15: 56–73.

Barnard, Chester I. (1938), The Function of the Executive, Cambridge, MA: Harvard University Press.

Barzel, Yoram (1989), Economic Analysis of Property Rights, Cambridge: Cambridge University Press.

Blau, Peter M. (1956), Bureaucracy in Modern Society, New York: Random House, Inc.

Boisot, Max (1998), Knowledge Assets: Securing Competitive Advantage in the Information Economy, Oxford: Oxford University Press.

Burns, Tom and G.M. Stalker (1961), The Management of Innovations, London: Tavistock Publications.

Casson, Mark (1994), ‘Why are Firms Hierarchical?’, International Journal of the Economics of Business, 1: 47–76.

Cheung, Stephen N.S. (1970), The Structure of a Contract and the Theory of a Non-exclusive Resource’, Journal of Law and Economics, 11: 49–70.

Cheung, Stephen N.S. (1983), ‘The Contractual Nature of the Firm’, Journal of Law and Economics, 26: 1–22.

Coase, Ronald H. (1937), ‘The Nature of the Firm’, in Nicolai J. Foss (ed.), (1999), The Theory of the Firm: Critical Perspectives in Business and Management, Vol. II. London: Routledge.

Coase, Ronald H. (1991). ‘The Nature of the Firm: Origin, Meaning, Infl uence’, in Oliver E. Williamson and Sidney G. Winter (eds), The Nature of the Firm, Oxford: Oxford University Press.

Conger, J. and R. Canungo (1998), ‘The Empowerment Process: Integrating Theory and Practice’, Academy of Management Review, 13: 471–82.

Demsetz, Harold (1988), ‘The Theory of the Firm Revisited’, Journal of Law, Economics, and Organization, 4: 141–61.

Demsetz, Harold (1991), ‘The Theory of the Firm Revisited’, in Oliver E. Williamson and Sidney G. Winter (eds), The Nature of the Firm, Oxford: Oxford University Press.

Demsetz, Harold (1995), The Economics of the Business Firm: Seven Critical Commentaries, Cambridge: Cambridge University Press.

Fama, E.F. and M.C. Jensen (1983), ‘Separation of Ownership and Control’, Journal of Law and Economics, 26: 301–25.

Foss, Kirsten (2001), ‘Organizing Technological Interdependencies: A Coordina-tion Perspective on the Firm’, Industrial and Corporate Change, 10 (1): 151–78.

Foss, Nicolai J. (2001), ‘Coase versus Hayek: Authority and Firm Boundaries in the Knowledge Economy’, LINK Working Paper (downloadable from http://www.cbs.dkllink).


Galbraith, J.R. (1974), ‘Organization Design: An Information Processing View’, Interfaces, 4: 28–36.

Ghoshal, Sumantra, Peter Moran and Luis Almeida-Costa (1995), ‘The Essence of the Mega Corporation: Shared Context, not Structural Hierarchy’, Journal of Institutional and Theoretical Economics, 151: 748–59.

Grandori, Anna (2001), Organizations and Economic Behaviour, London: Routledge.

Harrison, Richard T. and Claire M. Leitch (2000), ‘Learning and Organization in the Knowledge-Based Information Economy: Initial Findings from a Participa-tory Action Research Case Study’, British Journal of Management, 11: 103–19.

Hart, Oliver (1991), ‘Incomplete Contracts and the Theory of the Firm’, in O.E. Williamson and S.G. Winter (eds), The Nature of the Firm: Origins, Evolution and Development, Oxford: Oxford University Press.

Hart, Oliver (1995), Firms, Contracts, and Financial Structure, Oxford: Oxford University Press.

Hart, Oliver (1996), ‘An Economist’s View of Authority’, Rationality and Society, 8: 371–86.

Hart, Oliver and John Moore (1990), ‘Property Rights and the Nature of the Firm’, Journal of Political Economy, 98: 1119–58.

Hodgson, Geoff (1998), Economics and Utopia, London: Routledge.Holmstrom, Bengt (1999), ‘The Firm as a Subeconomy’, Journal of Law, Economics,

and Organization, 15: 74–102.Holmstrom, Bengt and Paul Milgrom (1994), ‘The Firm as an Incentive System’,

American Economic Review, 84: 972–91.Jensen, Michael C. and William H. Meckling (1992), ‘Specifi c and General

Knowledge and Organizational Structure’, in Lars Werin and Hans Wijkander (eds), Contract Economics, Oxford: Blackwell.

Kreps, David (1990), ‘Corporate Culture and Economic Theory’, in James E. Alt and Kenneth A. Shepsle (eds), Perspectives on Positive Political Economy, Cambridge: Cambridge University Press.

Liebeskind, Julia Porter, Amalya Lumerman Oliver, Lynne G. Zucker and Marilynn B. Brewer (1995), ‘Social Networks, Learning, and Flexibility: Sourcing Scientifi c Knowledge in New Biotechnology Firms’, Cambridge: NBER Working Paper No. W5320.

Masten, Scott (1991), ‘A Legal Basis for the Firm’, in Oliver E. Williamson and Sidney G. Winter (eds), The Nature of the Firm: Origins, Evolution, and Development, Oxford: Oxford University Press.

Matusik, Sharon F. and Charles W.L. Hill (1998), ‘The Utilization of Contingent Work, Knowledge Creation, and Competitive Advantage’, Academy of Management Review, 23: 680–97.

Miller, Gary (1992), Managerial Dilemmas, Cambridge: Cambridge University Press.

Minkler, Alanson P. (1993), ‘Knowledge and Internal Organization’, Journal of Economic Behavior and Organization, 21: 17–30.

Mintzberg, Henry (1983), Structures in Five, Englewood Cliff s: Prentice-Hall.Simon, Herbert A. (1947), Administrative Behavior, New York: The Free Press.Simon, Herbert A. (1951), ‘A Formal Theory of the Employment Relationship’,

in H.A. Simon (1982) (ed.), Models of Bounded Rationality, Cambridge: MIT Press.


Simon, Herbert A. (1961), ‘The Architecture of Complexity’, Proceedings of the American Philosophical Society, 156: 467–82.

Simon, Herbert A. (1991), ‘Organizations and Markets’, Journal of Economic Perspectives, 5: 25–44.

Thompson, James D. (1956), ‘Authority and Power in “Identical” Organizations’ American Journal of Sociology, 62: 290–301.

Vandenberghe, Ann-Sophie and Jacques Siegers (2000), ‘Employees versus Independent Contractors for the Exchange of Labor Services: Authority as Distinguishing Characteristic?’, Paper for 17th Annual Conference on the European Association of Law and Economics, Gent, 14–16 September.

Weber, Max (1946), Essays in Sociology, translated and edited by H.H. Gerth and C.W. Mills, New York: Oxford University Press.

Weber, Max (1947), The Theory of Economic and Social Organization, translated by A.M. Henderson and Talcott Parsons; edited by Talcott Parson, New York: Oxford University Press.

Williamson, Oliver E. (1985), The Economic Institutions of Capitalism, New York: The Free Press.

Zucker, Lynne (1991), ‘Markets for Bureaucratic Authority and Control: Information Quality in Professions and Services’, Research in the Sociology of Organizations, 8: 157–90.

104

6. Competence and learning in the experimentally organized economy1

Gunnar Eliasson and Åsa Eliasson

1. INFORMATIONAL ASSUMPTIONS FOR A THEORY OF INDUSTRIAL DEVELOPMENT

The single most important empirical assumption in economic theory concerns the totality of all possible states the economy can be in. This universal state space of the model of the economy, or what we call the business opportunities space (Eliasson, 1990a) sets the limits both of what actors can do, and of how informed about their environment they can be. We introduce the knowledge based economy (Eliasson 1987b, 1990a, b; OECD, 1996), which establishes as a necessary assumption an economic opportunities space of immense complexity that is theoretically impos-sible to comprehend more than fractionally from any one place. Each actor understands only a miniscule fraction of the total opportunities space. Together, however, the understanding of all actors in the markets is much larger, but still only encompasses a fraction of the whole, and the tacitness of their knowledge or competence eff ectively restricts expansion of that understanding through collective coordination. Attempts to speed up exploration of the opportunities space through rivalrous competition in markets encounter rapidly escalating information and communications (transactions) costs in the short run. Such exploration, however, off ers opportunities for learning and increases the number of creative encoun-ters such that the opportunities space expands, and probably faster than it is being explored. Hence, we may all be growing increasingly ignorant of what is theoretically possible to know about. This information paradox, or what we will call the Särimner eff ect,2 is a fundamental, sustained and distinguishing property of the experimentally organized economy (EOE) and the rational basis for the existence of a competence bloc (Eliasson and Eliasson, 1996).

Competence and learning in the experimentally organized economy 105

1.1 The Grossly Ignorant Actor

Under these circumstances ‘boundedly rational’ (Simon, 1955) and myopic, bordering on grossly ignorant actors will constantly partici-pate in a positive sum economic game, trying to orient themselves in the immense opportunities set, their success rates depending on their competence to discover and capture the new opportunities, constantly making more or less serious mistakes in the process (Day et al., 1974). Hence, each decision can be seen as a more or less well designed ‘eco-nomic experiment’ to be tested in the market in confrontation with all other actors. Both successful and mistaken experiments involve elements of learning and the creation of new combinations on which competitors can set up new economic experiments. Economic mistakes (to be defi ned below) then defi ne the ultimate transactions cost (Eliasson and Eliasson, 2005). Economic mistakes escalate if actors intensify their exploration of the state space to capture larger profi ts, and the reason is the problem, recognized already by Wicksell (1923) and repeated by Arrow (1959), that it is diffi cult, and perhaps impossible, to model the simultaneous determination of structures (quantities or organizations) and prices. In the highly non-linear micro (fi rm) based macro model that we will refer to below as an approximation of the theory of the EOE, both quanti-ties and prices become increasingly destabilized as actors are pushed on by market competition, and prices become increasingly unreliable signals to guide quantity adjustments towards a both unreachable and perhaps indeterminate equilibrium. The economy will constantly be in a ‘Heisenbergian fl ux’ (see also Eliasson, 1991a, and Eliasson et al., 2005). Since experimentation aimed at exploring and learning about state space also expands the opportunities space through learning and the creation of new opportunities (innovation), we have to reckon with the possibil-ity that the total economic opportunities space is not only immense but also indeterminate at any point in time. Demsetz (1969) recognized this possibility when criticizing the neoclassical doctrine as being subject to a ‘Nirvana fallacy’ in the sense that the best of all worlds that the neoclass-ical economists can calculate within their model make them miss the possibility of even better worlds beyond their prior assumptions.3 One property of the EOE to be explained below also is that there always exist better allocations than the current one, meaning that the economy will always be operating (far) below its production frontiers.4 Referring to Demsetz, Kirzner (1997) observed that economies will always fall short of their potential, a potential, Kirzner argued, that could be approached with the help of discoverers/entrepreneurs.


1.2 The Särimner Eff ect

The informational assumptions are a distinguishing feature of both the neo-classical model and the theory of the EOE. They are embodied in the state space assumed for the model. The assumptions of the neoclassical model are normally tailored such that (1) a unique optimum exists, (2) it can be identi-fi ed and (3) it can be reached at zero or negligible transactions costs.

Convexity assumptions of some sort, of course, have to be imposed to prevent the model economy from exploding or ceasing to exist altogether. The assumption of strict convexity and continuous derivatives ensures that a unique cost minimum, profi t maximizing optimum exists.5 Finding it is always possible if the information and communications (transactions) costs needed are (assumed to be) zero or negligible. If transactions costs are large they will have to infl uence both the position and the existence of the optimum as conventionally defi ned. And in reality the transactions costs are very large (Eliasson et al., 1985, p. 53; Eliasson, 1990a, b; Bergholm and Jagren, 1985; Pousette and Lindberg, 1986; Wallis and North, 1986).

The informational assumptions of the EOE are embodied in its state space or in what we call its business opportunities space. Initially it is assumed to be large and populated with grossly ignorant actors posi-tioned somewhere inside it. The problem, however, is that it has to stay that way for ever for the theory of the EOE not to converge upon the neoclassical or WAD (Walras–Arrow–Debreu) model. Large transactions costs in the form of business mistakes that escalate the closer you get to a stationary process keep the operating domain of the EOE away from the optimum.6 This is, however, not suffi cient for the economy to grow. For endogenous growth to occur this situation has to be for ever maintained and the actors kept grossly ignorant about circumstances that are critical for their business. This can be solved mathematically (Eliasson, 1990b, pp. 46 ff ) by assuming the opportunities space to be initially suffi ciently large, and that it grows from being explored through innovative discovery and learning. This establishes the positive sum game of the EOE that we call the Särimner eff ect, and the possibility (the paradox) that in a dynamic EOE actors may grow increasingly ignorant about what is important to them because the business opportunities space grows faster than it can be explored and learned about.

In this experimentally organized economy (Eliasson, 1991a, 1996) eco-nomic growth will be shown to be moved by innovative project creation and competitive selection, or the Schumpeterian creative destruction process of Table 6.1.7 The capacity of the economic system to identify winners and carry them on to industrial scale production and distribution determines economic growth. Hence, the dynamic effi ciency of selection becomes


critical for economic growth. The organization of effi cient selection and the defi nition of effi ciency in an EOE with no exogenous equilibrium are dealt with in competence bloc theory, of which the theory of the fi rm can be seen as a special case. Competence bloc theory is presented in this chapter as an integral part of the EOE.

Access to the business opportunies space is regulated by institutions that determine the incentives for actors to look for business opportunities there and to compete. In fact, the competence of a fi rm or an actor is best characterized as in Table 6.2.

1.3 Competence Specifi cation of the Firm in the EOE – the Tacit Dimension

Table 6.2 represents a typical situation of a fi rm in reality and in the EOE (Eliasson, 1996, p. 56, 1998a, p. 87). First, no actor, including government,

Table 6.1 The four investment and divestment mechanisms of Schumpeterian creative destruction and economic growth

Innovative entryenforces (through competition)ReorganizationRationalization

orExit (shut down)

Source: ‘Företagens, institutionernas och marknadernas roll i Sverige’, Appendix 6 in A. Lindbeck (ed.), Nya villkor för ekonomi och politik (SOU, 1993, p. 16) and G. Eliasson (1996, p. 45).

Table 6.2 Competence specifi cation of the experimentally organized fi rm

Orientation1. Sense of direction (business intuition)2. Risk willing

Selection/Flexibility3. Effi cient identifi cation of mistakes4. Eff ective correction of mistakes

Operation/Effi ciency 5. Effi cient coordination6. Effi cient learning feedback to (1)

Source: Eliasson, G. (1996, p. 56).


can survey the entire economic opportunities space from one point. The assumptions of the economic opportunities space make it impossible for each actor to be more than fractionally aware of its interior structures and contents and notably about what now and then may become critical for survival. Hence, business mistakes will be made by all actors all the time. Such business mistakes, however, also involve learning about the opportu-nities space and should be regarded as a normal cost for economic develop-ment, a transactions cost (Eliasson and Eliasson, 2005). One common form of learning is being confronted with a superior competitor and understand-ing that for better business solutions than one’s own are possible. Second, some actors may hit upon the absolutely best solution by chance, but they will never know, and nobody else will either. Hence, third, the economy will always be operating far below its production possibilities frontier, a viola-tion of a standard assumption of neoclassical theory. In fact, such frontiers may even be indeterminate.

Fourth, as a business actor you must always believe in your proposed business experiment. If not, you cannot act decisively and forcefully in dynamically competitive markets. Fifth, however, whatever you have invented, you know one thing with almost certainty: there will be many potential solutions that are much better. Therefore, sixth, you have to rec-ognize that among your many competitors you cannot be alone with such a good idea as yours. The business fi rm has to act decisively and prematurely on the basis of the competent judgment of its top competent team (intui-tion, Eliasson, 1990a) before somebody else has acted successfully. Each new solution, therefore, has the character of a business experiment, and the competence of a business fi rm is well categorized in Table 6.2. The fi rm in the EOE needs a good business intuition (orientation), it needs to be able to identify and correct mistakes (fl exibility) and it needs to be operation-ally effi cient. This is a tall order and all three categories of competencies can neither be embodied in one individual nor be assumed to be more than fractionally communicable between participating actors. They are tacit in the sense of limited communicability (Eliasson, 1990a). This problem of multidimensionality and tacitness of competence characteristics is solved in practice through organization, coordination being achieved to the extent possible by the top competent team of the organization and/or through competition in markets (Eliasson, 1976, 1990a). As demonstrated by Coase (1937), the mix between hierarchy and market is determined by the relative transactions costs of coordination within and between hierarchies, tacitness being one factor pushing for market solutions, dynamic effi ciency another (see below).

The competence and behavioral characteristics of the fi rm follow directly from the assumptions made about the business opportunities space and the


EOE. It is no coincidence, however, that development of this basic idea has grown upon us during many interactions with business fi rms (begin-ning with Eliasson (1976) and currently during a study in progress on the economics of health care (Å. Eliasson, 2007; Eliasson and Eliasson, 2007)) and frequent participation in internal fi rm educational programs with the purpose of relating the internal life of fi rms to their external economic environment. In such reality settings you are in a hopeless situation trying to explain how the mainstream neoclassical model can be of any use.

1.4 Macro Dynamics through Experimental Selection – Going from Micro to Macro

When something radically new is introduced, it almost always occurs through the launching of a new product, the establishment of a new divi-sion or through the entry of a new fi rm. A new product may be a com-plement to existing products or a substitute, in the latter case subjecting existing producers to competition and forcing them to reorganize and/or rationalize, or die (exit). When a competitor introduces a radically new product, a fi rm often cannot cope with the new situation through reorganization, because it is staff ed with the wrong human capital. It then has to contract or shut down, and possibly recruit new personnel to establish a new fi rm. The entry process, hence, is critical for long-term economic growth, pushing performance of the entire industry upwards through the four creation and selection mechanisms (‘investments and disinvestments’ of Table 6.1). But entry will not result in growth unless accompanied by a viable exit process.8 If superior entrants and success-fully reorganizing fi rms (the ‘winners’) are supported by the market and allowed to force inferior fi rms to exit, growth will follow. Hence, the major information cost, again, is business mistakes in the form of ‘lost winners’ and the losses of inferior fi rms that are allowed to continue for too long.9 Under normal circumstances, therefore, an economy will be reasonably fully employed and the thrust of fi rm turnover will be to real-locate resources. Hence, the question of the employment contribution of new fi rm formation and self-employment is normally irrelevant, and if asked (as in Blanchfl ower, 2004) it has to be related to a well-defi ned unemployment situation.10

The optimum or the benchmark for effi ciency measurements will be obtained by minimizing the cost of committing the two types of errors in Table 6.3a. Under the opportunities space assumption of the EOE there will always exist better allocations than the current one.11 This optimum, hence, does not exist under the assumptions of the EOE and a unique effi ciency reference cannot be determined. With lost winners defi ning the


potential of the economic system not reached, optimum performance cannot be determined even theoretically, since the benchmark for deter-mining effi ciency cannot be identifi ed. One aspect of effi ciency, however, is to make sure that the customer, or demand, is a determining force in setting directions. The optimal organization of the economy, hence, is determined by a delicate balancing of decentralized incentives, information processing and competition in markets (Eliasson and Taymaz, 2000). For a successful economic outcome actors and resource providers have to be competently guided and disciplined by signals from the ultimate end users, the customers. The competence bloc performs those functions of guidance and resource provision.

2. COMPETENCE BLOC THEORY

Critical competencies are tacit in the sense of being limitedly communi-cable12 (Eliasson, 1990a). While the nature of tacit knowledge cannot be represented analytically, it can be functionally defi ned. A competence bloc lists the minimum number of actors with such competence that are needed to successfully create, identify, select and commercialize new business ideas and to carry them on to industrial scale production and distribution, that is, to initiate and develop a new industry (Eliasson and Eliasson, 1996). This also confi rms that the competence bloc has a pronounced end product or market defi nition, as distinct from a technical defi nition, making the customer the ultimate arbiter of economic value.

2.1 The Nature of Business Competence and the Effi ciency of Project Selection

Competence bloc theory is structured around three nodes of actors; the customers, the technology suppliers and the commercializers. In between them there are three markets: the markets for products, innovations and factors of production (see Figure 6.1). The commercialization process in turn is intermediated through authority within hierarchies and through three types of asset markets – the venture capital market, the private

Table 6.3a The dominant selection problem

Error Type I: Losers kept too longError Type II: Winners rejected

Source: G. Eliasson and Å. Eliasson (1996).


equity market and the market for strategic acquisitions – and through the regular stock markets. Trading in intangible knowledge assets, however, poses particular property rights problems that we come back to in Section 3 below.

The diff erent functions of the actors in the competence bloc can be iden-tifi ed as relay stations in the innovation, creation and commercialization processes and are needed as a minimum. Sometimes they are all internal-ized within one hierarchy or in a planned economy. This was almost the case for IBM during its heyday in the 1970s (Eliasson, 1996, pp. 175ff ). IBM was then to a large extent its own customer in intermediate products. Normally, however, most functions are carried out by specialized actors in the market; call them fi rms. The determination of the mix between market and hierarchy within the competence bloc not only becomes a Coasian (1937) type dynamic theory of the fi rm but also determines the dynamic effi ciency of the entire economy (Eliasson and Eliasson, 2005).

The effi ciency of project selection in terms of identifying and support-ing (commercializing) winners and forcing the exit of losers, as interme-diated through the competence bloc (fi rst crudely sketched in Eliasson and Eliasson, 1996), determines the effi ciency of resource allocation and growth through the Schumpeterian creative destruction process of Table 6.1. Effi cient selection in the EOE is defi ned as the ‘minimizing’ of the economic incidence of two types of errors (Table 6.3a), that is, keeping losers on for too long and ‘losing the winners’. This minimization can, however, only be performed analytically if the business opportunities

1. Competent customers

Markets forinnovation

Strategic

Tactial

Operational

Choose

Coordinate

Manufactureand subcontract

Commercializing agents

2. Innovators(technology supply)

3. Entrepreneurs

4. Venture capitalists

5. Exit market agents

6. Industrialists

Source: Eliasson (2005a, p. 255).

Figure 6.1 Decision structure of the competence bloc


space of the model can be made fully transparent from at least one central point, all information and communications costs kept very small, and all winners identifi ed. The presence of critical tacit and incom-municable knowledge in the decision process makes this theoretically impossible in the EOE. The optimal coordination of production has to be organized as an endogenously determined combination of hierarchies (fi rms) and markets.

The limits of hierarchies are determined where the costs on the margin in the form of lost winners exceed the gains in coordination costs achieved through expanding the hierarchy. The mechanism behind this determina-tion is simple. Centralizing knowledge at one point and ordering outcomes top-down through authority is limited (1) to the part of knowledge that can be coded as information and interpreted centrally at a determinable transactions cost and (2) by the reach of and power to impose authority (compare Simon, 1945; Williamson, 1975; Foss, Chapter 5 in this volume). If that centralization is extended to the parts of the total knowledge (prob-ably the most important parts) that are tacit and not communicable to the intelligence center (the corporate headquarters) of the fi rm, an error bias in the central analysis will be introduced because some of it will be misin-terpreted along the way, owing to a lack of receiver competence (Eliasson, 1976). This can be shown to reduce the total knowledge that enters each decision. Distributing tacit knowledge (or human or team embodied com-petencies) over the market, on the other hand, can be shown to maximize the exposure of each project to a competent evaluation, and minimize a transactions cost measure that includes an economic value of the loss of winners (Eliasson and Eliasson, 2005).13 Competence bloc theory, hence, is an organizational solution to the effi cient allocation of tacit, human embodied competencies on business problems.

If economic competence consists signifi cantly of tacit knowledge, the same characteristics must also apply to consumer choices. If consumer choices are experience based ‘tacit knowledge’, exhibiting preferences for novelty, convergence on an exogenous optimum can in no way be guar-anteed (Day, 1986b). The (also experience based) competence of all actors of the competence bloc to visualize the unpredictable change in consumer preferences, however, should to some extent keep the economic system from becoming erratic, as proposed by Georgescu-Roegen (1950) and modeled by Benhabib and Day (1981).

The competence of the customers to appreciate quality and the com-petence of fi rms to produce new qualities go hand in hand. The perhaps most important quality demanded in an advanced market, furthermore, is product or quality variation.14 Only the customers can individually decide which variant they prefer. One critical function of the competence bloc,


hence, is to make sure that customers’ preferences and competencies fi lter down to the actors in the competence bloc that create and select innova-tions. Customer receiver competence (Eliasson, 1990a) is decisive for the existence of a market. With no customer receiver competence for sophisti-cated products there will be no market for the same products.

2.2 The Actors in the Competence Bloc

First, the products created and chosen in the experimental process never get better than what customers (item 1 in Table 6.3b) are capable of appre-ciating and willing to pay for. The long-term direction of technical change, therefore, is always set by the customers. Sophisticated customers defi ne a competitive advantage of a sophisticated industry.15 Without competent customers there are no sophisticated markets. This is so even though the innovator, entrepreneur or industrialist may take the initiative to launch a new sophisticated product. But quite often the customer takes the initia-tive. Technological development, therefore, requires a sophisticated cus-tomer base, capable of appreciating new products (Eliasson and Eliasson, 1996). The more advanced and radically new the product technologies, the more important customer quality becomes.

In one sense, the customer analysis of competence bloc theory opens up the Keynesian macro demand schedule. But as you peek inside that ‘black box’ you will fi nd that the customer dynamic of the competence bloc has little to do with Keynesian demand. The actors of the competence bloc contribute (commercial) competence in the technological choice process. They accept or reject products off ered to them in the market, thereby signaling what they want. But customers may also be directly involved in some phases of the development of the product. This is normally the case when it comes to very advanced and complicated products such as military and commercial

Table 6.3b Actors in the competence bloc

1. Competent and active customers

Technology supply2. Innovators who integrate technologies in new ways

Commercializion of technology3. Entrepreneurs who identify profi table innovations4. Competent venture capitalists who recognize and fi nance the entrepreneurs5. Exit markets that facilitate ownership change6. Industrialists who take successful innovations to industrial scale production

Source: G. Eliasson and Å. Eliasson (1996).


airplanes (Eliasson, 1995, 2001b). This fact also serves as a rationale for competent purchasing and acquisitions, including public purchasing in areas where goods and services are supplied by public authorities.

Second, technology supply is internationally available, but the capac-ity to receive it and make a business of it requires local competence. Part of this receiver competence (Eliasson, 1987b, 1990a, 1996, pp. 8, 14) is the ability to create new winning combinations of old and new technolo-gies (innovation). A rich and varied supply of subcontractor (technology) services is part and parcel of the innovation process and the competence bloc. The innovation gate into the competence bloc (item 2 in Table 6.3b), hence, is served by many technologies, or technological systems to use the terminology of Carlsson (1995), that are integrated innovatively.

Third, technology supply is not synonymous with industrial progress or growth. In between comes the competence to commercialize new technolo-gies, a far more resource demanding activity than innovation. So between innovation supply and commercialization, representing the demand for innovations, we fi nd the market for innovations in which winners and losers are sorted out (see Figure 6.1). The problem with new growth theory and evolutionary theory is that they do not distinguish between the innovator, the entrepreneur and the competent venture capitalist, and hence do not model that sorting process.

Commercialization competence is experience based and, hence, more nar-rowly defi ned than the creative innovation supply process (Eliasson and Eliasson, 2005).16 As a consequence there will always (ex defi nitione) be a loss of winners along the way. By explicitly modeling the commercialization process we break the linearity between innovation supply and economic growth commonly entered as a prior assumption in the diff erent versions of new growth theory and in evolutionary theory. We fi nd that increasing supplies of technology do not lead to faster growth and that growth can be radically increased on a sustainable basis under improved commerci-alization at given technology supplies.17 First among the commercializing agents come the entrepreneurs. The task of the entrepreneur is to identify commercial winners among the suppliers of technically defi ned innovations and to get his/her choice of technology on a commercial footing.18 The understanding may be of a long-run nature, or more temporary in the sense that entrepreneurs may have to reconfi gure their thoughts soon, or make a business mistake (see Table 6.2). The main thing is that the entrepreneur acts on the perceived economic opportunity (entre prendre in French).

A brief sidestep here. New growth theory is a sub-branch of neoclassical theory and neoclassical theorists tend to order their assumptions such that entrepreneurship has no economic role beyond innovation or technol-ogy supply. Innovation supply, furthermore, is commonly represented


as drawings from a lottery, the participation in which is free of charge. This process, furthermore, is staged within a rational expectations setting or as a stationary process such that the diff erences between ex ante plans and ex post outcomes cancel out in expectation over time as white noise. A stochastic exogenous equilibrium can be defi ned. On this the Swedish economic tradition, heralded by the Stockholm School economists, takes a contrary position, more in line with Austrian economics and us (see Eliasson, 1992, p. 256).19 The enormous and varied expanses of the business opportunities space of the EOE, which keeps opening up new vistas as a consequence of its exploration by the entrepreneurs, make the assumptions of the mainstream neoclassical model empirically unreasonable. The math-ematical reason is that the underlying structures of the assumed stationary process change constantly, radically redefi ning the distributions.20 This means that the standard assumptions of statistical learning do not hold.21 Learning under item 6 in Table 6.2 becomes unreliable.

The entrepreneur, however, rarely has resources of his own to move the business project forward. He, therefore, (fourth) needs funding from an industrially competent venture capitalist, that is, a provider of risk capital, capable of understanding innovators of radically new technology and able to identify business needs and provide context. The money is the least important thing. What matters (Eliasson and Eliasson, 1996; Eliasson, 1997b, 2005c) is the competence to understand and identify winners and, hence, provide reasonably priced equity funding.22 There is an asymmetry problem here that relates to the risk willingness item in Table 6.2. The entrepreneur believes s/he has understood the business situation (business intuition, item 1).23 S/he therefore considers the risks low and is willing to take them on. The outsider, for instance the venture capitalist, does not have the same insight, and therefore considers the same situation more, usually much more, uncertain.

Implicit in this statement is that the industrially incompetent venture capitalist does not understand the project and, therefore, charges an unreasonable (to the entrepreneur) price for his/her services. The supply of industrially competent venture capitalists is extremely scarce (Eliasson, 1997b, 2005c). They constitute the critical link in the overall selection process and, if lacking in performance, this is liable to result in the ‘loss of winners’.

The issue of competent venture capital has long been politically sensi-tive in some European countries since it signals the need for privately rich and industrially competent people to move new industry formation. Such signals run counter to political ambitions to even out income and wealth distributions. The low rate of new entry in continental European countries, including Sweden (Braunerhjelm, 1993), can have two explanations; low


entrepreneurial competence or lack of competent venture capital. The argument in Sweden for a long time focused on the lack of entrepreneurial spirit. It was often heard from banking circles that there was plenty of money but very few good projects to invest in. The most credible expla-nation, however, (Eliasson and Eliasson, 1996; Eliasson, 2005c) has been lack of industrially competent bankers and venture capitalists. Without a rich variety of such fi nancial competence, you will not see many entrepre-neurs. Hence, the venture capitalist and his escape (exit) market (fi fth) are the most important incentive supporting actors. With no understanding venture capitalists the price of new capital will be prohibitively high, or funding will not be forthcoming, and winners will be ‘lost’. With badly functioning exit markets the incentives for venture capitalists will be small and, hence, also for the entrepreneurs and the innovators.

With the rapid securitization of the global fi nancial system the markets for ownership or corporate control have gained in importance (Rybsczynski, 1993). New actors have emerged trading in risks and the exit markets have gained in sophistication and importance as private equity investors with the capacity to mobilize very large fi nancial resources have entered the scene. In growing markets for strategic acquisitions small high technology fi rms and large industrial fi rms (item 6) are trading in knowl-edge assets.

Sixth, and fi nally, therefore, when the selection process has run its course and a winner has been identifi ed, a new type of industrial competence is needed to take the innovation on to industrial scale production and distri-bution. We cannot tell in advance what the formal role of the industrialist is (CEO, chairman of the board, an active owner, or other). He or she fi gures in the competence bloc on account of his or her capacity to contribute functional competence. The capacity to identify, select and move winners to industrial scale production is the most important growth promoting property of the competence bloc. It defi nes a competitive advantage of an economy. This innovative dynamic is what endogenizes growth in the theory of the EOE.

Vertical completeness of the competence bloc, hence, is a necessary requirement for the viable incentive structures that guarantee increas-ing returns to a continued search for winners, that is, for new industry formation. The extreme diversity of the opportunities space of the EOE means that the competence needed to identify winners cannot be speci-fi ed in advance. Hence, an effi cient project identifi cation and selection in the competence bloc requires that a large number of each type of actor in the competence bloc be present, so that if one actor does not understand there will be others who might. Such horizontal diversity in competence is a necessary condition for maximum exposure of each project to a competent


evaluation. Vertical completeness and horizontal diversity make the com-petence bloc complete. Seeing to it that the competence blocs are complete must, therefore, be the prime task of industrial policy (Eliasson, 2000). None of the ‘pillars’ (the actors) of the competence bloc can be missing, or the whole incentive structure will fail to develop.

In the EOE a premium is placed on fl exibility. Actors all the time have to take premature decisions on scant and unreliable information. As a consequence they constantly commit more or less serious business mistakes and have to be prepared to change their minds constantly. Flexibility in the EOE is achieved in three ways. First, and most important, is to have the right business idea (item 1 in Table 6.2). Second, and decisive when you are on the wrong track, is early identifi cation and correction of mistakes (items 3 and 4). Third, when the fi rst two criteria fail, the competence bloc enforces fl exibility through exit by withdrawing support. But this occurs only after the project has been exposed to a varied and maximum com-petent evaluation, thus minimizing the risk of losing a winner. The more widely distributed over the market the competence bloc, the more fl exible the allocation process.

When vertical completeness and suffi cient horizontal variety have been achieved, critical mass has been reached. Then:

(1) Increasing returns to continued search for resources prevail. The loss of winners is minimized.

(2) Competition among all actors in the competence bloc for the gains that otherwise will be lost as lost winners ensures that less competent actors exit.

The competence bloc will now function as an investment attractor such that new entry takes place in such a way that the competence bloc benefi ts from the new entrants, but (because of competition) only new entrants that contribute to the competence bloc enter and/or survive.

The competence bloc then functions as an industrial spillover generator and will begin to develop endogenously through its internal momentum (critical mass). We have a positive sum game. These spillovers will diff use along many paths and both further reinforce the internal development forces of the bloc and contribute serendipitously to other related and unre-lated industries. Endogenous growth will occur (Eliasson, 1997a).

2.3 The Theory of the Firm Revisited

Compared with the internal project evaluation in a large fi rm, project evaluation over the competence bloc draws much larger direct transactions


costs, since the evaluation is done in a distributed fashion involving many independent actors in the market. Narrowing down the evaluation to an internal procedure within a hierarchy, therefore, may lower direct transac-tions costs, but the more narrow evaluation also raises the risk of losing winners and therefore both raises total transactions costs and lowers the effi ciency of project selection. Hence, a relevant analysis of the optimal organization of production has to include the loss of winners as a transac-tions cost (Eliasson and Eliasson, 2005).24

A large fi rm, such as IBM in the 1980s, internalized most of its compe-tence bloc. For a long time IBM was unable to get out of its mainframe mentality due to lack of internal experience from the new industrial markets of distributed computing and came close to disaster during the last years of the 1980s, during its attempts to transform itself away from mainframe technology. Business history is full of near losses based on narrow-minded internal business judgment in large hierarchies, the only ones that can be identifi ed (Eliasson, 1996).

The contrary solution with all functions distributed over the market, however, has other problems, even though Fama (1980) argued that there was no need for an entrepreneur in economic theory, since the services of the entrepreneur could always be rented in the market. Most of economic literature and debate, furthermore, neglects the commercialization compe-tence altogether and we have a neo-Schumpeterian literature that comes very close to neoclassical literature in presenting the fi rm or an industry as a direct linear R&D, technology supply and growth machine. And socialist thinking on the matter has been even more negligent in understanding any merit in the competitive functions of markets.

Competence bloc theory has no role to play under the assumptions of the static general equilibrium (neoclassical) model in which all knowledge has the qualities of codable and communicable information that can be centralized to one place for a complete overview. Competence bloc theory, however, has a decisive role to play in project selection within and between hierarchies when tacit, non-codable knowledge embodied in indi-viduals and hierarchies has to be mobilized within hierarchies and through markets. Once this has been said, we know that the nature and distribu-tion of knowledge determine the ‘optimal’ combination of hierarchies and markets through which the total knowledge base can be mobilized for particular business problems. We have a dynamic version of Coase’s (1937) theory of the fi rm in which the hierarchical structure of markets is not only endogenized as an equilibrium property but also changes as a consequence of the constantly ongoing project selection in an experimentally organized economy.

In that setting the critical task of the top competent team of a fi rm


is to identify the best organizational structure for the particular busi-ness problem at hand (Eliasson, 1990a). Since that structure constantly changes, new organizational structures constantly have to be identifi ed. This is not at all easy and management mistakes are constantly made. Hence, learning feedback (item 6 in Table 6.2) is limited. Such limited learning possibilities are typical of the theory of the EOE. It has also been a typical experience in the recent ongoing outsourcing business culture (Eliasson, 2005d). A systematic test of the limitations of business learning is presented in Eliasson (2005b) in which business planning and manage-ment methods during and between the three periods of post-World War II development are compared: (1) the pre-oil crisis steady state experience of 1969–75, (2) the post-oil crisis sobering-up through most of the 1990s and (3) the rapid emergence of globally distributed production from the mid 1990s, blurring the notion of the fi rm/hierarchy to be managed. In general, management experience did not carry over reliably between the periods, and business mistakes occurred on a grand scale during those transitions. Furthermore, distributed production and unstable hierarchies undermine the whole notion of the central authority of a corporate headquarters (Eliasson, 2005a, Chapter V). Top-down push-through of orders does not work when the hierarchy responds by changing its structure. This problem of endogenous hierarchies or organization poses the same analytical problem as that discussed already by Wicksell (1923), mentioned above, and provides a rational argument for integrating the theory of organiza-tion with that of allocation.25

Through the early 1990s Sweden featured an extreme concentration of historically grown and successful large-scale and international manufac-turing companies. For decades this was considered synonymous with an equally rich endowment of business leadership competence that could be carried over from generation to generation (Eliasson, 1990b, 2005b). To some extent this was of course true, but this competence had been acquired in traditional mature industries that innovate slowly. The management of innovation in the new type of industries like biotech is radically diff er-ent from that in mature industries such as engineering. Experience is that leadership competence acquired in traditional industries is of limited use in the radically new industries and sometimes outright dangerous to apply. The learning feedback (item 6 in Table 6.2) is correspondingly unreliable (Eliasson, 2005b). For some reason the large Swedish companies exhibited unusual organizational innovation capacity coming out of the oil crises of the 1970s and carrying Swedish manufacturing growth in the 1980s, but it was concluded in Andersson et al. (1993) that this was a story that no one should count on to be repeated. True enough. When Swedish industry entered the ‘New Economy’ of globally distributed production during the


second half of the 1990s the giants began to stumble, presumably because the earlier management experience was not a reliable management guide into the future (Eliasson, 2005a). Presumptuous big business leaders who enter radically new businesses with an air of authority, therefore, may even be a negative factor for development. Hence, Swedish manufacturing industry perhaps no longer features the rich and varied competence blocs needed for the effi cient project selection that worked well for decades but does not seem to work at all that well as we currently attempt to enter a radically New Economy.

3. INSTITUTIONS, TRADE IN PROPERTY RIGHTS AND SOCIAL CAPITAL/COMPETENCE

Project selection over markets always involves trade in knowledge assets between actors in a competence bloc. Competence bloc theory explains those transactions explicitly, why large resources (transactions costs) are needed and why transactions competence is critical for dynamically effi cient allocations in the theory of the EOE, but not in the mainstream neoclassical model. We therefore have to relate the competence bloc theory just introduced back into the economic environment of the EOE.

3.1 Institutions, Incentives and Competition

None of the economic dynamics discussed so far can occur without a minimum of supporting institutional infrastructure needed to establish the required tradability in knowledge assets. Institutions defi ne incentives to act on business opportunities that fuel competition. In general, if com-plete contracts that defi ne the transfer of ownership associated with trade cannot be formulated, trade will be correspondingly limited and under-supply will follow (Williamson, 1985). Resource allocation in the experi-mentally organized economy involves trade in intellectual property rights across the competence bloc, notably throughout the markets for venture capital, the exit markets and the markets for strategic acquisitions on its commercialization side. For the full macroeconomic growth potential of innovative project creation and selection in Table 6.1 to be realized, this trade will have to occur with a minimum of risks and transactions costs. For this to be realized property rights to these intellectual assets have to be well defi ned. Those property rights are not needed for the same transac-tions to be organized within a hierarchy. For all projects to be exposed to a maximum competent evaluation over the competence bloc, effi cient prop-erty rights of this kind thus have to be in place. Only then will the relative


superiority of the broad market exposure of competence on each project be fairly compared with the narrow exposure within a hierarchy, and the key factor will be the tradability of intangible knowledge in fi nancial asset markets (Eliasson and Wihlborg, 2003).

3.2 Social Capital/Competence

Competence bloc theory has been shown to be needed to explain the effi -ciency of project selection in the EOE. Under the informational assump-tions of the EOE, the selection and allocation processes supporting stable economic growth at the macro level become unpredictable and socially demanding at local micro levels. This suggests that the overriding welfare and policy concern should be to design institutions that support the ability of individuals to cope with environmental unpredictability and the expo-sure to arbitrary change that increases with economic effi ciency and growth (Day, 1986a, 1993, notably p. 77; Eliasson, 1983, 1984). This suggests that what we call social capital should be defi ned in terms of how it supports individuals’ ability to cope (Eliasson, 2001a)26 and in doing so we also recognize that the support of social capital development becomes a criti-cal part of both industrial and social policy, to be elaborated in another context (Eliasson, 2000, 2006b).

4. ENDOGENOUS GROWTH THROUGH COMPETITION – THE DYNAMICS OF GOING FROM MICRO TO MACRO AND BACK

Endogenous growth in the theory of the EOE is explained with reference to the Schumpeterian creative destruction process of Table 6.1 and how it works in the Swedish micro-to-macro model (Eliasson, 1991a, 1996, p. 45). Competition there keeps the entrepreneurial and commercialization processes in motion and the Särimner innovation/learning eff ect keeps the investment opportunities space expanding to sustain growth. But compe-tition has to be activated. It may be prohibited or regulated. Actors may enjoy monopoly conditions and earn more by colluding than competing. Incentives to compete are therefore not suffi cient to activate competition.

The rational foundation of endogenous growth through the Schumpeterian creative destruction process of Table 6.1, therefore, has to explain why each actor in a market of the EOE constantly has to innovate and improve its performance, or perish. The reason is that each actor lives in a constant state of fear of being overrun by new and old competition. The opportunities space or the Särimner eff ect combined with unrestricted


free entry keeps experimental exploration of the opportunities space active as a necessary means of survival. The rationale for this is that each actor is constantly challenged in the market, from above by superior (‘more profi table’) competitors that can pay more for factors of production or lower prices, and from below by inferior fi rms that have to innovate and leapfrog superior competitors to prevent them from competing them out of business.

The Salter curves (1960) in Figures 6.2a and 6.2b illustrate how this endogenization of growth has been specifi ed in the Swedish micro-to-macro model (Eliasson, 1977, 1985). The Salter curves rank fi rms (or divisions of large fi rms) in Swedish manufacturing industry in 1983 and 1990 by returns to capital and labor productivity. It is part of the micro database used to initiate simulations on the Swedish micro-to-macro model. The spread of the Salter distributions illustrate the great opportunities for improvement

Cumulative capital stock (percent)

Rat

e of

ret

urns

(pe

rcen

t)

0–25–20–15–10–505

1015

2520

1983

303540455055

10 20 30 40 50 60

1990

70 80 90 100

Note: The vertical columns show the position on the Salter curve of one fi rm, the width measuring its size in percent of total industry capital.

Source: MOSES Database (1992) and updatings.

Figure 6.2a Salter curve distribution of rates of return on total capital in Swedish manufacturing in 1983 and 1990


in macro performance through reallocation of resources. This reallocation can be set in motion through competition.

The fi rm is indicated by a column, its width measuring its size in percent of the total capital or the total employment of the fi rm population. This fi rm is challenging the less profi table or productive fi rms to the right, by being able to outbid them for resources. But these fi rms being challenged also challenge the fi rms to their left by attempting to leapfrog them through innovation and investment.

The Salter curves in Figures 6.2a and 6.2b are snapshots at one point in time. At each point in time entrants wait behind the scene. New entrants on average exhibit lower performance than incumbents, but the spread is much wider. This is empirically well established. Hence, there will always be a few winners in the entry fl ow of fi rms that survive and grow into major players, provided the local commercialization competence is suffi cient. The underperforming entrants are sooner or later competed out of business.

Note: The columns indicate the same fi rm as in the previous fi gure, the size now being measured in percent of total employment. The shaded areas measure unused capacity for each fi rm.

Source: MosesDataBase (1992) and updatings.

Figure 6.2b Salter curves of labor productivities in Swedish manufacturing in 1983 and 1990

Percent

Labour productivity

00

50

100

150

200

250

300

350

400

500

450

1983

550

600

650

10 20 30 40 50 60

1990

70 80 90 100


The outcome for the fi rm indicated in Figure 6.2a in 1990 has been an improvement in profi tability, but a loss in ranking. As a consequence the Salter curves ranking productivities in Figure 6.2b shift outward. Growth occurs.

For this competitive process, spurred by fear, to be sustained and result in sustainable growth it is, of course, possible to introduce the standard stochastic innovation lottery that are commonly found in the new growth models. We have done that in the Swedish micro-to-macro model (Hanson, 1986, 1989; Taymaz, 1991), but a more satisfactory solution is to model the innovation process explicitly by making it possible for fi rms to learn to upgrade their productivity from superior competitors when exploring the opportunities space and to model creative encounters during that explora-tion which create new and superior combinations that in turn expand the opportunities space and open possibilities for others to discover and learn from, and so on. This Särimner creativity process has been based on the Ballot and Taymaz (1998) genetic learning mechanism in the model. The distributions that defi ne innovation supplies will then be endogenously determined, explicitly derived and deterministic, and not entered as an assumed stochastic process as in Aghion and Howitt (1992, 1998), Pakes and Ericson (1998) and Nelson and Winter’s (1982) evolution ary model (on the last see in particular Winter, 1986). Competence bloc theory in combination with Ballot and Taymaz (1998) has made it possible to avoid the less satisfactory approach of modeling the source of economic devel-opment as draws of productivity gains in a given lottery (from a stationary distribution) that represents the combined outcome of innovators, entre-preneurs and venture capitalists, a lottery that does not even charge you for your participation and that is presumably organized by the state. Our approach is still a crude approximation of what we aim for, focusing on the role of industrially competent venture capital provision (Ballot et al., 2006), but it is suffi cient to overcome the logical but false linearity between innovation supply (through the lottery) and economic growth27 that short-circuits the commercialization process in new growth literature.

5. ALLOCATION AND ECONOMIC GROWTH

The Schumpeterian creative destruction process of Table 6.1 does not guarantee macroeconomic growth. Destruction may dominate over crea-tion for considerable time and with permanent long-term consequences. Reorganization of fi rms (item 2) might very well be guided by caution and mistaken perceptions and result in a general contraction of output among fi rms. The main thing is that institutions are organized such that


winners are created and identifi ed in the competence bloc and carried on to industrial scale production and distribution, while losers are pushed out. In the Swedish micro-to-macro model (Eliasson, 1991a), choice algo-rithms determine the decisions of individual fi rms in this respect. Since the competence bloc not only creates, identifi es and selects winners but also supports winners by directing (fi nancial) resources to them, this also illustrates the role of the competence bloc, not only as a creator and an identifi er of winners but also as an organizer of the allocation of tacit, human embodied knowledge, and as a fi nancial resource provider. Since all actors in the competence bloc embody a rich variety of tacit competen-cies, the competence bloc becomes an allocator of its own competence. The competence bloc becomes the core resource allocator in an EOE.

5.1 Technological Diff usion

The diff usion of new technology as directed by the competence bloc occurs along six distinct channels (Table 6.4): (1) when people with competence move over the labor market, (2) through the entry of new fi rms when people with competence leave established fi rms, (3) through mutual learn-ing between subcontractors and the systems coordinator, (4) when a fi rm strategically acquires other fi rms to integrate their particular knowledge with its own competence base, (5) when competitors imitate the products of successful and leading fi rms (the ‘Japanese approach’), and (6) through organic growth of and learning in incumbent fi rms.

Items 3 and 4 are particularly important for the advanced mature indus-trial economies. Some of them, but not all of them, have the capacity to develop and produce very advanced and technologically complex systems products such as aircraft, submarines and large trucks. Such products embody so many diff erent technologies that change rapidly and so many

Table 6.4 New technology is diff used

1. when people with competence move (labor market)2. through new establishment by people who leave other fi rms (innovation

and entrepreneurship)3. when subcontractors learn from the systems coordinating fi rm, and vice

versa (competent purchasing)4. through strategic acquisitions of small R&D intensive fi rms (strategic

acquisitions)5. when competitors learn from technological leaders (imitation)6. through organic growth and learning in incumbent fi rms

Source: G. Eliasson (1995).


specialist components that it is impossible to develop and produce them within one company. Both development and manufacturing have to be distributed over subcontractors in the market (Eliasson, 2001b). All actors have to understand how to operate as a sophisticated participant in the ‘whole’. Hence, mutual learning within such a systems product complex is an important part of the technological diff usion process of the very advanced economies.28

One observation can be made from a close study of Table 6.4: effi cient diff usion of new technology requires eff ective market support, notably in the labor market (item 1 in Table 6.4) but also in the venture capital market and the markets for mergers and acquisitions (M&A) (Eliasson and Eliasson, 2005). Completeness of the competence bloc again becomes a critical requirement for the introduction of radically new technology. Effi cient diff usion is also a necessary condition for spillovers and compe-tence bloc development, but it is not suffi cient. For new technology to be introduced in production receiver competence (Eliasson, 1987b, 1990a) is needed. Entrepreneurial and venture capital competencies are part of this, but the general and rapid introduction of new technology also requires a varied and competent labor force at all levels (workers, engineers, manag-ers and executive people).

5.2 The Informational Assumptions Revisited

The distinguishing features of the theory of the EOE hinge on its informa-tional assumptions. Under the assumptions of the EOE, tacit knowledge or competence in the sense of limited communicability can be shown to exist (Eliasson, 1990a). The coordination of actors guided by tacit competencies can never be perfect. It is costly, the largest cost being not resources used directly in information processing but the ‘profi ts lost’ when business mis-takes are being committed.29 It also involves the selection and coordination of tacit competencies for the same coordination, a task that unavoidably leads to an infi nite regress and no determinate best or optimum outcome (read exogenous equilibrium).

A competence bloc, hence, can also be defi ned as an organization of institutions and actors with (tacit) competence such that incentives and competition contribute to as effi cient an allocation of total resources as is possible. This includes the allocation of the tacit competencies embodied in the actors. Hence, under the informational assumptions of the EOE, the best possible allocation cannot be determined. There will always be unknown better allocations. This conclusion only requires our assumption of an immense, non-transparent business opportunities space, an assump-tion30 that places us, and the EOE, in the early Austrian tradition of Carl


Menger (1871), who emphasized the ignorance of actors and the frequent incidence of business mistakes (Alter, 1990).

It may be considered presumptuous to modify the standard informa-tional assumptions of mainstream economic theory such that the all-powerful mathematical tool of calculus appears to be rendered useless. But there are good reasons. First, it is easy to quote a massive empirical evidence in favor of the assumptions of the EOE. Second, attempts to penetrate the inner mechanics of new growth models, such as Pakes and Ericson (1998), that are based on a Walrasian, Arrow and Debreu (1954) type equilibrium platform tell a story that is not less complex than the story of the EOE or the mathematics of its model approximation, MOSES. A possible objection might, however, be (third) that, even though not correct, the informational assumptions of the neoclassical model work well as a reasonable approximation and generate good predictions in the spirit of Friedman (1953). Of what? It is diffi cult to fi nd good examples. However, fourth, a reasonable argument for being methodologically con-servative is that one might as well keep the familiar mathematical tool box until someone has come up with something better (Clower, 1986).31 Now, that has been done and it appears that the properties of the theory of the EOE that are obtained after some marginal modifi cations of the assump-tions of the WAD model require simulation mathematics in order to be satisfactorily investigated.

Removing the devotion to calculus in economics would therefore help make way for the powerful simulation tool and more relevant economic theory, a prediction voiced about fi fty years ago by Koopmans (1957, p. 174).

NOTES

1. This chapter merges the theory of the experimentally organized economy (EOE) (Eliasson, 1991a) with that of competence blocs (Eliasson and Eliasson, 1996). The theory of the EOE has grown out of many years of experimenting with the Swedish micro-to-macro model (Eliasson, 1977, 1991a; Albrecht et al., 1992, Ballot and Taymaz 1998) to the extent that the model should now be seen – as will be explained in the text – as a quantitative approximation of the theory of the EOE. As such the text is very empirical and based as well on several hundred interviews carried out by one of us, or the two of us together, and on analyses of data assembled for the model (see Albrecht et al., 1992). To keep the theory of the EOE and its model approximation in the sustain-able Austrian/Schumpeterian state that distinguishes it from the standard neoclassical model, certain assumptions relating to the nature of the state space of the model (its size, complexity and heterogeneity) have to be made.

2. From the pig in the Viking sagas that was eaten for supper in Valhalla, only to return the next day to be eaten again, and so on. The diff erence from the situation in Valhalla, which we make a point of, is that the opportunities space not only stays large; it grows from being explored. We are confronted with a positive sum game. By being concerned


not only with the neoclassical problem of how to manage a given endowment of scarce resources but also with the Aristotelian problem of how resources are created, we face the positive sum game of the EOE that endogenizes growth (Eliasson, 1987a, p. 28f, 1990b, p. 46f). During the year 2007, when the 300th year birthday of the Swedish botanist Linnaeus was celebrated, it felt appropriate to mention (Frankelius, 2007) that Linnaeus, in his until now not translated (from Latin), but frequently quoted, main work Systema Naturae Sive Regna Tria Naturae Systematice Proposita Per Classes, Ordines, Genera & Species (Leyden, 1735), considered economics to be the greatest of all sciences. And his concern was to create economic value through exploration of, and discoveries in, nature.

3. The Nirvana complex has a long history. In his Candide (1759) Voltaire poked fun at Leibnitz’s vision of the best of all worlds. Leibnitz, by the way, laid the foundation of the mathematics used by neoclassical economists today.

4. To the extent that they can be determined. 5. Hence, starting from any place, as long as you walk uphill you will eventually reach the

peak. There are hundreds of mathematical algorithms that approximate that task in economic modeling.

6. This feature of the EOE has been investigated on a quantitative micro-based macro model of the EOE in which structures and prices are simultaneously determined in an interactive fashion, using up transactions resources in the process. In static equilibrium that can only be achieved under very restrictive assumptions, notably zero transactions costs, duality prevails and prices refl ect exactly quantities, and vice versa. The further away from ‘static equilibrium’, the more unreliable prices are as signals of future optimum quantities and the more frequent and larger the mistakes. If you push the economy closer to an approximate equilibrium the entire model structure is destabilized and eventually collapses (Eliasson 1991a; Eliasson et al., 2005).

7. The theory of the EOE is not structured as a mathematical model. The micro assump-tions, therefore, can only be linked to macro through verbal reasoning. The Swedish micro(fi rm)-to-macro model (Eliasson, 1991a), on the other hand, is a mathematical model that can be said to approximate the EOE. Common to both, and the source of endogenous growth, is the dynamic of the Schumpeterian creative destruction process of Table 6.1. In fact, the theory of the EOE was inspired by experimenting with that model (Eliasson, 1987a).

8. This reasoning can be nicely illustrated using a Salter (1960) curve; see Figures 6.2a and b in Section 4, and Eliasson (1991a, 1996, p. 44f). This is also the way growth occurs in the Swedish micro-to-macro model (Eliasson, 1991a), which is based on empirically deter-mined Salter curves and fi rms constantly shifting position on the Salter curves as they are induced by incentives and pushed by competition. Competence bloc theory explains how this competitive process can be organized diff erently and more effi ciently. It is particularly important to observe that innovative entry subjects incumbent fi rms to competition and forces them to respond. Their response in the form of reorganization and rationalization may mean either expansion or contraction, depending upon incentives embodied in the institutions of the economy and the individual competence capital of fi rms.

9. Growth through competitive experimental selection through innovative entry is explic-itly modeled in the micro-based macro Model Of the Swedish Economic System (MOSES) (Eliasson, 1977, 1985, 1991a) in which learning costs through business mis-takes are explicit. The early, simple expectations functions have been complemented with genetic learning mechanisms in Ballot and Taymaz (1998).

10. Which may be a relevant situation for continental Europe where such a large number of people are prematurely retired, on sick leave or outright unemployed, but much less in the US where growth to a larger extent occurs through the reallocation of employed people (G. Eliasson, 2006a, b).

11. This is the case in the Swedish micro-to-macro model (Eliasson, 1977, 1991a) which approximates the EOE. Whatever output trajectory over the long run that you simulate, you can be fairly sure that better outcomes exist.


12. Most defi nitions of tacit knowledge are much broader than this. For our purpose, however, this narrow defi nition is suffi cient. A broader defi nition will only strengthen the empirical implications of our analysis.

13. Note that this is the exact opposite conclusion to the standard view of the Walras–Arrow–Debreu model, in which the Walrasian superauctioneer is assumed to be capable of achieving a complete and costless central overview of the entire economic landscape, and to identify the one superior position of the economy that is there by assumption for instance as modeled by Malinvaud (1967).

14. This constitutes a second information paradox of economics referred to earlier, namely that we are becoming less and less informed about what is becoming more and more important, namely the quality of inputs and outputs (Eliasson, 1990b, p. 16).

15. As pointed out already by Burenstam-Linder (1961). Burenstam-Linder, however, carried out his argument in terms of static international trade theory and called it com-parative advantage.

16. On this Granstrand and Sjölander (1990a, b) observe that a broad internal technology base makes the fi rm more effi cient in acquiring and implementing new complementary knowledge, for instance through the acquisition of innovative technology fi rms.

17. This was a property of the Swedish micro-to-macro model as early as Eliasson (1979, 1981) when the property was ‘discovered’ as part of a crudely modeled commercializa-tion process.

18. This distinction between the innovator and the entrepreneur originated in Mises (1949). Schumpeter was not clear on this and often used the term innovator to signify what we mean by an entrepreneur. With our defi nition we do not need the third concept, the inventor, which Schumpeter frequently used to emphasize the technical dimensions of an innovation.

19. Stationarity means that distributions have constant (over time) mean and variation. A stationary process will keep generating data such that the parameters of the statisti-cally defi ned entrepreneur will eventually be known for sure with any precision desired. Besides being an absurd representation of ‘the entrepreneur’, it has been demonstrated that such statistical learning requires a hopelessly narrow specifi cation of the state space, expressed as a stationary process, to allow any learning at all (Lindh, 1993).

20. In a huge Las Vegas gaming hall à la Rothschild (1974) the analogy would be each actor rushing around between the one-armed bandits using his or her personal criteria to change console. Even though each bandit is governed by a particular stationary process for ever, the activity going on in the entire gaming hall (the ‘market’) cannot be approximated by a stationary process that does not change over time.

21. Which are identical to those of rational expectations and effi cient market theory. Antonov and Trofi nov (1993) demonstrate how simple statistical learning through fore-casting a non-linear environment with linear economic prediction models of a standard Keynesian or neoclassical type produces worse macroeconomic outcomes than the use of completely ad hoc individual experience-based relationships.

22. The venture capitalists also contribute managerial, fi nancing and marketing competence through their network, but this comes after the ‘understanding’. Such services are nor-mally available in the market and, consequently, are less critical.

23. Knight (1921) would say that the entrepreneur converted an uncertain situation into a situation of calculable risks. See also LeRoy and Singell (1987) and Eliasson (1990a).

24. Even though the economic value of the loss of winners is indeterminate in the theory of the EOE. This is no problem in the neoclassical model in which business mistakes and loss of winners do not exist by assumption, with the possible exception of random losses in stochastic equilibrium models which are outright irrelevant in an industrial economics analysis.

25. An anonymous referee pointed out that this problem of ‘seemingly endless reorgani-zation’ and unpredictable market behavior has already been discussed by Ilinitch et al. (1996). It would, however, mean a new chapter to address also the problem of


‘hypercompetition’ here. True, however, and in keeping with my argument, standard neoclassical theory has little to say on this.

26. This is a more narrow defi nition of social capital than the broad, empirically diffi cult categories of Coleman (1988), Putnam (2000), Ritzen (2001) and Woolcock (2001). We are, in fact, much closer in defi nition to the abilities to cope that Wolfe and Haveman (2001) relate to educational capital.

27. Which by accident has its origin in Schumpeter (1942). Schumpeter was a great admirer of Walras. He sometimes ‘pedagogically’ began his argument by making the innovator an exogenous disturber of the Walrasian equilibrium. By postulating that once success-ful the routine R&D/innovation department of a fi rm would then for ever make that fi rm the dominant player in its market, Schumpeter invoked the perennial problem of economies of scale in the Walrasian model. This corner solution troubled Marshall. He invented the concept of an ‘industrial district’ where the scale economies originated in the district as a whole rather than with individual fi rms. Romer (1986, 1990) formulated this mathematically in macro under the title of ‘new growth theory’, however without quoting Marshall.

28. It is also obvious that as long as the distributed production system remains within the borders of a nation it off ers a protective shield against foreign ‘competitive imitation’ (under item 5).

29. In fact, accounting for the implicit cost of ‘lost profi ts’ is what distinguishes our model from the standard neoclassical model. For an interesting early discussion of this see Dahlman (1979).

30. Note that Arrow and Debreu (1954), pioneering modern mathematical economics, carefully crafted their assumptions to avoid this result. There is no mention of Austrian economics and Hayek (1937, 1940, 1945), which, though by far the most penetrating treaties on information economics at the time, are not even quoted.

31. In the discussion of Herbert Simon’s presentation at the conference on ‘The Dynamics of Market Economies’, organized by the IUI, 1983 (see Day and Eliasson, 1986, pp. 42ff ).

BIBLIOGRAPHY

Aghion, Philippe and Peter Howitt (1992), ‘A Model of Growth through Schumpeterian Creative Destruction’, Econometrica, 60 (2) (March), 323–51.

Aghion, Philippe and Peter Howitt (1998), Endogenous Growth Theory. Cambridge, MA and London: The MIT Press.

Albrecht, James W., Pontus Braunerhjelm, Gunnar Eliasson, Jörgen Nijlsson, Tomas Nordström and Erol Taymaz (1992), MOSES Database, Research Report No. 40, Stockholm: IUI.

Alter, Max (1990), Carl Menger and the Origins of Austrian Economics, San Francisco and Oxford: Westview Press.

Andersson, Thomas, Pontus Braunerhjelm, Bo Carlsson, Gunnar Eliasson, Stefan Fölster, Lars Jagrén, Eugenia Kazamaki Ottersten and Kent Rune Sjöholm (1993), Den långa vägen – Om den ekonomiska politikens begränsningar och möj-ligheter att föra Sverige ur 1990-talets kris, Stockholm: IUI.

Antonov, M. and G. Trofi nov (1993), ‘Learning through Short-Run Macroeconomic Forecasts in a Micro-to-Macro Model’, Journal of Economic Behavior and Organization, 21 (2), June.

Arrow, Kenneth J. (1959), ‘Toward a Theory of Price Adjustment’, in M. Abramovitz et al., (eds), The Allocation of Economic Resources, Stanford, CA: Stanford University Press.


Arrow, Kenneth J. and Gerard Debreu (1954), ‘Existence of Equilibrium for a Competitive Economy’, Econometrica, 22 (3) (July), 265–90.

Ballot, Gérard and Erol Taymaz (1998), ‘Human Capital, Technological Lock-in and Evolutionary Dynamics’, in G. Eliasson and C. Green (eds), The Microeconomic Foundations of Economic Growth, Ann Arbor: The University of Michigan Press, Stockholm: City University Press.

Ballot, Gerard, Gunnar Eliasson and Erol Taymaz (2006), ‘The Role of Commercialization Competence in Endogenous Economic Growth’, paper presented to the International J.A. Schumpeter Society 11th ISS Conference, Nice-Sophia Antipolis, 21–24 June.

Benhabib, Jess and Richard H. Day (1981), ‘Erratic Accumulation’, Economics Letters 6:113–17.

Bergholm, F. and L. Jagrén (1985), ‘Det utlandsinvesterande företaget – en empirisk studie’, in G. Eliasson, F. Bergholm, E.C. Horwitz and L. Jagrén, De Svenska Storföretagen – en studie av internationaliseringens konsekvenser för den svenska ekonomin, Stockholm: IUI.

Blanchfl ower, David G. (2004), ‘Self-employment: More may not be Better’, Swedish Economic Policy Review, 11 (2) (Fall):15–94.

Braunerhjelm, Pontus (1993), ‘Nyetablering och Småföretagande i Svensk Industri’, in Andersson et al. (1993), Chapter 4.

Braunerhjelm, Pontus (ed.) (2001), Huvudkontoren fl yttar ut, Stockholm: SNS förlag.

Buigues, Pierre, and Alexi Jacquemin and Jean-Francois Marchipont (2000), Competitiveness and the Value of Intangible Assets, Cheltenham, UK and Northampton, MA, USA: Edward Elgar.

Burenstam-Linder, S. (1961), An Essay on Trade and Transformation, Uppsala: Almqvist & Wiksell.

Carlsson, Bo (ed.) (1989), Industrial Dynamics – Technological Organizational, and Structural Changes in Industries and Firms, Boston/Dordrecht/London: Kluwer Academic Publishers.

Carlsson, Bo (ed.) (1995), Technological Systems and Economic Performance: The Case of Factory Automation, Boston/Dordrecht/London: Kluwer Academic Publishers.

Carlsson, Bo (ed.) (1997), Technological Systems; Cases, Analyses, Comparisons, Boston/Dordrecht/London: Kluwer Academic Publishers.

Clower, Robert W. (1986), ‘Discussion’, in Day, Richard H. and G. Eliasson (eds), The Dynamics of Market Economies, Stockholm: IUI, and Amsterdam: North-Holland, pp. 42ff .

Coase, R.H. (1937), ‘The Nature of the Firm’, Economica, New Series, IV (13–16) (Nov.), 386–405.

Coleman, J.S. (1988), ‘Social Capital in the Creation of Human Capital’, American Journal of Sociology, 94, Suppl., S95–S120.

Dahlman, C.J. (1979), ‘The Problem of Externality’, Journal of Law & Economics, 22 (1), 141–62.

Dahmén, E. (1950), Svensk industriell företagarverksamhet (Entepreneurial Activity in Swedish Industry, 1919–39), Band 1 and 2, Stockholm: IUI. Also published in 1970 by the American Economic Association Translation Series under the title Entrepreneurial Activity and the Development of Swedish Industry, 1919–1939.

Day, Richard H. (1986a), ‘Disequilibrium Economic Dynamics: A


Post- Schumpeterian Contribution’, in R.H. Day and G. Eliasson (eds) (1986).

Day, Richard H. (1986b), ‘On Endogenous Preferences and Adaptive Economiz-ing’, in R.H. Day and G. Eliasson (eds) (1986).

Day, Richard H. (1993), ‘Bounded Rationality and the Coevolution of Market and State’, in Day, Eliasson and Wihlborg (eds) (1993).

Day, Richard H. and Gunnar Eliasson (eds) (1986), The Dynamics of Market Economies, Stockholm: IUI; Amsterdam: North-Holland.

Day, Richard H., Gunnar Eliasson and Clas Wihlborg (eds) (1993), The Markets for Innovation, Ownership and Control, Stockholm: IUI; Amsterdam: North-Holland.

Day, Richard H., S. Morley and K.R. Smith (1974), ‘Myopic Optimizing and Rules of Thumb in a Micro-Model of Industrial Growth’, American Economic Review (March).

Demsetz, H. (1969), ‘Information and Effi ciency: Another Viewpoint’, Journal of Law and Economics, 12 (April), 1–22.

Eliasson, Åsa (2007), ‘The Markets for Bioterror and Biodisaster Protection’, prelim-inary manuscript for the study, ‘The Markets for Security Products and Disaster Preparedness’, Stockholm: Swedish Academy of Engineering Sciences (IVA).

Eliasson, Gunnar (1976), Business Economic Planning – Theory, Practice and Comparison, London, New York etc.: Wiley & Sons.

Eliasson, Gunnar (1977), ‘Competition and Market Processes in a Simulation Model of the Swedish Economy’, American Economic Review, 67 (1), 277–81.

Eliasson, Gunnar (1983), ‘On the Optimal Rate of Structural Adjustment’, in G. Eliasson, M. Sharefkin and B.-C. Ysander (eds) (1983), Policy Making in a Disorderly World Economy, Stockholm: IUI.

Eliasson, Gunnar (1984), ‘Micro Heterogeneity of Firms and Stability of Growth’, Journal of Behavior and Economic Organization, 5 (3–4), (Sept.–Dec.), 249–98 (also in R.H. Day and G. Eliasson (eds), 1986).

Eliasson, Gunnar (1985), The Firm and Financial Markets in the Swedish Micro-to-Macro Model – Theory, Model and Verifi cation, Stockholm: IUI.

Eliasson, Gunnar (1987a), Technological Competition and Trade in the Experiment-ally Organized Economy, Research Report No. 32, Stockholm: IUI.

Eliasson, Gunnar (1987b), ‘The Knowledge Base of an Industrial Economy’, in Eliasson and Ryan, The Human Factor in Economic and Technological Change, OECD Educational Monograph Series No. 3.

Eliasson, Gunnar (1990a), ‘The Firm as a Competent Team’, Journal of Economic Behavior and Organization, 13 (3), 275–98.

Eliasson, Gunnar (1990b), ‘The Knowledge-Based Information Economy’, in G. Eliasson et al. (1990), The Knowledge Based Information Economy, Stockholm: IUI, Chapter 1.

Eliasson, Gunnar (1991a), ‘Modeling the Experimentally Organized Economy’, Journal of Economic Behavior and Organization, 16 (1–2), 153–82.

Eliasson, Gunnar (1991b), ‘The International Firm: A Vehicle for Overcoming Barriers to Trade and a Global Intelligence Organization Diff using the Notion of a Nation’, in L.-G. Mattson and B. Stymne (eds) (1991).

Eliasson, Gunnar (1991c), ‘Financial Institutions in a European Market for Execu-tive Competence’, in C. Wihlborg, M. Fratianni and T.D. Willett (eds) (1992).

Eliasson, Gunnar (1992), ‘Business Competence, Organizational Learning, and Economic Growth: Establishing the Smith–Schumpeter–Wicksell (SSW)


Connection’, in F.M. Scherer and M. Perlman (eds), Entrepreneurship, Technological Innovation, and Economic Growth. Studies in the Schumpeterian Tradition, Ann Arbor: The University of Michigan Press.

Eliasson, Gunnar (1995), En teknologigenerator eller ett nationellt prestigeprojekt? – Exemplet svensk fl ygindustri (A technology generator or a national prestige project? – The Swedish aircraft industry), Stockholm: City University Press.

Eliasson, Gunnar (1996), Firm Objectives, Controls and Organization – the Use of Information and the Transfer of Knowledge within the Firm, Boston/Dordrecht/London: Kluwer Academic Publishers.

Eliasson, Gunnar (1997a), ‘General Purpose Technologies, Industrial Competence Blocs and Economic Growth’, in B. Carlsson (ed.) (1997).

Eliasson, Gunnar (1997b), ‘The Venture Capitalist as a Competent Outsider’, mimeo, INDEK, KTH, IEO R: 1997-06, Stockholm.

Eliasson, Gunnar (1998a), ‘Competence Blocs and Industrial Policy in the Knowledge Based Economy’, OECD Science, Technology, Industrial (STI) Revue.

Eliasson, Gunnar (2000), ‘Industrial Policy, Competence Blocs and the Role of Science in the Economic Development’, Journal of Evolutionary Economics, 10, 217–41.

Eliasson, Gunnar (2001a), ‘The Role of Knowledge in Economic Growth’, in Helliwell, John (ed.) (2001).

Eliasson, Gunnar (2001b), Advanced Purchasing, Spillovers, Innovative Pricing and Serendipitous Discovery, paper presented at the E.A.R.I.E. 2001 Conference in Dublin 30 Aug.–Sept.

Eliasson, Gunnar (ed.), (2005a), The Birth, the Life and the Death of Firms – the Role of Entrepreneurship, Creative Destruction and Conservative Institutions in a Growing and Experimentally Organized Economy, Stockholm: The Ratio Institute.

Eliasson, Gunnar (2005b), ‘The Nature of Economic Change and Management in a New Knowledge Based Information Economy’, Information Economics and Policy, 17, 428–56.

Eliasson, Gunnar (2005c), ‘The Venture Capitalist as a Competent Outsider’, in G. Eliasson (2005a) Chapter 4.

Eliasson, Gunnar (2005d), ‘Insourcing of Production from Foreign Subsidiaries or Subcontractors – an Empirical Study of Swedish Firms’, paper prepared for Invest in Sweden Agency (ISA), Stockholm; can be downloaded at www.isa.se/kostnadellerkompetens.

Eliasson, Gunnar (2006a), ‘From Employment to Entrepreneurship’, Journal of Industrial Relations, 48 (5), 633–56.

Eliasson, Gunnar (2006b), ‘Policies for a New Entrepreneurial Economy’, paper presented to the International J.A. Schumpeter Society 11th ISS Conference, Nice-Sophia Antipolis, 21–24 June.

Eliasson, Gunnar and Åsa Eliasson (1996), ‘The Biotechnological Competence Bloc’, Revue d’Economie Industrielle, 78 (4), 7–26.

Eliasson, Gunnar and Åsa Eliasson (2005), ‘The Theory of the Firm and the Markets for Strategic Acquisitions’, in Cantner, U., Dinopoulos, E. and Lanzilotti, R.F. (eds), Entrepreneurship: The New Economy and Public Policy, Berlin, Heidelberg and New York: Springer.

Eliasson, Gunnar and Åsa Eliasson (2007), ‘Competence in Health Care’, KTH, Stockholm, mimeo.

Eliasson, Gunnar and Erol Taymaz (2000), ‘Institutions, Entrepreneurship, Economic Flexibility and Growth – Experiments on an Evolutionary Model’,


in U. Cantner, H. Hanush and S. Klepper (eds) (2000), Economic Evolution, Learning and Complexity – Econometric, Experimental and Simulation Approaches, Heidelberg: Physica–Verlag.

Eliasson, Gunnar and Clas Wihlborg (2003), ‘On the Macroeconomic Eff ects of Establishing Tradability in Weak Property Rights’, Journal of Evolutionary Economics, 13, 607–32.

Eliasson, Gunnar, Fredrik Bergholm, Eva Christina Horwitz and Lars Jagrén (1985), De svenska storföretagen – en studie av internationaliseringens konsekven-ser för den svenska ekonomin (The Giant Swedish Groups – a Study of the Consequences of Internationalization for the Swedish Economy), Stockholm: IUI.

Eliasson, Gunnar, Dan Johansson and Erol Taymaz (2005), ‘Firm Turnover and the Rate of Growth’, in Eliasson (2005a), Chapter 6.

Fama, E.F. (1980), ‘Agency Problems and the Theory of the Firm’, Journal of Political Economy, 88 (2) (April), 288–307.

Frankelius, Per (2007), Linne i Nytt Ljus: Den första översättningen av Systema naturae samt ny analys av Linnes perspektiv (with a translation from Latin by Bertil Alden), Malmö: Liber.

Friedman, M. (1953), ‘The Methodology of Positive Economics’, in M. Friedman (ed.), Essays in Positive Economics, Chicago, IL: The University of Chicago Press.

Georgescu-Roegen, Nicholas (1950), ‘The Theory of Choice and the Constancy of Economic Laws’, The Quarterly Journal of Economics, 44, 125–38.

Granstrand, Ove and S. Sjölander (1990a), ‘Managing Innovation in Multi-Technology Corporations’, Research Policy, 19 (1) (Feb.), 36–60.

Granstrand, Ove and S. Sjölander (1990b), ‘The Acquisition of Technology and Small Firms by Large Firms’, Journal of Economic Behavior and Organization, 13 (3), 367–86.

Hanson, K.A. (1986), ‘On New Firm Entry and Macro Stability’, in The Economics of Institutions and Markets, IUI Yearbook 1986–1987, Stockholm: Industriens Utredningsinstitut (IUI).

Hanson, K.A. (1989), ‘Firm Entry’, in J.W. Albrecht et al., MOSES Code, Stockholm: Industriens Utredningsinstitut (IUI).

Hayek, Friedrich A. von (1937), ‘Economics and Knowledge’, Economica, 4, 33–54.

Hayek, Friedrich A. von (1940), ‘Socialist Calculation’, Economica, 7 (26), 125–49.Hayek, Friedrich A. von (1945), ‘The Use of Knowledge in Society’, American

Economic Review, 35 (4) (Sept.), 519–31.Helliwell, John (ed.) (2001), The Contribution of Human and Social Capital to

Sustained Economic Growth and Well-being, HRDC, Canada.Ilinitch, Anne, Richard D’Áveni and Arie Lewin (1996), ‘New Organizational

Forms and Strategies for Managing in Hypercompetitive Environments’, Organization Science, 7 (3), Special Issue 4, Part 1 of 2: Hypercompetition (May–June), 211–20.

Johansson, Dan (2001), ‘The Dynamics of Firm and Industry Growth – the Swed ish Computing and Communications Industry’, Doctoral thesis, Department of the Organization and Management, KTH, Stockholm.

Kirzner, Israel M. (1997), ‘Entrepreneurial Discovery and the Competitive Market Process: An Austrian Approach’, Journal of Economic Literature, 35 (1), 60–85.

Knight, F. (1921), Risk, Uncertainty and Profi t, Boston: Houghton-Miffl in.


Koopmans, Tjalling C. (1957), Three Essays on the State of Economic Science, New York: McGraw Hill Book Company Inc.

LeRoy, S.F. and L.D. Singell, Jr (1987), ‘Knight on Risk and Uncertainty’, Journal of Political Economy, 95 (2) (April), 394–406.

Lindh, T. (1993), ‘Lessons from Learning to have Rational Expectations’, in R.H. Day, G. Eliasson, C.G. Wihlborg (eds), The Markets for Innovation, Ownership and Control, Stockholm: IUI; Amsterdam, London, New York, Tokyo: North-Holland.

Malinvaud, E. (1967), ‘Decentralized Procedures in Planning’, in E. Malinvaud and M.O.L. Bacharach (eds) (1967).

Malinvaud, E. and M.O.L. Bacharach (eds) (1967), Activity Analysis in the Theory of Growth and Planning, London: Macmillan.

Marshall, Alfred (1890), Principles of Economics. London: Macmillan.Marshall, Alfred (1919), Industry and Trade, London: Macmillan.Mattson, L.-G. and B. Stymne (eds) (1991), Corporate and Industry Strategies for

Europe, Amsterdam: North-Holland.Menger, Carl (1871), Grundsätze der Volkswirtschaftslehre, Vienna: Wilhelm

Braumüller.Mises, Ludwig von (1949), Human Action, Chicago: Contemporary Books.MOSES Database (1992), see Albrecht et al. (1992).Nelson, R.R. and Winter, S.G. (1982), An Evolutionary Theory of Economic

Change, Cambridge, MA: Harvard University Press.Nyström, Kristina (2006), Entry and Exit in Swedish Industrial Sectors, Jönköping:

Jönköping International Business School Dissertation Series No. 032.OECD (1996), The Knowledge Based Economy, Paris.Pakes, Ariel and Richard Ericson (1998), ‘Empirical Implications of Alternative

Models of Firm Dynamics’, Journal of Economic Theory, 79 (1), 1–45.Pousette, Tomas and Thomas Lindberg (1986), ‘Tjänster i Produktionen och

Produktionen av Tjänster’, in Eliasson, Gunnar, Bo Carlsson, Enrico Deiaco, Thomas Lindberg and Tomas Pousette, Kunskap, Information och Tjänster – en studie av svenska industriföretag, Stockholm: IUI

Putnam, R. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster.

Ritzen, Jozef M. (2001), ‘Social Cohesion, Public Policy, and Economic Growth: Implications for OECD Countries’, in Helliwell (2001).

Romer, P.M. (1986), ‘Increasing Returns and Long-Run Growth’, Journal of Political Economy, 94 (5) (Oct.), 1002–37.

Romer, P.M. (1990), ‘Endogenous Technological Change’, Journal of Political Economy, 98, (5) pt. 2, S71–102.

Rothschild, M. (1974), ‘A Two-Armed Bandit Theory of Market Pricing’, Journal of Economic Theory, 9 (2) (Oct.), 185–202.

Rybsczynski, T.M. (1993), ‘Innovative Activity and Venture Financing: Access to Markets and Opportunities in Japan, the US and Europe’, in Day, Richard H., G. Eliasson and C. Wihlborg (eds) (1993), The Markets for Innovation, Ownership and Control, Stockholm: IUI, and Amsterdam: North-Holland.

Salter, W.E.G. (1960), Productivity and Technical Change, Cambridge, MA: Cambridge University Press.

Schumpeter, Joseph (1911) (English edition 1934), The Theory of Economic Development, Harvard Economic Studies, Vol. XLVI, Cambridge, MA: Harvard University Press.


Schumpeter, Joseph (1942), Capitalism, Socialism and Democracy, New York: Harper & Row.

Simon, Herbert A. (1945), Administrative Behavior, New York: Macmillan.Simon, Herbert A. (1955), ‘A Behavioral Model of Rational Choice’ Quarterly

Journal of Economics, 69, 99–118.Smith, Adam (1776), An Inquiry into the Nature and Causes of the Wealthy of

Nations, reprinted New York: Modern Library, 1937.Taymaz, Erol (1991), MOSES on PC: Manual, initialization, and calibration,

Stockholm: IUI.Wallis, John and Douglass North (1986), ‘Measuring the Transaction Sector in

the American Economy’, in Engerman, Stanley and Gallman (eds), Long Term Factors in American Economic Growth, Chicago: The University of Chicago Press.

Walras, L. (1874), Elements d’économie politique pure. English translation of 1926 edition: Elements of Pure Economics, or the Theory of Social Wealth, London: Allen and Unwin, 1954.

Wicksell, Knut (1923), ‘Realkapital och kapitalränta’, Ekononmisk Tidskrift, Häfte 5–6, 145–80.

Wihlborg, C., M. Fratianni and T.D. Willett (eds) (1992), Financial Regulation and Monetary Arrangements after 1992, Amsterdam: Elsevier Science Publishers B.V.

Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications: A Study in the Economics of Internal Organization, New York: Free Press.

Williamson, Oliver E. (1985), The Economic Institutions of Capitalism, New York/London: The Free Press.

Winter, S.G. (1986), ‘Schumpeterian Competition in Alternative Technological Regimes’, in Day, Richard H. and G. Eliasson (eds), The Dynamics of Market Economies, Stockholm: IUI, and Amsterdam: North-Holland, Chapter 8.

Wolfe, Barbara and Robert Haveman (2001), ‘Accounting for the Social and Non-Market Benefi ts of Education’, in Helliwell (2001).

Woolcock, Michael (2001), ‘The Place of Social Capital in Understanding Social and Economic Outcomes’, in Helliwell (2001).

PART III

Investments and the legal environment

139

7. Corporate governance and investments in Scandinavia – ownership concentration and dual-class equity structure*Johan E. Eklund

1. INTRODUCTION

In essence, the corporate governance system in a country is the institutional framework that supports the suppliers of fi nance to corporations and enables fi rms to raise substantial amounts of capital (Shleifer and Vishny, 1997).1 By protecting suppliers of capital and safeguarding property, sound governance systems facilitate mobilization and allocation of capital to useful investments. Corporate governance systems are of outmost impor-tance for the allocation of capital to its highest value use. It can be argued that the corporate governance system in a country determines the speed of structural change and economic development by aff ecting allocation and reallocation of capital. Therefore the crucial question is whether the cor-porate governance system induces managers of corporations to make good value enhancing investments decisions, or not. In particular, the ownership concentration and composition appear to matter for fi rm performance, as shown by Morck et al. (1988).2 This chapter looks at corporate governance and the rate of return on corporate investments in Scandinavia. The struc-ture of ownership and its eff ects on performance are examined.

Taking an outsider’s view of Scandinavia, the corporate governance sys-tems in the Scandinavian countries, Sweden, Finland, Norway and Denmark, arguably display more similarities than diff erences. The countries share a number of important features that unify them in comparison with other countries. It has, for example, been hypothesized that the common origin of the legal systems in Scandinavia is still refl ected in the quality of corporate governance (La Porta et al., 1997). Furthermore, Scandinavian fi rms are typi-cally controlled by a dominant owner and only a small minority of fi rms are characterized by dispersed ownership structure. Finally, the Scandinavian countries can also be said to have a common political orientation, with strong

140 Investments and the legal environment

social democratic traditions (for example, Högfeldt, 2004), which, according to Roe (2003), matters for corporate governance.

Such apparent homogeneity of the Scandinavian countries in combina-tion with the importance of well functioning corporate governance systems motivates a comparison of corporate returns and ownership structure in Scandinavia. The purposes of this chapter are therefore the following. First, the returns on investments made by the largest fi rms in Scandinavia are assessed. Second, the eff ect of ownership structure on investment decisions is examined as a factor explaining variation in performance and in returns on investments. Finally, and perhaps most importantly, the chapter analyses how deviations from the one-share-one-vote principle aff ect this ownership–performance relationship. Outright expropriation of corporate assets and investor funds by managers is likely to be small in developed economies, such as the Scandinavian ones. Over-investment in pursuit of ends other than profi t maximization and misallocation of assets is more likely to be a problem.

The chapter is organized in six sections. Relevant literature on invest-ments, corporate governance and ownership is reviewed in Section 2. In Section 3 the method is derived and the data are described. In Section 4, the return on corporate investments in Scandinavia is assessed. The fi fth section examines how ownership and the extensive use of dual-class shares aff ect investment decisions. Section 6 provides the conclusions.

2. CORPORATE CONTROL AND INVESTMENT

Neoclassical investment theory suggests that investments are expanded up to the point where the expected marginal rate of return equals the oppor-tunity cost of capital. This condition would be satisfi ed in a friction-free world without any informational asymmetries, agency problems or trans-action costs. Capital would fl ow automatically to the most effi cient use and thereby guarantee that welfare is maximized. However, in the modern corporation, with its separation of owners and fi nanciers from the manage-ment, there arises a set of agency problems that can cause investment deci-sions to deviate from what is predicted in neoclassical models (see Mueller (2003) for a review of investment theories).

Berle and Means (1932) were the fi rst to call attention to the potential agency costs.3 They argued that corporate ownership in large listed fi rms would become dispersed up to a point where professional managers would become unaccountable to the shareholders. Later, Jensen and Meckling (1976) provide a more theoretical underpinning to the linkages between agency costs and ownership structure. Jensen and Meckling analyse how the interests of utility maximizing owner-managers and minority shareholders

Corporate governance and investments in Scandinavia 141

diverge as ownership structure becomes more dispersed. Their basic argument is that the owner-manager will not bear the full cost of on-the-job consump-tion.4 Potential minority investors will realize this and subsequently the share price will refl ect the divergence of interests between owner-managers and minority shareholders. Arguably the confl ict of interests becomes more severe as the equity stake of owner-managers decreases. Jensen and Meckling (1976) argue that investors with high stakes will also have incentives to maximize fi rm value. This is referred to as the incentive eff ect. Hypothesis 1 is therefore:

H1 Ownership concentration will improve investment performance.

In this view, agency costs increase as ownership is diluted and becomes dis-persed. However, not all have seen the separation of ownership and control as a potential problem, where the counter-hypothesis is that control and ownership separation may improve allocation. Thorstein Veblen (1921), for example, argues that this separation would lead to the control being turned over from ‘monopoly’-seeking owners/businessmen to growth and effi ciency-seeking management. Veblen claims for example that if

industry were completely organized as a systemic whole, and were then managed by competent technicians with an eye single to maximum production of goods and services; instead of, as now, being manhandled by ignorant business men with an eye single to maximum profi ts; the resulting output of goods and serv-ices would doubtless exceed the current output by several hundred per cent. (Veblen, 1921)

Recognizing that owner-managers are also guided by utility maximiza-tion and not pure profi t maximization, Demsetz (1983) argues that it is not clear that diff usion of ownership will automatically have a detrimental eff ect. In fact, it has been argued that as the stake of owner-managers increases, so does their ability to misallocate resources (Stulz, 1988). This eff ect is referred to as the entrenchment eff ect (see Morck et al., 1988, and Stulz, 1988). Morck et al. (1988) fi nd a non-monotonic relationship between ownership and Tobin’s q. They fi nd that performance initially increases with ownership concentration, then declines and fi nally increases again, which is consistent with an entrenchment eff ect. McConnell and Servaes (1990) fi nd similar results.5 Expecting a managerial entrenchment eff ect leads to the second hypothesis:

H2 Ownership concentration will have a non-linear eff ect on performance.

The generality of the Berle and Means (1932) observation is, however, empirically challenged. Looking at ownership structure around the world,


most corporations have concentrated ownership and are controlled by families (Morck et al., 2005; La Porta et al., 1999). Faccio and Lang (2002) study the ownership in Europe and fi nd that corporations are predominantly controlled by families in continental Europe. This control is achieved without corresponding capital by means of primarily three diff erent control enhancing mechanisms (CEM): vote-diff erentiation of shares, pyramid ownership and cross-holdings. This means that the divi-sion between what Berle and Means (1932) call ‘nominal ownership’ and the corporate control is further enhanced by separating the capital stake and the voting power, making it possible for a small group of investors, often the founding family, to maintain control of the fi rm.

Burkart and Lee (2008) review the theoretical literature on one-share-one-vote arguing that there are both positive and negative eff ects associated with dual-class shares. Adams and Ferreira (2008) review the empirical literature and fi nd the empirical evidence inconclusive.6 Bebchuk et al. (1999), on the other hand, argue that these control mechanisms distort the incentives of the controlling owners and there-fore potentially may cause a sharp increase in agency costs. When the incentives are distorted, this may potentially have a negative impact on the optimal choice of investments, scope of the fi rm and transferral of control. Separation of control rights and cash-fl ow rights not only alters the control structure of the corporation but also changes the incentives of owner-managers. An eff ect one can expect from the separation of cash-fl ow and control rights is that the positive incentive eff ect will be weakened whereas the entrenchment eff ect will be enhanced. From this, hypothesis 3 follows:

H3 Control mechanisms such as dual-class equity structure will weaken the incentive and enhance the entrenchment eff ect.

Using a market-to-book measure of Tobin’s q, Claessens et al. (2002) fi nd evidence that is consistent with this hypothesis. They examine a large number of fi rms in East Asia and fi nd that cash-fl ow rights are positively correlated with performance. However, control rights in excess of cash-fl ow rights have a negative eff ect on fi rm value. A large number of studies also establish a link between ownership structure and concentration on the one hand and performance on the other. Countries with weaker inves-tor protection tend to have a more concentrated ownership structure (see for example La Porta et al., 1997). In fact, the two most common ways of dealing with the agency aspects of corporate governance are, according to Shleifer and Vishny (1997), fi rst, legal and regulatory protection of investor and minority rights, and second, large and concentrated owners.


2.1 Corporate Governance in Scandinavia

The corporate governance systems in Scandinavia have some unique fea-tures that change the prediction of the Jensen and Meckling model. Like most fi rms in continental Europe, the Scandinavian fi rms very often have controlling owners that have maintained their control even when their capital stake has declined and the fi rms have grown. Most European coun-tries allow at least one of the three principal instruments for enhancing ownership control: cross-holdings, pyramid ownership and vote-diff eren-tiation (Söderström et al., 2003).

In particular, the extensive use of vote-diff erentiated shares has had a substantial impact on the way in which the ownership structure has evolved in Scandinavia. In Norway about 14 percent of listed fi rms use dual-class shares, in Denmark and Finland more than 30 percent, and in Sweden it is as high as 55 percent (Bøhren and Ødegaard, 2006; Söderström et al., 2003). Many countries in Europe do not allow for dual-class share systems, so this is one of the prominent distinguishing features of the corporate governance systems in Scandinavia. The frequent use of dual-class shares, with strong separation of voting rights and equity claims, has produced very strong and stable ownership structures in Scandinavia (Högfeldt, 2004; Henrekson and Jakobsson, 2006). By using vote-diff erentiation, the founding families may retain control of fi rms even with a very small equity share. Most fi rms in Scandinavia have a single controlling owner and very few fi rms are characterized by dispersed ownership. Bennedsen and Nielsen (2005) report signifi cant diff erences in the frequency of control mechanisms for a sample of 4096 European fi rms (see Table 7.1).

Cronqvist and Nilsson (2003) examine a large sample of Swedish listed fi rms and fi nd that controlling owners have a negative eff ect on Tobin’s average q. These controlling owners are also more likely to use control mech-anisms. Maury and Pajuste (2004) examine a sample of Finnish fi rms and show that a more uniform distribution of votes among large block holders is positive for fi rm valuation. They also fi nd that divergence between cash-fl ow rights and control rights have a negative eff ect on fi rm value. An additional consequence of the strong separation of ownership claims and control is that the so-called market for corporate control (Manne, 1965) virtually does not exist in Scandinavia. Successful hostile bids are therefore very rare.

Supposed advantages of strong and stable owners provide the under-pinning argument for the Scandinavian legislation that allows for vote-diff erentiation of share and pyramid ownership. In this chapter, ownership concentration is measured as the share of capital and votes controlled by the largest owner (CR1 & VR1) and the fi ve largest owners (CR5 & VR5). About 40 percent of the fi rms in the aggregate Scandinavian sample


separate control and cash-fl ow rights; see Table 7.2. The ownership data have been collected from the annual reports for each fi rm.

Ownership concentration is very high in Scandinavian listed fi rms, espe-cially compared with those in the Anglo-Saxon countries. Demsetz and Lehn (1985) examine the ownership structure in 511 large US fi rms. They report that, on average, the fi ve largest owners together hold 24.8 percent and the top 20 shareholders 37.7 percent. Frequently 20 percent is assumed to be more than enough to control a fi rm (Morck et al., 2005).

La Porta et al. (1997) have hypothesized that the legal origin of a country determines the effi ciency of the country’s fi nancial system. Scandinavia can in this respect be regarded as being relatively homoge-neous. Scandinavia has a long tradition of cooperation in drafting new legislation (Carsten, 1993). Interestingly, there are still important diff er-ences with respect to deviations from the one-share-one-vote principle. Denmark, Finland and Sweden all allow dual-class shares. In Norway deviations from the proportionality principle need government approval (Faccio and Lang, 2002).

Table 7.1 Corporate control mechanisms in European countries

Dual-class shares Pyramid structures Cross-holdings

Sweden 0.62 0.27 0.01Switzerland 0.52 0.06 0.00Finland 0.44 0.07 0.00Italy 0.43 0.25 0.00Denmark 0.29 0.17 0.00UK 0.25 0.22 0.00Ireland 0.25 0.18 0.00Austria 0.23 0.26 0.01Germany 0.19 0.24 0.03Norway 0.11 0.33 0.02France 0.03 0.15 0.00Belgium 0.00 0.27 0.00Portugal 0.00 0.13 0.00Spain 0.00 0.16 0.00European average 0.24 0.20 0.01Scandinavian average 0.37 0.21 0.01

Note: The fi gures represent the percentage of fi rms that use dual-class shares, pyramid structures and cross-holdings, respectively.

Source: Bennedsen and Nielsen (2005).


3. METHODOLOGY

This chapter applies a method developed by Mueller and Reardon (1993) to assess the rate of return on investments. The measure produced is a marginal version of Tobin’s q. Tobin’s q is defi ned as the market value of a fi rm over the replacement cost of its assets, which translates to the average return on total assets. The marginal version of Tobin’s q, on the

Table 7.2 Ownership concentration in Scandinavia (2004)

All fi rms

Mean Std. dev. Min Max No. fi rms Skewness

Capital share one owner, CR 1

23.5 15.5 0.4 82.4 214 0.90

Capital share fi ve owners, CR 5

44.8 19.6 1.5 95.1 214 0.33

Voting rights one owner, VR 1

29.4 19.7 0.4 89.3 211 0.89

Voting rights fi ve owners, VR 5

52.0 22.6 1.5 96.5 211 0.08

Vote-diff erentiated fi rms



23.5 13.7 2.9 60.4 90 0.70


47.4 19.0 9.4 93.8 90 0.43


35.8 20.3 4.6 89.3 88 0.73


64.8 19.8 18.6 96.5 87 −0.33

Firms with one-share-one-vote



23.2 16.7 0.4 82.4 124 1.01


42.9 19.9 1.5 95.1 124 0.32


23.2 16.7 0.4 82.4 124 1.01


42.9 19.9 1.5 95.1 124 0.32


other hand, measures the return on investments, or the marginal return on capital relative to the cost of capital (Mueller, 2003). This is in eff ect a measure of what Tobin (1982) calls the ‘functional form’ of stock market effi ciency.7 Marginal q is also a more appropriate measure of performance since average q contains infra-marginal returns.8

Marginal q can be derived from the simple insight that any investments should ex ante be evaluated against the discounted present value of future cash fl ows that the investment generates. Obviously, only projects that have a positive net present value should be carried out. Consider an invest-ment, It, made by a fi rm in period t. This investment generates cash fl ows, CFt1j in j periods. The present value, PVt, of this cash fl ow is as follows:

PVt 5 an

j51CFt1 j/ (1 1 rt)

j (1)

where rt is the discount rate. Note that the present value is the discounted expected value of future cash fl ows. This equation can be expressed in the fol-lowing way, where it can be regarded as a quasi-permanent rate of return:

PVt 5 Itit/rt (2)

For investments to be effi cient from a shareholder perspective the invest-ment being considered must generate future cash fl ows which, discounted to the present value, equal or exceed the investment cost.

The ratio i/r is essentially a marginal version of Tobin’s q (Mueller, 2003) which measures the return on a marginal investment, and will there-fore henceforth be referred to as qm. Equation (2) can be rearranged and expressed as follows:

PVt

It5 it/rt 5 qm,t (3)

For investments to be meaningful, we must have that PVt $ It. This implies that qm $ 1. If fi rms are investing at qm 5 1, investments are effi cient. This implies that there are no further profi table investment opportunities (see Figure 7.1). Whereas if qm , 1, fi rms are receiving a return on their invest-ments that is less than the cost of capital, which can only be interpreted as over-investment and a managerial failure of some sort.

At the end of period t the market value of a fi rm may be decomposed into the market value in period t 2 1 (Mt-1), the present value of investments made in period t (PVt), the change in market value of the old capital stock (dt), and an error term for the errors the market may make in its evaluation of the fi rm (mt).9

Mt ; Mt21 1 PVt 2 dtMt21 1 mt (4)


By replacing Mt-1 in equation (4) in each subsequent period, the following expression is obtained:

Mt1n 5 Mt21 1 an

i50PVt1 i 2 a

n21

150dt1 iMt1 i 1 a

n

i50mt1 i (5)

In a single period, the error in the market’s evaluation of the fi rm can be sub-stantial, assuming effi cient markets: E(mt) 5 0 and E(mt, mt21) 5 0, which implies E(g n

i50mt1 i) 5 0. Thus, as n grows the last term will approach zero. From equation (3) we get the following expression:

qm 5

an

i50qm,t1 iIt1 i

an

i50It1 i

5

an

i50PVt1 i

an

i50It1 i

(6)

Using equation (5) this expression can be formulated in the following way:

qm 5(Mt1n 2 Mt21)

an

i50It1 i

1

an

i50dt1 iMt1 i21

an

i50It1 i

2

an

i50mt1 i

an

i50It1 i

(7)

This can be used to calculate a weighted average qm for each fi rm.10

Assuming that qm and d both are constant over time and across fi rms, we can use equation (4) to estimate qm and d directly. Taking equation (4) and subtracting Mt-1 from both sides we get:

Itqm* = 1

r, i

i

r

qm > 1 > qm* qm< 1 < qm*

Figure 7.1 Marginal rate of return on capital, i, cost of capital, r, and marginal q


Mt 2 Mt21 5 2dMt21 1 qmIt 1 mt (8)

Dividing by Mt-1 we normalize the equation and get the following relation-ship that can be empirically estimated:

Mt 2 Mt21

Mt215 2d 1 qm

It

Mt211

mt

Mt21 (9)

Mueller and Reardon’s (1993) methodology can be applied to test the agency hypotheses. In contrast to the average Tobin’s q, this method meas-ures the marginal return on investments, which makes it more appropriate when testing the agency hypotheses.

To study the eff ects of ownership structure or various institutional factors on investment decisions, measures of ownership may be added as interac-tion terms with It/Mt21 in equation (9). If interaction terms are added, the functional form will be: Y 5 a 1 b1X 1 b2XZ, and qm is the economic interpretation of the marginal eff ect, �Y/�X 5 b1 1 b2Z. This method has been applied by Gugler and Yurtoglu (2003) and by Bjuggren et al. (2007). The equations estimated have the following functional form:

Mt 2 Mt21

Mt215 2d 1 b1

It

Mt211 b2Z1

It

Mt211 . . . 1 bi11Zi

It

Mt211 ei

(10)

where the Z’s denote the explanatory variables. Thus, the marginal eff ect, qm, of equation (10) is:

qm 5 b1 1 b2Z1 1 . . . 1 bi11Zi (11)

The total market value of a fi rm is defi ned as the total number of out-standing shares times the share price at the end of year t, plus total debt. Investments are approximated as:

I 5 After tax profi t – Dividends 1 Depreciation 1 DEquity

1 DDebt1 R&D 1 Advertising & Marketing

The market and accounting data have been collected from Compustat Global database.11 The fi rms included were listed at one of the four stock exchanges in Scandinavia (Copenhagen Stock Exchange in Denmark, Helsinki Stock Exchange in Finland, Oslo Stock Exchange in Norway and Stockholm Stock Exchange in Sweden) between 1998 or 1999 and 2005, in total 292 fi rms (2004 observations). All fi gures have been adjusted by


harmonized consumer price indexes to 2005 constant prices. The indexes used have been compiled by Eurostat. Naturally, the standard caveats apply to the data.

To use equation (7) to calculate qm, it is also necessary to determine the size of the deprecation rate d. That is, the rate at which the value of the fi rm’s assets is declining over time. According to Mueller and Reardon (1993), most estimates are around 10 percent. Naturally, the actual depre-ciation rate varies across fi rms and industries, depending on the durability of employed assets. Even within fi rms, we have reason to believe that the depreciation rate diff ers across the capital stock.

Equation (9) has the advantage that no assumption regarding the size of d is necessary. In an empirical estimation of equation (9) the intercept (d) will capture the depreciation rate plus any systematic changes in market valuations of the stock of old capital. The estimated d has no bearing on the interpretation of qm.

4. CORPORATE RETURN IN SCANDINAVIA

This study covers 292 large Scandinavian fi rms that are listed at one of the four stock exchanges. This accounts for about 40 percent of all listed fi rms. In 2004, the top 100 of these 292 fi rms (25 largest in each country) accounted for approximately 42 percent of the total stock market capitali-zation (33 percent of GDP).12 The fi rms approximately follow a rank–size distribution, where the second largest fi rm is about half the size of the largest.13

As a fi rst step, equation (7) is used to calculate a qm for each individual fi rm. For Scandinavia, the estimated average marginal q, excluding the upper 95 percentile and the lower 5 percentile, is 1.19. This means that during the period 1999–2005 the Scandinavian fi rms had an average return on investments that was 19 percent above the cost of capital. However the median qm is 1.03, which implies a return that is 3 percent above cost of capital. Neither the average qm nor the median qm give any reason to believe that Scandinavian fi rms are under-performing. This is based on the assumption that the depreciation rate was 10 percent per annum. Equation (7) is sensitive to the choice of depreciation rate. Consequently, a more rapid deprecation will translate into a higher qm, all else equal.

Investments as defi ned in this chapter can be negative. This will be the case if a fi rm is making losses that are larger in absolute terms than new equity and debt. It is not meaningful to ask what the returns on investment are if investments are negative. Nor does equation (7) make any sense when


Table 7.3 Cumulative distribution of marginal q

Denmark

Range of qm 1999 2000 2001 2002 2003 2004 2005

qm ≥ 2.00 10 9 6 2 4 5 51.50 ≤ qm < 2.00 3 2 3 4 3 4 81.00 ≤ qm < 1.50 3 3 3 5 3 6 60.50 ≤ qm < 1.00 10 14 12 9 11 13 130.00 ≤ qm < 0.50 11 19 23 26 25 23 22−0.50 ≤ qm < − 0.00 9 5 5 8 8 5 3−1.00 ≤ qm < − 0.50 4 4 1 0 2 1 0qm < − 1.00 4 4 6 6 4 4 2Number of fi rms 54 60 59 60 60 61 59Number of qm ≥ 1 16 14 12 11 10 15 19Number of qm < 1 38 46 47 49 50 46 40

Finland

Range of qm 1999 2000 2001 2002 2003 2004 2005

qm ≥ 2.00 20 17 9 11 13 12 151.50 ≤ qm < 2.00 5 1 9 5 5 4 41.00 ≤ qm < 1.50 9 6 2 9 10 16 220.50 ≤ qm < 1.00 8 11 17 16 15 11 90.00 ≤ qm < 0.50 3 13 10 8 8 9 3−0.50 ≤ qm < − 0.00 3 3 4 1 1 1 1−1.00 ≤ qm < − 0.50 0 3 2 3 1 2 2qm < − 1.00 3 4 5 5 6 4 3Number of fi rms 51 58 58 58 59 59 59Number of qm ≥ 1 34 24 20 25 28 32 41Number of qm < 1 17 34 38 33 31 27 18

Norway

Range of qm 1999 2000 2001 2002 2003 2004 2005

qm ≥ 2.00 23 19 13 8 12 17 231.50 ≤ qm < 2.00 5 6 3 4 6 5 71.00 ≤ qm < 1.50 7 10 15 10 13 15 90.50 ≤ qm < 1.00 5 3 7 15 7 3 20.00 ≤ qm < 0.50 0 4 2 3 4 4 2−0.50 ≤ qm <− 0.00 1 3 1 2 0 0 0−1.00 ≤ qm < − 0.50 1 0 3 0 0 0 1qm < − 1.00 2 2 3 6 4 1 1Number of fi rms 44 47 47 48 46 45 45Number of qm ≥ 1 35 35 31 22 31 37 39Number of qm < 1 9 12 16 26 15 8 6


investments are negative or equal to zero. Accordingly, such fi rms have been excluded from Table 7.3.

As can be seen from Table 7.3, returns on investments are approximately normally distributed around a mean of 1 in all of the four Scandinavian countries, except Norway. As the estimated qms are cumulated over 1999 to 2005 the distribution seems to become more concentrated around 1.

There are a few extreme values that have a large impact on the average qm across fi rms. These are typically smaller fi rms that, for some reason, have either a very high return on invested capital or a massive loss in market value. There are several plausible explanations for these extreme values. Firms may for example introduce radical innovations that do not require any substantial investments but nevertheless substantially increase fi rm value. Average qm is, for this reason, also calculated excluding 5 percent in both ends of the distribution.

Dropping 5 percent in both ends of the distribution, the average qm is 0.76 for Denmark, 1.27 for Finland, 1.83 for Norway and 1.11 for Sweden. The median qm is 0.57 for Denmark, 1.18 for Finland, 1.86 for Norway and 0.85 for Sweden. If the assumption that d 5 10 percent is approximately correct, this means that all four of the Scandinavian countries, with the exception of Denmark, have average returns equal to or above the cost of capital.

Bjuggren and Wiberg (2008) fi nd that qm is sensitive to stock market swings and that, depending on period selection, the qm may either be over- or under-estimated. The choice of a 5–6 year period, which approximately

Table 7.3 (continued)

Sweden

Range of qm 1999 2000 2001 2002 2003 2004 2005

qm ≥ 2.00 36 26 14 5 12 15 181.50 ≤ qm < 2.00 4 3 7 2 5 5 121.00 ≤ qm < 1.50 12 16 9 11 12 15 140.50 ≤ qm < 1.00 4 20 25 27 26 28 280.00 ≤ qm < 0.50 23 26 31 36 32 30 29−0.50 ≤ qm < − 0.00 6 4 9 6 9 7 3−1.00 ≤ qm < − 0.50 0 4 3 6 3 3 1qm < −1.00 3 8 10 13 10 6 2Number of fi rms 88 107 108 106 109 109 107Number of qm ≥ 1 52 45 30 18 29 35 44Number of qm < 1 36 62 78 88 80 74 63

Note: Assuming d 5 10 percent.


coincides with the average length of a business cycle, reduces this problem.

Table 7.4 reports the value of marginal q for the ten largest fi rms in each Scandinavian country. The fi rst two columns report the total market values in 2005 and 1998 adjusted to 2005 constant prices. Total investments made during this period are reported in column 3. Since equation (7) is sensitive to choice of depreciation, qm has been calculated assuming 5, 10 and 15 percent depreciation of old capital (columns 4, 5 and 6). Furthermore, the implicit d can be calculated from equation (7) by assuming that qm 5 1. This implicit depreciation rate is reported in column 7.

A few fi rms in Table 7.4 have implicit deprecation rates that are nega-tive, which indicates that these fi rms all had returns in excess of their cost of capital. The dominant fi rm in Finland, Nokia, for example, has performed well over a long period and consequently has a qm around or slightly above 1. This can be compared with the Swedish telecom fi rm Ericsson, one of Nokia’s main competitors. Ericsson seems to have a lower qm given any depreciation rate, but remains approximately equal to 1. It is plausible to assume that the diff erences in returns can be attributed to dif-ferences in performance, since Nokia and Ericsson can be assumed to have approximately the same depreciation rate. The dominant fi rm in Denmark, Möller-Maersk, with its high marginal q, appears to be under-investing. Finally, the dominating Norwegian fi rm Norsk Hydro seems to have a marginal q approximately equal to 1.

Assuming that the marginal rate of return (qm) and the depreciation rate (d) are the same across companies and over time, these can be esti-mated by equation (9). Since the data consist of a cross-sectional time series, a fi xed eff ect model is used (industry and time fi xed eff ects model). The stock market may fail to make a correct valuation in a single period, but assuming effi cient markets, this error will approach zero as the time span increases. To take the possibility of market errors into account, time dummies were used in the estimations. Both industry and time dummies are restricted to sum to zero, so that the eff ects are measured as the devia-tion from the average depreciation rate. The results are reported in Table 7.5.

In order to remove outliers, some of the observations have been removed from the data set. The absolute deviation between the dependent variable and the explanatory variable, |(Mt – Mt-1)/Mt-1 – It/Mt-1|,14 has been used to identify outliers. Observations that had an absolute deviation above 2 (41 observations) were removed. This captures, for example, fi rms that have large swings in market value without corresponding changes in invest-ments. The excluded fi rms are predominantly found among relatively small high-tech fi rms within the biotechnology and ICT sectors. Bjuggren

153

Tab

le 7

.4

Ten

larg

est c

ompa

nies

in e

ach

Scan

dina

vian

cou

ntry

(20

05)

12

34

56

7

Com

pany

M

2005

aM

1998

a∑

INV

aq m (d

5 5

%)b

q m (d 5

10%

)bq m (d

5 1

5%)b

dc

Den

mar

kA

.P. M

ÖL

LE

R –

MA

ER

SK47

702.

972

11.4

2743

3.4

1.66

81.

860

2.05

2−

0.12

4T

DC

13

932.

115

676.

012

303.

00.

244

0.63

01.

016

0.14

8N

OV

O N

OR

DIS

K13

124.

187

52.7

1715

0.9

0.48

10.

706

0.93

20.

048

CA

RL

SBE

RG

68

42.0

4706

.619

592.

30.

199

0.28

90.

378

0.11

3H

. LU

ND

BE

CK

39

04.8

2607

.8d

5400

.00.

515

0.79

01.

065

0.13

8D

AN

ISC

O

5105

.637

28.0

6866

.80.

395

0.59

00.

784

0.20

5W

ILL

IAM

DE

MA

NT

HO

LD

ING

31

65.8

962.

123

03.6

1.28

81.

620

1.95

10.

007

CO

LO

PLA

ST

2486

.813

26.3

1998

.20.

869

1.15

71.

444

0.07

3C

OPE

NH

AG

EN

AIR

POR

TS

2479

.614

41.4

811.

01.

870

2.46

03.

049

−0.

024

DE

SA

MM

EN

SLU

TT

EN

DE

2411

.329

2.3

1215

.61.

972

2.20

12.

430

−0.

162

Fin

land

NO

KIA

6486

1.2

6844

5.3

1496

6.5

0.61

91.

293

1.96

60.

078

STO

RA

EN

SO

1510

8.7

1326

8.3

3364

.90.

690

1.21

41.

739

0.08

0U

PM-K

YM

ME

NE

1396

4.4

1136

1.0

1818

.20.

650

1.07

31.

495

0.09

1M

ET

SO40

61.7

1415

.516

770.

90.

429

0.53

70.

644

0.31

5SA

NO

MA

-WSO

Y40

50.2

789.

665

8.2

2.07

22.

517

2.96

3−

0.07

0M

-RE

AL

3914

.629

33.6

1602

7.0

0.47

60.

752

1.02

90.

014

RA

UT

AR

UU

KK

I33

04.8

1953

.914

06.0

1.80

62.

396

2.98

6−

0.01

8W

AR

TSI

LA

2752

.516

62.5

1287

.01.

770

2.38

12.

993

−0.

013

TIE

TO

EN

AT

OR

2739

.722

98.9

2135

.40.

968

1.61

52.

263

0.05

2Y

IT C

OR

P25

89.4

409.

866

2.4

3.53

03.

902

4.27

5−

0.28

9

154

Tab

le 7

.4

(con

tinue

d)

12

34

56

7

Com

pany

M

2005

aM

1998

a∑

INV

aq m (d

5 5

%)b

q m (d 5

10%

)bq m (d

5 1

5%)b

dc

Nor

way

NO

RSK

HY

DR

O24

942.

512

014.

919

283.

60.

986

1.30

21.

617

0.05

2O

RK

LA

9539

.457

29.7

3582

.91.

698

2.33

22.

967

−0.

005

NO

RSK

E S

KO

GIN

DU

STR

IER

5043

.323

55.1

1477

.61.

573

1.59

91.

624

−1.

050

HA

FSL

UN

D31

35.1

1156

.518

23.6

1.44

81.

811

2.17

4−

0.01

2F

RE

D. O

LSE

N E

NE

RG

Y

2152

.964

1.2

610.

73.

019

3.56

34.

107

−0.

136

SCH

IBST

ED

1900

.911

21.6

495.

62.

462

3.35

24.

241

−0.

032

DN

O18

67.4

46.8

474.

23.

971

4.10

44.

236

−1.

071

TO

MR

A S

YST

EM

S10

56.9

1399

.733

6.8

3.67

74.

130

4.58

4−

0.24

5F

AR

STA

D S

HIP

PIN

G95

0.9

263.

679

1.9

1.11

11.

355

1.59

80.

027

Swed

enE

RIC

SSO

N

4936

7.6

4966

1.1

6241

1.5

0.35

50.

714

1.07

30.

140

VO

LV

O24

244.

618

240.

932

494.

60.

376

0.56

80.

760

0.21

3H

& M

HE

NN

ES

& M

AU

RIT

Z21

244.

015

067.

114

736.

60.

825

1.23

21.

638

0.07

1A

TL

AS

CO

PCO

1259

9.3

5329

.911

514.

70.

850

1.06

91.

287

0.08

4SC

A11

652.

275

53.9

8847

.40.

860

1.25

71.

654

0.06

8SA

ND

VIK

1103

8.2

5502

.112

743.

40.

648

0.86

21.

076

0.13

2SC

AN

IA92

62.2

6128

.111

352.

30.

528

0.78

01.

032

0.14

4E

LE

CT

RO

LU

X74

26.0

9994

.222

765.

30.

019

0.15

00.

281

0.42

4SE

CU

RIT

AS

6787

.254

21.3

7191

.60.

525

0.86

01.

195

0.12

1SK

F59

40.4

2348

.379

75.8

0.59

10.

731

0.87

20.

196

SOL

STA

D O

FF

SHO

RE

623.

313

9.2

675.

00.

922

1.11

11.

301

0.07

1

Not

e:

a Mill

ion

euro

s, 20

05 c

onst

ant p

rices

; b qm

cal

cula

ted

assu

min

g th

at d

is 5

, 10

and

15 p

erce

nt, r

espe

ctiv

ely;

c dep

reci

atio

n ra

te c

alcu

late

d gi

ven

qm 5

1; d m

arke

t val

ue 1

999.


and Wiberg (2008) have shown that the marginal q measure is sensitive to swings in valuation of new high-tech fi rms.

The regressions in Table 7.4 were estimated with diff erent intercepts, d, for the diff erent countries; these were, however, insignifi cant and were therefore dropped out of the regression. In order to test for country eff ects, country dummy variables were interacted with It/Mt-1. These too were estimated under the restriction to sum to zero so that the country eff ects measure the deviation from the average Scandinavian marginal q. The Scandinavian average reported in Table 7.5 is signifi cantly below 1. Marginal q is 0.66 for Denmark, 1.07 for Norway, 0.93 for Finland and 0.77 for Sweden. These fi ndings seem to corroborate previous estimates of marginal q for the Scandinavian countries.

In a large cross-country study, Gugler et al. (2002) found similar estimates for Scandinavia. Between 1985 and 2000, they estimated 0.65 for Denmark, 0.96 for Finland, 1.04 for Norway and 0.65 for Sweden. Bjuggren et al. (2007) have also estimated an average qm of 0.65 for Sweden. The fi ndings reported in Table 7.4 are, in other words, consistent with previous estimates for Finland and Norway. Gugler et al. (2002) have esti-mated the Scandinavian average at 0.78. Their fi ndings support the legal

Table 7.5 Average qms in Scandinavia (1999–2004)

Dependent variable: (Mt 2 Mt-1)/Mt-1

Constant (5 d) −0.034** (−2.25)

−0.039** (−2.52)

It / Mt-1 0.868*** (27.79)

0.794*** (34.06)

Denmark* It / Mt-1 −0.205*** (−5.32)

Norway* It / Mt-1 0.244*** (4.87)

Finland* It / Mt-1 0.057(0.88)

Sweden* It / Mt-1 −0.097***(−2.85)

No. obs. 1963 1963No. fi rms 292 292R2 0.48 0.47R2-adjusted 0.46 0.45F-value 32.45 32.78

Note: *** indicates signifi cance at 1 percent, ** at 5 percent and * at 10 percent level; t-values in brackets.


origin hypothesis. Anglo-Saxon countries perform best with qm 5 1.02. Average qm for Germanic and French origin is 0.74 and 0.59, respectively.

However, there is considerable variation in the returns in all four Scandinavian countries, where a large number of fi rms deviate from the average marginal return on investments. This can have several causes; it might for example be plausible to believe there are industry diff erences. This is supported by the variation of the implicit deprecation rates in Table 7.4.

In the following section the relationship between ownership concentra-tion, separation of cash-fl ow rights and control rights, and performance is examined.

5. CORPORATE RETURN AND OWNERSHIP STRUCTURE

In this section, equation (10) is used to test the eff ects of ownership concen-tration and separation of control from cash-fl ow rights on performance. As measures of ownership concentration, the share of capital (cash-fl ow rights) held by the largest owner (CR1) and the fi ve largest (CR5) are used. Control rights are measured by the share of votes (control rights) held by the largest (VR1) and fi ve largest owners (VR5). Dummies are used to control for dual-class shares. In the sample, 49 percent of the fi rms use a dual-class share structure. Matching accounting and market data with the ownership data leaves 142 fi rms out of 292. Correlations are reported in Table 7.6.

Naturally all ownership variables display high and signifi cant correla-tions. Sales are negatively correlated with all ownership variables, but weaker for VR1 and VR5 than for CR1 and CR5. In other words, owner-ship concentration measured by cash-fl ow rights is inversely related to fi rm size. This means that controlling owners remain in large fi rms by resorting to dual-class equity structure. It is also interesting to note that investments are signifi cantly correlated with control rights and vote-diff erentiation, but not with cash-fl ow rights.

In order to control for unobserved, time-invariant heterogeneity across fi rms, a fi xed eff ect model with fi rm and time eff ects is applied. The time fi xed eff ect is motivated by the effi cient markets hypothesis; a fi rm may, in any single period, be under- or over-valued but over time this error is expected to be zero. The fi rm fi xed eff ect controls for diff erences in depre-ciation rates across fi rms and industries.

To identify non-linear eff ects on performance, the ownership variables are also estimated in quadratic and cubic form. In Tables 7.7a and b, merely


3 out of 24 estimated ownership parameters are signifi cant. However, a deviation from one-share-one-vote creates large negative eff ects. Firms with only a single class of equity do not signifi cantly under-perform, that is, qm is not diff erent from 1, whereas fi rms that rely on dual-class equity shares on average have a return on dual-class shares that is 30 percent below the opportunity cost of capital.

The fact that vote-diff erentiation has a signifi cant negative eff ect on fi rm performance indicates that the ownership–performance relationship may diff er between fi rms with one class of shares and those having separated cash-fl ow rights and control rights. This negative eff ect increases in equa-tions A to G when the ownership variables are added. One possible inter-pretation is that the ownership variables are picking up a positive incentive eff ect. This, in turn, suggests that the ownership eff ects diff er between the two categories of fi rms.

In Table 7.8 ownership variables are interacted with the dummy variable for dual-class share structure (1 for vote-diff erentiation and zero for single-class share structure). Diff erent specifi cations of the functional form have been estimated. The results are relatively robust with respect to choice of fi xed eff ect, random eff ect or simply pooled OLS model. A previous study has also found estimates of qm to be stable to model specifi cation (Bjuggren et al., 2007).

The results are robust with respect to the choice between simple pooled OLS with year dummies, fi xed eff ect model with year and fi rm eff ects, and random eff ects model. The estimates are robust with respect to model spec-ifi cation. Since the number of fi rms with available ownership data is limited

Table 7.6 Correlation matrix

Sales It/Mt-1 Mt-Mt-1/Mt-1 CR1 CR5 VR1 VR5

Sales 1It/Mt-1 0.012 1Mt-Mt-1/Mt-1 −0.043 0.422* 1CR1 −0.088* 0.069 0.033 1CR5 −0.224* 0.068 0.022 0.847* 1VR1 −0.031 0.118* 0.019 0.812* 0.710* 1VR5 −0.119* 0.102* 0.014 0.678* 0.835* 0.817* 1Vote-diff eren-tiation

0.082* 0.071* −0.049 −0.053 0.033 0.310* 0.422*

Note: * indicates signifi cance at 5 percent.

158

Tab

le 7

.7a

Con

cent

ratio

n of

cas

h-fl o

w ri

ghts

and

per

form

ance

Dep

ende

nt v

aria

ble:

(Mt-M

t-1)/

Mt-

1

Equ

atio

n A

Equ

atio

n B

Equ

atio

n C

Equ

atio

n D

Equ

atio

n E

Equ

atio

n F

Equ

atio

n G

Con

stan

t, (5

d)−

0.08

8***

(−3.

82)

−0.

088*

** (−

3.84

)−

0.08

7***

(−3.

78)

−0.

087*

** (−

3.78

)−

0.08

7***

(−3.

80)

−0.

088*

** (−

3.82

)−

0.08

7***

(−3.

79)

I t/M

t-1

0.9

29**

* (1

4.95

) 0

.982

***

(9.4

1) 0

.711

***

(4.6

1) 0

.734

***

(3.2

7) 0

.853

***

(6.6

8) 0

.948

***

(4.1

4) 0

.583

(1.5

8)D

ual-c

lass

Sha

res

−0.

312*

** (−

3.67

)−

0.32

7***

(−3.

71)

−34

0***

(−3.

86)

−0.

335*

** (−

3.55

)−

0.30

7***

(−3.

60)

−0.

304*

** (−

3.56

)−

0.34

4***

(−3.

78)

CR

1−

0.00

2 (−

0.63

) 0

.026

** (2

.16)

0.0

22 (0

.69)

CR

12−

0.00

1**

(−2.

38)

−0.

000

(−0.

27)

CR

13−

0.00

0(−

0.14

)C

R5

0.0

02 (0

.68)

−0.

003

(−0.

33)

0.0

31(1

.07)

CR

520.

000

(0.5

0)−

0.00

1(−

1.17

)C

R53

0.0

00 (1

.27)

No.

obs

.79

479

479

479

479

479

479

4N

o. fi

rms

142

142

142

142

142

142

142

F-v

alue

12.6

912

.39

12.3

212

.03

12.4

012

.11

11.8

8R

20.

420.

420.

420.

420.

420.

420.

42A

vera

ge q

m0.

785

0.79

00.

894

0.89

00.

796

0.76

90.

702

Dua

l-cla

ss q

m0.

617

0.61

40.

711

0.71

00.

630

0.60

50.

516

Sing

le-c

lass

qm

0.92

90.

941

1.05

11.

044

0.93

70.

909

0.86

0

Not

e:

*, *

* an

d **

* in

dica

te si

gnifi

canc

e at

10,

5 a

nd 1

per

cent

resp

ectiv

ely;

t-va

lues

in b

rack

ets.

159

Tab

le 7

.7b

Con

cent

ratio

n of

con

trol

/vot

ing

righ

ts a

nd p

erfo

rman

ce

Dep

ende

nt v

aria

ble:

(Mt-M

t-1)/

Mt-

1

Equ

atio

n B

Equ

atio

n C

Equ

atio

n D

Equ

atio

n E

Equ

atio

n F

Equ

atio

n G

Con

stan

t, (5

d)−

0.08

8***

(−3.

80)

−0.

085*

** (−

3.69

)−

0.08

4***

(−3.

63)

−0.

085*

** (−

3.72

)−

0.08

9***

(−3.

86)

−0.

090*

** (−

3.90

)I t/

Mt-

1 0

.911

***

(9.5

6) 0

.818

***

(6.0

2) 0

.709

***

(3.6

0) 0

.702

***

(5.5

1) 0

.966

***

(4.2

9) 0

.649

*(1

.82)

Dua

l-cla

ssSh

ares

−0.

317*

** (−

3.62

)−

0.31

7***

(−3.

61)

−0.

331*

** (−

3.69

)−

0.39

0***

(−4.

19)

−0.

396*

** (−

4.26

)−

0.40

4***

(−4.

33)

VR

1 0

.001

(0.2

4) 0

.007

(0.9

9) 0

.022

(1.0

7)V

R12

−0.

001

(−0.

96)

−0.

001

(−0.

92)

VR

13 0

.000

(0.7

6)V

R5

0.0

05**

(2.0

3)−

0.00

7 (−

0.82

) 0

.021

(0.8

0)V

R52

0.0

00 (1

.42)

−0.

001

(−0.

92)

VR

53 0

.000

(1.1

5)N

o. o

bs.

794

794

794

794

794

794

No.

fi rm

s14

214

214

214

214

214

2F

-val

ue12

.38

12.1

211

.86

12.5

412

.32

12.0

8R

20.

420.

420.

420.

420.

420.

42A

vera

ge q

m0.

812

0.83

60.

836

0.77

40.

728

0.73

0D

ual-c

lass

qm

0.61

20.

657

0.66

50.

564

0.51

40.

512

Sing

le-c

lass

qm

0.92

90.

988

0.98

20.

954

0.91

00.

916

Not

e:

*, *

* an

d **

* in

dica

te si

gnifi

canc

e at

10,

5 a

nd 1

per

cent

resp

ectiv

ely;

t-va

lues

in b

rack

ets.


to 143, the fi rm eff ects capture possible industry eff ects. Consequently, all equations have been estimated with two-digit industry SIC codes.

The stock market may be under- or over-estimated in any single period, but for a longer period of time the expected error in stock market evalu-ations is zero, E(mt) 5 0. To control for this possibility, annual dummy variables are included and estimated under the restriction that they summarize to zero. Annual deviations in stock market evaluations are therefore measured as deviations from the average. To control for the possibility that the Scandinavian countries have systematic diff erences in returns, country dummies are also included. These are also estimated under the restriction that they summarize to zero, so that any deviation is measured as the deviation from the Scandinavian average. Time, industry and country eff ects are not reported.

Hypotheses 1 and 2 (H1 and H2) cannot be rejected. For all measures of ownership concentration (CR1, CR5, VR1 and VR5) a positive non-linear relationship is found. In fi rms with one-share-one-vote, increasing

Table 7.8a Dual-class shares, ownership and performance


Equation H Equation I

Constant, (5d) −0.087***(−3.80)

Constant, (5d) −0.084***(−3.67)

It/Mt-1 0.521***(3.61)

It/Mt-1 0.404***(2.60)

CR1 0.053***(3.79)

VR1 0.063***(4.34)

CR12 −0.001***(−3.93)

VR12 −0.001***(−4.43)

CR1*Vote diff erentiation −0.045***(−4.43)

VR1*Vote diff erentiation

−0.053***(−5.07)

CR12*Votediff erentiation

0.001***(3.79)

VR12*Votediff erentiation

0.001***(4.67)

No. obs. 794 No. obs. 794No. fi rms 142 No. fi rms 142F-value 12.27 F-value 12.35R2 0.42 R2 0.43Average qm 0.933 Average qm 0.993Dual-class qm 0.652 Dual-class qm 0.721Single-class qm 1.199 Single-class qm 1.226

Note: *, ** and *** indicate signifi cance at 10, 5 and 1 percent, respectively; t-values in brackets.


ownership has a positive but marginally diminishing eff ect on performance. For the two concentration measures of the single largest owner, CR1 and VR1, a quadratic form gives the best fi t, whereas for the concentration of the fi ve largest owners, CR5 and VR5, a cubic form provides the best fi t.

The average qm for single class equity fi rms lies between 0.95 and 1.23.15 In fi rms with vote-diff erentiated shares, the eff ects are similar but much weaker. By comparing the parameters in equations H and I in Tables 7.8 a and b, one can see that in vote-diff erentiated fi rms the positive eff ect of ownership concentration is signifi cantly lower than in other fi rms. The average qm’s estimate for fi rms with dual-share class structure lies between 0.65 and 0.78 (equations H and I respectively). As in equation A, fi rms with dual-class shares seem to be investing at approximately 30 percent below

Table 7.8b Dual-class shares, ownership and performance


Equation H Equation I

Constant, (5d) −0.091***(−3.93)

Constant, (5d) −0.092***(−3.99)

It/Mt-1 0.336(0.89)

It/Mt-1 0.050(0.12)

CR5 0.067**(2.06)

VR5 0.089***(2.64)

CR52 −0.002**(−2.31)

VR52 −0.002***(−2.84)

CR53 0.000***(2.45)

VR53 0.000***(2.94)



−0.094***(−3.89)

CR52*Vote diff erentiation 0.002***(2.82)


0.003***(3.55)



−0.000***(−3.37)

No. obs. 794 No. obs. 794No. fi rms 142 No. fi rms 142F-value 11.58 F-value 11.91R2 0.42 R2 0.43Average qm 0.808 Average qm 0.889Dual-class qm 0.656 Dual-class qm 0.784Single-class qm 0.952 Single-class qm 0.978

Note: *, ** and *** indicate signifi cance at 10, 5 and 1 percent, respectively; t-values in brackets.


their cost of capital. From equations H and I in Tables 7.8 a and b, it is clear that the separation of cash-fl ow and control rights reduces the posi-tive eff ect of ownership and enhances the entrenchment eff ect. Examining listed fi rms in Sweden, Bjuggren et al. (2007) found similar negative eff ects of vote-diff erentiation and positive eff ects of ownership concentration on investment performance of fi rms.

Controlling for ownership characteristics and dual-class equity weakens the country eff ects (not reported here), but remains signifi cantly negative for Sweden and positive for Norway. However the eff ects of ownership and deviations from one-share-one-vote cut across national boundaries.

The intercept will, as discussed in Section 2, capture both the deprecia-tion rate and any systematic changes in market evaluations. In equations A to I, the intercept is estimated at approximately 9 percent. This is a reason-able estimate and is in line with previous estimates of depreciation rates. Furthermore, the intercept does not aff ect the marginal eff ect, and is thus not of any importance for the interpretation of the results.

6. CONCLUSIONS

This chapter examines the linkage between corporate investments, returns and ownership structure in the Scandinavian countries. Marginal q is used as performance measure. Marginal q measures the marginal return on capital relative to its cost. This return to cost of capital ratio (r/i 5 qm) is a measure of what Tobin (1982) labelled the functional effi ciency of capital markets. When studying fi rm performance, this method has some clear advantages over the conventional market-to-book measures of Tobin’s average q.

Few Scandinavian fi rms can be characterized as having dispersed own-ership as described by Berle and Means (1932). Vote-diff erentiation is a common tool for creating and maintaining strong and concentrated own-ership structures. Scandinavian fi rms make more frequent use of control mechanisms than fi rms in comparable countries. On average the largest owner holds more than 20 percent of the capital (CR1) and close to 30 percent of the voting rights (VR1).

The hypothesis that ownership concentration improves resource alloca-tion is supported in this chapter. The eff ect of ownership on investment performance is, however, found to be non-linear: cubic or quadratic in form. This is consistent with the entrenchment hypothesis. Strong support of the hypothesis that control mechanisms are detrimental to fi rm perform-ance is also found.

Ownership concentration is found to have a non-linear eff ect on fi rm performance. This is consistent with previous studies that fi nd both


positive incentive eff ects and negative entrenchment eff ects of ownership concentration. For fi rms with one-share-one-vote ownership has a positive impact but marginally diminishing, whereas for fi rms controlled by dual-class shares this eff ect is weaker. These fi rms have a systematically worse performance than other fi rms. Dual-class shares drive a wedge between cash-fl ow rights and control rights. Not only does this change the control structure, but it also changes the incentive structure. Firms with only one equity class are, on average, investing effi ciently, whereas fi rms with dual-class equity structure are over-investing. The separation of cash-fl ow rights and control rights reduces the positive incentive eff ect and enhances the negative entrenchment eff ect. By impairing capital reallocation, corporate control mechanisms are in the long run harmful for industry dynamics and economic renewal.

Vote-diff erentiation creates massive entrenchment eff ects and destroys large values. In the long run, they are likely to harm the functional effi -ciency of the Scandinavian capital markets. On average, ‘entrenched’ fi rms have returns on investments that are approximately 30 percent below their cost of capital.

Diff erences in investment performance across fi rms can largely be explained by diff erences in ownership structure and, in particular, to what extent corporate control is upheld by dual-class equity structure. Separation of cash-fl ow rights from control appears to distort the incentives of the controlling owner by signifi cantly reducing the incentive eff ect.

NOTES

* Acknowledgments: Financial support from the Ratio Institute and Sparbankernas fors-kningsstiftelse is gratefully acknowledged. This chapter was initiated during a six-month visit to George Mason University, and subsequently benefi ted a great deal from valuable comments and suggestions given by a large number of people. In particular, I am grate-ful for comments provided by Robin H. Hansson. Furthermore, I greatly appreciate valuable discussions during workshops and seminars held at Jönköping International Business School. In particular, I thank Åke E. Andersson, Tom Berglund, Per-Olof Bjuggren, Börje Johansson, Agostino Manduchi, Dennis C. Mueller, Ajit Singh, Steen Thomsen and Daniel Wiberg for valuable insights and helpful comments. Naturally, any remaining errors are my own.

1. For a review of the corporate governance literature see, for example, Shleifer and Vishny (1997), Morck et al. (2005), Mueller, (2003) and Denis and McConnell (2003).

2. There is a large literature on ownership and fi rm performance/value emanating from the work of Morck et al. (1988). See e.g. McConnell and Servaes (1990). For critique see Demsetz and Lehn (1985).

3. Agency costs are costs that arise from the principal–agent problem, that is, divergence of managerial objectives from the objectives of shareholders.

4. Jensen and Meckling (1976) also point out that the most serious problem of not


having equity claims is probably that the incentive to seek new profi table investment opportunities and engage in innovative eff orts will fall.

5. Cho (1998) criticizes these fi ndings and shows that market value aff ects ownership con-centration. See also Loderer and Martin (1997).

6. In an external study commissioned by the European Commission the proportionality between ownership and control of listed fi rms in the EU is examined. Among other things this study reports the results from a survey sent to institutional investors with €4.9 trillion of assets under management. A clear majority of the investors expect a discount of 10 to 30 percent of the share price of fi rms using CEM. See for further details ECGI (2007).

7. Functional stock market effi ciency is related to but diff erent from the standard term market effi ciency. Functional effi ciency refers to the way in which capital markets are allocating resources to the most effi cient usage (Tobin, 1982). Morck et al. (2005) survey a literature that shows how the functional effi ciency of capital markets depends on the structure and composition of corporate control.

8. When fi rms are price takers and perfectly competitive, marginal q and average q will be equal. Firms with market power will have a higher average q. For a derivation of the relationship between average q and marginal q, see Hayashi (1982).

9. If the market makes errors in their valuation of the fi rm, the error component, m, may contain a revaluation factor in the following period.

10. See Mueller and Reardon (1993) for a description of the methodology and an account of the properties of qm.

11. Accounting data and market prices have been collected from Standard & Poor’s Compustat Global Database, 2006 version. The following variables have been collected from Compustat (mnemonics in brackets): after-tax profi t (IB), depreciation (DP) dividends (DVT), total debt (DT), research and development (XRD), market price (MKVAL), advertising and marketing expenditures (XSGA), D equity (SSTK minus PRSTKC).

12. These 292 fi rms represent all non-fi nancial fi rms for which suffi cient ownership informa-tion was available. In 2004, there were a total of 796 listed fi rms in Scandinavia (194 in Denmark, 143 in Finland, 177 in Norway and 282 in Sweden).

13. The formula, Mi 5 M 1/i where M 1 is the largest fi rm and i the fi rm rank, approximates the size distribution of the fi rms in the sample.

14. In practice this excludes variables that have missing observations or contain accounting errors. Observations that were excluded were only among the small fi rms in the sample. Using a robust estimation technique yields consistent results.

15. The marginal eff ects have been calculated based on the average ownership concentration in the data set (CR1 5 22.24, CR5 5 44.52, VR1 5 28.58 and VR5 5 52.85).

REFERENCES

Adams, R. and Ferreira, D. (2008), ‘One Share–One Vote: The Empirical Evidence’, Review of Finance, 12 (1), 51–91.

Bebchuk, L., Kraakman, R. and Triantis, G. (1999), ‘Stock Pyramids, Cross-Ownership, and Dual-Class Equity: The Creation and Agency Costs of Separating Control from Cash Flow Rights’, NBER Working Paper, No. 6951.

Bennedsen M.B. and Nielsen, K.M. (2005), ‘The Principle of Proportionality: Separating the Impact of Dual Class Shares, Pyramids and Cross-ownership on Firm Value Across Legal Regimes in Western Europe’, Centre for Industrial Economics Discussion Papers, University of Copenhagen.

Berle, A.A. and Means, G. (1932), The Modern Corporation and Private Property, New York: The Macmillan Company.


Bergström, C. and Rydqvist, K. (1990), ‘Ownership of Equity in Dual-class Firms’, Journal of Financial Economics, 14, 255–69.

Bjuggren, P.-O. and Wiberg, D. (2008), ‘Industry Specifi c Eff ects in Investment Performance and Valuation of Firms: Marginal q in a Stock Market Bubble’, Empirica, 35 (3), 279–91.

Bjuggren, P.-O., Eklund, J.E. and Wiberg, D.K. (2007), ‘Ownership Structure, Control and Firm Performance: the Eff ects of Vote Diff erentiated Shares’, Applied Financial Economics, 17, 1323–34.

Burkart, M. and Lee, S. (2008), ‘One Share – One Vote: The Theory’, Review of Finance, 12 (1), 1–49.

Bøhren, Ø. and Ødegaard, A.B. (2006), ‘Governance and Performance Revisited’, in Gregouriu, G. and Ali, P. (eds), International Corporate Governance after Sarbanes-Oxley, Wiley.

Carsten, G. (1993), ‘Europäische Integration und Nordische Zusammenarbeit auf dem Gebiet des Zivilrechts’, Zeitschrift für Europäisches Privatrecht, 335.

Cho, M.-H. (1998), ‘Ownership Structure, Investments, and the Corporate Value: An Empirical Analysis’, Journal of Financial Economics, 47 (1), 103–21.

Claessens, S., Djankov, S., Fan, J.P.H. and Lang, L.H.P. (2002), ‘Disentangling the Incentive and Entrenchment Eff ects of Large Shareholdings’, Journal of Finance, 57 (6), 2741–71.

Cronqvist, H. and Nilsson, M. (2003), ‘Agency Costs of Controlling Minority Shareholders’, Journal of Financial and Quantitative Analysis, 38 (4), 695–707.

Demsetz, H. (1983), ‘The Structure of Ownership and the Theory of the Firm’, Journal of Law and Economics, 26 (2), 375–90.

Demsetz, H. and Lehn, K. (1985), ‘The Structure of Corporate Ownership: Causes and Consequences’, Journal of Political Economy, 93 (6), 1155–77.

ECGI (2007), ‘Report on the Proportionality Principle in the European Union’, external study commissioned by the European Commission, European Corporate Governance Institute (ECGI).

Faccio, M. and Lang, L.H.P. (2002), ‘The Ultimate Ownership of Western European Corporations’, Journal of Financial Economics, 65 (3), 365–95.

Gugler, K. and Yurtoglu, B.B. (2003), ‘Average q, Marginal q, and the Relation between Ownership and Performance’, Economics Letters, 78 (3), 379–84.

Gugler, K. Mueller, D.C. and Yurtoglu, B.B. (2002), ‘Corporate Goverance, Capital Market Discipline and Return on Investments’, mimeo, University of Vienna.

Hayashi, F. (1982), ‘Tobin’s Marginal q and Average q: A Neoclassical Interpretation’, Econometrica, 50 (1), 213–24.

Henrekson, M. and Jakobsson, U. (2006), Den svenska modellen för företagsägande och företagskontroll vid skiljevägen, Institutet för näringslivsforskning, Policy Paper No. 1, Stockholm.

Högfeldt, P. (2004), ‘The History and Politics of Corporate Ownership in Sweden’, NBER Working Paper No. 10641.

Jensen, J. and Meckling, W. (1976), ‘Theory of the Firm: Managerial Behavior, Agency Costs and ownership structure’, Journal of Financial Economics, 3 (4), 305–60.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R.W. (1997), ‘Legal Determinants of External Finance’, Journal of Finance, 52 (3), 1131–50.

La Porta, R., Lopez-de-Silanes, F. and Shleifer, A. (1999), ‘Corporate Ownership around the World’, Journal of Finance, 54 (2), 471–517.

Loderer, C. and Martin, K. (1997), ‘Executive Stock Ownership and Performance: Tracking Faint Traces’, Journal of Financial Economics, 45 (2), 223–55.


McConnell, J.J. and Servaes, H. (1990), ‘Additional Evidence on Equity Ownership and Corporate Value’, Journal of Financial Economics, 27 (2), 595–612.

Manne, H. (1965), ‘Mergers and the Market for Corporate Control’, Journal of Political Economy, 75 (2), 110–20.

Maury, B. and Pajuste, A. (2004), ‘Multiple Large Shareholders and Firm Value’, SSRN working paper series.

Morck, R., Shleifer, A. and Vishny, R. (1988), ‘Management Ownership and Market Valuation: An Empirical Analysis’, Journal of Financial Economics, 20 (1), 293–316.

Morck, R., Wolfenzon, D. and Yeung, B. (2005), ‘Corporate Governance, Economic, Entrenchment, and Growth’, Journal of Economic Literature, 43 (3), 655–720.

Mueller, D.C. (2003), The Corporation; Investments, Mergers and Growth, London: Routledge.

Mueller, D.C. and Reardon, E.A. (1993), ‘Rates of Return on Corporate Investment’, Southern Economic Journal, 60 (2), 430–53.

Roe, M.J. (2003), Political Determinants of Corporate Governance, Oxford: Oxford University Press.

Shleifer, A. and Vishny, R.W. (1997), ‘A Survey of Corporate Governance’, Journal of Finance, 52 (2), 737–83.

Söderström, H.T. (ed.), Berglöf, E., Holmström, B., Högfeldt, P. and Milgrom, E.M. (2003), Ägarmakt och omvandling – Den svenska modellen utmanad, Ekonomirådets rapport 2003, SNS, Stockholm.

Stulz, R.M. (1988), ‘Managerial Control of Voting Rights, Financing Policies and the Market for Corporate Control’, Journal of Financial Economics, 20, 25–54.

Tobin, J. (1982), ‘On the Effi ciency of the Financial Systems’, Lloyd’s Banking Review, 153 (July), 1–15.

Veblen, T. (1921), The Engineers and the Price System, New York: Viking.

167

8. The cost of legal uncertainty: the impact of insecure property rights on cost of capitalPer-Olof Bjuggren and Johan E. Eklund*

1. INTRODUCTION

In the required rate of return on investments a special risk premium labeled institutional risk should be included. Diff erent countries represent institu-tional risks for investors. It is various risk factors tied to the institutional framework that give the rules of the game facing investors. The rules can be of a supportive nature or make long-term investments hazardous due to the lack of secure property rights. It is more or less evident that investors must use a higher discount rate in evaluations of investments in countries where property rights are weakly protected.

How to account for the political risk has not received much attention in the fi nance literature, even though the concept ‘political risk’ is sometimes men-tioned. However, a formal treatment is not off ered. An exception is Faure and Skogh (2003), who have shown how ‘political risk’ can be incorporated in investment analysis. It is their approach that has inspired this chapter.

But it is one thing to propose the necessity to add an ‘institutional risk premium’ and another thing to prove the existence of such risk premiums and estimate their size. One reason why few attempts have been made so far is the diffi culty of empirically quantifying and pricing institutional risk. The purpose of this chapter is to show the existence of property risk pre-miums and measure their magnitude.

As measures of institutional risk we use two indexes; one on property right protection provided by the Heritage Foundation, and one on investor right protection provided by the Political Risk Group. With these indexes we use a type of CAPM (capital asset pricing model) to estimate the eff ect of property rights uncertainty on the cost of capital.

The chapter starts, in Section 2, with a discussion of what is meant by institutional risk. Section 3 provides the theoretical underpinning to rational investment decisions. How to calculate capital costs and risk premiums for


assets with political risk is the subject matter of Section 4. The data used in our empirical analysis are presented in Section 5. The empirical fi ndings are analysed in Section 6. The chapter ends with conclusions in Section 7.

2. POLITICAL RISK, PROPERTY RIGHTS, CONTRACTS AND LEGAL UNCERTAINTY

Growth, prosperity and welfare are a function of the investments made in a country. Investments add to the capital stock which is a basic produc-tion factor. However, investments are risky. A cost in the form of a capital outlay is paid today, while the benefi ts represented by positive net cash fl ows lie in the future. It is like a deferred exchange where a payment is made today in return for enhanced future consumption. When the future unfolds it might turn out that the products produced by the new capital (the investment) cannot be sold at profi table prices.

This is a risk that every entrepreneur has to face. But in addition to this risk there might be an institutional risk caused by insecure property rights and a defective judicial system that makes it diffi cult to enforce contracts in an eff ective way. Secure property rights and contract enforcement were put forward long ago by David Hume and Adam Smith as two of the most important institutional factors for the prosperity of a nation. According to Kasper and Steit (1999, p. 20) David Hume and Adam Smith stressed three institutions of fundamental importance for progress and welfare: ‘the guarantee of property rights, the free transfer of property by volun-tary contractual agreement, and the keeping of promises made’. In other words, secure property rights, freedom of contracts and enforcement of agreements are basic cornerstones in the quest for prosperity.

Secure property rights, freedom of contracts and enforcement of agree-ments are basic parts of the institutional framework within which an economy is organized. It is the task of the state to develop a well function-ing and adequate institutional framework through formal rules. According to North (1990, p. 47) ‘formal rules include political (and judicial) rules, economic rules, and contracts’. By economic rules North means property rights, which are defi ned as ‘the bundle of rights over the use and the income to be derived from property and the ability to alienate an asset or a resource’ (p. 47). The link to investments is clear as investments mean the creation of new assets. Furthermore, North ascertains a hierarchical order between rules in the sense that ‘the rules descend from polities to property rights to individual contracts’ (p. 52). According to, for example, North (1990) and Williamson (2000) these institutions are very stable over time.

Hernando de Soto (2000) has argued powerfully that the varying degree

The cost of legal uncertainty 169

to which countries succeed to support capital formation and accumulation is to be found in the legal structure of the property rights system of the western world. De Soto claims that:

When advanced nations pulled together all the information and rules about their known assets and established property systems that tracked their economic evolution, they gathered into one order the whole institutional process that underpins the creation of capital. If capitalism had a mind, it would be located in the legal property system. (de Soto (2000) p. 65)

Furthermore, de Soto argues that informal institutions are far less impor-tant than formal institutions in this respect.

In other words it is through polity that a nation can infl uence the insti-tutional framework to stimulate investments and growth. The institutional framework might be of a kind that makes investors feel secure that no one else will appropriate the fruits of their investments or the framework might be one that makes investors think that there is a risk that someone else will reap the benefi ts. If the property rights are insecure, long-term investments will be hampered and come at the cost of lower welfare.

3. RISK, RETURN, PORTFOLIO THEORY AND INVESTMENT

Conventional investment theory holds that investors will evaluate alterna-tive investments on the basis of the net present value (NPV). According to the NPV rule investments should be evaluated according to the expected cash fl ows (CF) minus the expected investment costs (I). Only projects with expected positive net present values should be initiated (NPV 5 PV − I). When calculating the present value of cash fl ows generated by an invest-ment one uses a discount factor, 1/(1 1 r), where r is the discount rate. The discount rate is also referred to as the required rate of return or as the cost of capital. For risk-free projects the required rate of return equals the risk-free interest rate, which also serves as a reference when valuing risky assets. The PV of future cash fl ows is calculated as follows:1

PV 5 aT

t51

CFt

(1 1 r) t (1)

The discount rate will depend on the riskiness of the future cash fl ows. If the risk is that the actual cash fl ow will be lower than predicted, the dis-count rate will be accordingly higher than for a more certain future cash fl ow. As the discount rate, r, increases the present value declines, hence the aggregate number of investments also declines.


The crucial question is therefore how to determine the size of the dis-count rate and the risk associated with various assets. We argue that the discount rate can be broken down to a multitude of components, among which property rights protection is one. Accordingly the discounted rate r in the formula above should be:

r 5 rf 1 RPo 1 RPp (2)

where rf is the risk-free interest rate, RPp the risk premium associated with weak property rights protection, and RPo a ‘general’ risk premium.2 Note that the ‘general’ risk premium can presumably be broken down further into diff erent risk factors.3

The conventional capital asset pricing model (CAPM) makes a distinc-tion between diversifi able fi rm specifi c risk and non-diversifi able systemic risk (see Section 3). The specifi c risk can be diversifi ed away. It is only the remaining non-diversifi able risk that matters for the pricing of the asset (that is, investors are only compensated for the systematic non- diversifi able risk). As a consequence, the return r that investors require depends on the systematic risk of the investment (see Section 4).

On a global capital market even much of the country specifi c risk can be diversifi ed away. One exception is, however, the institutional risk of insecure property rights and contracts. This is a risk associated with the institutional framework of the rules of a country. This institutional framework is made up of both informal rules like norms, customs, tradi-tion and religion and formal rules like property and contract laws and the enforcement of these rules. It is, as pointed out above, primarily the formal rules that polity can exert an infl uence on. To change informal rules is a much tougher task. According to North (1990) and Williamson (2000) this institutional framework changes very slowly over time. Even changes in the formal rules (like property rights rules) and their enforcement tend to take decades to implement. As an investor you are more or less stuck with the institutional framework for a considerable time. The prospects of balancing changes in the security of property rights by having an inter-national portfolio are small. Hence, insecure property rights represent a truly systematic risk that, according to the theory (see next section), will increase the cost of capital (the required return on investment, r). The rise in the cost of capital will, as shown in standard investment theory, decrease investment.4

According to the CAPM the expected return on a security can be calculated as:

E( ri) 5 rf 1 bi (E(rm) 2 rf) (3)


where bi measures the sensitivity of a security to market risk (systemic risk) and rm the return on a market portfolio m. The model holds that the expected rate of return should equal the risk-free interest rate plus a risk premium that varies with b. The expected rate of return, E(ri), that is obtained is simply the discount rate r used in equation (4) to calculate the present value of future cash fl ows. The term E(rm) − rf is the market price of risk for effi cient portfolios (see for example Elton and Gruber, 1995). In this model it is only one factor, the market portfolio, that matters in calculation of the risk premium.

This standard CAPM can be extended into a so-called multi-beta CAPM by including other factors that infl uence the size of the risk premium. Merton (1973) was among the fi rst to include a number of uncertainty factors that could infl uence the price of an asset. Friend et al. (1976) used the market portfolio infl ation uncertainty as an additional factor that determined cost of capital. We follow their approaches and add legal uncertainty in the form of the degree of security of property rights as a risk factor.

4. ESTIMATIONS OF RISK AND RETURN

Roll (1977) has in a seminal article levelled criticism at a general equilib-rium model of the form of the CAPM. One of Roll’s points is that all assets in the entire world should be included in the market portfolio m. However, in the empirical literature national stock indices like Standard and Poor’s 500 and the New York Stock Exchange index are used as market portfolios (see, for instance, Elton and Gruber (1995) for examples). This is clearly an incorrect measure according to Roll’s critique. A step towards a more ‘correct’ market portfolio measure is to use an index containing all securi-ties in the entire world. Such an index for traded corporate shares is the Morgan Stanley world market index. That is the index that will represent the market portfolio in the present study.

Estimation of beta and the risk premium can be made according to a two-pass procedure.5 In a fi rst-pass regression, time series analysis is used to estimate bi in equation (4).

rti 5 ai 1 bi 3 rmt 1 eit (4)

The estimated betas (bi) are then used in a second-pass regression

ri 5 ai 1 RPm 3 bi 1 ei (5)


where ai is an estimate of the risk-free interest rate, ri is the average return of a security or portfolio of securities, RPm is the risk premium of the market portfolio and bi is the estimated beta-coeffi cient.

Since the institutional framework infl uencing investments tends to be extremely stable over time, the institutional risk stemming from fl aws in the protection of property is not diversifi able and thus is systematic. However, the stability of these institutional types of risk factors off ers an empirical challenge to estimation. An ideal way of estimating various risk factors’ infl uence on return is the so-called arbitrage pricing theory (APT). The APT model comes from an entirely diff erent tradition from the CAPM.6 In contrast to the CAPM, the APT model emphasizes surprise eff ects; that is, sudden unexpected changes in macro variables can be used to explain return. This is not an option in the present case. The stability of institutions means that there is virtually no variation across time so one needs to focus on the cross-sectional variation. As a consequence it is not possible to use indexes of institutional risk factors in the fi rst-pass estimation procedure.

A number of tests of the CAPM using the two-pass procedure have been performed. Several of these indicate that a two-factor model (a multi-CAPM) can be used.7 Consequently, there is some empirical support for use of a model where legal uncertainty as well as a market portfolio is included as a factor in the calculation of cost of capital. In that case the second-pass regression will contain a second factor representing institu-tional risk and look like:

ri 5 ai 1 RPm 3 bi 1 RPp 3 Institutional Risk 1 ei (6)

where RPp is the right risk premium due to insuffi cient safeguarding of property and investor rights.

5. DATA

To calculate the impact of institutional risk on risk premium and the required return on investments stock exchange data, risk data and data of a global market portfolio are needed. Table 8.1 shows the data we have used for these variables.

Monthly stock market indexes compiled by Morgan Stanley are used to calculate the rate of return on national stock markets and the world market portfolio.8 The stock market indexes cover a ten-year period, 1995 to 2005 (129 months more exactly), are expressed in US dollars, and are corrected for dividends. This ensures that the indexes are consistently defi ned and include all relevant returns. As a proxy for the market portfolio the world stock


market index from Morgan Stanley is used, which includes 49 developed and emerging country market indexes. These are the countries examined.

As measures of institutional risk the Heritage Foundation index of prop-erty rights protection (PRP) and the index of investor rights protection (IRP) provided by the International Country Risk Guide (ICRG) are used. The Heritage index ranges from 1 to 5, where 1 indicates strong protection of property rights. It is an annual index which is available for the period 1995 to 2005. The property right index is an assessment of the quality of contract enforcement, legal protection of property, existence of corruption in the judicial system and the probability of expropriation. The ICRG index ranges from 1 to 12, where 1 indicates strong protection. This is a monthly index that here is used for the period 1995 to 2005. This index also measures factors that have to do with protection of property rights and enforcement of contracts. (For a further description of the two indexes used here, refer to the Heritage Foundation and the Political Risk Group.)

In Table 8.2 the 49 developed and emerging countries that are included in Morgan Stanley’s world market are shown. Three variables are pre-sented. The fi rst variable, R2-values from fi rst-pass regression, shows the proportion of the national rate of return that can be explained by variation

Table 8.1 Description of the variables

Country stock market indexes

Measures the stock price performance including dividends. Expressed in US dollars. Source: Morgan Stanley

World market index

Measures the stock price performance including dividends for 49 developed and developing countries. Source: Morgan Stanley

Property right protection (PRPit)

Assessment of the protection and certainty of property rights. Annual index ranging from 1 to 5, with a higher number meaning weaker protection of property rights. Source: Heritage Foundation Index of Economic Freedom

Investor right protection(IRPit)

Investor profi le. Assessment of a number of factors infl uencing the risk of investments. Monthly index ranging from 1 to 12, with a higher number meaning stronger protection of investors. The index assesses contract viability, risk of expropriation, payment delays and profi t repatriation. Source: International Country Risk Guide

Returns rit

rmt

Diff erent forms of return are calculated. Monthly stock market return on country level. National stock market indexes corrected for dividends and in US dollars are used. Source: Morgan Stanley World market return calculated with monthly world market index. Source: Morgan Stanley


Table 8.2 Country data for 1995–2005 (129 months)

Developed countries

R2-values from 1st-pass

regression

Property right protection

(PRPit) (average)*

Investor right protection

(IRPit) (average)*

Average rate of return

Australia 0.53 1.00 7.52 0.128Austria 0.22 1.00 9.16 0.122Belgium 0.40 1.00 8.68 0.127Canada 0.68 1.00 8.52 0.160Denmark 0.47 1.00 8.21 0.150Finland 0.40 1.00 8.74 0.221France 0.69 2.00 8.44 0.125Germany 0.65 1.00 8.50 0.105Greece 0.20 2.36 6.89 0.164Hong Kong 0.38 1.00 7.63 0.108Ireland 0.48 1.00 8.87 0.107Italy 0.39 2.00 7.99 0.131Japan 0.37 1.36 8.49 0.011Netherlands 0.68 1.00 8.94 0.106New Zealand 0.30 1.00 8.47 0.106Norway 0.48 1.18 8.23 0.140Portugal 0.32 2.00 7.89 0.107Singapore 0.37 1.00 8.79 0.045Spain 0.59 2.27 8.96 0.178Sweden 0.60 1.64 8.12 0.172Switzerland 0.45 1.27 8.95 0.124United Kingdom 0.69 1.00 8.85 0.103United States 0.87 1.00 8.96 0.121

Emerging economies


regression

Property right protection (average)

Investor right protection(average)


Argentina 0.16 2.73 5.56 0.165Brazil 0.43 3.00 6.06 0.200Chile 0.38 1.00 7.73 0.083China 0.17 4.00 6.68 0.017Colombia 0.05 3.36 6.34 0.179Czech Republic 0.09 2.00 8.52 0.205Egypt 0.04 3.09 6.31 0.274Hungary 0.26 2.00 7.75 0.287India 0.10 3.00 6.36 0.108Indonesia 0.16 3.45 6.12 0.101Israel 0.33 2.00 6.92 0.135


of the rate of return in a world market portfolio. It is evident that there is a clear diff erence between developed and emerging countries in this respect. In most cases the stock exchanges in developed countries are more infl u-enced by variations in the world market portfolio than those in emerging countries are. The very low R2-values for many countries could be seen as an indication that there is a need for more factors than our market portfo-lio to explain the rate of return on assets in diff erent countries.

The average value on the Heritage index of property rights for the exam-ined period diff ers also a lot for the countries. With a few exceptions the developed countries have more secure property rights (lower values). With regard to average rate of return there are not similar systematic diff erences between the two types of countries. One observation is that there is more variation in the rate of return for emerging than for developed countries. The negative rate of return for the Philippines is troublesome from a theo-retical perspective. According to the CAPM model this rate of return would indicate a negative risk-free interest rate, which is not possible. We will therefore exclude the Philippines when the risk premiums are calculated.

Table 8.3 is a table of summary statistics that confi rms the picture given.


Emerging economies


regression

Property right protection (average)

Investor right protection(average)


Jordan 0.01 2.36 6.76 0.158South Korea 0.24 1.27 7.85 0.149Malaysia 0.14 2.45 7.26 0.044Mexico 0.44 2.91 7.72 0.175Morocco 0.00 3.09 6.84 0.115Pakistan 0.02 3.18 4.94 0.141Peru 0.12 3.45 6.12 0.156Philippines 0.20 2.73 6.32 −0.061Poland 0.27 2.27 7.34 0.166Russia 0.23 3.36 5.59 0.413South Africa 0.35 2.91 7.55 0.114Taiwan 0.26 2.09 7.03 0.022Thailand 0.25 1.36 9.06 0.028Turkey 0.26 2.36 6.43 0.322Venezuela 0.09 3.45 5.28 0.172

Note: * The two indexes are inverse to each other. For the PRP index a low value is good, whereas for the IRP index a high value is good.

176

Tab

le 8

.3

Sum

mar

y st

atis

tics f

or th

e ag

greg

ates

of d

evel

oped

and

em

ergi

ng e

cono

mie

s 199

5–20

05 (

129

mon

ths)

R2 f

rom

1st

-pas

s re

gres

sion

Prop

erty

righ

t pr

otec

tion

Inve

stor

righ

t pro

tect

ion

Rat

e of

retu

rn

Dev

elop

ed

econ

omie

sE

mer

ging

ec

onom

ies

Dev

elop

ed

econ

omie

sE

mer

ging

ec

onom

ies

Dev

elop

ed

econ

omie

sE

mer

ging

ec

onom

ies

Dev

elop

ed

econ

omie

sE

mer

ging

ec

onom

ies

Mea

n0.

487*

0.19

4*1.

307*

2.64

9*8.

426*

6.78

6*0.

124

0.14

9St

anda

rd

devi

atio

n0.

170

0.12

80.

473

0.74

90.

556

0.98

40.

043

0.10

1

Min

imum

0.2

01

16.

894.

940.

011

-0.0

61M

axim

um0.

870.

442.

364

9.16

9.06

0.22

10.

413

Cou

nt23

2623

2623

2623

26

Not

e:

* in

dica

tes t

hat z

-tes

t sho

ws s

igni

fi can

tly d

iff er

ent m

eans

at l

ess t

han

5 pe

r cen

t lev

el.


The world market portfolio is a signifi cantly better explanatory factor of rate of return in developed than in emerging countries. Property rights protection is signifi cantly higher in developed countries. Also, property rights protection and the rate of return show a much higher variation in emerging countries.

Finally, Table 8.4 shows that the correlation between the variables is especially high between property rights and R2-values. This result does also point to an interpretation that in countries with insecure property rights the market portfolio alone does not explain as much of the rate of return as in countries with secure rights. And for obvious reasons the property right and investor right indexes are highly correlated.

To test whether there is a signifi cant link between the R2-values from the fi rst-pass regression and the institutional framework we regress the indexes on the R2-values. We fi nd a highly signifi cant link in both cases and the explanatory power is relatively high. The results are reported in Table 8.5. These results suggest that in countries with weak institutional protection of property rights the systematic economic factors infl uencing the world market return have less explanatory power. In comparison, Roll (1988) has examined the R2s for individual stock in the US and found that there is no systematic diff erence in the explanatory power across industries. This seems to suggest that the CAPM provides a less adequate tool for under-standing asset pricing in less developed fi nancial markets. Therefore we believe that the R2-values can be used as proxies for fi nancial development; countries in which the CAPM displays a lower explanatory power might be interpreted as less developed fi nancially.

6. MODELS AND RESULTS

In the fi rst step monthly data were used to estimate the fi rst-pass regression as in equation (4). In the second-pass regression the average Heritage and International Country Risk Guide indexes of property rights protection were included:

Table 8.4 Correlation matrix

R2 Property rights

Investor rights

Average returns

R2 1Property rights −0.652 1Investor rights 0.675 −0.827 1Average returns −0.086 0.205 −0.254 1


ri 5 ai 1 RPm 3 bi 1 RPPRP 3 PRPi 1 ei (7a)

ri 5 ai 1 RPm 3 bi 1 RPIRP 3 IRPi 1 ei (7b)

where PRPi is the average value of the Heritage Foundation property right index and IRPi is the International Country Risk Guide index of investor right protection for a country i. We identify the Philippines as an outlier and consequently exclude it from our regression.

Signifi cant coeffi cients, RPPRP for the average value of the Heritage index and RPIRP for the average value of the ICRG index, indicate that the market portfolio is not the only important explanatory variable in calcula-tions of a risk premium.

The result of the estimation of the second-pass equation is shown in Table 8.6. The security of property rights is important. The coeffi cient for PRP is almost signifi cant at the 5 percent level and the coeffi cient for IRP is signifi cant at 5 percent. A higher degree of insecurity is consistent with a higher cost of capital.

With the conventional CAPM we fi nd that the world risk-free interest rate, (a), was on average 8 percent. However, assuming that the countries with the best values on the property right index (PRP 5 1) and the investor protection index (IRP 5 9.16) have virtually no uncertainty with regard to property rights, we fi nd lower risk-free interest rates. Using the best values of the two risk factors we estimate the risk-free rate to be 5.7 percent for the property right index and 4.8 percent for the investor right index. The

Table 8.5 First-pass R2-values and protection of property

Estimation method

Property right protection(PRP)

Investor right protection(IRP)

OLS OLS

Dependent variable: R2

(fi rst-pass) Coeffi cient t-value Coeffi cient t-value

Property rights (PRPi ) 20.147† 25.84*Investor rights (IRPi) 0.122† 6.22*R2 0.42 0.45Adj. R2 0.41 0.44F-value 34.1 38.7No. observations 49 49

Note: * indicates signifi cance at 1 percent. † The diff erences in sign and intercept are due to the fact that the two indexes are inverse to each other (see Table 8.1).

179

Tab

le 8

.6

Ris

k pr

emiu

m fa

ctor

s: p

rope

rty

righ

ts a

nd in

vest

or p

rote

ctio

n

Est

imat

ion

met

hod

Prop

erty

righ

t pro

tect

ion

(PR

P)In

vest

or ri

ght p

rote

ctio

n (I

RP)

Con

vent

iona

l CA

PM

OL

SO

LS

OL

S

Dep

ende

nt v

aria

ble:

r i

Coe

ffi ci

ents

t-va

lues

Coe

ffi ci

ents

t-va

lues

Coe

ffi ci

ents

t-va

lues

Inte

rcep

t (a

)0.

036

1.06

0.24

03.

36*

0.08

2.92

*Pr

oper

ty ri

ghts

PR

Pi

0.02

1†

1.98

**–

––

–In

vest

or ri

ghts

IRP

i−

0.02

1†

−2.

41*

––

b0.

059

2.55

*0.

056

2.45

*0.

062.

47*

R2

0.1

90.

220.

12A

dj. R

2 0

.15

0.18

0.10

F-v

alue

5.1

86.

276.

08N

o. o

bser

vatio

ns 4

848

48

Not

e:

* an

d **

indi

cate

sign

ifi ca

nce

at 5

per

cent

and

10

perc

ent r

espe

ctiv

ely.

† T

he d

iff er

ence

s in

sign

and

inte

rcep

t are

due

to th

e fa

ct th

at th

e tw

o in

dexe

s are

inve

rse

to e

ach

othe

r (se

e T

able

8.1

).


infl uence of security of property rights is signifi cant whichever of the two indices we use. The traditional beta stays approximately the same whatever index is used. This suggests robustness in the result that security of prop-erty rights does matter for the cost of capital.

Estimated risk premiums for property right protection and investor right protection (plus the estimated risk-free interest rate) are reported in Table 8.7. We fi nd an average 3 percent diff erence in interest rates, which can be explained by diff erences in institutional quality. A z-test shows that these diff erences are signifi cant. The general beta does not diff er in a signifi cant fashion between developed and emerging countries. Hence, the develop-ing countries have a much lower risk premium on investments due to the systematic risk that the institutional framework represents. Improvement in the institutional framework will probably be conducive to more invest-ments and higher welfare.

For the conventional CAPM Ramsey’s regression specifi cation error test (RESET) indicated a problem of omitted variables (F-test 8.65), which supports the inclusion of further variables in the model. This is consistent with the fact that the R2 and R2-adjusted almost double with the inclusion of the two indexes.

7. CONCLUSIONS

David Hume and Adam Smith stressed the importance of secure property rights for prosperity and growth. A link in the form of a formal treatment of how secure property rights lead to lower cost of capital and thereby more investment has not been established. Portfolio theory in corporate fi nance theory provides such a link. In portfolio theory as represented by CAPM and APT, systematic risk and surprise changes in fundamentals are determinants of cost of capital.

The rationale is that an investor can get rid of unsystematic risk through portfolio diversifi cation. A time perspective is used where unsystematic risk is avoided by combining assets that show diff erent patterns of change over time, implying a correlation of less than 1. Insecure property rights have to do with what institutional framework a country has. The institutional framework changes slowly over time. Hence it is diffi cult to diversify away from the systematic risk that the institutional framework represents.

The specifi cities of institutional frameworks of countries make it impossi-ble to use traditional APT and CAPM methodologies to estimate the impact of insecure property rights on cost of capital. There is not much variation over time in the institutional frameworks. The variation is between coun-tries. This makes it diffi cult to apply APT methodology. Instead, a multi-


Table 8.7 Risk-free rate plus risk premiums

Developed economies

Property right premium (PRP)

+ a

Investor right premium (IRP)

+ a

bi^

Australia 0.058 0.084 0.87Austria 0.058 0.050 0.61Belgium 0.058 0.060 0.80Canada 0.058 0.063 1.11Denmark 0.058 0.070 0.84Finland 0.058 0.059 1.62France 0.079 0.065 1.07Germany 0.058 0.064 1.26Greece 0.087 0.097 0.95Hong Kong 0.058 0.082 1.20Ireland 0.058 0.056 0.85Italy 0.079 0.074 0.94Japan 0.066 0.064 0.87Netherlands 0.058 0.055 1.08New Zealand 0.058 0.064 0.81Norway 0.062 0.069 1.07Portugal 0.079 0.076 0.82Singapore 0.058 0.058 1.15Spain 0.085 0.054 1.14Sweden 0.072 0.072 1.42Switzerland 0.064 0.054 0.79United Kingdom 0.058 0.056 0.77United States 0.058 0.054 1.00Averages 0.064* 0.065* 1.00

Emerging economies

Argentina 0.095 0.125 1.12Brazil 0.101 0.114 1.85Chile 0.058 0.080 1.02China 0.122 0.102 1.14Colombia 0.108 0.109 0.52Czech Republic 0.079 0.063 0.63Egypt 0.103 0.109 0.46Hungary 0.079 0.079 1.30India 0.101 0.108 0.65Indonesia 0.110 0.113 1.46Israel 0.079 0.097 1.07Jordan 0.087 0.100 0.15


factor CAPM approach has been used with a world market portfolio as one systematic factor and indices over secure property rights and contracts. The security of property rights and contractual agreements is introduced in a second-pass regression that looks at diff erences between countries.

If a CAPM type of analysis is used an indication of higher risk premi-ums in countries with insecure property rights is also found. We fi nd that a signifi cant part of the required rate of return in developing countries can be explained by weak institutional protection of property and contracts. Our two measures of the institutional safeguarding of property rights seem to capture the same eff ect and indicate that institutional risk needs to be included in capital asset pricing. The conventional single-factor CAPM seems to be less useful in countries where the institutional framework is weak.

NOTES

* Acknowledgements: Financial support from Föreningssparbankernas Forskningsstiftelse to Johan Eklund’s dissertation work is gratefully acknowledged. A research grant from the Ratio Institute and the Marcus and Amalia Wallenberg Memorial Fund Foundation is also gratefully acknowledged. Furthermore we acknowledge valuable comments


Emerging economies

South Korea 0.064 0.077 1.59Malaysia 0.089 0.089 0.94Mexico 0.099 0.080 1.44Morocco 0.103 0.098 0.06Pakistan 0.105 0.138 0.41Peru 0.110 0.113 0.69Philippines 0.095 0.109 1.06Poland 0.085 0.088 1.37Russia 0.108 0.124 2.13South Africa 0.099 0.083 1.11Taiwan 0.081 0.094 1.10Thailand 0.066 0.052 1.63Turkey 0.087 0.107 2.15Venezuela 0.110 0.131 1.01

Averages 0.093* 0.099* 1.08

Note: * indicates that the z-test shows signifi cantly diff erent means at 1 percent.


provided by Åke E. Andersson, Göran Skogh and participants at the 2005 conference of the European Association of Law and Economics.

1. The present value of future cash fl ows from all investments made by a fi rm is equal to the value of a fi rm set by the capital markets.

2. A political risk premium is proposed by Faure and Skogh (2003).3. It can in line with the arbitrage pricing theory be argued that the discount factor RP0 can

be broken down further into a multitude of components (see, for example, Ross, 1976). The problem is however to isolate the diff erent factors that infl uence the rate of return.

4. See, for example, Mueller (2003) for a pedagogical exposition of the relationship between cost of capital and investment.

5. See, for example, Elton and Gruber (1996, Chapter 2) for a description of the procedure.

6. The wide use of and support for the CAPM can, according to Roll and Ross (1980), be ascribed to the empirical regularity of common variation of securities, and according to the CAPM this common variation can be ascribed to a single factor, plus an error term. Even though the rationale behind the CAPM is, as noted earlier, based on the separability of systemic non-diversifi able risk and non-systemic diversifi able risk, Roll and Ross argue that

There are two major diff erences between the APT and the original Sharpe ‘diagonal’ model, a single factor generating model which we believe is the intuitive grey eminence behind the CAPM. First, and most simply, the APT allows for more than just one generating factor. Second, the APT demonstrates that since any market equilibrium must be consistent with no arbitrage profi ts, every equilibrium will be characterized by linear relationship between each asset’s expected return and its return’s amplitudes, or loadings, on the common factors. (Roll and Ross, 1980, p. 1074)

7. See, for example, Sharpe and Cooper (1972), Douglas (1968) and Black et al. (1972).8. MSCI total return indexes with gross dividends.

REFERENCES

Black, F., Jensen, M.C. and Scholes, M. (1972), ‘The Capital Asset Pricing Model:Some Empirical Tests’, in M.C. Jensen (ed.), Studies in the Theory of Capital

Markets, New York: Praeger.De Soto, H. (2000), The Mystery of Capital – Why Capitalism Triumphs in the West

and Fails Everywhere Else, New York: Basic Books.Douglas, G. (1968), Risk in the Equity Markets: An Empirical Appraisal of Market

Effi ciency, Ann Arbor, MI: University Microfi lms, Inc.Elton, E.J. and Gruber, M.J. (1995), Modern Portfolio Theory and Investment

Analysis, 5th edn, New York: John Wiley & Sons, Inc.Faure, M. and Skogh, G. (2003), The Economic Analysis of Environmental Policy

and Law – an Introduction, Cheltenham, UK and Northampton, MA, USA: Edward Elgar.

Friend, I., Landskroner, Y. and Losq, E. (1976), ‘The Demand for Risky Assets and Uncertain Infl ation’, Journal of Finance, 31, 1287–97.

Kasper, W. and Streit, M.E. (1999), Institutional Economics – Social Order and Public Policy, Cheltenham, UK and Northampton, MA, USA: Edward Elgar.

Merton, R.C. (1973), ‘An Intertemporal Capital Asset Pricing Model’, Econom-etrica, 41, 867–87.


Mueller, D.C. (2003), The Corporation – Investments, Mergers and Growth, London: Routledge.

North, D.C., (1990) Institutions, Institutional Change and Economic Performance, Cambridge and New York: Cambridge University Press.

Roll, R. (1977), ‘A Critique of the Asset Pricing Theory’s Tests: Part I: On Past and Potential Testability of the Theory’, Journal of Financial Economics, 4, 129–76.

Roll, R. (1988), ‘R2’, Journal of Finance, 43 (2), 541–66.Roll, R. and Ross, S.A. (1980), ‘An Empirical Investigation of the Arbitrage

Pricing Theory’, Journal of Finance, 35, 1073–1103.Ross, S. (1976), ‘The Arbitrage Theory of Capital Asset Pricing’, Journal of

Economic Theory, 13 (3), 341–60.Sharpe, W.F. and Cooper, G.M. (1972), ‘Risk-Return Class of New York Stock

Exchange Common Stocks, 1931–1967’, Financial Analysts Journal, 28, 46–52.Williamson, O.E. (2000), ‘The New Institutional Economics: Taking Stock,

Looking Ahead’, Journal of Economics Literature, 38 (3), 595–613.

185

9. The stock market, the market for corporate control and the theory of the fi rm: legal and economic perspectives and implications for public policySimon Deakin and Ajit Singh1

1. INTRODUCTION

In this chapter we consider the relationship between shareholder value, the market for corporate control, and economic and legal theories of the fi rm. We argue that, contrary to current conventional wisdom, support for an active market for corporate control is neither a core principle of company law nor an essential ingredient of fi nancial and economic development. Indeed, the opposite could well be the case – an important element of reform should be the prevention of the emergence of a market for corpo-rate control. The absence of such a market in coordinated market systems, such as Germany and Japan during their modern economic development, was not an evolutionary defi cit but an eff ective and positive institutional arrangement. The unravelling of that arrangement, which is currently being encouraged by regulatory changes and which some commentators see as a necessary part of adjustment to globalization, has the potential to destabilize existing production regimes, although so far it has failed to have this eff ect. Aspects of a regulatory regime favourable to the market for corporate control, such as a mandatory bid rule, have been adopted in many developing economies over the past decade; however, in the absence of countervailing power for employees of the kind found in most European systems, it is possible that these developments could impose signifi cant economic and social costs associated with restructuring.

The chapter is ordered as follows. Section 2 outlines the relationship between shareholder value and the core principles of company law in the common law and civil law worlds. It is argued here that company law systems, regardless of legal origin, tend to recognize the principle of


managerial autonomy from shareholder control, and provide managers with discretion to run the company in such a way as to maximize returns to all stakeholders and not simply shareholders. Section 3 turns the focus to takeover regulation, which we suggest is the pivotal factor that has led to the prioritization of shareholder interests in common law systems. We explore the diff erent approach to takeover bids in civilian regimes and discuss the implications of attempts to encourage a market for corporate control through regulatory changes in continental Europe and Japan, before looking at the adoption of similar regulatory moves in emerging and transition systems.

Having reviewed the common law and civil law approaches to the market for corporate control, we turn in Section 4 turn to economic analysis. The takeover mechanism plays a pivotal role in many branches of economic theory including the theory of the fi rm, the theory of industrial organiza-tion and welfare economics. It is also critical to an analysis of economic and industrial policy. We, however, confi ne ourselves in this chapter to analysing the relationship between the market for corporate control and the theory of the fi rm. This subject is examined in Section 4, together with a discussion of alternative views on the effi ciency of takeovers, and the need or otherwise for government regulation. Section 5 contains a discussion of the pricing process and the takeover mechanism on the stock market. It provides empirical evidence on aspects of the market for corporate control. Section 6 briefl y concludes.

2. SHAREHOLDER VALUE AND THE LEGAL CONCEPTION OF THE FIRM IN THE COMMON LAW AND CIVIL LAW

The idea that the managers of private-sector companies should act as the agents of shareholders is a focal point of the contemporary corporate gov-ernance debate. According to this view, the pursuit of shareholder value is the single ‘corporate objective function’ which drives organizational and allocative effi ciencies (Jensen, 2001). Its infl uence is increasingly felt in civilian systems that have, up to now, enjoyed a diff erent tradition, and its adoption in transition economies and in the developing world is on the policy agenda there. This apparent convergence is occurring in large part as a result of ‘the recent dominance of a shareholder-centred ideology of corporate law among the business, government and legal elites in key com-mercial jurisdictions’ (Hansmann and Kraakman, 2001: 439).

Too close a focus on the supposed effi ciency of prevailing institutions is liable to make us forget the often tortuous and uneven path by which they

The stock market and market for corporate control 187

came to acquire their apparent dominance. Britain’s industrial revolution took place during a period when few businesses enjoyed limited liability. In the US, many states allowed personal claims to be brought against shareholders for corporate debts late into the nineteenth century and some, including California, into the twentieth. Yet corporate law scholars today assert that limited liability and the partitioning of corporate from per-sonal assets are essential parts of the legal ‘bedrock’ supporting enterprise (Hansmann and Kraakman, 2001). This view arguably ascribes ‘survival value’ to institutions whose endurance may have more to do with histori-cal contingency than effi ciency (Deakin, 2003). Is the same true of today’s norm of shareholder value?

It is surprisingly diffi cult to fi nd support within core company law for the notion of shareholder primacy. It cannot be found by referring to the rhetorical claim, associated with today’s pension funds and other institu-tional investors, that shareholders ‘own the company’. No legal system acknowledges the claims that shareholders ‘own the company’. If we understand the company to be the fi ctive legal entity which is brought into being through the act of incorporation, it is not clear in what sense such a thing could be ‘owned’ by anyone. But, more pertinently, nor does the ownership of a share entitle its holder to a particular segment or portion of the company’s assets, at least while it is a going concern (see Parkinson, 2003).

The law on directors’ duties is no more helpful. In the English-law based common law systems, with only a few exceptions, directors’ fi duci-ary interests of loyalty and care are owed to the company, not directly to the shareholders. In practice, the company’s ‘interests’ will often be syn-onymous with those of its members, that is, the shareholders. However, shareholders are not entitled to engage directly in the management of the enterprise; this is the responsibility of the board. According to Delaware corporate law, ‘the business and aff airs of every corporation . . . shall be managed by or under the direction of a board of directors’ (see Millon, 2002: 92). Many of the formative cases of English company law, dating from the nineteenth and early twentieth centuries, make the same point (see Davies, 1997: 183–8).

Company law does not say anything about the level of returns to which shareholders are entitled, nor about the time scale over which their expectations are to be met. This ambiguity enabled the Company Law Review Steering Committee, in the review of UK company law which was concluded in 2002, to express its support for the idea of ‘enlightened shareholder value’. This implies ‘[a]n obligation on directors to achieve the success of the company for the benefi t of the shareholders by taking proper account of all the relevant considerations for that purpose’ including


‘a proper balanced view of the short and long term, the need to sustain eff ective ongoing relationships with employees, customers, suppliers and others; and the need to maintain the company’s reputation and to con-sider the impact of its operations on the community and the environment’ (Company Law Review Steering Group, 2000: 12; see also Company Law Review Steering Group, 2001: 41).

The Steering Group regarded its proposal as a compromise between the ‘enlightened shareholder value’ position and a ‘pluralist’ point of view which would have seen management as having multiple commitments to a range of stakeholder groups. The Steering Group accepted the position of agency theory that making management formally accountable to a diverse body of stakeholders might limit the eff ectiveness of managerial deci-sion-making and blur lines of accountability. Nevertheless, the Steering Group’s proposal was based on the proposition that ‘companies should be run in such a way which maximizes overall competitiveness and wealth and welfare for all’ (Company Law Review Steering Group, 2000: 14–15, emphasis added). The means chosen to achieve this end were the ‘inclusive duty’ and ‘broader accountability’:

The proposed statement of directors’ duties requires directors to act in the collective best interests of shareholders, but recognizes that this can only be achieved by taking due account of wider interests. The transparency element provides the information needed to underpin this approach to governance. Just as importantly, we believe that [a] wider reporting requirement – particularly for large companies – will be an important contribution to competitiveness. Companies are increasingly reliant on qualitative and intangible, or ‘soft’ assets such as the skills and knowledge of their employees and their corporate reputa-tion. The reporting framework must recognize this and ensure that companies provide the market and other interests with the information they need to understand their companies’ business and assess performance. (Company Law Review Steering Group, 2000: 14–15).

Section 172 of the Companies Act 2006, which is headed ‘Duty to promote the interests of the company’, now provides that:

(1) A director of a company must act in the way he considers, in good faith, would be most likely to promote the success of the company for the benefi t of its members as a whole . . .

(3) In fulfi lling the duty imposed by this section a director must (so far as reason-ably practicable) have regard to –(a) the likely consequences of any decision in the long term,(b) the interests of the company’s employees,(c) the need to foster the company’s business relationships with suppliers, customers and others,


(d) the impact of the company’s operations on the community and the environment,(e) the desirability of the company maintaining a reputation for high standards of business conduct, and(f) the need to act fairly as between the members of the company.

Such a position is by no means as contrary to US corporate law as many writings on that system might make it seem. According to Leo Strine, a leading Delaware company law judge, writing extra-judicially but express-ing a view which is arguably quite compatible with the general thrust of Delaware law (as opposed to corporate governance practice) on the issue of shareholder value,

Most American workers obtain the bulk of their wealth from their labor and even most top American managers can trace their wealth (including the equity they have accumulated) to their labor as executives. Therefore, both manage-ment and labor might be thought to have more concern than trust fund babies or investment bankers do for the continued ability of American corporations to support domestic employment. Likewise both management and labor are likely to view a public corporation as something more than a nexus of contracts, as more akin to a social institution that, albeit having the ultimate goal of produc-ing profi ts for stockholders, also durably serves and exemplifi es other societal values. In particular, both management and labor recoil at the notion that a cor-poration’s worth can be summed up entirely by the current price equity markets place on its stock, much less that the immediate demands of the stock market should thwart the long-term pursuit of corporate growth. (Strine, 2007: 4)

The idea that the company is an organization or ‘entity’ with a distinct set of interests above and beyond those of all the stakeholder groups com-bined is even more clearly articulated in the civil law systems. These recog-nize the ‘enterprise’ as a legal form that corresponds to the organization. This is distinct from the concept of the ‘company’ that essentially describes a set of claims to income streams and property rights. The explicit recogni-tion of the company’s organizational dimension has implications for the way in which stakeholder interests are regarded, as this quotation from the Viénot report on French corporate governance recognizes:

In Anglo-Saxon countries the emphasis is for the most part placed on the objec-tive of maximising share values, whilst on the European continent and France in particular the emphasis is placed more on the human assets and resources of the company . . . Human resources can be defi ned as the overriding interest of the corporate body itself, in other words the company considered as an autonomous economic agent, pursuing its own aims as distinct from those of its sharehold-ers, its employees, it creditors including the tax authorities, and of its suppliers and customers; rather, it corresponds to their general, common interest, which is that of ensuring the survival and prosperity of the company. (Viénot, 1995, cited in Alcouff e and Alcouff e, 1997)


Various versions of this concept of the ‘company interest’ have been expressed in continental European law and practice since the period of industrialization in the second half of the nineteenth century; it was articulated in the ‘communitarian’ concept of the enterprise advanced by the German jurist Otto von Gierke in the 1890s, and in the corporatist model popularized by the industrialist and politician Walther Rathenau in the 1920s. Although German corporate law scholarship has been increas-ingly infl uenced by agency theory since the 1990s, a signifi cant strand of it remains sceptical of the US-inspired model, and theories based on the varieties of capitalism approach, stressing the importance of fi rm-specifi c human capital in coordinated market systems, have recently been deployed to explain and defend the core civilian model (see Gelter (2007) on Germany, and Rebérioux (2007) on France).

3. THE LEGAL REGULATION OF TAKEOVER BIDS: COMMON LAW AND CIVILIAN APPROACHES2

3.1 The Origins of Takeover Regulation

The vital factor in institutionalizing the shareholder value norm in the common law or ‘liberal market’ systems has been the encouragement given to the hostile takeover bid by regulatory changes which have often been in tension with the core principles of company law. Takeover regulation is a comparatively recent phenomenon. Following the economic depression of the inter-war years, there was intense discussion of the responsibilities of companies, the role of the fi nancial system, and the need for public regulation of the economy. The solution argued for by Adolf Berle, in particular in his debate with E. Merrick Dodd in the mid-1930s, was to reverse the ‘separation of ownership and control’ (which at that point was a comparatively new development in the US: see Hannah, 2007) by return-ing control to the shareholders (see Dodd, 1932; Berle, 1932). But this route was thought to be impractical during a period when it was generally considered that the large enterprise was the norm, family ownership was in decline, and further dispersion of ownership could be expected. Rather, the outcome was a combination of managerial and public control of the corporation, much as Dodd had advocated, and as Berle came belatedly to accept (Berle, 1962; see Ireland, 1999).

From the mid-1970s onwards, the intellectual climate in Britain and North America began to turn away from public and social control of industry, with results which are now familiar: the privatization of the large, state-owned corporations and utilities, and the so-called deregulation of


many areas of economic life. As part of this policy turn, and as a result of the changes to industry and the economy that fl owed from it, the corporate governance debate was relaunched. Corporate fi nance scholars argued that, in place of the state, the market should provide the principal mecha-nism for controlling the managers of large corporations. These authors, in common with Berle and Means, argued that dispersed share ownership – the fracturing of share capital among hundreds, sometimes thousands, of individual holdings – freed management from direct supervision by inves-tors. The solution, however, lay in the activation of the very instrument which the legislation of the New Deal era in the United States, and of the post-war regulatory state in many other countries, had sought to constrain in the interests of economic stability: the capital market.

The mechanism by which this was achieved was the hostile takeover bid. By off ering to buy shares in a company at a premium over the existing stock market price, so-called corporate raiders or predators could obtain control of the enterprise, remove the existing managerial team, and install one of their own. If the shareholders had no greater interest in the company than the fi nancial investment represented by their shares, they could be induced to sell in return for the premium off ered by the raider, in particular if they felt that the incumbent managerial team was not looking after their interests. For the bidder, the cost of mounting the bid and buying out the shareholders could be recouped, after the event, by disposing of the com-pany’s assets to third parties. If the company had not been well run before, these assets would, by defi nition, be worth more in the hands of others. Thus the hostile takeover bid performed a number of tasks. It empowered shareholders, who now had a means to call management to account if it was underperforming. Conversely, the hostile takeover disciplined mana-gerial teams, who knew that their jobs and reputations were on the line if a bid was mounted. In addition, it provided a market-led mechanism for the movement of corporate assets from declining sectors of the economy to more innovative, growing ones. That, at least, was the theory.

The rise of the hostile takeover can be traced back to the late 1950s and early 1960s in the UK and the US. There had always been mergers and acquisitions of fi rms; what was relatively new was the idea of a bid for control directed to the shareholders, over the heads of the target board. In the inter-war period, incumbent boards would ‘just say no’ to unwelcome approaches from outsiders, often without even informing shareholders that a bid was on the table (Hannah, 1974; Njoya, 2007). At this stage, accounting rules had not evolved to the point where companies were under a clearly enforceable obligation to publish objectively verifi able fi nancial information. This changed in the post-war period as a consequence of the legal and accounting changes that were put into place in both Britain and


America by way of response to the fi nancial crises of the 1930s. Greater transparency made it easier for unsolicited bids to be mounted and more diffi cult for incumbent boards to resist them. Institutional protection for minority shareholders followed, with the adoption in Britain in 1959 of the Bank of England’s ‘Notes on Reconstructions and Amalgamations’ and, in 1968, the City Code on Takeovers and Mergers; 1968 was also the year in which the US Congress adopted the Williams Act, instituting a system of regulation for hostile tender off ers for US listed companies.

3.2 The US Model

The Williams Act sets time limits on tender off ers and requires bidders with 5 per cent of a company’s stock to disclose their holdings and to give an indication of their business plan for the company, but it does not explicitly rule out two-tier or partial bids as it does not contain a mandatory bid rule along the lines of the City Code. It regulates fraudulent activity, broadly defi ned, but does not place target directors under a clear-cut duty of care to provide independent fi nancial information to shareholders in the way that the Code does. At state level, US courts have accepted that, under the ‘busi-ness judgment’ rule, target directors can take steps to resist a hostile takeo-ver where they act in good faith and with ‘reasonable grounds for believing that a danger to corporate policy and eff ectiveness existed’; specifi cally, they can take into account the ‘inadequacy of the price off ered, nature and timing of the off er, questions of illegality, the impact on “constituencies” other than shareholders (i.e., creditors, customers, employees, and perhaps even the community generally), the risk of non-consummation, and the quality of the securities being off ered in exchange’.3 The Delaware courts have nevertheless vacillated between an ‘auction rule’ which would require the board to take steps to maximize shareholder returns in the event of a proposed change of control,4 and the ‘just-say-no’ defence under which the target board would be ‘obliged to charter a course for the corporation which is in its best interests without regard to a fi xed investment horizon’ without being ‘under any per se duty to maximize shareholder value in the short term, even in the context of a takeover’.5

US corporate law has permitted the growth of a battery of anti-takeover defences of the type that are virtually never adopted in the case of publicly quoted companies in Britain. Shark repellents enable the composition of the board to be structured so as to make it diffi cult for an outsider to gain control. For example, company byelaws may stipulate that directors are elected for three-year terms, with only part of the board coming up for renewal each year. It is also common for byelaws to prohibit greenmail (a raider forcing the target board to buy back its shares at a premium), to


prevent shareholders from voting by proxy, to lengthen the gaps between general meetings, and to require supermajority decisions to change these and other rules. Poison pills – various devices whereby insider shareholders acquire rights which are triggered when a hostile takeover bidder makes its entry – have been permitted, including the fl ip-in (where if the raider increases its shareholding above a certain level the target board declares a high dividend for existing shareholders, or existing shareholders are given the right to buy additional stock at half the market price) and the fl ip-over (shareholders get the right to buy stock of the new parent at a 50 per cent discount).

In addition, most states have enacted anti-takeover statutes that enable companies to adopt internal rules aimed at fending off hostile bids. The fi rst wave of such statutes was ruled unconstitutional in Edgar v. MITE Corp.6 on the grounds that the Williams Act preempted them. In this case, an Illinois statute that extended the bid timetable beyond that set by the Williams Act and gave state-level competition authorities the right to nullify off ers was struck down. However, a second generation of ‘share control’ statutes (providing the incumbent shareholders with the power to decide on whether a raider with a controlling stake should retain the voting rights of its shares) was upheld in CTS Corp. v. Dynamics Corp.7 at around the same time as the Supreme Court gave a restrictive ruling to the Williams Act, deciding that it did not bar defensive actions such as ‘crown jewel’ options or sales.8 This simultaneously limited the scope of federal regulation and opened the way for further pro-defensive laws at state level. A ‘third generation’ of state laws followed, which, broadly speaking, validated various poison pill defences and introduced ‘constituency’ or stakeholder provisions into the defi nition of directors’ fi duciary duties.

These ‘stakeholder statutes’ vary in strength. Certain of them include provisions for profi t disgorgement, whereby the state imposes a high tax on any short-term profi ts by raiders acquiring shares in connection with a bid, for example as a result of greenmail (selling their shares back to the company at a premium to the market). Modifi cations to directors’ duties include provisions to the eff ect that fi duciary duties are owed solely to the corporation and that no party, not even the shareholders, can enforce them directly; that directors may consider the long-term interests of the corporation; and that the directors need not regard any one constituency’s interest as dominant. The stakeholder statutes, together with the adoption of poison pills by a majority of large public corporations, are credited with having helped to restrict the number and volume of takeovers at the end of the 1980s: by the mid-1990s, over two-thirds of large US public corpo-rations had adopted poison pills, and acquisitions of public corporations, which had been running at over 400 per annum in the late 1980s, had fallen


to half that fi gure (Useem, 1996: 27–8). However, the stakeholder statutes did little to defl ect the wider impact of shareholder pressure on corporate management, which today increasingly takes the form of pressure from activist hedge funds and private-equity-led restructurings, and which has been refl ected in continuing high levels of lay-off s and restructurings (Uchitelle, 2006).

3.3 The British Model

The City Code, like the Williams Act, dates from the late 1960s but, unlike the US measure, it did not until recently have statutory backing. The Panel on Mergers and Takeovers, a self-regulatory body set up by the fi nancial and legal professions and fi nancial sector trade associations based in the City of London, had no direct legal powers of enforcement. Its provisions were strictly observed, however, since UK-based fi nancial and legal pro-fessionals who were found to have breached the Panel’s rulings could be barred from practising as a consequence. As a result of the adoption by the European Union of the Thirteenth Company Law Directive,9 the Panel has recently acquired a statutory underpinning,10 but the substance of the Code remains essentially the same as it was before, and it continues to be based on the Panel’s deliberations and rulings. The expectation of both the UK government and the Panel is that the implementation of the Directive will not have a major impact on the Panel’s mode of operation.11

The City Code refl ects the strong infl uence of institutional shareholder interests within the UK fi nancial sector, and their capacity for lobbying to maintain a regulatory regime, which operates in their favour (Deakin and Slinger, 1997; Deakin et al., 2003). Its most fundamental principle is the rule of equal treatment for shareholders: ‘all holders of the securities of an off eree company of the same class must be off ered equivalent treatment’.12 This is most clearly manifested in the Code’s ‘mandatory bid’ rule which requires the bidder, once it has acquired 30 per cent or more of the voting rights of the company, to make a ‘mandatory off er’ granting all share-holders the chance to sell for the highest price it has paid for shares of the relevant kind within the off er period and the preceding 12-month period.13 Partial bids, involving an off er aimed at achieving control through pur-chasing less than the total share capital of the company, require the Panel’s consent, which is only given in exceptional circumstances.14 During the bid, information given out by either the bidder or the target directors must be made ‘equally available to all off eree company shareholders as nearly as possible at the same time and in the same manner’.15

The Code also imposes on target directors a series of specifi c obliga-tions that can be thought of as clarifying their duty to act bona fi de in the


interests of the company but in some respects extend this duty. The target directors must fi rst of all obtain competent, independent fi nancial advice on the merits of the off er,16 which they must then circulate to the sharehold-ers with their own recommendation.17 Any document issued by the board of either the bidder or the target must be accompanied by a statement that the directors accept responsibility for the information contained in it.18 While the point is not completely clear, the likely eff ect of this is to create a legal duty of care owed by the directors to the individual shareholders to whom the information is issued (and not to the company as is the case with their general fi duciary duties).19

All this places the directors of the target in the position of being required to give disinterested advice to the shareholders on the merits of the off er, and makes it more diffi cult for them to resist a bid simply on the grounds that it would lead to the break-up of the company. In a case where the board considers that a hostile bid would be contrary to a long-term strategy of building up the company’s business in a particular way, it can express this opinion, but it must be cautious in doing so, since it still has a duty to provide an objective fi nancial assessment of the bid to the shareholders. In the case of the takeover of Manchester United FC by the US business-man Malcolm Glazer in 2005, the board took the view that Glazer’s off er, because it would impose a high debt burden on the company, was not in its long-term interests. However, the board was also aware that the off er could well be regarded as a fair one, since it was by no means clear that the shareholders would not be better off by accepting it. The board issued this statement:

The Board believes that the nature and return requirements of [the proposed] capital structure will put pressure on the business of Manchester United . . . The proposed off er is at a level which, if made, the Board is likely to regard as fair . . . If the current proposal were to develop into an off er . . . the Board considers that it is unlikely to be able to recommend the off er as being in the best interests of Manchester United, notwithstanding the fairness of the price.

Following this statement, a majority of the shareholders accepted Glazer’s bid.

General Principle 9 of the Code used to state that ‘it is the shareholders’ interests, taken as a whole, together with those of employees and credi-tors, which should be considered when the directors are giving advice to shareholders’. This provision, like section 309 of the Companies Act 1985, was less signifi cant in practice than it appeared to be on paper, since it provided no basis on which employees or creditors, who had (and have) no standing before the Takeover Panel, can challenge a board’s decision. Case law on fi duciary duties from the 1980s also suggested that, during a


contested takeover, only the interests of the shareholders could be taken into account.20 As a result it probably matters little that, following recent revisions to the Code, the relevant General Principle, which is based on the parallel provisions of the European Union’s Thirteenth Company Law Directive, now simply states that ‘the board of an off eree company must act in the interests of the company as a whole and must not deny the holders of securities the opportunity to decide on the merits of the bid’.21

The Code used to require the bidder to state, in its off er document, ‘its intentions regarding the continuation of the business of the off eree company; its intentions regarding any major changes to be introduced in the business, including any redeployment of the fi xed assets of the off eree company; the long-term commercial justifi cation for the proposed off er; and its intentions with regard to the continued employment of the employ-ees of the off eree company and its subsidiaries’. This meant little in practice; it simply required the bidder to issue a general statement of its intentions, which generally took the form of a standard-term or ‘boilerplate’ provi-sion in off er documents. However, as a result of changes made to the Code following the implementation of the Thirteenth Directive, more prescrip-tive provisions concerning the potential impact of takeovers on employees have been introduced. The bidder must now provide detailed information on its strategic intentions with regard to the target, possible job losses, and changes to terms and conditions of employment,22 and the target must give its views, in the defence document, on the implications of the bid for employment.23 Breach of these provisions is a criminal off ence. They also have potentially signifi cant implications for employees’ consultation rights under labour law.24 In addition, employee representatives of the target have the right to have their views of the eff ects of the bid on employment included in relevant defence documents issued by the target.25

The Code contains extensive provisions controlling the use of defences against hostile bids (or ‘frustrating measures’ in the terms used by the Thirteenth Directive). Once an off er is made or even if the target board has reason to believe that it is about to be made, the target board cannot issue new shares; issue or grant options in respect of any unissued shares; create securities carrying rights of conversion into shares; sell, dispose or acquire assets of a material amount, or contract to do so; or ‘enter into contracts otherwise than in the ordinary course of business’.26 General company law is also relevant here. The ‘proper purposes’ doctrine prevents the board issuing shares for the purpose of forestalling a hostile takeover, even well in advance of any bid being made.27 Other advance anti-takeover defences, such as the issue of non-voting stock or the issuing of new stock to friendly insiders, have been discouraged by a combination of the Listing Rules of the London Stock Exchange and institutional shareholder pressure. Protection


of pre-emption rights, or the rights of existing shareholders to be granted preference when new stock is issued, is recognized by legislation28 as well as by guidelines issued by stock exchange and fi nancial industry bodies.29 The issue of non-voting stock is permissible under general company law, but is vigorously opposed in practice by institutional shareholders.30

Thus the Takeover Code, taken in conjunction with related aspects of company law, can be seen to provide strong protection for the interests of the target shareholders; that, after all, has been its main purpose (see Johnston, 1980). An important side eff ect of this protection, however, is to encourage hostile takeover bids by placing limits on the defensive options available to the target management. An incumbent management is not required to be completely passive, and is permitted to put a case in its own defence, but opportunities for defence only arise in the context of an over-riding responsibility to see that the shareholders’ interests are safeguarded. The eff ect is not far removed from that of an ‘auction rule’ which requires the incumbent management to extract the highest possible price for the target shareholders, if necessary by making it possible for rival off ers to be made. The entry of second bidders is facilitated by the bid timetable imposed by the Code and by the eff ective ban on two-tier and partial bids which might otherwise be used to strong-arm the target shareholders into accepting the terms of the fi rst bid.

These regulatory features might be thought to deter bids, by increas-ing the risk that either the target shareholders or any second bidder will free-ride on the eff orts of the initial bidder. However, the possibility of free-riding by the shareholders is alleviated by the right of the bidder com-pulsorily to purchase the last 10 per cent of shares in the event of taking control; in the language of the European Directive, a ‘squeeze-out’ rule (on its economic eff ects, see Yarrow, 1985). Other factors which serve to reduce the risk of an initial bid failing due to free-rider eff ects are the concentration of voting shares in most UK publicly quoted companies in the hands of a relatively small number of institutional shareholders (so reducing the number of shareholders who need to be persuaded to sell) and the right of an initial bidder to raise its off er price during the bid period (thereby enabling it to over-bid a second bidder). While there may, then, be a certain screening-out of partial bids which, given their oppressive nature, are arguably not effi ciency-enhancing in any event (see Yarrow, 1985), the eff ect of the Code is to reduce the autonomy enjoyed by the management of the target company in relation to its shareholders and thereby to limit the defensive options it has available to it.

This suggestion is borne out by empirical research carried out in Cambridge in the late 1990s (Deakin et al., 2003). In this study, the objec-tive was to construct a sample of bids which contained examples of both


hostile and agreed bids and of cross-border bids by UK companies and for UK companies mounted during the period 1993–96. In interviews we put this question to company directors, lawyers, merchant bankers, insti-tutional shareholders and employee representatives:

Did directors’ duties to consider interests of creditors and employees as well as those of shareholders aff ect the preparations for, the conduct of and the after-math of the bid?

On the central question of directors’ duties, the response was almost invariably that, while directors might consider employees’ and creditors’ interests, the outcome of a bid was determined by shareholder value. Shareholder value took precedence over all other considerations. The responses to the question are separated out below by group, with advis-ers fi rst, followed by directors, institutional investors, and employee representatives.

A typical comment from an adviser was as follows:

Directors do consider employees’ interests, but no one really knows what that means. At the margin the touchy-feely things matter, but the board of directors, faced with two people off ering £1 and £1.10, must go for the higher. The deci-sion, of course, is not usually put like that, but I don’t know of any cases where employees’ interests have come fi rst.

Employees were only mentioned out of lip service to the obligation of the off eror company to state its intentions with regard to employment:

Directors’ duties to consider other interests are rarely an issue unless the company is near to insolvency. These clauses together are a bit of a sop. Rule 24 of the Code requires a statement of intentions towards employees, which always gets reduced to the standard phrase: ‘the bidder will ensure that all rights of the target employees will be met in full.’ Sometimes people do say more – sometimes a target will screw a stronger statement out of the bidder. And where companies intend not to make redundancies, they will tend to say it.

More pithily, we were told: ‘much is spoken about directors’ duties to employees, but it is rarely relevant’, and ‘the Takeover Code and Companies Acts just muddle these issues up: directors have to recommend “the deal” when they are really just recommending the price.’

Directors told us that their focus was on the fi nancial aspects of a bid:

The one thing that [our merchant bankers] kept saying was that ‘you have to be sure that when you say that a price is inadequate, you mean it and can back it up.’ Were we advised that we could take into account the interests of the


company as a whole? No – the primary advice was that ‘there is a price at which you have to say yes.’

In particular, non-executive directors were identifi ed as advocates for the shareholder interest, even where this meant dismembering the corporate enterprise:

Were we advised of our legal obligations to our shareholders? Yes – there was lots of advice. One of the non-executive directors did push us hard to consider closure and selling up as an option to get maximum shareholder value (about fi ve years before the bid).

Institutional investors likewise thought that directors should focus on shareholder concerns. One was ‘happy with the idea that directors owe duties to “the company”’ but was of the view that ‘during a bid, especially, the directors understand this as being a duty to shareholders’. Another considered that for directors to perform according to their fi duciary duties, ‘they had to show that it was in the interests of shareholders to sell’. The pursuit of stakeholder interests was not seen as a viable alternative to shareholder value:

It is hard to make a case that [the duty to further the interests of the company as a whole] aff ected the bid greatly. In principle a defending company might put employees’ interests before those of shareholders but they are basically serving shareholders’ interests fi rst. If directors have a duty, it is to ensure that employ-ees have marketable skills. I see directors’ duties to employees as being more like pension rights protection than long-term employment safeguards.

Employee representatives were less clearly opposed to bids than might have been thought. Hostile bids were sometimes seen as shaking up incum-bent managerial teams with which the employees had little by way of common interest. Hence employee representatives commented unfavour-ably on the tendency of target directors to be excessively well rewarded, even before bids, in pay and share options, and on the negative eff ect that this had on the workforce. Particular criticism was reserved for the practice of linking managerial remuneration to the number of workers dismissed:

The other thing that caused trouble was the directors’ incentives schemes. They had a bonus system which had work completed according to certain targets divided by the number of staff that they employed to do it. So what they did was to sack a lot of staff , and employed outside contractors, to fulfi l their conditions and increase their bonuses.

None of the employee representatives were convinced that a higher commitment from management to consultation would have materially


aff ected the bids in which they were involved. In part this was out of a frank recognition that the decision was in the hands of shareholders and hence was ‘purely a commercial thing’. The priority was to keep lines of communication open after the bid in an attempt to avoid compulsory redundancies and smooth the way of the new owners. This was a typical comment:

We take the view now that we’re not going to be able to prevent [the takeover] – so we try to get the best deal we can. Given the current industrial relations climate, I don’t think that even a ‘requirement to consult’ would make much diff erence.

For target directors, the nature of the advice received was of paramount importance. During bids, they saw their duty in terms of maximizing the potential value of the company as a fi nancial asset of the shareholders. This obligation stood before any requirement to consult employees, to con-sider their interests, or to further the interests of the company as a whole. Even outside the bid period, the perceived ‘duty’ to focus on shareholder value could lead a non-executive director to see it as their role to force management to consider closing down the enterprise. Correspondingly, institutional investors applauded directors who saw their responsibilities in these terms.

The attitudes of employee representatives are best described as prag-matic. They expected little from target managers whose interests were seen to be tied up with share options and remuneration packages that would leave them better off whatever the outcome of the bid. There was no expec-tation of consultation with the target management, and no prospect of it making a diff erence to the outcome of the bid if it did take place. By con-trast, the intervention of bidders could be seen in a positive light, particu-larly where there had already been a breakdown of trust with incumbent management. Informal links could be established with the bidder at an early stage, and a relationship constructed with a view to the future, even though it was recognized on both sides that the most immediate issue was likely to be the management of redundancies.

In 1974 Leslie Hannah wrote that the takeover bid had ushered in ‘an economic system whose logic is still being developed and is still only imper-fectly understood’ (Hannah, 1974). We now see more clearly what kind of system it is. The takeover revolution was a catalyst for a raft of other meas-ures and devices aimed at ensuring that managers of large corporations acted fi rst and foremost in the interests of their shareholders. However, it is important to stress that, even in the UK context, the current focus on shareholder value is therefore the consequence not of the basic company law model but of those institutional changes which have occurred in capital


markets and securities law with increasing rapidity, in particular since the early 1980s, namely the rise of the hostile takeover bid and the increasing use of share options and shareholder value metrics. Thus the contemporary ‘norm’ or reference point of shareholder primacy is the result of a mix of institutional changes, the emergence of new forms of self-regulation and soft law, and shifts in corporate culture.

3.4 The Civil Law Model: Mainland Europe

In the civil law world, there has, until recently, been no equivalent to the rules on takeover regulation that are found in common law systems. This is not to say that there is no record, historically, of hostile takeover activ-ity in civil law countries; as Hannah (2007) points out, takeovers by share purchase did take place in Germany in the early years of the twentieth century. However, for much of the twentieth century, they were actively suppressed, particularly in the post-World War II period when they were seen as incompatible with economic reconstruction in Germany, France and Japan. The growth of corporate cross-shareholdings and the rise of bank-led governance in these systems led to stabilization of share owner-ship, but also to the sterilization of the external capital market as a mecha-nism for controlling management.

A major change appeared to be about to take place in the early 2000s as a result of the adoption of the Thirteenth Directive in the EU and changes in the Japanese system which encouraged the revival of hostile takeover activity, but a closer inspection also shows that there has been resistance to attempts to institutionalize a market for corporate control. The fi rst signifi cant document in the current round of initiatives was the report of the High Level Group of Experts on takeover bids, published in October 2002. This argued that what the EU needed was ‘an integrated capital market’ in which ‘the regulation of takeover bids [would be] a key element’ (High Level Group, 2002a: 18). The report noted that ‘the extent to which in a given securities market takeover bids can take place and succeed is determined by a number of factors’, including general or structural factors aff ecting fi nancial markets, and company-specifi c factors such as rules of company law and articles of association aff ecting voting rights, protection of minority shareholders, and the legitimacy of takeover defences. It then observed that ‘there are many diff erences between the Member States in terms of such general and company specifi c factors’, with the result that the EU lacked a ‘level playing fi eld’.

The substantive content of state-level company laws was also an issue for the High Level Group. The essence of the problem was that the laws of most member states did not suffi ciently conform to a model of corporate


governance in which managers understood their principal duty to be to return value to shareholders, and in which takeovers played a crucial dis-ciplinary role in reminding them of this obligation:

Actual and potential takeover bids are an important means to discipline the management of listed companies with dispersed ownership, who after all are the agents of shareholders. If management is performing poorly or unable to take advantage of wider opportunities the share price will generally under-perform in relation to the company’s potential and a rival company and its management will be able to propose an off er based on their assertion of their greater compe-tence. Such discipline of management and reallocation of resources is in the long term in the best interests of all stakeholders, and society at large. These views also form the basis for the Directive. (High Level Group, 2002a: 19)

The High Level Group could not have been clearer: they were proposing a measure based on the standard fi nance theory or ‘principal–agent’ view of the role of hostile takeover bids in enhancing shareholder value. The assertion that managers are ‘after all’ the agents of shareholders is one based on a particular economic-theoretical position, and has no grounding in the legal conceptions of the company that the High Level Group might have looked for in the laws of the member states. Even UK company law does not go this far; it has not followed the Delaware practice of sometimes referring to duties owed by directors to the shareholders rather than to the company as a separate entity. Be that as it may, it was very largely to the UK that the EU experts looked to fi ll out the content of the Directive. Even more so than its many predecessors, this draft of the Thirteenth Directive drew on the model of the City Code on Takeovers and Mergers, a text notable, as we have seen, for the high level of protection it gives minority shareholders and for its restriction of poison pills and other anti-takeover defences that US law, which is otherwise takeover-friendly, by and large allows (see Deakin and Slinger, 1997).

The High Level Group’s second report, in November 2002, struck a similar note in stressing the role of non-executive directors in monitoring management, which is a feature of British and American practice, but is relatively underdeveloped in other member states:

Good corporate governance requires a strong and balanced board as a monitor-ing body for the executive management of the company. Executive managers manage the company ultimately on behalf of the shareholders. In companies with dispersed ownership, shareholders are usually unable to closely monitor manage-ment, its strategies and its performance for lack of information and resources. The role of non-executive directors in one-tier board structures and supervisory directors in two-tier board structures is to fi ll this gap between the uninformed shareholders as principals and the fully informed executive managers as agents by monitoring the agents more closely (High Level Group, 2002b: 59).


Again, the standard fi nance or ‘principal–agent’ model was stressed, and a feature of the British and American systems was presented as if it had universal validity. Features of national systems that did not conform to the principal–agent approach, such as the distinctive role of worker directors and community representatives in two-tier boards, were simply shoehorned into the supposedly universal model. The High Level Group’s second report set out a series of objectives for reform of corporate govern-ance (among other things) which refl ected this point of view, and which were then incorporated into the Commission’s Action Plan on company law, with eff ect from 2003.31

What happened next, and in particular the fate of the Thirteenth Directive, is instructive. Although the Directive was eventually adopted, in 2004,32 this was only after a series of compromises had been agreed, which considerably diluted the draft presented by the Commission in 2002. Contrary to the expectation that the Directive would roll out a liberal-market model of takeover regulation along similar lines to those of the UK’s City Code on Mergers and Takeovers, in its fi nal form it allows member states to retain laws which permit multiple voting rights and limit shareholder sovereignty in various ways, such as allowing anti-takeover defences to be put in place in advance of bids.

Some of the derogations in the Thirteenth Directive are transitional; its general thrust is in favour of the principle of one-share-one-vote, and proportionality between investment risks and decision-making powers is clear. However, rather than impose a single model on member states, the Directive can be seen as setting out an ‘experimentalist’ framework for law-making at state level. This was far from being its original objective. Nevertheless, the result of the rough-hewn compromises which informed the fi nal text of the Directive is that the liberalization of takeover rules can be achieved in one of several diff erent ways, which may take into account specifi c features of the legal and institutional environments of the diff erent member states.

Both Germany and France have taken advantage of the derogations in the Directive. In Germany, the supervisory board of a listed company has the power to authorize poison-pill-like defences. In France the board of directors can issue warrants granting new stock to existing shareholders in the face of a hostile takeover bid, subject only to majority shareholder approval at an ordinary meeting. Germany is moving in the direction of a one-share-one-vote rule but this principle is not recognized by most large listed companies in France. Cross-shareholdings in both countries are not as strong as they were. In France, over 40 per cent of shares in the top 40 listed companies (the CAC 40) are now held by overseas pension and mutual funds (mainly based in the UK and the US), a considerable shift


from just a decade ago. However, in neither country has a market for cor-porate control to match the British or American model yet emerged.

Another signifi cant feature of the Thirteenth Directive is that it enabled the reformed takeover rules to make provision for information and consul-tation of employees. An element of employee consultation was present in earlier drafts of this Directive, and the provisions on this issue which were included in the fi nal text are not especially far-reaching, and do not go as far as the laws of a number of member states. However, the Thirteenth Directives set a pattern, in that mandatory employee consultation provi-sions were then included in other company law directives, including the directive on cross-border mergers, as well as the Societas Europaea (or ‘European company’) measures (where again there has been a long debate on this issue). This illustrates the complexities involved in translating the principal–agent model of corporate governance into specifi c legal provi-sions. The fi nance theory espoused by the High Level Group fi nds no room for managerial engagement with employees on issues of corporate govern-ance, regarding it as a qualifi cation of the principle of shareholder-based control of the fi rm. However, the issue of employee involvement is una-voidable when it comes to legislating at EU level. This is not just because organized labour interests have numerous possibilities for presenting their view when directives are being formulated, but also because the principle of employee consultation in the event of corporate restructurings has come to be recognized, over several decades, as an important point of reference within the EU legal order, as it is embodied in numerous labour law direc-tives as well as in the EU Charter of Fundamental Rights. Thus the inclu-sion of employee voice rights in the new EU takeover regime is consistent with the wider structure of EU law in the company and labour law fi elds, although the extent to which these rights provide real countervailing power to that of the capital markets remains to be seen.

3.5 The Civil Law: Japan

Most large Japanese enterprises are listed companies with (by international standards) a relatively high degree of dispersed ownership. In the immedi-ate post-war decades, cross-shareholdings were common, and indeed were actively deployed as means of limiting the infl uence of foreign investors. Between the mid-1960s and the mid-1970s the ‘stable shareholding ratio’ across the listed company sector as a whole, including cross-shareholdings, rose from 47 per cent to 62 per cent (Miyajima and Kuruoki, 2005: 5–6). However, the ratio had declined again to 45 per cent by 1993 and was only 24 per cent in 2003. Cross-shareholdings of the traditional type represented only 7.6 per cent of the total in 2003 compared with 17.6 per cent in 1993


(NLI Research, 2004). Foreign shareholdings have risen from 11.9 per cent of the market in 1996 to 26.7 per cent in 2005 (National Stock Exchanges, 2006). In 2006 around 8 per cent of the fi rst (main) section of the Tokyo stock market, 196 companies in total, were more than 30 per cent owned by overseas investors (TSE, 2007: 4).

At the same time, large Japanese companies continue to stress their role as social institutions or ‘community fi rms’ which provide stable employ-ment to a core of long-term employees, in return for a high level of commit-ment and identifi cation with the goals of the fi rm. This tension between the legal form of the enterprise and its changing ownership structure, on the one hand, and its aspect as a social institution, on the other, has recently been thrown into sharp relief by a series of hostile takeover bids.

The most controversial of these involved the planned takeover of Nippon Broadcasting System (NBS) by the Internet service provider Livedoor, which was launched in February 2005 (see Whittaker and Hayakawa, 2007). NBS had a cross-shareholding agreement with Fuji Television Ltd, which in turn dominated a corporate group, the Fuji-Sankei media con-glomerate. Livedoor’s intentions were widely interpreted as being based on ‘greenmail’. When NBS attempted to issue new stock in order to dilute Livedoor’s holdings and frustrate its bid, the courts declared the move unlawful. In granting Livedoor an injunction, the Tokyo District Court ruled as follows:

It is inappropriate for the board of directors of a publicly listed company, during a contest for control of the company, to take such measures as the issue of new shares with the primary purpose of reducing the stake held by a particular party involved in the dispute, and hence maintain their own control. In principle the board, which is merely the executive organ of the company, should not decide who controls the company, and the issuing of new shares, etc., should only be recognized in special circumstances in which they preserve the interests of the company, or the shareholders overall.

When this judgment was appealed, eventually, to the High Court, it was upheld:

The issue of new shares, etc., by the directors – who are appointed by the share-holders – for the primary purpose of changing the composition of those who appoint them clearly contravenes the intent of the Commercial Code and in principle should not be allowed. The issue of new shares for the entrenchment of management control cannot be countenanced because the authority of the directors derives from trust placed in them by the owners of the company, the shareholders. The only circumstances in which a new rights issue aimed pri-marily at protecting management control would not be unfair is when, under special circumstances, it aims to protect the interests of shareholders overall.


However, the High Court also ruled that defensive measures would be potentially legitimate in four situations: greenmail, asset stripping, a lever-aged buy-out, and share manipulation. This was an approach based in part on the jurisprudence of the Delaware courts (Milhaupt, 2006). Unable to make a new rights issue, NBS instead lent shares, minus voting rights, to two friendly parties, and Livedoor subsequently agreed to drop its bid. It sold its shares in NBS to Fuji Television, with Fuji Television, in its turn, buying around 12 per cent of the shares in Livedoor.

Around the same time, the economics ministry (METI) and the Ministry of Justice (MOJ) issued takeover guidelines that drew in part on the report of METI’s Corporate Value Committee (CVC). The report of the CVC refers to the concept of ‘corporate value’ in the following terms:

The price of a company is its corporate value, and corporate value is based on the company’s ability to generate profi ts. The ability to generate profi ts is based not only on managers’ abilities, but is infl uenced by the quality of human resources of the employees, their commitment to the company, good relations with suppli-ers and creditors, trust of customers, relationships with the local community, etc. Shareholders select managers for their ability to generate high corporate value, and managers respond to their expectations by raising corporate value through creating good relations with various stakeholders. What is at issue in the case of a hostile takeover is which of the parties – the bidder or the incumbent management – can, through relations with stakeholders, generate higher corporate value.

The Guidelines (METI and MOJ, 2005) take a more shareholder- orientated view, referring to corporate value as ‘attributes of a corporation, such as earnings power, fi nancial soundness, eff ectiveness and growth poten-tial, etc., that contribute to shareholder interests’. However, they also rec-ommend giving scope for companies to put anti-takeover defences in place to deal with what could be regarded as opportunitistic or predatory bids. In 2006 a new law, the Financial Instruments and Exchange Law, amending basic securities legislation, came into eff ect. This introduced a version of the mandatory bid rule: a party purchasing 10 per cent of a company’s stock over a three-month period would be required to make a public tender off er or be limited to holding no more than one-third of the company’s issued share capital. In 2006 changes to company law came into eff ect that formally allowed companies to put in place anti-takeover defences. These include the powers to issue special class shares with limited voting rights or which can be compulsorily repurchased by the company (thereby depriving a potential bidder of its stake), to make rights issues which exclude a bidder, and to issue golden shares which confer certain rights such as the power to appoint directors or restrain voting rights. The latter type of provisions requires two-thirds majority support from existing shareholders.

The response to these developments has been complex and multi-layered


(see Whittaker and Hayakawa, 2007; Buchanan and Deakin, 2007). On the one hand, a large number of companies have put in place takeover defences. By February 2007, 197 listed companies had announced anti-takeover strategies of various kinds (Nikkei, 2007). Some large companies, such as Toyota, have strengthened intra-group cross-shareholdings in an attempt to defl ect Livedoor-type bids, and others, such as the three main steel producers, have announced anti-takeover defence pacts.

At the same time, there has been some resistance to the growing use of anti-takeover defences. One of the main institutional investor bodies, the Pension Fund Association (PFA), has made clear its opposition to takeover defences that do not have the approval of a simple majority of shareholders. The Tokyo Stock Exchange (TSE) has also been hostile to poison pill type defences, seeing them as a barrier to stock market trans-parency and to accountability. In March 2006 the TSE amended its own guidelines to allow golden shares, after the main employers’ federation, the Keidanren, criticized the Exchange’s previous opposition to this type of arrangement, but the TSE guidelines continue to stress the need for major-ity shareholder approval, in line with the PFA position. Further evidence of growing shareholder pressure comes in the form of dividend increases, which in a number of cases can be traced to activist shareholder pressure in the companies concerned.

Having said that, the current position of Japanese law is a long way from the model of the City Code. Notwithstanding the introduction of a version of the mandatory bid rule, the Japanese position is closer to Delaware law, which permits poison pills, but with a clearer authorization for takeover defence in the face of ‘greenmail’ or asset restructuring. The concept of ‘corporate value’ is being distinguished from the US-inspired ‘shareholder value’ in contemporary debates. Managerial practice, too, continues to prioritize the model of the community fi rm, with only a few exceptions. This statement, made to one of the present authors by the president of a large company in the course of empirical research on Japanese corporate governance during the autumn of 2006 (see Buchanan and Deakin, 2007), is typical of current attitudes:

I’m not quite sure whether shutting out these sorts of opportunities [i.e. bid approaches] can really be called ‘corporate defence’. However – this is a Japanese sort of environment – the fact is that 6,000 people are working in our group and hitherto they have always had a great feeling of confi dence and attachment towards the management. Accordingly, with regard to philosophy, even if for the sake of argument someone were to appear with a philosophy that was even more elevated than ours, I would be very worried and doubtful as to whether these employees who are currently contributing their confi dence and attachment to us would continue to do so in the same way for them.


3.6 Takeover Regulation in Emerging and Transition Systems

In the economies of the common law world, there is growing evidence of shareholder rent extraction. A curious eff ect of successive takeover waves from the 1970s onwards is that, in Britain and America, the net contribu-tion of new equity to the fi nancing of the corporate sector as a whole has become negative. This is the result of share buy-backs and the ‘retirement’ of capital following mergers. The phenomenon has led to questioning of the sustainability of the current model from within the business school community, as in Allen Kennedy’s afterword to his 2000 book The End of Shareholder Value:

How many companies would spend their wealth on stock buyback programs if their objective was to create wealth? How many companies would see fi t to cut R&D expenditures if their objective was to build wealth? How many companies would cavalierly shed long-term, loyal employees, their heads crammed full of information valuable to the company, if their objective was to create wealth?

In a similar vein, Marjorie Kelly (2002), writing in the pages of the Harvard Business Review, argued that:

stock-market investors have become, collectively, an extraordinarily unpro-ductive force in business. Indeed, for the last two decades, their contribution to corporations has been literally negative . . . it’s wrong to shovel money out to shareholders in ever larger scoops and force other stakeholders to pay the price.

The shareholders’ role is no longer simply to supply fi nance to companies. Most trades of shares in listed companies consist of movements from one shareholder to another with no new capital being supplied to the company. Rather, as agency theory prescribes, the function of shareholders is to dis-cipline corporate management. Thanks to the takeover revolution and the changes associated with it, the managers of listed companies must maintain shareholder approval. If they do not, they face the prospect of a takeover bid. In practice, this means that companies have to satisfy, on a continuing basis, shareholders’ expectations for high rates of return on equity. If they can do this, a rising share price becomes an asset in its own right, which can be deployed to fund growth through acquisitions (Millon, 2002).

The position is, however, diff erent in civil law systems and in the devel-oping world. The net contribution of equity capital has been positive in those systems that have not followed the Anglo-American shareholder primacy norm: mainland Europe, Japan, and developing world economies such as Brazil and India (Singh et al., 2002).


Is this going to change as a result of shifts in takeover regulation in devel-oping and transition economies? One of the central features of the British (and now, to a degree, the EU) model is the mandatory bid rule. This is at the core of the City Code system, which aims to protect the right of minor-ity shareholders to access, in proportion to their holdings, the surplus gen-erated by a takeover bid. It is an important stimulus to the fragmentation and dispersion of ownership while also discouraging the construction of cross-shareholdings. Infl uenced by a mixture of British and EU practice, many systems have adapted a version of the mandatory bid rule in the past ten years as part of a general realignment of takeover regulation in favour of the protection of minority shareholder interests: in 1987 in Malaysia, 1994 in India, 2000 in Pakistan, 2000 in Chile, 2002 in Argentina, 2005 in Mexico (Siems, 2007).

A similar trend can be observed in transition systems as a result of the adoption of the Thirteenth Company Law Directive (Commission, 2007). Two important aspects of the Directive are the ‘board neutrality rule’, which limits the scope for takeover defences both ex ante and during a bid, and the ‘breakthrough rule’ under which poison pills and golden shares can be overridden during a bid. Most western European systems have taken up the opportunity provided by the Directive to derogate from both these rules, but the rate of take-up of derogations is lowest in the Central and Eastern European (CEE) countries which constitute the ‘accession’ member states: the board neutrality rule has been adopted in the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Slovakia and Slovenia, and the breakthrough rule has been adopted in Estonia, Latvia and Lithuania. Of the CEE accession states, only Poland has followed the lead of Germany and other western European countries in rejecting both the breakthrough rule and the board neutrality rule. This suggests that the approach which began in the City Code is on the way to becoming a global standard. Is this in the long-run interests of developing and transi-tion systems?

4. THE ECONOMIST’S VIEW OF THE MARKET FOR CORPORATE CONTROL

From a discussion of the alternative legal approaches to the question of takeover bids, we now turn to economic analysis. In terms of the language of the agency theory and the theory of asymmetric information, the central issue may be stated in the following terms. The modern corporations are, it is suggested, characterized by serious principal–agent problems, particu-larly between shareholders (principals) and managers (agents); asymmetric


information between the two; and incomplete contracting. The operation of these factors provides analytical justifi cation for the proposition that managers have scope to pursue their own ends.

One branch of the literature has pointed to the diffi culties involved in limiting managerial discretion through a more rigorous analysis of corpo-rate governance, that is, the internal governance mechanisms of the cor-poration, including shareholder voting and the eff ectiveness of the board of directors. The alternative means available to the shareholders to limit such discretion is to attempt to align managers’ interests with their own by devising appropriate incentive contracts. All such eff orts, however, have a cost (the so-called ‘agency cost’). Within the framework of these concepts, the takeover selection argument can then be deployed in two ways. The strong form would suggest that only fi rms that are able to devise and imple-ment optimal incentive contracts will be selected for survival; others will be taken over. In a weaker form, this theory would propose that, while in the real world substantial agency costs are inevitable and it is not always possible to design satisfactory incentive contracts, the takeover mechanism nevertheless helps to reduce agency costs and thus promote economic effi ciency. The implication of this weaker proposition is that, other things being equal, the greater the agency costs, the more likely it is that the fi rm will be taken over.

In this paradigm, in normative terms, the free operation of the takeover mechanism can benefi t society through two distinct channels: (a) the threat of takeovers can discipline ineffi cient managements and reduce ‘agency costs’; (b) even if the fi rms are working effi ciently, takeovers may lead to a reorganization of their productive resources and thereby enhance share-holder value.

There is a sharp dispute between industrial organization economists and specialists in fi nance about the effi ciency of mergers and takeovers. The former believe that takeovers do not lead to increased effi ciency; at best they are neutral, but most likely they reduce effi ciency. The industrial organization economists use accounting data to arrive at this conclusion (Scherer, 2006; Mueller, 2003). This leads these economists to advocate regulation of mergers since they are likely to enhance the monopoly power of the amalgamating fi rms without, on average, increasing their effi ciency. In contrast, the fi nance specialists believe, on the basis of stock market data and events studies methodology, that mergers enhance economic and social effi ciency. These scholars are therefore opposed to regulation of mergers. A leading exponent of this view is Professor Jensen (2005). He and his col-leagues regard anti-takeover legislation, which, as mentioned before, many individual American states have instituted in reaction to the successive huge merger waves of the last three decades, as being misconceived and promoted


by special interests. Some scholars belonging to this school would go even further. They not only oppose any new anti-merger regulatory measures but also suggest that the extant institutional obstacles to takeovers should be eliminated. There are at present a number of stock exchange provisions both in the US and the UK whose main purpose is to aff ord protection to minority shareholders and to ensure ‘fair play’ and transparency in share transactions connected with the takeover process. For example, in the UK, a corporate raider is obliged to disclose its stake in the victim company after it has purchased 3 per cent of the victim’s shares. Moreover, as we have seen, the raider is required to make a full cash bid for all the shares of the company after it has purchased 30 per cent of the victim’s stock. The exponents of the market for corporate control believe that such regulations constitute imperfections in the free functioning of the market; they are, therefore, ipso facto ineffi cient, and hence should be removed.

This very positive fi nance specialists’ view of the market for corporate control is vigorously contested by industrial organization economists. Their reservations are best conveyed by a critical analysis of the US busi-ness model of shareholder wealth maximization subject to the constraints of liquid stock markets (including the takeover mechanism). This model is being promoted for emerging countries by the IMF and the World Bank, and indeed is recommended as a universal standard for the whole world by these institutions and orthodox policy makers. Ironically, as we shall see below, this model has risen from the shadow of strong criticism in the early 1990s, when it was held responsible for the sluggishness of the US economy, to its current acclaim for having engineered the US lead in the information and technology revolution and for fostering faster growth in the US economy.33 The extent to which the US corporate model facilitates technological dynamism is perhaps the central issue in any assessment of its merits.

However, it is now accepted that since 1995 there has been an increase in the US economy’s long-term rate of growth by perhaps as much as one percentage point, from 2.5 to nearly 3.5 per cent per annum. This strong performance is attributed by leading scholars such as Jorgenson (2001, 2003) to the US lead in information technology. This success, in turn, is attributed by many economists and, particularly, by the media to the pivotal role played by US stock markets and venture capital markets in fi nancing this technological revolution.34 It is suggested that not all econo-mies with stock markets are able to achieve these feats. Black and Gilson (1998) have argued that other advanced economies such as Germany have tried to imitate the US venture capital market but have not been success-ful. The American success is due in part to its having a highly effi cient and eff ective market for corporate control. This allows a timely ‘exit option’,


making it possible for the American-type venture market to fl ourish. There are also other advantages attributed to the US stock market, such as the widespread use of stock options in technology industries that bring indi-vidual managers’ incentives in line with corporate objectives.

Larry Summers (1998, 1999), who in the past has been critical of the short-term focus of the stock market, has changed his mind. He now sug-gests that the increasing stock market pressure for performance has played a key role in the US economic success of the last decade. Further, the huge investment in new technology fi rms in the US during the technology boom of the 1990s, despite their zero or negative short-term profi ts, is regarded as an obvious refutation of the short-termism alleged by critics of stock markets (however, see below).

Nevertheless, taking into account the above facts, a critical examination of the functioning of the stock market in the last ten years raises the following questions. Does the experience of the last decade warrant a complete reversal of the conclusions reached by Michael Porter and his colleagues in 1992? Does the so-called ‘new’ US economy constitute a conclusive proof of the superiority of the country’s fi nancial system over all others? Is there adequate analysis and empirical evidence to indicate that the Anglo-Saxon model of corporate governance outlined above is the one that all countries, including developing ones, should adopt? Singh et al. (2005) have carried out a detailed analysis of these issues and they report the following conclusions:

The experience over the last decades in the US capital markets pro- ●

vides little justifi cation for revising the unfavourable 1992 verdict of Michael Porter and his colleagues, although the reasons for this are not necessarily the same now as they were then.Instead of maximizing shareholder wealth, developing country com- ●

panies should pay no attention to their market valuations. Rather, they should pursue their traditional objective of increasing market share or corporate growth within the overall framework of the coun-try’s industrial policy.The stock market based model of shareholder wealth maximization ●

does not represent the ‘end of history’ or the epitome of corporate law as some suggest.

The main reasons for these conclusions lie in the severe defi ciencies of two market processes which are central to the effi cient operation of stock markets: fi rst the pricing process and second the market for corporate control. It will further be appreciated that the last decade of applause for the US stock market must at least be tempered by the fact that during this period there was not only a boom but also a very signifi cant bust. The


NSDAQ index of share prices of new technology companies is still well below half the value that it reached at its peak in 2000.

5. STOCK MARKET PRICES AND THE MARKET FOR CORPORATE CONTROL

5.1 The Pricing Process on the Stock Market

It will be observed that during the last two decades the orthodox effi cient markets hypothesis concerning share prices has suff ered fundamental set-backs. These are specifi cally due to the following events: (a) the 1987 US stock market crash, (b) the meltdown in the Asian stock markets in the late 1990s and (c) the bursting of the technology stocks bubble in 2001. Following Tobin (1984) a useful distinction may be made between two kinds of effi ciency of stock markets. First, there is ‘information arbitrage effi ciency’ (IAE), which ensures that all information concerning a fi rm’s shares immediately percolates to all stock market participants, ensuring that no participant can make a profi t on such public information. Second, there is ‘fundamental valuation effi ciency’ (FVE), whereby share prices accurately refl ect a fi rm’s fundamentals, that is, its long-term expected profi tability (Tobin, 1984). The growing consensus view is that stock market prices may at best be regarded as effi cient in the fi rst sense above (IAE), but are far from being effi cient in the economically more important second sense (FVE) (Singh, 1999). This point hardly needs labouring today in the light of the burst of the technology bubble in leading stock markets in 2001 and almost two decades of stock market stagnation and decline in Japan. It will be diffi cult to preach an EMH gospel to citizens in Thailand and Indonesia, who suff ered a virtual meltdown of their stock markets during the Asian crisis of 1997–99 (see further Singh et al., 2005).

5.2 The Market for Corporate Control as an Evolutionary Mechanism35

There are good theoretical reasons as well as a large body of empirical evidence to suggest why the markets for corporate control in advanced countries, including the UK and the US, do not work at all well. A central point of this research is that the takeover selection process does not simply punish poor performance and reward good performance. The evidence indicates that selection in this market does not take place entirely on the basis of performance but much more so on the basis of size. A large rela-tively ineffi cient fi rm has a greater chance of survival than a small effi cient fi rm (Singh, 2008).


Further, there are good theoretical reasons as well as empirical evidence for suggesting that takeovers may lead to ‘short-termism’, and/or specula-tive buying and selling of shares. In addition, more broadly, they may result in economic rewards being given for fi nancial engineering rather than for entrepreneurial eff ort in improving products and cutting costs. Empirical research indicates that the takeover disciplinary process is very noisy and is often arbitrary and haphazard (Ravenscraft and Scherer, 1987; Scherer, 1998, 2006; Tichy, 2001; Singh, 2000). The defi ciencies of the pricing and takeover processes are compounded in the case of developing countries because of their regulatory defi cits and the relative immaturity of their stock markets. Singh (1998) argues for restrictions on the development of a market for corporate control for these countries. Rather, he suggests that developing countries should fi nd cheaper and less haphazard mechanisms to change managements than the above stock market process.

5.3 The Technology Boom, the Mispricing of Shares and the Market for Corporate Control

It is generally accepted that there was a widespread mispricing of shares during the technology boom of 1995–2000. There was also a huge over-investment in technology companies. Importantly, in addition to the fore-going, there was evidence of signifi cant resource misallocation through the working of the market for corporate control. In essence, grossly overpriced technology companies bought up underpriced old economy companies to the detriment of both and to the detriment of social welfare. Jensen (2003) drew attention in this context to the case of Nortel, a large US company that between 1997 and 2001 acquired 19 companies at a price of US$33 billion. Many of these acquisitions were paid for in Nortel shares whose value had skyrocketed during that period. When the company’s price fell 95 per cent in the technology stock burst, all the acquisitions had to be written off . Jensen observed, ‘Nortel destroyed those companies and in doing so destroyed not only the corporate value that the acquired companies – on their own – could have generated but also the social value those companies represented in the form of jobs, products and services.’ (p. 15)

Although Jensen suggests various ways of reducing the mispricing of shares, in Keynesian analysis such mispricing is inherent in any asset valu-ation pricing process via the stock market. In this paradigm, stock market players base their investment decisions not on fundamentals but on specu-lative and gambling considerations. With such pricing, shareholder wealth maximization is clearly not a useful objective for corporate managers who have the fi rm’s interest in view. Kay (2003) therefore rightly suggests that corporate managers should pay no attention to the stock market at all.


Indeed, the creation of shareholder value should not be a corporate goal. The Keynesian view of pricing process is supported by a large body of analytical and empirical studies: see, for example, Shiller (2000, 2004) and Shleifer (2000).

6. CONCLUSION

In orthodox economic analysis, the market for corporate control is thought to be the evolutionary endpoint of stock market development. This propo-sition has been seriously questioned in this chapter from the perspective of both legal and economic analysis.

Takeovers are a very expensive way of changing management. There are huge transaction costs associated with takeovers in countries like the US and the UK, which hinder the effi ciency of the takeover mechanism (Peacock and Bannock, 1991). Given the lower income levels in develop-ing countries, these costs are likely to be proportionally heavier in these countries. It may also be observed that highly successful countries overall, such as Japan, Germany and France, have not had an active market for corporate control and have thus avoided these costs, while still maintain-ing systems for disciplining managers. Signifi cantly, the lack of a market for corperate control has not imposed any great hardship on these econo-mies as their superior long-term economic record, say over the last 50 or 100 years, compared with that of Anglo-Saxon countries, indicates. Furthermore, there is no evidence that corporate governance necessarily improves after takeovers. This is for the simple reason that not all takeo-vers are disciplinary; in many of them the acquiring fi rm is motivated by empire-building considerations or indeed by asset-stripping.

In summary, contrary to current conventional wisdom, an active market for corporate control is not an essential ingredient of either company law reform or fi nancial and economic development. The economic and social costs associated with restructuring driven by hostile takeover bids, which are increasingly seen as prohibitive in the liberal market economies, would most likely harm the prospects for growth in developing and transitional systems. Developing countries simply cannot aff ord the burden of the extremely expensive, and hit and miss system of management change that takeovers represent.

The following argument might be raised against the claims that we have made here: if two mechanisms were available, an internal one based on corporate governance and an external one represented by takeovers, why not use them both to improve corporate performance? One immediate dif-fi culty with this argument is that acquiring companies may themselves be


empire builders rather than disciplining shareholder value maximizers, as was noted above. It was also seen that, at a more macro-economic level, the takeover mechanism may subvert capitalist values by rewarding fi nancial engineering rather than enterprise. In a survey carried out by Cosh, Hughes and Singh in the 1980s, it was found that 60 per cent of the time of the chief executives of Britain’s top companies was spent on roadshows to investors, rather than promoting new products or reducing costs, the essential tasks of enterprise. The authors also found that a great deal of time was spent by the chief executives and fi nancial directors in either avoiding takeovers or trying to take over other companies themselves (Cosh et al., 1990).

Perhaps our suggestion that developing country corporations should pay no attention to their market valuations is somewhat extreme. But our essential argument here is that stock market valuations are a highly inaccu-rate guide to the fundamental valuations of companies. This is especially so in developing countries, where theory predicts that the share price volatility is even greater than in advanced countries. With the kind of meltdown in share prices observed in East Asia during the crisis years of 1997–99, it was scarcely useful to ask corporations to judge their performance by changes in share prices. As mispricing of shares cannot be forecast with any accu-racy, and as historical evidence suggests that such mispricing may continue over long periods of time, it does not auger well for companies to use stock market values as the main criterion for judging success or failure.

Finally, it could be argued against us that all we have off ered here is a critique, when what is needed is practical answers to the question of how to design institutions for the market for corporate control. A clear con-clusion of our argument is that the mandatory bid rule and other similar aspects of the UK model, which are now being very widely exported around the world, will not aid the cause of economic development; we do not favour these rules. This does not mean that takeovers should be entirely unregulated – far from it. Not only developing countries but also those in Continental Europe, which have long operated without a market for corporate control, should seek alternative institutional mechanisms for disciplining errant managements rather than adopting the Anglo-Saxon takeover mechanism. Instead of concentrating on shareholder value, these countries should be actively promoting new institutional mechanisms for inclusive development of the company and its diverse constituencies.

NOTES

1. This chapter was originally a paper presented at the conference on ‘The Economics of the Modern Firm’, University of Jönköping, 21–22 September 2007. We are very grate-ful for comments received at the conference and from a referee.


2. This section updates, in part, material fi rst set out in Deakin and Slinger (1997) and Deakin et al. (2003), and draws on Deakin (2009).

3. Unocal v. Mesa Petroleum 493 A.2d 946, 955 (1985). 4. Revlon Inc. v. McAndrews & Forbes Holdings Inc. 506 A.2d 173 (1986); Paramount

Communications Inc. v. QVC Network Inc. 637 A.2d 34 (1994). 5. Paramount Communications Inc. v. Time Inc. 571 A.2d 1140 (1989). On Delaware’s ‘zig-

zags’, see Roe (1993) and Blair (1995: 220–22). 6. 457 US 624 (1985). 7. 481 US 69 (1987). 8. Schreiber v. Burlington Northern Inc. 475 US 1 (1985). 9. Directive 2004/25/EC of the European Parliament and the Council of 21 April 2004 on

Takeover Bids, L 142 Offi cial Journal of the European Union 30.4.2004).10. The Takeovers Directive (Interim Implementation) Regulations 2006 (SI 2006/1183),

which came into force on 20 May 2006, provide a statutory basis for the Panel’s operation and empower it to issue rules on takeover bids. These Regulations have more recently been superseded by the relevant provisions of the Companies Act 2006.

11. See DTI (2005) and Takeover Panel (2005).12. City Code, General Principle 1.1.13. Ibid., rule 9. See also Companies Act 1985, s. 430A providing a statutory right to sell

where the bidder and its associates control 90 per cent in value of the relevant shares; s. 428 grants the bidder a right of compulsory purchase of the last 10 per cent of the shares.

14. City Code, rule 36.15. City Code., rule 20.1.16. Ibid., rule 3.1.17. Ibid., rule 25.1(a).18. Ibid., rule 19.2.19. A claim in tort might well be made out notwithstanding the restrictive decision of the

House of Lords (on auditor liability) in Caparo Industries plc v. Dickman [1990] 2 AC 6, and it is also possible that directors who provide misleading advice on the sale of shares may commit a breach of statutory duty actionable by the shareholders: Gething v. Kilner [1972] 1 All ER 1166.

20. Heron International Ltd. v. Grade [1983] BCLC 244.21. General Principle 3.22. City Code, rule 24.1.23. Ibid., rule 25.1(b).24. See below, Section 3.25. City Code, rule 30.2(b). This is however subject to the target board receiving the

employee representatives’ views in good time, which may not always be straightforward. See Takeover Panel (2006: 32–3) for discussion.

26. City Code, rule 21.27. See Howard Smith Ltd. v. Ampol Petroleum Ltd. [1974] AC 821, discussed by Parkinson

(1995: 143).28. Companies Act 1985, ss. 85–9.29. These Guidelines were fi rst issued on 21 October 1987 by the International Stock

Exchange’s Pre-emption Group, consisting of members of the ISE and offi cers of the principal representatives of institutional shareholders, namely the Association of British Insurers and the National Association of Pension Funds. Under Guideline 1.2, the Investment Committees of the ABI and the NAPF agreed to advise their members, under normal circumstances, to approve resolutions for annual disapplication of pre-emption rights, as long as the non-pre-emptive issue did not exceed 5 per cent of the issued ordinary share capital as shown in the most recent published accounts of the company.

30. Guidelines published by the Institutional Shareholders Committee (a body represent-ing a number of fi nancial industry interests and trade associations) in December 1991,


‘The Responsibilities of Institutional Shareholders in the UK’, stated that ‘institutional shareholders have for many years been opposed to the creation of equity shares which do not carry full voting rights and have sought the enfranchisement of existing restricted voting or non-voting shares’ (para. 3).

31. See High Level Group (2002a: 10–12). On the Action Plan, and its development since 2002, see Commission (2003) and the company law website of the Internal Market Directorate: http://ec.europa.eu/internal_market/company/index_en.htm.

32. Directive 2004/25/EC.33. Porter (1992) reported on the fi ndings of a US Blue Ribbon Commission (comprising

22 leading US economists including Larry Summers) on the country’s business model and the associated system of allocating capital. The Commission made serious criticisms of America’s capital markets, indicating that they were misallocating resources and jeopardizing the American position in the world economy. It is indeed true that the US economy stagnated between 1973 and 1995, registering hardly any overall increase in productivity growth.

34. This is not necessarily Professor Jorgenson’s view. He attributes the rapid uptake of information technology in the US to the sharp fall in the price of semiconductors as a result of increased competition. This in turn arose from a reduction in the product cycle from three to two years.

35. This section is based on and updates some of Singh’s previous contributions in this area including Singh (1992, 2000, 2006).

BIBLIOGRAPHY

Alcouff e, A. and Alcouff e, C. (1997), ‘Control and executive compensation in large French companies’, Journal of Law and Society, 24: 85.

Berle, A. (1932), ‘For whom corporate managers are trustees: a note’, Harvard Law Review, 45: 1365.

Berle, A. (1962), ‘Modern functions of the corporate system’, Columbia Law Review, 62: 443.

Black, B. and Gilson, R. (1998), ‘Venture capital and the structure of capital markets: banks versus stock markets’, Journal of Financial Economics, 47: 243–77.

Blair, M. (1995), Ownership and Control: Rethinking Corporate Governance for the Twenty-First Century, Washington, DC: Brookings Institution.

Blair, M. and Stout, L. (1999), ‘A team production theory of corporate law’, Virginia Law Review, 85: 247–328.

Buchanan, J. and Deakin, S. (2007), ‘Japan’s paradoxical response to the new “global standard” in corporate governance’, ECGI Working Paper No. 87/2007, August.

Commission (2003), Modernising Company Law and Enhancing Corporate Governance in the European Union – A Plan to Move Forward, COM(2003) 584 fi nal, Brussels: European Commission.

Commission (2007), ‘Report on the implementation of the Directive on Takeover Bids’, Brussels, 21.02.2007, SEC(2007) 268.

Company Law Review Steering Group (2000), Modern Company Law for a Competitive Economy: Developing the Framework, London: DTI.

Company Law Review Steering Group (2001), Modern Company Law for a Competitive Economy: Final Report, Volume 1, London: DTI.

Cosh, A., Hughes, A. and Singh, A. (1990), ‘Takeovers and short-termism:


analytical and policy issues in the UK economy’, Industrial Policy Paper No. 2, London: IPPR.

Davies, P. (1997), Gower’s Principles of Modern Company Law, London: Sweet and Maxwell.

Deakin, S. (2003), ‘Evolution for our time: a theory of legal memetics’, Current Legal Problems, 55: 1–42.

Deakin, S. (2009), ‘Refl exive governance and European company law’, European Law Journal, 15 (2), 224–45.

Deakin, S. and Slinger, G. (1997), ‘Hostile takeovers, corporate law and the theory of the fi rm’, Journal of Law and Society, 24: 124–150.

Deakin, S., Hobbs, R., Nash, D. and Slinger, G. (2003), ‘Implicit contracts and cor-porate governance: in the shadow of the City Code’, in Campbell, D., Collins, H. and Wightman, J. (eds) Implicit Dimensions of Contracts, Oxford: Hart, 289–331.

Dodd, E.M. (1932), ‘For whom are corporate managers trustees?’, Harvard Law Review, 45: 1145–63.

DTI (2005), Company Law Implementation of the European Directive on Takeover Bids, London: Department of Trade and Industry.

Gelter, M. (2007), ‘A theory on the impact of shareholder control and ownership concentration on corporate stakeholders: or, team production meets compara-tive corporate governance’, paper presented at LSA conference, Berlin.

Hannah, L. (1974), ‘Takeover bids in Britain before 1950: an exercise in business “pre-history”’, Business History, 16: 65–77.

Hannah, L. (2007), ‘The divorce of ownership and control from 1900: recalibrating imagined global historical trends’, Business History 49: 404–38.

Hansmann, H. and Kraakman, R. (2001), ‘The end of history for corporate law’, Georgetown Law Journal, 89: 439–68.

High Level Group (2002a), Report of the High Level Group of Company Law Experts on Issues Related to Takeover Bids, Brussels: European Commission.

High Level Group (2002b), Report of the High Level Group of Company Law Experts on a Modern Regulatory Framework for Company Law in Europe, Brussels: European Commission.

Ireland, P. (1999), ‘Back to the future: Adolf Berle, the Law Commission and direc-tors’ duties’, Company Lawyer, 20: 203–11.

Jensen, M. (2001), ‘Value maximization, stakeholder theory, and the corporate objective function’, Journal of Applied Corporate Finance, 14: 8–21.

Jensen, M.C. (2003), ‘The agency costs of overvalued equity’, Munich Economic Summit, 2–3 May, CESifo Forum, 4 (2), 14–16.

Jensen, M. (2005), A Theory of the Firm: Governance, Residual Claims, and Organizational Forms, Cambridge, MA: Harvard University Press.

Johnston, A. (1980), The City Take-over Code, Oxford: OUP.Jorgenson, Dale W. (2001), ‘Information technology and the U.S. economy’,

American Economic Review, 91 (1), 1–32.Jorgenson, Dale W. (2003), ‘U.S. growth resurgence’, CES Ifo Economic Studies,

49 (1), 27–47.Kay, John (2003), ‘Challenging the claims for the role of capital markets’, Munich

Economic Summit 2–3 May, CESifo Forum, 4 (2), 17–20.Kelly, M. (2002), ‘The incredibly unproductive shareholder’, Harvard Business

Review, January.Kennedy, A. (2000), The End of Shareholder Value: Corporations at the Crossroads,

Cambridge, MA: Perseus.


METI and MOJ (2005), (Guidelines Regarding Takeover Defence for the Purposes

of Protection and Enhancement of Corporate Value and Shareholders’ Common Interests), Tokyo: (METI) and (Ministry of Justice).

Milhaupt, C.J. (2006), ‘The rise of hostile takeovers in Japan’, Zeitschrift für Japanisches Recht (Journal of Japanese Law), 11: 199–204.

Millon, D. (2002), ‘Why is corporate management obsessed with quarterly earnings and what can be done about it?’, George Washington Law Review, 70; 890–920.

Miyajima, H. and Kuruoki, F. (2005), ‘The unwinding of cross-shareholding: causes, eff ects, and implications’, in Aoki, M., Jackson, G. and Miyajima, H. (eds), Corporate Governance in Japan: Institutional Change and Organizational Diversity, RIETI Discussion Paper Series 05-E-006.

Mueller, D. (2003), ‘The fi nance literature on mergers: a critical survey’, in M. Waterson (ed.), Competition, Monopoly and Corporate Governance, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 161–205.

National Stock Exchanges (2006), (Results of Shareholding Distribution Survey for 2005), Tokyo, Osaka,

Nagoya, Fukuoka, Sapporo: (National Stock Exchanges).Nikkei (2007), (‘Introduction of anti-takeover

strategies: close to 200 companies’), Tokyo: (Nihon Keizai Shimbun).

Njoya, N. (2007), Property in Work: The Employment Relationship in the Anglo-American Firm, Aldershot: Ashgate.

NLI Research (2004), (Cross-Share-holding Survey 2003), Tokyo: (NLI Research).

Parkinson, J. (1995), Corporate Power and Responsibility: Issues in the Theory of Company Law, Oxford: OUP.

Parkinson, J. (2003), ‘Models of the company and the employment relationship’, British Journal of Industrial Relations, 41: 481–509.

Peacock A. and Bannock, G. (1991), Corporate Take-overs and the Public Interest, Aberdeen: Aberdeen University Press for the David Hume Institute.

Porter, M. (1992), ‘Capital disadvantage: America’s failing capital investment system’, Harvard Business Review, 70, (September–October), 65–82.

Ravenscraft, D.J. and Scherer, F.M. (1987), Mergers, Sell-off s and Economic Effi ciency. Washington DC: Brookings Institution.

Rebérioux, A. (2007), ‘Does shareholder primacy lead to a decline in managerial accountability?’, Cambridge Journal of Economics, 31: 507–24.

Roe, M. (1993), ‘Takeover politics’, in M. Blair (ed.), The Deal Decade, Washington DC: Brookings Institution.

Scherer, F.M. (1998), ‘Corporate take-overs: the effi ciency argument’, Journal of Economic Perspectives, 2 (1) (Winter), 69–82.

Scherer, F.M. (2006), ‘A new retrospective on mergers’, Review of Industrial Organization, 28: 327–41.

Schleifer, A. and Summers, L. (1988), ‘Breach of trust in hostile takeovers’, in Auerbach, A. (ed.) (1988), Corporate Takeovers: Causes and Consequences, Chicago: University of Chicago Press, pp. 33–56.

Schleifer, A. (2000), Ineffi cient Markets, Oxford: Oxford University Press.Shiller, Robert J. (2000), Irrational Exuberance, Princeton: Princeton University

Press.


Shiller, Robert J. (2004), ‘Figuring out fi nancial markets: more psychology than economics?’, interview with Robert Shiller, IMF Survey, 12 April, pp. 111–12.

Siems, M. (2007), ‘Shareholder protection around the world (‘Leximetric II’)’, CBR Working Paper No. 359, Centre for Business Research, University of Cambridge.

Singh, A. (1992), ‘Corporate take-overs’, in J. Eatwell, M. Milgate and P. Newman (eds), The New Palgrave Dictionary of Money and Finance, London: Macmillan, pp. 480–86.

Singh, A. (1998), ‘Liberalisation, the stock market and the market for corporate control: a bridge too far for the Indian economy?’, in Ahluwalia, I.J. and Little, I.M.D. (eds), India’s Economic Reforms and Development: Essays for Manmohan Singh, Oxford: OUP, 169–96.

Singh, A. (1999), ‘Should Africa promote stock market capitalism?’, Journal of International Development, 11: 343–65.

Singh, A. (2000), ‘The Anglo-Saxon market for corporate control: the fi nancial system and international competitiveness’, in Candace Howes and Ajit Singh (eds), Competitiveness Matters: Industry and economic performance in the US, Ann Arbor: University of Michigan Press, 89–105.

Singh, A. (2006), ‘Stock Market and Economic Development’, in Clark, David Alexander (ed.), The Elgar Companion to Development Studies, pp. 584–90, Cheltenham, UK and Northampton, MA, USA: Edward Elgar.

Singh, A. (2008), ‘Stock Markets in Low and Middle Income Countries’, presented at the Workshop on Debt, Finance and Emerging Issues in Financial Integration, 8–9 April, United Nations Headquarters, New York.

Singh, Ajit, Singh, Alaka and Weisse, B. (2002), ‘Corporate governance, com-petition, the new international fi nancial architecture, and large corporations in emerging markets’, Centre for Business Research Working Paper no. 250, University of Cambridge.

Singh, A., Glen, J., Zammit, A., De-Hoyos, R., Singh, Alaka and Weisse, B. (2005), ‘Shareholder value maximization, stock market and new technology: should the US corperate method be the universal standard?’, Centre for Business Research Working Paper no. 315, University of Cambridge.

Strine, L. (2007), ‘Toward common sense and common ground? Refl ections on the shared interests of management and labor in a more rational system of corporate governance’, Discussion Paper No. 585, 05/2007, John M. Olin Center for Law, Economics and Business, Harvard University.

Summers, L.H. (1998), ‘Opportunities out of Crises: Lessons from Asia’, remarks to the Overseas Development Council from the Offi ce of Public Aff airs, 19 March.

Summers, L.H. (1999), quoted in ‘Winning ways: ready bucks and a fl air of risk’, Financial Times, 14 December.

Takeover Panel (2005), Explanatory Paper: Implementation of the European Directive on Takeover Bids, London: Takeover Panel.

Takeover Panel (2006), The Implementation of the Takeovers Directive Statement by the Panel and the Code Committee following the External Consultation on Process on PCP 2005/5, London: Takeover Panel.

Tichy, G. (2001), ‘What do we know about success and failure of mergers?’, Journal of Industry, Competition and Trade, 1 (4), 347–94.

Tobin, J. (1984), ‘On effi ciency of the fi nancial system’, Lloyds Bank Review, 1–15 July.


TSE (2007), TSE-Listed Companies White Paper of Corporate Governance 2007, Tokyo: Tokyo Stock Exchange, Inc.

Uchitelle, L. (2006), The Disposable American: Layoff s and their Consequences, New York: Knopf.

Useem, M. (1996), Investor Capitalism, New York: Basic Books.Viénot, M. (1995), ‘Rapport sur le conseil d’administration des sociétés cotées’,

Revue de droit des aff aires internationales, 8: 935–45.Whittaker, H. and Hayakawa, M. (2007), ‘Contesting “corporate value” through

takeover bids in Japan’, Corporate Governance: An International Review, 15: 16–26.

Yarrow, G. (1985), ‘Shareholder protection, compulsory acquisition and the effi -ciency of the takeover process’, Journal of Industrial Economics, 34: 3–16.

PART IV

The board, management relations and ownership structure

225

10. Institutional ownership and dividendsDaniel Wiberg*

1. INTRODUCTION

During the late 1990s fi rms’ dividend payout ratios reached unprecedented low levels despite high earnings and price-to-dividend ratios. Recently however, with a continuing institutionalization of capital, dividend payout ratios have soared. At present many multinational fi rms pay out special dividends and buy back shares on a scale previously unseen. What role does the increasing institutionalization of capital play in this development? This chapter addresses this issue by investigating the eff ect of institutional ownership on dividend changes.

A large body of research exists on how corporate ownership structure infl uences fi nancing, investments and dividend decisions. The relationship between management ownership and dividend policy has been especially well documented (see, for example, Rozeff , 1982; Jensen et al., 1992; Eckbo and Verma, 1994; Moh’d et al., 1995). The link between institu-tional investors’ ownership and dividend policy is, however, somewhat neglected (for dividends decisions see Short et al. (2002) and Gugler and Yurtoglu (2003)). This lack of research is remarkable since there has been such an increase in the importance and presence of these types of investor in recent decades. Although studies exist they are predominantly done on US or UK data (for example, Short et al., 2002) which, although central, fail to provide comprehensive insights when the institutional framework is diff erent from what is usually referred to as the Anglo-Saxon corporate governance system. In Continental Europe and Scandinavia the general corporate governance structure is characterized by a much more concen-trated ownership, often in combination with control instruments such as dual-class shares and pyramidal ownership structures. The Swedish corpo-rate governance system is particularly interesting from this point of view, since it allows for the use of both vote-diff erentiated shares and corporate pyramid structures, which have jointly produced a remarkably persistent and concentrated ownership structure.

226 The board, management relations and ownership structure

The purpose of this chapter is to investigate the impact of ownership on dividends. In particular, institutional ownership, and its relation to dividends, is considered in the context of an earnings trend model. This model allows both for partial adjustments of dividends to changes in earnings and for trends in the fi rms’ dividend behaviour. By examining Swedish listed fi rms the chapter also provides empirical evidence on the eff ects of control instruments such as dual-class shares on dividend policies.

In line with the assumption that institutional investors may play a moni-toring role, mitigating agency problems related to separation of ownership and control, the results show that institutional ownership has a positive eff ect on dividend payout policies. The relation is found to be positive but diminishing, which supports previous research concerning non-linearity and ownership structure. The chapter also provides empirical support for a negative impact of dual-class shares on dividends. The result, in line with agency cost theory, is that control instruments, such as vote-diff erentiated shares, induce investors to demand higher levels of dividends as compensa-tion for the increased agency costs.

Utilizing a panel data methodology which accounts for fi rm-specifi c eff ects and time eff ects, unobservable heterogeneity is controlled for. Furthermore, the chapter contributes to the literature by looking par-ticularly at the Swedish case. In fact, Sweden is a very interesting case because it is a civil law country which, according to La Porta et al. (1999), has weaker protection of minority owners than common law countries such as the UK and the US. Opportunistic behaviour of the controlling owners is therefore more likely vis-à-vis minority owners (Miguel et al., 2004; Pindado and de la Torre, 2006). By European standards Sweden also has a vital capital market with a substantial part of the stock market equity controlled by both foreign and domestic institutional investors.

The chapter is organized as follows. Section 2 continues with a discus-sion about the possible relations between institutional ownership and dividend policy. In particular, the importance of agency confl icts and sig-nalling is discussed. The statistical models for dividend payout behaviour are provided in Section 3, together with defi nitions of the variables used in the regressions. Summary statistics and ownership concentration by dif-ferent types of owners in the sample fi rms are examined in Section 4. The empirical method, estimation results and analysis are provided in Section 5. Conclusions end the chapter in Section 6.

Institutional ownership and dividends 227

2. OWNERSHIP AND CORPORATE GOVERNANCE

Given the divergence of ownership and control in listed fi rms, shareholders cannot perfectly control the managers’ actions in the strict interest of the shareholders. Hence principal–agent problems arise. Managers may divert funds in their own interest at the expense of the shareholders (Williamson, 1963, 1964). This diversion of funds, usually referred to as managerial discretion, may include expropriation1 or diversion of cash fl ows to unprof-itable projects. It might be that these alternative investments provide a positive return. In relation to the shareholders’ cost of capital, however, the return is too low and, therefore, in terms of shareholder value maximiza-tion, it is unprofi table (Mueller, 2003).

With a separation of votes from capital, as in many fi rms in Sweden, agency costs might be substantial for the minority shareholders. A key feature in any corporate governance system is therefore the legal protection of minority shareholders. The eff ectiveness of the corporate governance system however, may also require the presence of large investors other than the controlling owner(s) or management2 (La Porta et al., 2000; Burkart et al., 1997). They can infl uence the managers to distribute profi ts to the shareholders, thus limiting the recourses available for managerial discre-tion. The downside to large investors of this kind is of course that they might just as well override the interest of minority shareholders (La Porta et al., 1999). Indeed, Morck et al. (1988) fi nd that profi tability is higher for fi rms with shareholders that have up to 5 per cent ownership; beyond that, profi tability drops. This pattern indicates that larger block-holding inves-tors might seek to generate private benefi ts of control that are not shared by minority shareholders.

A constraint on institutional investors is that they are often limited, either by regulation or by a desire to maintain liquidity, to holding a rela-tively small ownership stake in the fi rm’s equity (Davis and Steil, 2001). Indeed, in Sweden mutual funds, which constitute the largest part of the institutional owners, are regulated by the mutual funds act of 2004.3 In this act it is stipulated that no single mutual fund can invest more than 5 per cent of its capital in a single equity issuer. The presence of institutional investors in the ownership structure of fi rms might nevertheless infl uence managers to be more focused on shareholder value maximization. It is also likely that this relationship between institutional ownership and dividend payout is non-linear (Miguel et al. 2004; Bjuggren and Wiberg, 2008; Bjuggren et al., 2007). That is, although the eff ect in general might be posi-tive it is most likely marginally diminishing. A non-linear relation between ownership and dividend also indicates that the direction of causality goes from ownership to dividends and not the opposite.


2.1 Institutional Ownership and Dividends

The increasing number of institutional investors and their growing domi-nance as owners has had a substantial infl uence on corporate governance (for extended discussion of agency costs and institutional owners see Davis and Steil, 2001). Compared with Anglo-Saxon countries such as the US and the UK, Continental European and Scandinavian fi rms pay out relatively little in dividends or via repurchase of shares (La Porta et al., 2000), despite high profi tability and a very mature corporate structure. One principal reason for the low levels of dividends in Sweden is the tax system, which persistently disfavours dividends in favour of investments made with retained earnings (Högfeldt, 2004; Henrekson and Jakobsson, 2006). A stated purpose of this tax policy is to foster so-called long-term investment. The eff ect, however, is that substantial funds have been made available for managers to invest with little or no scrutiny from the external capital market.

So, even if high desired levels of dividends can be seen as a sign of ‘short-termism’ in the institutional owners’ attitudes (see, for example, Hutton 1995; Haskins 1995), it might just as well be an eff ect of these owners’ attempts to reduce the free cash fl ow available to management.

2.2 Taxation Arguments

Institutional owners might prefer dividends for other reasons as well. First of all, many institutional owners are tax-exempt with regard to dividends, and might thus prefer dividends to capital gains. In Sweden the majority of institutional owners are in fact tax-exempt mutual fund companies and insurance companies that manage pensions and other types of savings on behalf of the general public. Foreign ownership on the Swedish Stock Exchange is also predominantly made up of these types of institution.

The Swedish corporate taxation system is a classical company tax system in which the companies are taxed separately from their sharehold-ers. While fi rms pay a fl at4 rate of corporation tax on their profi ts, indi-viduals pay a slightly higher dividend gains tax on dividend incomes. The dividend gains tax is higher than the corporate tax rate, and individual owners might thus prefer to postpone taxes rather than pay a dividend tax immediately. Mutual fund companies and similar institutional investors are, however, tax-exempt in the sense that they do not pay tax on incomes received as dividends. The eff ect of this system is of course that indi-viduals and company owners might prefer retained earnings and capital gains, whilst tax-exempt institutional owners are either neutral or positive towards dividends.


A related issue is the need of many institutional owners for funds on an ongoing basis. That is, institutions invest in order to provide returns to fund their liabilities. Regardless of the tax bias in favour of dividends, institutions can therefore not rely entirely on capital gains to fund their activities, and hence they require dividends. For institutional owners as a group, and particularly in the case of Sweden, a positive relation to divi-dend payout must consequently be expected.

2.3 Agency Arguments

A second reason for why institutional owners in particular might favour dividends over reinvestments within the fi rm is that they might serve to curb the agency problems between controlling owners/managers and the minority shareholders, as suggested by Jensen (1986). Again, by high divi-dend payout ratios less funds are available for managerial discretion, and more funds will be allocated through the external capital market subject to market scrutiny.

Empirically the predictions of agency theories on dividend payout (Rozeff , 1982; Easterbrook, 1984; Jensen, 1986; Eckbo and Verma, 1994) support a positive association between dividends and institutional owner-ship. The prediction is basically that dividends substitute for poor monitor-ing by the fi rms’ shareholders. Institutional owners might act as infl uential principals who are able to impose their preferred payout policy upon fi rms. The result is less cash available within the fi rm for managerial discretion and a somewhat mitigated agency problem.

Based on the arguments above, Zeckhauser and Pound (1990) suggest that institutional owners might act as a substitute monitoring device, which would also reduce the need for external monitoring by the capital markets. However, the well-known incentives for institutional shareholders to free-ride on monitoring activities suggests that institutional shareholders are in fact unlikely to provide direct monitoring themselves.

2.4 Signalling Arguments

A third reason why institutional owners might favour dividends is the poten-tial information asymmetries that exist between owners and managements. Given these asymmetries and the equity market’s preference for liquidity, dividends can act as a signal about the future prospects of the fi rm.

A way for managements to signal their private information regarding the future earnings of the fi rm would be through dividends (Bhattacharya, 1979, 1980; Miller and Rock, 1985). A somewhat alternative hypothesis is put forward by Zeckhauser and Pound (1990). They argue that the presence


of large outside shareholders, such as institutions, can act as a signal of the fi rm’s good performance. The presence of such shareholders might therefore lessen the use of dividends as a signalling device. This would then to some extent change Zeckhauser and Pound’s (1990) agency theory prediction. It is however unclear in what way institutional shareholders would act as a signal of future prospects. Is it a signal of reduced agency costs due to monitoring of the institutional shareholders? According to the free-rider arguments mentioned before, probably not. The alternative is then that the institutional shareholders have some superior information regarding the future prospects of the fi rm. Although this explanation has some appeal, there is little evidence to support this scenario. Insider laws may, for instance, make institutional shareholders very careful in handling this type of information (if they get hold of it to start with). Also, the rapid increase of indexation, especially with respect to institutional sharehold-ings, implies that the presence of an institutional shareholder might not necessarily mean that the particular institution believes that the fi rm has better than average prospects (Short et al., 2002). While possible, the notion that dividends and institutional shareholders may act as substitut-ing devices is not very convincing. The expected results with respect to the relationship between institutional ownership and dividends in terms of signalling would subsequently be mixed as well.

These three main considerations, taxation, agency costs, and signalling, are now summarized in order to construct empirically testable hypotheses regarding the association between ownership and dividends.

2.5 Summary and Hypotheses

The association between ownership and dividends seems to depend cru-cially on three factors related to the corporate governance system. The fi rst is the consideration of taxes. In a country like Sweden, with a clas-sical company tax system, dividend payments are essentially taxed twice, both as profi ts within the fi rm and then as capital gains for the individual. Tax-exempt shareholders might for various other reasons (liabilities, and so on) prefer dividends to capital gains. Consequently one would expect a positive or at least neutral attitude to dividends relative to capital gains for this type of investor.

The second factor decisively linking the corporate governance system to dividends is agency problems related to the separation of ownership from control. In corporate governance systems, such as the Swedish, where ownership is further separated from control via control instruments, such as vote-diff erentiated shares, the agency confl icts described by Jensen and Meckling (1976) are aggravated. From this perspective infl uential


shareholders such as institutions may demand higher levels of dividends in order to force fi rms to go to the capital market for external funding, and hence be subject to monitoring by the external market, a notion that would hold particularly when there is a separation between ownership in terms of capital and control. The reduced levels of cash fl ow will thus mitigate the free cash fl ow problem as described by Jensen (1986) and thus lead to less ineffi ciency, in terms of managerial discretion. Based on the argu-ments of the agency theory, therefore, the hypothesized relation between institutional shareholdings and dividends is positive when capital rights are separated from control rights.

Research by, amongst others, Miguel et al. (2004), Pindado and de la Torre (2006) and Crutchely et al. (1999) has shown that the relationship between dividends and institutional ownership is non-linear, and margin-ally diminishing. Although positive, the impact of increasing ownership leads to a convergence of the monitoring and entrenchment eff ects. This notion has also been widely supported by previous literature (Morck et al., 1988; McConnell and Servaes, 1990; Gedajlovic and Shapiro, 1998). One would therefore expect that any impact of institutional ownership on dividend policy is positive but diminishing.

The causal relation between dividends and ownership in terms of signal-ling is, as mentioned, more complex, if existent. A distinct empirically test-able hypothesis of this relationship is thus hard to formulate. Based on this and the arguments above about taxation concerns and the agency theory, hypotheses 1a and 1b are formulated.

Hypothesis 1a Institutional shareholdings have a positive eff ect on divi-dend changes.

Hypothesis 1b Institutional shareholdings have a positive but marginally diminishing eff ect on dividend changes

Hypotheses 1a and 1b are expected to hold for ownership in terms of both votes and capital.

As the agency problems related to the separation of ownership from control would be aggravated by the use of control instruments such as vote-diff erentiated shares, institutional owners and outside investors will demand higher dividends where such control instruments are in place. A positive relationship can therefore be expected between dividend changes and vote-diff erentiated shares. The next hypothesis is therefore:

Hypothesis 2 The use of vote-diff erentiated shares has a positive relation to dividend changes.


Again, this relationship is expected to hold for ownership in terms of votes as well as capital.

As the current period’s earnings are of primary importance to any eventual dividend payout, an earnings component will be incorporated in the estimated dividend model, as suggested by Fama and Babiak (1968). The interpretation of this component is straightforward: higher earnings mean more funds available for dividends and consequently a positive impact on dividend changes can be expected. To control for the previous period’s earnings, an earnings trend component will also be included in the model.

In addition to earnings another variable which must be controlled for is the previous period’s dividends. The parameter estimate of this variable represents the speed of adjustment of dividends to new levels of earnings and is thus expected to be negative, meaning that there is some reluctance to change dividends immediately in response to changes in earnings (Short et al., 2002).

3. METHOD AND VARIABLES

To test the relation between institutional ownership and dividends, a partial adjustment model which accounts for earnings trends is used. The model is modifi ed by interacted shareholdings of the diff erent ownership types. A similar approach used by Short et al. (2002) is limited to using interactive dummy variables due to the lack of ultimate ownership data. In this chapter, however, the continuous shareholdings of the diff erent owner-ship categories, focusing on institutional ownership, are used.

Following Short et al. (2002) the derivation of the model is based on four related models for the dividend–earnings relation; the full and partial adjustment models by Lintner (1956), the Waud model (1966) and the earnings trend model by Fama and Babiak (1968).5

3.1 The Modifi ed Earnings Trend Model

Assuming that for any year, t, the target level of dividend D* for fi rm i at time t is related to the long-run expected earnings, E*ti, of fi rm i at time t earnings, by a desired payout ratio, r:

D*ti 5 rE*ti (1)

Based on the Waud model (1966) it is further assumed that the formation of expectations follows an adoptive expectation process of the form:


E*ti 2 E*(t21)i 5 d(Eti 2 E*(t21)i) (2)

Then if ownership structure, by for example institutions (Inst representing the ownership of institutional investors), alters the desired payout ratio (r) fi rms would have another D*ti, so the model becomes:

D*ti 5 rEti 1 rIEti 3 Inst (3)

where rI is the impact on the fi rm’s dividend payout policy related to insti-tutional ownership.

This earnings generating process can then be combined with the adjust-ment models of dividends developed by Lintner (1956). The partial adjustment model in particular assumes that, in any given year, the fi rm adjusts only partially to the target dividend level, as follows:

Dti 2 D(t21)i 5 a 1 c(D*ti 2 D(t21)i) (4)

where a is a constant representing the resistance to change dividends, and c is the ‘speed of adjustment’ coeffi cient which represents management’s reluctance to adjust the dividends to the new target level immediately. With the target dividend level D* for fi rm i at time t, as in equation (1), we can substitute in equation (4) and get the following model:

Dti 2 D(t21)i 5 a 1 c(rE*ti 2 D(t21)i) 1 mti (5)

where the term uti is the usual residual term.So far the specifi cation has yielded a partial adjustment model. But one

would also like to consider that earnings can follow a fi rm’s specifi c trend or process (Fama and Babiak, 1968). Assume that the specifi c profi t gen-erating process, for fi rm i at time t, is of the form:

Eti 5 (1 1 g)E(t21)i (6)

where g is an earnings trend factor. If the fi rms’ ownership structures also have a signifi cant infl uence on the earnings of the fi rms it seems reason-able to assume a possible diff erence in the earnings trend factor. The profi t generating process thus becomes:

Eti 5 E(t21)i 1 gE(t21)i 1 gIE(t21)i 3 Inst (7)

It is then possible to combine the Waud model’s adoptive expectation process in equation (2) with the partial adjustment model of equation (4) to get:


Dti 2 D(t21)i 5 a 1 c(r(d(Eti 2 E*(t21)i) 1 E*(t21)i) 2 D(t21)i) 1 mti (8)

Assuming that there is full adjustment of dividends to the expected change (c 3 d 5 1), and partial adjustment to the reminder, equation (8) can be rearranged and reduced. The reduced and empirically testable model accounting for both trends in earnings and adjustments to target dividend levels, equation (9), is consequently:

Dti 2 D(t21)i 5 a 1 rcEti 1 rg(1 2 c)E(t21)i 1 rgI(1 2 c)E(t21)i

3 Inst 2 cD(t21)i 1 uit (9)

Note that the term Inst is an example of an interaction term made up of an ownership variable (institutional ownership). In the same way other owner-ship variables can be tested by inserting another interaction term made up of the relevant ownership variable (for example, VotDiff , which is a dummy of vote-diff erentiated shares interacted with previous period’s earnings).

3.2 Variables

All data on the fi rms’ book values and earnings are provided by the Compustat-Global database. The period covered is 1996 until 2005. The time period in the regressions is 1997–2005, due to the fi rst diff erence in the dependent variable. Financial fi rms are removed from the sample, due to the particular nature of their investments. The ownership data are provided by Ownership and Power in Sweden,6 which is a unique database covering ownership structure, on a yearly basis, for all fi rms listed on any of the three major lists at the Stockholm Stock Exchange.

All aspects considered, the setup requirements produced a sample of 189 Swedish quoted fi rms. The sample fi rms correspond to an aggregate share of more than 90 per cent of the total market capitalization at the Stockholm Stock Exchange, and approximately 80 percent of the total Swedish export value.

The variable institutional ownership is made up of the aggregate owner-ship controlled by institutions, in terms of both cash fl ow rights (IC) and vote rights (IV). The same notation applies for foreign ownership (FC) and (FV) and so on; see Table 10.2. The group institutional investors consist of banks, pension and mutual funds, insurance companies and endowment foundations.7 The diff erent ownership categories and how they are defi ned and grouped are summarized in Table 10.1.

Table 10.2 provides a list of the variables used in the descriptive statistics and the regressions, together with their defi nitions.


4. DESCRIPTIVE STATISTICS AND OWNERSHIP CONCENTRATION

Before continuing to the estimation results, a more thorough assessment of the descriptive statistics is warranted. Descriptive statistics for the vari-ables in the regressions are provided in Table 10.3. In addition to the vari-ables used in the regressions, statistics of the fi rms’ sales/turnover, R&D expenses, and working capital are provided in Table 10.3. Also, descriptive statistics of the fi ve largest owners in terms of capital share (C5) and votes (V5) are included in the table. All fi gures, both in the descriptive statistics and in the regressions, have been defl ated to 2006 price level.

It is interesting to note that in the sample fi rms the largest shareholder on average controls 34.84 per cent of the votes (V1); see Table 10.3. This con-centrated ownership is remarkable, not only because of the concentrated ownership compared with other European and Anglo-Saxon countries, but also because of the relative size of the Swedish fi rms in the sample (mean sales 11 231.43 million SEK8). The sample of fi rms is therefore consistent with the view that the Swedish economy to a large extent is dominated by

Table 10.1 Ownership categories

Owner type Defi nition

Private All shares controlled by individuals as well as other fi rms. The private owner can be either the founder of the fi rm or an investor who has acquired control.

Foreign These owners can be institutions as well as individuals since it is hard to separate these two groups with certainty.

Institutional All shares controlled by Swedish fi nancial institutions belong to this category. In all cases the institution belongs to one of the three following types.Insurance companyInsurance company-controlled shares are all fi rms that have an insurance company as their largest owner. Note however that mutual funds belonging to an insurance company make a separate group of controlling owners.Mutual fundAs the name indicates, all shares controlled by a mutual fund; a fund can belong to a bank, an insurance company or the state-owned pension funds.FoundationThis category includes foundations donated by private individuals as well as, for example, various types of profi t-sharing funds and pension funds tied to individual companies.


closely held, relatively large, often old industrial and multinational fi rms (Agnblad et al., 2001; Högfeldt, 2004; Henrekson and Jakobsson, 2005).

When considering cash fl ow rights (C1), the share controlled by the largest owner is on average 23.77 per cent, substantially lower than the vote rights (V1534.84 per cent), but still remarkably high in an international comparison. The median values for these two variables also support this notion, that the single largest owner controls the fi rm to a large extent by vote-diff erentiated shares (median C1520.50 per cent and median V1531.30 per cent).

For the foreign and institutional owners cash fl ow rights seem to be more important than control, in line with the expectation. The ownership of vote rights for foreign and institutional owners (FV518.12 per cent

Table 10.2 Variables

Variable name Defi nition

Dti Total amount of dividends paid by fi rm i in period t (million SEK)

Dti-D(t-1)i Change in total amount of dividends paid by fi rm i between periods t-1 and t.

Prstkcti Purchase of fi rm i stocks by fi rm i in period t (million SEK)TPayti Total payout, dividends and repurchase of shares, by fi rm i

in period tTPayti-TPay(t-1)i Change in total payout, dividends and repurchase of shares,

by fi rm i between periods t-1 and t.Et Earnings, calculated as net profi ts from ordinary trading

activities after depreciation and other operating provisions (million SEK)

Et-1 Earnings of fi rm i in period t-1C1 Share of capital owned by the largest owner (cash-fl ow

rights), per centV1 Vote rights controlled by the largest owner (control rights),

per centFC Share of capital owned by foreign investors, per centFV Vote rights controlled by foreign investors, per centIC Share of capital owned by institutional investors, per cent.IV Vote rights controlled by institutional investors, per centVoteDiff Dummy variable for vote-diff erentiated shares, 1 if dual-

class shares, 0 if one-share-one-voteSalesEmployed

Total sales (million SEK)Total number of persons employed by the fi rm i in time t

R&D-exp Research and development expenses if reported (millions SEK)

WCap Working capital (millions SEK)


and IV511.14 per cent) is substantially below the level of cash fl ow rights (FC520.59 per cent and IC514.11 per cent). For both ownership types the diff erence is around 3 per cent, which supports the assumption that the two ownership types are in fact very similar. That is, the majority of the foreign owners are in fact institutions. The incentive structure and the infl uence of ownership on the performance should therefore be similar for foreign and institutional owners, as expected by hypotheses 1a and 1b.

Dividing the sample according to whether or not the fi rms have vote-diff erentiated shares reveals some additional insights. Table 10.4 shows the descriptive statistics of the group of fi rms with only one type of share (one-share-one-vote). This group represents 37 per cent of the total sample of 189 fi rms, or 445 observations. It also seems that this group on average represents smaller fi rms, compared with the group of fi rms that have vote-diff erentiated shares described in Table 10.5.

Table 10.3 Descriptive statistics all fi rms

Mean Median Std. dev. Min Max Obs

Dt 261.26 9.29 789.66 0 7862 1190Dt-D(t-1) 31.74 0 395.72 −5169.44 4939.99 1190Prstkcti 28.10 0 318.10 0 6518.13 1190TPayti 289.36 9.51 895.15 0 8996.52 1190TPayti- TPay(t-1)i

31.74 0 395.15 −5169.44 4939.98 1190

Et 714.62 45.10 2797.37 −34 529.32 30 724.00 1190Et-E(t-1) 78.53 8.71 2038.43 −40 652.38 37 146.87 1190C1 23.77 20.50 15.16 1.00 74.50 1190V1 34.84 31.30 20.75 2.50 95.10 1190C5 47.01 45.9 18.40 6.40 97.60 1190V5 58.15 59.75 20.99 6.40 98.80 1190FC 20.59 16.20 17.47 0.00 79.60 1190FV 18.12 11.30 18.25 0.00 93.50 1190IC 14.11 11.5 12.28 0.00 54.90 1190IV 11.14 8.10 10.94 0.00 67.6 1190VoteDiff 0.62 1 0.48 0 1 1190Sales 11 231.43 1204.26 31 099.15 0.04 298 428.10 1190Employed 6.93 0.45 20.76 0.01 216.99 1190R&D-exp 406.07 0 3010.69 0 49 553.76 1190WCap 1954.50 166.15 8535.88 −10 884.00 110 201.90 1190

Note: All ownership variables, votes (V) and capital (C), are given in percentage. The vote-diff erentiation dummy variable (VoteDiff ) takes the value 1 if the fi rm has vote-diff erentiated shares, 0 otherwise. Sales are given in millions of Swedish kronor (SEK).


The group of fi rms with vote-diff erentiated shares consists of 745 observa-tions which represent 63 percent of the total number of fi rms in the sample. Looking at the fi gures for sales, R&D, and working capital, and comparing Tables 10.4 and 10.5, confi rms that the fi rms with vote-diff erentiated shares on average are larger than the fi rms without vote-diff erentiated shares.

The correlation between the diff erent variables is provided in Table A10.1 in Appendix 10.1. The correlations confi rm the negative relation-ship between both foreign ownership of capital and votes (FC and FV) and institutional ownership of capital and votes (IC and IV) relative to vote-diff erentiation. Also, a high correlation between dividends and earn-ings is evident, as expected.

Repurchase of shares (Prstkcti) only constitutes a fractional part of the total payout by the sample fi rms. Due to regulation, this way of distributing funds back to the shareholders has previously been closed for Swedish fi rms.

Table 10.4 Descriptive statistics fi rms without vote-diff erentiated shares


Dt 130.14 0 544.10 0 5656.38 445Dt-D(t-1) 18.55 0 186.74 −1006.33 2812.53 445Prstkcti 9.62 0 89.04 0 1158.50 445TPayti 139.76 0 555.14 0 5656.38 445TPayti- TPay(t-1)i

18.55 0 186.74 −1006.33 2812.53 445

Et 312.20 10.69 1588.51 −5823.43 17 972.37 445Et-E(t-1) 44.32 7.77 1210.15 −14052.01 18 860.71 445C1 22.09 19.4 13.78 2.50 74.50 445V1 22.09 19.4 13.78 2.50 74.50 445C5 43.91 42.3 17.63 6.40 89.20 445V5 43.91 42.3 17.63 6.40 89.20 445FC 22.39 18.2 17.94 0 77.00 445FV 22.39 18.2 17.94 0 77.00 445IC 14.03 11.3 12.01 0 54.90 445IV 14.03 11.3 12.01 0 54.90 445VoteDiff 0 0 0 0 0 445Sales 4646.75 650.63 12 093.59 0.05 87 661 445Employed 2.48 0.39 5.96 0.03 39.61 445R&D-exp 81.49 0 287.87 0 2875 445WCap 565.97 113.95 1832.96 −6236.19 13 727.85 445



The correlation matrix (Table A10.1, Appendix 10.1) nonetheless confi rms a positive correlation between institutional ownership and this type of payout. As few fi rms in the sample have made use of this method to distribute cash to the shareholders, the focus of this chapter is on dividend changes.

5. EMPIRICAL RESULTS AND ANALYSIS

In order to test if there is any linear relationship between institutional own-ership and dividends, the policy adjustment model is estimated with inter-action terms; in Table 10.6, Model 1. The estimation is made in the form of a pooled OLS, and ownership is measured as percentage of both votes and capital. The results are presented in Table 10.6., Models 1a and 1b. The results support hypothesis 1, with a positive eff ect of institutional ownership

Table 10.5 Descriptive statistics of fi rms with vote-diff erentiated shares


Dt 339.58 18.53 896.33 0 7862 745Dt-D(t-1) 39.62 0 478.83 −5169.44 4939.98 745Prstkcti 39.14 0 395.79 0 6518.13 745TPayti 378.72 18.54 1036.95 0 8996.52 745TPayti- TPay(t-1)i

39.62 0 478.83 −5169.44 4939.98 745

Et 954.99 69.40 3293.19 −34 529.32 30 724 745Et-E(t-1) 98.97 9.75 2401.13 −40 652.38 37 146.87 745C1 24.77 20.90 15.86 1 74.10 745V1 42.43 40.70 20.51 2.90 95.10 745C5 48.86 48.50 18.61 8.90 97.50 745V5 66.58 69.50 18.09 9.60 98.80 745FC 19.51 15.30 17.10 0 79.60 745FV 15.54 9.00 17.94 0 93.50 745IC 14.15 11.60 12.45 0 54.70 745IV 9.42 6.80 9.85 0 67.60 745VoteDiff 1 1 0 1 1 745Sales 15 164.57 1613.46 37 642.09 1.02 298 428.10 745Employed 9.58 0.97 25.47 0.01 216.99 745R&D-exp 599.97 0 3786.24 0 49 553.76 745WCap 2783.89 204.42 10 611.02 −10 884.00 110 201.90 745



on changes in dividends, for institutional ownership measured by votes. Although robust in terms of size and sign, the coeffi cient on institutional ownership is insignifi cant when ownership is measured in terms of capital. The estimated coeffi cient on previous periods’ dividends, Div(t-1), is negative and signifi cant, which suggests that the fi rms adjust dividends slowly to changes in earnings, which confi rms the fi ndings in Short et al. (2002).

In order to account for a potential non-linear eff ect of institutional ownership, another interaction term of squared institutional ownership and earnings is added (Grier and Zychowicz, 1994; Schooley and Barney, 1994; Crutchley et al., 1999); see Table 10.7, Models 2a and 2b. This allows for a marginally diminishing eff ect of institutional ownership on dividend changes. Pindado and de la Torre (2006) use a somewhat diff er-ent approach with optimal breakpoints of the value–ownership relation estimated in Miguel et al. (2004). As institutional ownership is measured as the aggregate ownership share by this type of investor, this specifi cation seems unwarranted. Each individual institutional owner has its specifi c breakpoint associated with its investment profi le and so on. Consequently only a diminishing eff ect of aggregate institutional ownership is tested.

Table 10.6 Pooled-OLS estimations: Model 1 linear and Model 2 non-linear institutional ownership (votes and capital)

Dependent variable(Divt-Div(t-1))

Model 1a(votes)

Model 1b (capital)

Model 2a(votes)

Model 2b(capital)

Et 0.1247*(26.69)

0.1248*(26.60)

0.1236*(26.30)

0.1233*(26.16)

E(t-1) −0.0612*(−5.21)

−0.0616*(−4.89)

−0.0777*(−5.49)

−0.1072*(−5.07)

E(t-1)*Inst 0.0007*(2.56)

0.0007(1.45)

0.0028*(2.71)

0.0055*(2.95)

E(t-1)*Inst2 −0.00005**(−2.09)

−0.0001*(−2.68)

E(t-1)*VoteDiff 0.0133(1.26)

0.0110(1.02)

0.0193***(1.76)

0.0112(1.05)

Div(t-1) −0.2122*(−10.30)

−0.2067*(−10.02)

−0.2224*(−10.52)

−0.2081*(−10.12)

constant 13.9757(1.48)

13.8101(1.46)

13.5043(1.43)

15.5981(1.65)

Number of obs51190Number of groups5189

R250.3990R2adj50.3965

R250.3967R2adj50.3942

R250.4012R2adj50.3982

R250.4004R2adj50.3973

Note: t-statistics are in parentheses. * denotes signifi cance at the 1% level, ** denotes signifi cance at 5%, and *** denotes signifi cance at the 10% level.


The estimates of the non-linear specifi cation of Model 2 again reveal a pos-itive and signifi cant relation between institutional ownership and changes in dividends. Correctly specifi ed, institutional ownership, both in terms of votes (Model 2a) and capital (Model 2b), is found positive and signifi cant. For ownership measured in votes the coeffi cient related to the use of vote-diff erentiated shares is also signifi cant. This suggests that fi rms using vote-diff erentiated shares have higher levels of dividends, confi rming hypothesis 2. The speed of adjustment coeffi cient, related to previous periods’ dividends, is again signifi cant and negative, as expected. The same holds for this period’s earnings. Consistent with the equality and stability conditions of the model, the estimated parameter for previous period’s earnings is negative.

As displayed by the descriptive statistics there are substantial size and scale eff ects in the sample of fi rms. For the OLS regression to produce effi -cient estimates under such conditions we need to control that the data are homoscedastic. The Breusch-Pagan/Cook-Weisberg test,9 however, reveals that the sample suff ers from severe heteroscedasticity, and consequently we cannot rely on the results of the OLS estimation for inference. To account

Table 10.7 Cross-sectional time-series FGLS estimations: Model 3 institutional ownership (votes and capital)


Model 3a(votes)

Model 3b(capital)

Et 0.0817*(23.21)

0.0821*(21.47)

E(t-1) −0.0412*(−6.15)

−0.0395*(−4.89)

E(t-1)*Inst 0.0007***(1.82)

0.0009(1.50)

E(t-1)*Inst2 −9.59e-06(−1.06)

−1.4e-05(1.08)

E(t-1)*VoteDiff 0.0244*(4.21)

0.0217*(3.47)

Div(t-1) −0.1878*(−8.79)

−0.1906*(8.57)

constant 7.1562*(7.75)

8.0771*(7.60)

Number of obs51190Number of groups5189

Notes: t-statistics are in parentheses. Panels: heteroscedastic. Correlation: panel-specifi c AR(1).* denotes signifi cance at the 1% level, ** denotes signifi cance at 5%, and *** denotes signifi cance at the 10% level.


for this heteroscedasticity in the data a GLS methodology is required. Utilizing both the cross-sectional and time-series properties of the data, an FGLS regression will allow heteroscedasticity in the panels (fi rms) as well as panel-specifi c correlation (AR(1)).

By including a time-specifi c dummy variable it is also possible to control for temporal eff ects, that is, the eff ects of macroeconomic variables that might infl uence the fi rms and their dividend behaviour, as well as their ownerships structures.

Table 10.7 provides the results of the FGLS estimations, where owner-ship is measured in terms of both votes (Model 3a) and capital (Model 3b). As expected, institutional ownership is found to have a signifi cantly positive eff ect on dividend payout, when ownership is measured in terms of votes, which support hypothesis 1. The presence of institutional owners is thus associated with positive dividend changes. For institutional owner-ship in terms of capital share, the results are insignifi cant but positive, as expected. The use of vote-diff erentiated shares is again found to be posi-tively related to dividend changes, in support of hypothesis 2. This relation holds for ownership measured in terms of both votes and capital.

In order to investigate the role of institutional owners in the context of the agency confl ict related to the separation of ownership and control, the sample of fi rms is separated into two groups depending on whether or not they have vote-diff erentiated shares. Naturally the interaction term with the dummy for vote-diff erentiation is taken out of the regressions, as it would have produced collienarity.

Table 10.8 presents the results from the FGLS estimations when the fi rms are dividend into groups depending on whether or not they have a vote-dif-ferentiated share structure, Model 3aI and 3bI (without vote-diff erentiated shares) and Model 3aII and 3bII (with vote-diff erentiated shares). The esti-mations are made for ownership in terms of both votes and capital.

As can be seen from Table10.8, comparing Model 3aI and 3bI with Model 3aII and 3bII, institutional ownership only has a positive eff ect on dividend changes if the fi rms have vote-diff erentiated shares. This means that fi rms that separate cash fl ow rights from control rights suff er more from agency problems, and that institutional owners require these fi rms to pay higher dividends in order to reduce the cash available for management. This result is in accordance with the predictions of the agency theory argu-ments (Rozeff , 1982; Easterbrook, 1984; Jensen, 1986; Eckbo and Verma, 1994; Zeckhauser and Pound, 1990). No signifi cance is found with respect to the non-linear parameter (E(t-1)*Inst2).

The coeffi cient of earnings in period t (Et) and in period t−1 (E(t−1)) is also signifi cant at the 1 percent level. Previous period’s dividend payout (Div(t-1)) is again signifi cant, both statistically and in real economic terms.


This indicates that the fi rms only partially adjust the dividends to meet changed target dividend levels.

A key assumption which must hold if the FGLS method is to provide reliable estimates is that the errors are randomly distributed. Most likely, the errors are in fact correlated with the regressors, or in other words there are individual fi rm eff ects. To test whether this is true, a fi xed-eff ects model, which allows not only time eff ects but also individual fi rm eff ects, is tested (Model 4 in Table 10.9). The Hausman test confi rms that the suspicion of individual eff ects and the Hausman-H0 of non-correlated errors can be soundly rejected.

As the Hausman test confi rms the existence of signifi cant fi rm eff ects correlated to the regressors, the fi xed-eff ects estimation method is appro-priate. Table 10.9 presents the results from this estimation with individual fi rm and time eff ects. As before, the estimation is made with ownership in terms of both votes (Model 4a) and capital (Model 4b).

Table 10.8 Cross-sectional time-series FGLS estimations: Model 3a fi rms with vote-diff erentiated shares, Model 3b fi rms without vote-diff erentiated shares


Model 3aI

(votes)Model 3bI

(capital)Model 3aII

(votes)Model 3bII

(capital)

Et 0.0528*(9.71)

0.0529*(9.71)

0.0840*(19.76)

0.0876*(18.41)

E(t-1) −0.0237*(−3.09)

−0.0235*(−3.07)

−0.0206*(−4.17)

−0.0215*(−2.58)

E(t-1)*Inst −0.0013(−1.48)

−0.0012(−1.57)

0.0012**(2.35)

−0.0015**(−1.97)

E(t-1)*Inst2 3.0e−05(1.52)

3.2e−05(1.60)

−9.15e−06(−0.74)

−2.9e−05(−1.60)

Div(t-1) −0.0613***(−1.80)

−0.0598***(−1.75)

−0.2090*(−7.97)

−0.2282***(−7.80)

constant 2.9757*(3.15)

2.9029*(3.11)

8.7769*(7.38)

9.9986*(6.49)

No. obs Model 4a 5 443A

No. groups Model 4a 5 85No. obs Model 4b 5 742B

No. groups Model 4b 5 116

Notes: t-statistics are in parentheses. Panels: heteroscedastic. Correlation: panel-specifi c AR(1).A 2 obs dropped because only 1 obs in group. B 3 obs dropped because only 1 obs in group.* denotes signifi cance at the 1% level, ** denotes signifi cance at 5%, and *** denotes signifi cance at the 10% level.


The results of the fi xed-eff ects estimation in Table 10.9 are highly signifi -cant. The coeffi cient of earnings in period t (Et) is signifi cant and positive, and that of earnings in t−1 (E(t-1)) is signifi cant and negative. As expected, there is a signifi cant earnings component related to dividends. The coef-fi cients related to dividends in previous period (Div(t-1)) are likewise again signifi cant and negative with respect to dividend change. Recall that this term represents the ‘speed of adjustment’ of dividend changes. The results for the estimation with institutional ownership in terms of both votes and capital share are in fact remarkably stable with regard to the size of the coeffi cients and so on. The elasticity of dividends with regard to changes in earnings is around 30 percent, which seems highly plausible. This again confi rms the robustness of the model formulation.

In both estimations vote-diff erentiated shares have a signifi cantly posi-tive eff ect on dividend changes. Again, this is an indication that investors demand higher dividends in fi rms which allow vote-diff erentiated shares. Hence hypotheses 1a, 1b, and 2 are corroborated.

Table 10.9 Fixed-eff ects estimations: Model 4 institutional ownership (votes and capital)


Model 4a(votes)

Model 4b(capital)

Et 0.1060*(7.86)

0.1065*(8.76)

E(t-1) −0.1100*(−2.85)

−0.1320*(−2.34)

E(t-1)*Inst 0.0066*(2.77)

0.0079**(2.23)

E(t-1)*Inst2 −0.0002*(−2.69)

−0.0002*(−2.76)

E(t-1)*VoteDiff 0.0774*(3.08)

0.0514*(2.17)

Div(t-1) −0.6094*(−2.96)

−0.5582*(−2.81)

Fixed eff ects signifi cant? Yes* Yes*Number of obs51190Number of groups5189

R2

within50.5054between50.4884overall50.1774

R2

within50.4913between50.4980overall50.1965

Notes: Robust t-statistics are in parentheses.* denotes signifi cance at the 1% level, ** denotes signifi cance at 5%, and *** denotes signifi cance at the 10% level.


As before, the sample of fi rms is separated into two groups depending on whether or not they have vote-diff erentiated shares. The interaction term made up of earnings and the dummy for vote-diff erentiation is taken out of the regressions, as it would produce collinearity. Table 10.10 provides the results for the fi xed-eff ects estimation with institutional ownership when the sample of fi rms is divided in two groups depending on whether or not they have vote-diff erentiated shares (Model 4aI and 4bI and Model 4aII and 4bII).

Looking at the results in Table 10.10, there is as expected a positive but non-linear relation between institutional ownership and dividend changes if the fi rms have a vote-diff erentiated share structure (Model 4aII and 4bII). Based on the arguments of Miguel et al. (2004) and the discussion about institutional owners’ incentives, hypothesis (1b) of non-linearity between institutional ownership and dividend behaviour was formulated. To control for this eventual non-linearity additional interaction terms of squared institutional ownership are added.10 The signifi cance of these parameters confi rms hypotheses 1a and 1b of a positive and diminishing eff ect of institutional ownership on dividend changes, for ownership in terms of both votes (Model 4aII ) and capital (Model 4bII ). For large inves-tors in general, Mork et al. (1988) fi nd that profi tability is higher for fi rms with shareholders that have up to 5 per cent ownership stakes; beyond that, profi tability drops (see Section 2 for further discussion).

As the sample is divided between fi rms with vote-diff erentiated shares (Model 4aI and 4bI) and fi rms without (Model 4aII and 4bII), the estimated parameter on previous periods’ earnings loses its signifi cance in the group of fi rms that have vote-diff erentiated shares (Model 4aI and 4bI).

The results for all the estimations are remarkably robust in terms of the sign and size of the coeffi cients. The pooled OLS results strongly support the results in the FGLS estimation. However, as there are signifi cant indi-vidual fi rm eff ects, the fi xed-eff ects method is more appropriate, although the FGLS results point in the same direction. Furthermore the use of institutional ownership measured continuously, and not simply by dummy variables related to fi xed levels of ownership percentages, provides a more thorough understanding of the non-linear relationship between ownership and dividend policies.

As much of the analysis is based on reported earnings, the usual caveats related to accounting fi gures apply. Ownership, however, is a very stable variable over time, even though institutional ownership belongs to the cat-egory of ownership that is perhaps most volatile. This and the inclusion of time and fi rm eff ects in the estimation give a good indication of the robust-ness in the results. All estimations have also been made with total payout.11 These results, although limited by the small number of fi rms involved in share repurchases in the sample, support the estimation results for dividends.

246

Tab

le 1

0.10

Fi

xed-

eff e

cts e

stim

atio

ns: M

odel

4a

fi rm

s with

out v

ote-

diff e

rent

iate

d sh

ares

, Mod

el 4

b fi r

ms w

ith v

ote-

diff e

rent

iate

d sh

ares

Dep

ende

nt v

aria

ble

(Div

t-D

iv(t

-1))

Mod

el 4

aI

(vot

es)

Mod

el 4

bI

(cap

ital)

Mod

el 4

aII

(vot

es)

Mod

el 4

bII

(cap

ital)

Et

0.12

85*

(3.7

3)0.

1285

*(3

.73)

0.09

88*

(6.7

9)0.

1011

*(7

.91)

E(t

-1)

−0.

0835

***

(−1.

71)

−0.

0835

***

(−1.

71)

−0.

0243

(−1.

01)

−0.

0737

(−1.

44)

E(t

-1)*

Inst

0.

0006

(−0.

11)

−0.

0006

(−0.

11)

0.00

69*

(2.6

1)0.

0079

**(1

.98)

E(t

-1)*

Inst

2 5.

37e-

06(0

.04)

5.38

e-06

(0.0

4)−

0.00

02**

(−2.

45)

−0.

0002

**(−

2.35

)D

iv(t

-1)

−0.

4287

***

(−1.

68)

−0.

4287

***

(−1.

68)

−0.

6594

*(−

2.76

)−

0.59

20**

(−2.

61)

Fix

ed e

ff ect

s sig

nifi c

ant?

Yes

*Y

es*

Yes

**Y

es**

No.

obs

Mod

el 4

a544

5N

o. g

roup

s Mod

el 4

a587

No.

obs

Mod

el 4

b574

5N

o. g

roup

s Mod

el 4

b511

9

R2

with

in5

0.69

61be

twee

n50.

5934

over

all5

0.08

27

R2

with

in5

0.69

61be

twee

n50.

5934

over

all5

0.08

27

R2

with

in5

0.49

90be

twee

n50.

3745

over

all5

0.19

02

R2

with

in5

0.48

15be

twee

n50.

3778

over

all5

0.21

78

Not

es:

Rob

ust t

-sta

tistic

s are

in p

aren

thes

es.

* de

note

s sig

nifi c

ance

at t

he 1

% le

vel,

** d

enot

es si

gnifi

canc

e at

5%

, and

***

den

otes

sign

ifi ca

nce

at th

e 10

% le

vel.


6. CONCLUSIONS

This chapter investigates the relationship between institutional ownership and dividends. To test this relationship a version of the so-called earnings trend model is utilized, with the inclusion of interaction terms made up of institutional ownership. Using a panel data methodology which accounts for fi rm-specifi c eff ects and time eff ects, unobservable heterogeneity is controlled for. Furthermore the relationship is tested by extending the investigation into a non-linear setting in which incentives, monitoring and agency-cost eff ects can be more accurately accounted for.

The results clearly show that institutional ownership, in terms of both votes and capital, where these two are separated, has a positive eff ect on dividend payout policies. So even if high desired levels of dividends can be seen as a sign of ‘short-termism’ (Hutton, 1995; Haskins, 1995), it might just as well be an eff ect of these owners’ attempts to reduce the free cash fl ow available to management, as argued by Jensen (1986). Institutional owners might thus play a monitoring role, and in doing so mitigate the problems associated with the separation of ownership and control in listed fi rms. The relation is found to be positive but diminishing, which supports previous research concerning the relation between dividends and owner-ship structure. The use of a comprehensive database covering institutional ownership continuously allowed for this additional test and also the rejection of other functional forms of the ownership–dividend relation-ship. Furthermore, and in line with expectations, earnings have a positive impact on dividend changes.

By examining Swedish listed fi rms, the chapter also provides empirical evidence on the eff ects of control instruments such as dual-class shares on dividend policies. The result, in line with agency cost theory, is that control instruments such as vote-diff erentiated shares induce investors to demand higher levels of dividends as compensation for the increased agency costs. This means that fi rms using this type of control instrument suff er more from subsequent agency problems.

NOTES

* Acknowledgments: Financial support from Sparbankernas Forskningsstiftelse to Daniel Wiberg’s dissertation work is gratefully acknowledged. A research grant from the Centre of Excellence for Science and Innovation Studies (CESIS), Royal Institute of Technology, Stockholm, is also gratefully acknowledged.

1. Beyond the obvious cases of theft, transfer pricing, and asset sales, expropriation may take the form of perquisites, high salaries, diversion of funds to pet projects, and general entrenchment even in cases in which the managers are no longer competent or qualifi ed to run the fi rm.


2. In this chapter managerial ownership is not considered. Ownership by the largest shareholder in terms of votes is thus considered in alignment with managerial ownership.

3. Law concerning investment funds; Swedish reference, SFS 2004:46; (following European Union Directive EGT L 375, 31.12.1985, s. 3, Celex 31985L0611).

4. In fact a myriad of diff erent tax rates are applied dependent on the type of fi rm, that is, limited liability, private, partnership, and so on. For the sake of brevity this discussion is not extended beyond this note, as it is far beyond the scope of this chapter to analyse the impact of various tax rates on dividends.

5. For extended discussion and derivation of the four models see Short et al. (2002). 6. SIS-Ägarservice. 7. Note that the typical Swedish ownership spheres, large-scale conglomerates combining

a number of control-enhancing mechanisms and often controlled by a foundation, are not included in this defi nition. The incentives of this type of owner are probably sub-stantially diff erent from those of what are usually referred to as institutional investors, that is, fi nancial intermediaries.

8. Approximately €1.2 billion, or $1.6 billion as of June 2007. 9. Breusch-Pagan/Cook-Weisberg test for heteroscedasticity H0: constant variance Variables: fi tted values of Divt-Div(t-1) Chi2(1) 5 171.96 Prob.chi2 5 0.0000 Breusch-Pagan/Cook-Weisberg test for heteroscedasticity H0: constant variance Variables: fi tted values of Et E(t-1) E(t-1)*Inst (votes) E(t-1)*Inst2 (votes) E(t-1)*VoteDiff

Dummy Div(t-1) Chi2(1) 5 171.96 Prob.chi2 5 0.0000 Each of these tests indicates that there is a signifi cant degree of heteroscedasticity in this

model. In order to get effi cient estimators and account for this heteroscedasticity GLS estimation is thus required.

10. A cubic specifi cation of the model has been tested but yields no signifi cant results.11. For the sake of brevity these results are available from the author upon request.

REFERENCES

Agnblad, J., Berglöf, E., Högfelt, P. and Svancar, H. (2001), Ownership and Control in Sweden: Strong Owners, Weak Minorities and Social Control. The Control of Corporate Europe, Oxford: Oxford University Press.

Bhattacharya, S. (1979), ‘Imperfect Information, Dividend Policy, and the Bird in the Hand Fallacy’, Bell Journal of Economics, 10, 259–70.

Bhattacharya, S. (1980), ‘Nondissipative Signalling Structure and Dividend Policy’, Quarterly Journal of Economics, 95, 1–24.

Bjuggren, P.-O. and Wiberg, D. (2008), ‘Industry Specifi c Eff ects in Investment Performance and Valuation of Firms – Marginal q in a Stock Market Bubble’, EMPERICA, 35 (3), 279–91.

Bjuggren, P.-O., Eklund, J. and Wiberg D. (2007), ‘Ownership Structure, Control and Firm Performance: The Eff ects of Vote Diff erentiated Shares’ Applied Financial Economics, 17, 1323–34

Burkart, M., Gromb, D. and Panunzi, F. (1997), ‘Large Shareholders, Monitoring and the Value of the Firm’, Quarterly Journal of Economics, 113, 693–728.

Crutchley, C., Jensen, M., Jahera, J. and Lang, L. (1999), ‘Agency Problems and the Simultaneity of Financial Decision Making: the Role of Institutional Ownership’, International Review of Financial Analysis, 8 (2), 177–97.


Davis, P. and Steil, B. (2001), ‘Institutional Investors’, Cambridge, MA: The MIT Press.

Easterbrook, F. (1984), ‘Two Agency-Cost Explanations of Dividends’, American Economic Review, 74, 650–59.

Eckbo, B. and Verma, S. (1994), ‘Managerial Shareownership, Voting Power, and Cash dividend Policy’, Journal of Corporate Finance, 1, 33–62.

Fama, E.F. and Babiak, S. (1968), ‘Dividend analysis: an empirical analysis’, Journal of the American Statistical Association, 63, 1132–61.

Gedajlovic, E. and Shapiro, D. (1998), ‘Management and Ownership Eff ects: Evidence from Five Countries’, Strategic Management Journal, 19, 533–53.

Grier, P. and Zychowicz, E. (1994), ‘Institutional Investors, Corporate Discipline and the Role of Debt’, Journal of Economics and Business, 46 (1), 1–11.

Gugler, K. and Yurtoglu, B. (2003), ‘Corporate Governance and Dividend Pay-out Policy in Germany’, European Economic Review, 47, 731–58.

Haskins, C. (1995), ‘A Company Should be Run Primarily in the Interest of its Shareholders’, Enterprise and Governance, The Proceedings of a Conference Held at the Institute of Directors on 24 October, IoD, London.

Henrekson, M. and Jakobsson, U. (2005), ‘The Swedish Model of Corporate Ownership and Control in Transition’, in Huizinga, H. and Jonung, L. (eds), Who Will Own Europe? The Internationalisation of Asset Ownership in Europe, Cambridge: Cambridge University Press.

Henrekson, M. and Jakobsson, U. (2006), ‘Den Svenska Modellen för Företagsä-gande och Företagskontroll vid Skiljevägen’, Institutet för Näringslivsforskning, Policy Paper, No. 1.

Hutton, W. (1995), The State We’re In, London: Jonathan Cape.Högfeldt, P. (2004), The History and Politics of Corporate Ownership in Sweden,

NBER Working Paper Series, 10641.Jensen, M. (1986), ‘Agency Costs of Free Cash Flow, Corporate Finance, and

Takeovers’, American Economic Review, 76, 323–9.Jensen, M. and Meckling, W.H. (1976), ‘Theory of the Firm: Managerial Behaviour,

Agency Costs and Ownership Structure’, Journal of Financial Economics, 3, 305–60.

Jensen, M., Solbreg, D. and Zorn, T. (1992), ‘Simultaneous Determination of Insider Ownership, Debt, and Dividend Policies’, Journal of Financial and Quantitative Analysis, 27 (2), 247–63.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R. (1999), ‘Corporate Ownership Around the World’, Journal of Finance, 52, 1131–50.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A. and Vishny, R. (2000), ‘Agency Problems and Dividend Policies around the World’, Journal of Finance, 55 (1), 1–33.

Lintner, J. (1956), ‘Distribution of Incomes of Corporations among Dividends, Retained Earnings and Taxes’, American Economic Review, 46, 97–113.

McConnell, J. and Servaes, H. (1990), ‘Additional Evidence on Equity Ownership and Corporate Value’, Journal of Financial Economics, 27, 595–612.

Miguel, A., Pindado, J. and de la Torre, C. (2004), ‘Ownership Structure and Firm Value: New Evidence from Spain’, Strategic Management Journal, 25, 1199–1207.

Miller, M. and K. Rock, (1985), ‘Dividend Policy under Asymmetric Information’, Journal of Finance, 40, 1031–51.

Moh’d, M., Perry, L. and Rimbey, J. (1995), ‘An Investigation of the Dynamic


Association Between Agency Theory and Dividend Policy’, Financial Review, 30 (2), 367–85.

Morck, R., Shleifer, A. and Vishny, R. (1988), ‘Management Ownership and Market Valuation: An Empirical Analysis’, Journal of Financial Economics, 20, 293–317.

Mueller, D. (2003), The Corporation, London: Routledge.Pindado, J. and de la Torre, C. (2006), ‘The Role of Investment, Financing and

Dividend Decisions in Explaining Corporate Ownership Structure: Empirical Evidence from Spain’, European Financial Management, 12 (5), 661–87.

Rozeff , M.S. (1982), ‘Growth, Beta and Agency Costs as Determinates of Dividend Payout Ratios’, Journal of Financial Research, 5, 249–59.

Schooley, D. and Barney, D. (1994), ‘Using Dividend Policy and Managerial Ownership to Reduce Agency Costs’, Journal of Financial Research, 17 (3), 363–73.

Short, H., Zhang, H. and Keasey, K. (2002), ‘The Link between Dividend Policy and Institutional Ownership’, Journal of Corporate Finance, 8, 105–22.

Waud, R. (1966), ‘Small Sample Bias due to Misspecifi cation in the “Partial Adjustment” and “Adapted Expectations” Models’, Journal of the American Statistical Association, (Dec.), 134–45.

Williamson, Oliver E. (1963), ‘Managerial Discretion and Business Behavior’, The American Economic Review, 53 (5), 1032–57.

Williamson, Oliver E. (1964), The Economics of Discretionary Behavior: Managerial Objectives in a Theory of the Firm, Englewood Cliff s, NJ: Prentice-Hall.

Zeckhauser, R. and Pound, J. (1990), ‘Are Large Shareholders Eff ective Monitors? an Investigation of Share Ownership and Corporate Performance’, in Hubbard, R.G. (ed.), Asymmetric Information, Corporate Finance and Investment, Chicago: University of Chicago Press, pp. 149–80.

251

APP

EN

DIX

10.

1

Tab

le A

10.1

C

orre

latio

n m

atri

x pa

irw

ise

corr

elat

ion

Var

iabl

eD

ivt

DD

iv

Prst

kcT

P tD

TP

Et

DE

C1

V1

C5

Div

t1.

000

DD

iv0.

518*

1.00

0Pr

stkc

0.15

2*0.

032

1.00

0T

Pay t

0.93

6*0.

450*

0.49

0*1.

000

DT

P0.

508

1.00

0*0.

032

0.46

0*1.

000

Et

0.77

8*0.

107*

0.21

6*0.

763*

0.48

1*1.

000

DE

0.19

7*0.

112*

0.06

6*0.

197*

0.51

8*0.

470*

1.00

0C

1−

0.02

10.

007

−0.

064*

−0.

042

0.01

0−

0.05

00.

018

1.00

0V

10.

034

−0.

033

−0.

067*

0.00

60.

010

−0.

009

−0.

017

0.78

0*1.

000

C5

−0.

092*

−0.

013

−0.

083*

−0.

111*

−0.

037

−0.

114*

−0.

063*

0.79

2*0.

660*

1.00

0V

50.

029

−0.

028

−0.

079*

−0.

002

−0.

011

0.01

70.

029

0.63

6*0.

823*

0.78

9*F

C0.

192*

−0.

001

0.07

2*0.

195*

0.06

0*0.

191*

0.04

8−

0.21

2*−

0.20

7*−

0.21

9*F

V0.

062*

0.01

80.

044

0.07

0*0.

043

0.09

5*0.

047

−0.

151*

−0.

269*

−0.

145*

IC0.

203*

0.04

20.

103*

0.21

6*0.

056

0.18

0*0.

029

−0.

216*

−0.

163*

0.19

7*IV

0.20

9*0.

040

0.15

1*0.

238*

0.05

60.

192*

0.01

9−

0.24

2*−

0.32

8*−

0.23

9*V

oteD

iff 0.

128*

−0.

016

0.04

50.

129*

0.02

60.

111*

0.01

30.

086*

0.47

3*0.

130*

Sale

s0.

734*

0.02

70.

176*

0.71

0*0.

150*

0.54

1*−

0.01

50.

115*

−0.

010

−0.

157*

Em

p0.

523*

0.11

3*0.

117*

0.50

3*0.

113*

0.43

3*0.

022

−0.

139*

−0.

014

−0.

143*

R&

D0.

409*

0.02

20.

060*

0.38

2*0.

022

0.17

5*−

0.11

0*−

0.11

9*−

0.00

4−

0.13

4*W

Cap

0.57

9*0.

149*

0.12

8*0.

556*

0.14

9*0.

356*

0.05

5−

0.13

3*−

0.01

5−

0.17

1*

252

Tab

le A

10.1

(c

ontin

ued)

Var

iabl

eV

5F

CF

VIC

IVV

oteD

iff Sa

les

Em

pR

&D

-exp

WC

ap

Div

t

DD

ivPr

stkc

TPa

y tD

TP

Et

DE

C1

V1

C5

V5

1.00

0F

C−

0.20

8*1.

000

FV

−0.

268*

0.92

0*1.

000

IC−

0.15

2*−

0.00

1*−

0.01

11.

000

IV−

0.32

1*0.

058*

0.06

4*0.

899*

1.00

0V

oteD

iff 0.

520*

−0.

080*

−0.

182*

0.00

5−

0.20

3*1.

000

Sale

s0.

039

0.24

5*0.

050

0.18

4*0.

208*

0.16

4*1.

000

Em

p0.

034

0.29

5*0.

108*

0.15

8*0.

154*

0.16

6*0.

743*

1.00

0R

&D

0.07

5*0.

164*

−0.

031

0.02

70.

033

0.08

3*0.

668*

0.39

5*1.

000

WC

ap0.

045

0.20

4*−

0.00

30.

097

0.11

7*0.

126*

0.79

0*0.

500*

0.90

2*1.

000

Not

e:

* C

orre

latio

n co

effi c

ient

sign

ifi ca

nt a

t the

5%

leve

l.

253

11. Contracting around ownership: shareholder agreements in France1

Camille Madelon and Steen Thomsen

1. INTRODUCTION

A wealth of studies in economics, strategy and fi nance have examined the relationship between corporate ownership structure and performance (Hill and Snell, 1988, 1989; Holderness and Sheehan, 1988; McConnell and Servaes, 1990; Gedajlovic and Shapiro, 1998, 2002; Thomsen and Pedersen, 2000; De Miguel et al., 2004; Anderson and Reeb, 2003; Villalonga and Amit, 2006). Other studies have examined the eff ect of ownership structure on strategic decisions (Amihud and Lev, 1981; Hill and Snell, 1988, 1989; Graves, 1988; Baysinger et al., 1991; Lane et al., 1998; Denis et al., 1997, 1999; Allen and Phillips, 2000; David et al., 2001; Hoskisson et al., 2002; Lee and O’Neill, 2003; Desai et al., 2004; Lerner and Rajan, 2006; Mathews, 2006). Overall, this literature fi nds that corporate ownership structures matter to company behavior and value creation (for example, Shleifer and Vishny, 1997). Yet, there is a distinction to be made between the publicly observable formal ownership structure and what we are tempted to call the real ownership structure, namely the allocation of control, cash fl ow, and transfer rights, which results from implicit or explicit contracting among the various owners. Take for example Publicis, the world’s fourth largest communication group. The formal ownership structure points to two large owners in 2006: Dentsu, a Japanese based communication group, with circa 15 per cent of the equity, and Mrs Badinter, the founder’s daughter with 10 per cent of the equity. However, a closer look reveals a diff erent picture. In 2002, Mrs Badinter and Dentsu signed a binding agreement in which Dentsu committed to act in unison with the founder’s family for decisions related to corporate strategy and board member nominations. While the formal structure indicates two blockholders with no majority power, the shareholder agreement reveals a dominant owner with a majority in terms of voting power. This case is not isolated and points to the importance of


an in-depth understanding of ownership arrangements when studying the relationship between ownership and fi rm outcomes. Real ownership struc-ture may supersede the formal structure.

Surprisingly, this ‘contracting around’ phenomenon has received almost no attention in the literature. We have no idea (1) why the owners of publicly listed fi rms would resort to shareholder agreements and (2) how this might infl uence company behavior and value creation. One stream of literature broadly termed fi nancial contracting has studied the deals between fund providers and those needing the funds (Hart, 2001), yet it has empirically focused on private companies and more particularly on the vertical relationships between venture capitalists and small entrepreneurial fi rms (Kaplan and Strömberg, 2003) in contrast to the horizontal contracts between shareholders. Another stream of literature tracks corporate stra-tegic alliances and ownership ties between companies (Allen and Phillips, 2000; Gomes-Casseres et al., 2006), but does not include contracts between shareholders, To our knowledge, there are no studies explicitly examin-ing the antecedents of agreements among shareholders in listed fi rms or their impact on value creation. In this chapter, we seek to take a fi rst step towards bridging this gap.

Given that the literature is still at an early stage, we chose to conduct an exploratory study. To this end, we performed a qualitative multi-case analysis on a selection of listed French fi rms with shareholder agreements. France off ers a unique empirical setting to study shareholder agreements for three main reasons. First, shareholder agreements are quite common among listed fi rms and equally so among the very largest fi rms (which are part of the benchmark CAC 40 stock index). Second, the French market authority (AMF) requires that owners publicly disclose their agreements. All contracts are stored in the AMF database and their detailed content is readily accessible. Third, because of the public nature of these agreements, we may study how markets react to the announcement of agreements, thus capturing a measure of value creation.

From our exploratory analysis, we propose several conditions under which shareholders are more likely to opt for agreements. They are linked respectively to the nature of ownership, the nature of the industry and the nature of the problems facing the fi rm. In these cases, we argue that the overall cost of signing and enforcing shareholder agreements will be off set by the benefi ts. In addition, we suggest that shareholder agreements are highly idiosyncratic and may create or destroy value, depending on the scope and nature of the agreement.

The remainder of the chapter is organized as follows. First, we lay out the background to this study by briefl y reviewing the existing arguments on shareholder agreements. Second, we describe the empirical setting and

Contracting around ownership 255

methods. Third, using an extended transaction cost framework we propose hypotheses on the antecedents and eff ects of shareholder agreements. Finally, we discuss our fi ndings and conclude.

2. BACKGROUND

2.1 Defi nition of Shareholder Agreements

Shareholder agreements can be broadly defi ned as written or unwritten contracts between shareholders. In this chapter, we choose to focus only on written contracts because unwritten contracts are hard to document. They involve rights or obligations beyond what is prescribed by law. Standard components include (Chemla et al., 2007):

(i) Options or limitations on the right to buy or sell shares, for example, preemption rights (call options at a specifi ed ‘fair’ price), collective action clauses, such as tag-along rights (the rights to go with other investors) or drag-along rights (the obligation to do so)

(ii) Control rights, for example, rights to appoint a member of the board or veto certain critical decisions or the obligation to vote with another shareholder

(iii) Cash fl ow rights, for example, catch-up clauses which specify rights to parts of the proceeds in case the company is sold to a third party

(iv) Procedures for dispute resolution in case of disagreement.

Just like other private contracts, shareholder contracts are enforceable in court.

One particular stream of literature, called ‘fi nancial contracting’, has been concerned with the fi nancial deals between fi nanciers and those needing the funds (Hart, 2001). Yet, it has mainly addressed fundamen-tal questions such as the nature of debt and equity or optimal capital structure. There has been very little empirical work done on the contracts themselves, and even less on the role of these contracts for strategy and management. Notable exceptions are Kaplan and Strömberg’s (2003) research on shareholder contracts in venture capital fi rms and Chemla and colleagues (2007) on closely held fi rms. Kaplan and Strömberg (2003) analyze the characteristics of the ‘real world’ contracts between venture capital fi rms and entrepreneurs with respect to board rights, liquidation rights, voting rights and cash fl ow rights. They fi nd those features to be in line with the existing fi nancial contracting theories such as principal–agent and control theories (for example, Aghion and Bolton, 1992) but with


much more internal complexity. Chemla and colleagues (2007) argue that the inclusion of strong clauses such as tag-along and drag-along clauses in the contracts protects investors from opportunism and thus ensures some effi cient ex ante investments in the fi rm.

2.2 Related Literature

Although there has been little attention to shareholder agreements in listed corporations in previous work, there are several interesting streams of research, which may contribute to theory development: general contract theory (for example, Shavell, 2004, Chapters 13–14), fi nancial contract-ing theory (Hart, 2001), transaction cost theory (Williamson, 1985, 2005), research on strategic alliances and joint ventures (Kogut, 1988; Mowery et al., 1996; Doz and Hamel, 1998; Gomes-Casseres et al., 2006), theories of relational contracting (Macaulay, 1963; Ellickson 1991; Mnookin and Komhauser, 1979; Zaheer and Venkatraman, 1995; Poppo and Zenger 2002; Carson et al., 2006), takeover theories (Scharfstein, 1988; Danielson and Karpoff , 1998; Mikkelson and Partch, 1997; Adams and Ferreira, 2007; Burkart and Lee, 2007) and macrostudies of law and fi nance (La Porta et al., 2000; Djankov et al., 2007). We draw on these contributions in the following.

3. EMPIRICAL SETTING AND METHODS

France off ers a unique empirical setting to study shareholder agreements. First, shareholder agreements are quite common among listed fi rms. Second, the French market authority requires that owners publicly dis-close their agreements. All contracts are stored in the market authority’s database and their detailed content is readily accessible. Third, because of the public nature of these agreements, we may study how markets react to the announcement of agreements, thus capturing a measure of value creation.

In this section, we briefl y describe the nature of shareholder agreements in French listed companies and explain the sources and methods used. We then provide a short description of each of the cases.

3.1 Shareholder Agreements in France

Shareholder agreements (pactes d’actionnaires) are relatively common among listed fi rms in France. Among the 749 listed fi rms on Euronext Paris,2 268 fi rms had had at least one shareholder agreement between 1997


and 2007. Over the same period, an average of 38 new shareholder agree-ments were announced each year (see Appendix 11.1). These contracts are prevalent across all sizes of listed fi rms including the very largest (CAC 40).

A typical contract is a 1–6-page document that specifi es the nature of the agreement between the signing shareholders. It starts by stating the names of the shareholders and their respective equity and voting shares. It goes on to describe the allocation of power between the parties, notably in terms of voting rights and allocation of board seats. It then describes the obligations and rights of the signing parties in case of events such as the nomination of new board members, corporate strategy decisions, takeo-vers and equity selling.

The shareholder agreements used by listed fi rms appear to be much less detailed than the ones used by private fi rms and in particular by private equity funds. As an illustration, take Legrand, one of the leading world-wide manufacturers of electrical equipment. In 2002, the company was sold to two private equity funds, KKR and Wendel; in 2006, part of the equity was fl oated, with KKR and Wendel keeping 59 per cent of the equity. The 2006 (publicly available) agreement between KKR and Wendel is a 6-page contract. We interviewed one of the directors at Wendel, who revealed that the initial contract between the two private equity funds and subsequent partners was more than a hundred pages long with a large amount of detail. Why? We hypothesize that complex contracts among shareholders in closely held fi rms substitute for the standard form contracts of listing requirements and securities law which listed companies implicitly rely on. In the private setting, there is a higher level of expropriation risk for share-holders, requiring in turn a tighter control of the allocation of power and cash fl ow rights between shareholders. In addition, in the absence of public information requirements, shareholders of private fi rms often add detailed specifi cations of the information they will get from management (such as the detailed monthly reports). Last but not least, shareholder agreements among private equity fi rms usually include the compensation packages of top management and profi t-sharing schemes.

Shareholder agreements among publicly traded companies are most often signed for periods ranging from two to fi ve years with tacit renewals for the shortest periods. Clauses usually specify how the parties can put a halt to the contract without suff ering penalties.

Shareholder contracts are offi cially enforceable in a commercial court (article I. 233-11 of commercial law). Yet, in case of disagreements, the contract parties rarely resort to court. They use private settlements. Why? According to the lawyers and contract parties we interviewed, private settlements are more discreet than going to court. Nothing is published.


The common practice is to go through what is called a ‘referee’ (arbitre). Referees are not judges; they are appointed by the chamber of commerce. They act as ‘go-betweens’ and ‘tension soothers’ to prevent escalation between the signing parties. While private settlements keep shareholders from the spotlights, they are reported to be very costly, suggesting in turn that the benefi ts of privacy are highly valued by the contract parties. In fact seeking private agreements rather than enforcing contracts through the courts is quite common for many types of contract (Macaulay, 1963). The parties bargain in ‘the shadow of the law’ (Mnookin and Komhauser, 1979).

3.2 Disclosure Requirements

The French market authority (Autorité des Marchés Financiers) requires that shareholders of listed companies disclose the details of the contracts they have signed with other shareholders (article I-233-11 of commercial law). There is an additional notifi cation when shareholders choose to act in concert (article I-233-10 of commercial law).3 Those breaking the law face the risk of being deprived of their shareholder rights. The detailed contracts of French listed fi rms are compiled in a database and made acces-sible to the general public on the AMF website (www.amf-france.org) for the period 1997–2007.

3.3 Sources and Methods

Faced with the need to develop new theory we decided on an explora-tive approach which derives theoretical propositions from case studies and related research. We chose multiple cases because more cases tend to increase the degree of generalizability of the results (Yin, 1984) through the use of a replication logic. The emerging theory is tested by confronting it with new cases which force it to distinguish between the common and idiosyncratic features of the cases. We followed an ‘abductive’ approach (Peirce, 1935) rather than a grounded approach (Glaser and Strauss, 1967) as we did not start with a blank slate but with some knowledge of share-holder agreements – although not directly related to our setting of listed fi rms – drawing on the theory of fi nancial contracting, transaction theory, general contract theory and the theory of relational contracting.

We focused on fi ve diff erent cases, all involving large listed fi rms. Whereas in a quantitative study the sampled fi rms need to be representative of the population, in a multiple-case-study design, fi rms are selected for theoretical reasons (Einsenhardt and Graebner, 2007). We purposely chose cases from diff erent settings. This selection was made after reading through


the 380 or so contracts available for the listed fi rms in the AMF database. We focus on the following fi ve cases, which are described in greater detail in the following section:

The Pernod Ricard agreement is an unbalanced contract between a ●

large founding family and a small minority investor.The Publicis deal features a large blockholder ready to give up ●

control to the founding family.The Club Med agreement is a complex contract between multiple ●

small owners with diff erent interests (institutional investors, a hotel group and a real estate fi rm).The Legrand agreement involves two large private equity funds ●

which initially took over the company privately.The Schneider Electric agreement reveals complex ownership link- ●

ages between banks, insurance companies and large non-fi nancial corporations.

For each case, we proceeded in the following way. We fi rst read the shareholder agreements in detail, classifying the content of the agreements into several categories (see Appendix 11.2). Second, we used a variety of secondary sources, including annual reports, analysts’ reports, press search and internet search to substantiate the context and content of the contracts. For data on ownership structure, we used the Dafsaliens database. Third, we carried out interviews with the fi rms’ signing parties and lawyers. At this point interviews are accepted and performed for two of the cases.

For the second research question on the value impact of shareholder agreements, we examined share price reactions around announcement dates (1/2 1 month and 1/2 2 months around the publication) using Datastream fi nancial data.

3.4 Description of the Cases

We now give a short description of each case covering key facts on the fi rm, the context to the shareholder agreement and the summarized content of the agreement (details are available in Appendix 11.2).

Pernod RicardPernod Ricard is the world’s second largest operator in wine and spirits (2005 sales: €3674 million; 2005 net profi t: €475 million). It owns brands like Chivas Regal, Ballentin’s, Malibu, Ricard and Mumm. The company originates from the 1975 merger between two traditional French com-panies, Pernod and Ricard (respectively held by the families Pernod


and Ricard). Between 1975 and 2001 it grew through both external and organic growth. In 2001, it bought a large part of the Seagram’s wine and spirits activities. In 2005 it acquired Allied Domecq in partnership with Fortune brands and became number two worldwide. In March 2006, a shareholder agreement was signed between the Ricard family (10 percent equity) and Kirin Corporation, Japan’s largest spirits operator (3 percent equity). In a nutshell, the agreement states that both shareholders should act in concert and that Kirin commits to vote in favor of the board rec-ommendations on a stated number of issues. The question is, why? In the following section, we examine why Kirin would want to enter such an agreement.

PublicisPublicis is the world’s fourth largest communication group (2005 sales: €4127 million; 2005 net profi t: €386 million), operating in Europe (40 percent sales with a leading market share) and the US (42 percent sales). Publicis was initially founded by Marcel Bleustein-Blanchet in 1926. The company acquired Saatchi and Saatchi (UK) in 2000, and subsequently Nelson Communication (US) in 2002. In 2002 (March), Publicis merged with Bcom 3, a large US communication network including Leo Burnett, D’Arcy and media buying company Starcom MediaVest. Bcom 3 was created in 2000 through the merger of the Leo Group and the MacManus Group with a capital investment from Dentsu, one of Japan’s largest communication companies. In May 2002, Elisabeth Badinter, Marcel Bleustein-Blanchet’s daughter, Publicis’ main shareholder (28 percent of the equity in 2001; 20 percent of the equity of the new ‘merged’ entity in 2002) and chairwoman, signed an agreement with Dentsu (18 percent equity). In this contract, Dentsu agrees to follow Elisabeth Badinter’s ‘voice’ on all major strategic issues including board nomination. It also commits not to sell its shares before 2012.

Club MedClub Med is one of the leading operators of holiday villages and tours (2005 sales: €1590 million; 2005 net profi t: €4 million). The company was founded in 1950 by two entrepreneurs, and grew mostly organically over the subsequent years. In 2004, Accor, one of the largest worldwide hotel chains acquired a 29 percent stake in Club Med from the Agnelli family (Italian family and former dominant owner) and from institutional inves-tor CDC. It became the dominant (so-called ‘reference’) owner. In 2006, after a CEO change, Accor partly exited Club Med and sold a portion of its shares to several investors: Icade (the real estate arm of French govern-ment-backed CDC), Air France Finance (fi nance arm of Air France) and


Fipar (CDC equivalent for Morocco). In June 2006, a shareholder agree-ment was signed between Accor and the investors. The newly formed coali-tion totaled 22 percent equity. In a nutshell, the investors committed to act in concert and support the management’s decisions. The agreement defi nes the right of the parties for board representation and in case of takeover.

LegrandLegrand is one of the leading worldwide manufacturers of electrical equipment (2005 sales: €3248 million; 2005 net profi t: €101 million). In 2001, Legrand was acquired by its competitor Schneider. However the deal was blocked by the anti-trust authorities and Schneider had to fi nd a new owner. In December 2002, Legrand was acquired through LBO by equity funds KKR and Wendel (investing arm of the Wendel family). In 2006 (April), circa 20 percent of Legrand’s equity was fl oated again on the Euronext stock exchange. In March 2006, KKR and Wendel (total-ing 59.1 percent equity and voting rights, with individual shares of 27.4 percent and a joint entity, Lumina Participation, owning 4.3 percent) published their new shareholder agreement in which they carefully allocate the fi rm’s board seats, cash fl ow rights, voting rights, equity shares and obligations.

Schneider ElectricSchneider Electric is one of the world’s leaders in the design and distribu-tion of electrical equipment (2005 sales: €11679 million; 2005 net profi t: €494 million). The company was founded in 1836 by the Schneider family and was initially focused on steel production. Over the years, it divested the steel business and invested in electricity-related activities. In 1999, it changed its name to Schneider Electric to signal its strong focus on electricity. Over the past decade, it has acquired a range of companies in various electricity related areas. Between 1993 and 2006, Schneider signed a number of agreements with French institutional investors (banks and insurance companies) and large French corporations. In 1993, it signed a contract with insurance companies (AXA, AGF), banks (Paribas, Société Générale) and a large energy company (Elf). In 1998, the contract was renewed between AGF (3.3 percent voting rights), AXA (9.6 percent voting rights) and BNP-Paribas (5.4 per cent voting rights). In 2002, the contract was modifi ed again (avenant), updating the respective shares of the signing parties. In 2002, AXA, AGF and BNP-Paribas broke the agreement. In 2006 (May), AXA and Schneider signed an agreement in which AXA committed to keep a certain number of shares in Schneider while Schneider reciprocally committed to maintain a minimum number of shares in Schneider.


4. ANTECEDENTS OF SHAREHOLDER AGREEMENTS

Why would shareholders in listed companies resort to shareholder agree-ments rather than use other types of governance mechanisms such as own-ership concentration or fi nancial hedging? Also, why would shareholders bother to enter into contracts rather than use the standardized provisions of company charters?

We need to understand the conditions under which the benefi ts of con-tracting between shareholders outweigh their costs. The costs of shareholder agreements span both transaction costs (such as the costs of negotiating, renegotiating, concluding and enforcing the contracts) (Williamson, 1979, 2005) and the costs of ‘lock in’ (shareholders agreeing to bind themselves in a contract and thus to lose some degree of fl exibility).

In the following sections we examine the determinants and eff ects of shareholder contracts in an extended transaction cost framework, which draws on advances in the theory of fi nancial contracting, contract theory, relational contracting, law and fi nance, research on strategic alliances and M&A (mergers and acquisitions) as well as Williamsonian transaction cost theory. The basic argument is that ownership and shareholder contracts are alternative means of exercising control and that both may under certain circumstances be more transaction cost effi cient than market solutions.

4.1 Nature of the Ownership

Prior research on governance mechanisms suggests that ownership con-centration is a way to address agency issues between owners and managers (Shleifer and Vishny, 1986, 1997; Holderness and Sheehan, 1988). With increased concentration comes the power and motivation to discipline managers. Yet, in some cases, shareholders may prefer not to increase their stake in the fi rm. They may be capital constrained or risk averse; alterna-tively, they may expect a control loss and lower effi ciency if they take over the fi rm from its present owners. Moreover, ownership concentration may not be necessary if owners can achieve their objectives by other (contrac-tual) means like shareholder agreements that possibly allow them to act in unison and thereby exercise eff ective control.

Our cases suggest that shareholder agreements are more prevalent for fi rms with intermediate levels of ownership concentration, that is, with no majority owner, yet with several large shareholders. Table 11.1 summarizes the fi ndings. In all of the cases, there are at least two large shareholders (with more than 3 percent equity) in the ownership base. As shown by the cases, the contract is made between two or more large shareholders.


A simple explanation is transaction costs of writing and enforcing con-tracts, which may be too large to be worthwhile for smaller shareholders. In addition, small shareholders would expect to gain relatively little by entering into such agreements, because they have limited bargaining power except for rare cases where they can infl uence the balance of power between competing blockholders. In fi rms with a highly fragmented ownership base, small shareholders have no interest in signing an agreement because the costs of joining forces outweigh the benefi ts. In fi rms with high levels of ownership (for example, majority ownership or dominant owner), share-holder agreements are less critical because the need to establish control is absent. However, dominant owners may be interested in entering agree-ments in a dynamic context when ownership changes are expected and when, for example, they plan to sell out part of their shares. In the Club Med case, Accor was initially the dominant owner with 28.9 percent of equity shares. In 2006, when it wished to partly exit its investment (down to an 11.4 percent share), it signed a contract with three institutional owners who, in addition to buying Accor’s pending shares, agreed to act in unison and create a ‘virtual dominant’ owner. Accordingly, we propose:

Proposition 1 Shareholder agreements are more likely to be found in companies with intermediate levels of ownership concentration.

Another striking feature is the identity of the owners entering the agree-ments. Findings are reported in Table 11.1. Signing owners are mostly family owners, corporations and active fi nancial investors such as private equity funds. In both Publicis and Pernod Ricard, the two signing parties are the founder’s family and a corporation. In Club Med, the coalition is led by hotel group Accor. In Legrand, the contract parties are the two private equity funds and their joint entity. As argued above, these owners may not have the incentive to increase their ownership share in the fi rm. Having a controlling stake can be too risky or too costly (for families such as Pernod who need external funding to sustain the strong external growth of the fi rm) or outside of their scope (for corporations such as Accor whose main business is hotels). The common characteristic of these share-holders is their long-term involvement in fi rms. Shareholder agreements are signed for an average of three to fi ve years and create some ‘lock-in’ for the signing parties. The cost of lock-in is likely to be much lower for shareholders with a long-term interest than for fi nancial investors – such as mutual funds – whose managers are evaluated on the short-term perform-ance of their funds (Graves, 1988; Hoskisson et al., 2002; Verstegen Ryan and Schneider, 2002, 2003) and who value fl exibility. Thus we make the following proposition:

264

Tab

le 1

1.1

Shar

ehol

der a

gree

men

ts a

nd th

e na

ture

of o

wne

rshi

p

Firm

Dat

eV

alid

ity p

erio

dC

ontr

act p

artie

sO

wne

rshi

p co

ncen

trat

ion1

Iden

tity

of m

ain

owne

rs

Pern

od

Ric

ard

2006

Mar

ch3.

5 ye

ars

(Mar

ch 2

006

to D

ec. 2

009)

Fou

nder

’s fa

mily

(R

icar

d) –

9.1

% –

an

d co

rpor

atio

n (K

irin

Inte

rnat

iona

l) –

3.6%

Inte

rmed

iate

● 3

ow

ners

with

mor

e th

an 3

% e

quity

sh

ares

(tot

al: 1

9.9%

)●

Aut

o-co

ntro

l (3.

3%)

● F

ound

er’s

fam

ily (R

icar

d): 9

.1%

2

● I

nstit

utio

nal i

nves

tor (

Fra

nklin

R

esou

rces

): 4.

0%●

Cor

pora

tion

(Kiri

n): 3

.6%

● I

nstit

utio

nal i

nves

tor (

CD

C):

3.2%

Publ

icis

2002

May

10 y

ears

(May

200

2 to

Ju

ly 2

012)

Fou

nder

’s fa

mily

(B

adin

ter)

– 1

9.7%

–

and

corp

orat

ion

(Den

tsu)

– 1

8.2%

Inte

rmed

iate

● 2

ow

ners

with

mor

e th

an 5

% e

quity

sh

ares

(tot

al: 3

7.8%

)●

Aut

o-co

ntro

l (6.

7%)

● F

ound

ing

fam

ily (E

. Bad

inte

r3 ): 1

9.7%

● C

orpo

ratio

n (D

ents

u): 1

8.2%

Clu

b M

ed20

06Ju

ne3

year

s (ta

cit

rene

wal

eve

ry

year

); no

lo

nger

val

id

if in

vest

ors

colle

ctiv

ely

own

less

than

15

%

Cor

pora

tion

(Acc

or)

– 11

.4%

and

sele

cted

in

stitu

tiona

l inv

esto

rs

(Fip

ar h

oldi

ng –

10%

–

and

Icad

e –

4%)

Inte

rmed

iate

● 6

ow

ners

with

mor

e th

an 3

% e

quity

sh

ares

(tot

al: 6

0.1%

)

● I

nstit

utio

nal i

nves

tor (

Ric

helie

u F

inan

ce):

26.4

%●

Cor

pora

tion

(Acc

or):

11.4

% (f

rom

28

.9%

in 2

005)

● I

nstit

utio

nal i

nves

tor (

Fip

ar h

oldi

ng):

10%

● F

amily

(Agn

elli

fam

ily v

ia R

olac

o): 4

.7%

● I

nstit

utio

nal i

nves

tor (

Icad

e): 4

%●

Ins

uran

ce c

ompa

ny (N

ippo

n lif

e): 4

%L

egra

nd20

06M

arch

Unt

il th

e da

te

of th

e fi r

st o

f th

ese

2 ev

ents

(1

) KK

R

Bet

wee

n pr

ivat

e eq

uity

fund

s (W

ende

l –

27.7

% –

and

KK

R

– 27

.7%

)

Inte

rmed

iate

● 5

ow

ners

with

mor

e th

an 3

% e

quity

sh

ares

(tot

al: 6

8.6%

)

● P

rivat

e eq

uity

fund

(Wen

del):

27.

7%●

Priv

ate

equi

ty fu

nd (K

KR

): 27

.7%

● P

rivat

e eq

uity

fund

(Lum

ina

Part

icip

atio

n4 ): 4

.3%

265

and

Wen

del

join

tly o

wn

less

than

33%

of

Leg

rand

’s eq

uity

, or

(2) o

ne o

f th

e 2

part

ies

indi

vidu

ally

ow

ns le

ss th

an

5%

and

thei

r joi

nt

vent

ure

Lum

ina

Part

icip

atio

n –

4.3%

● T

op m

anag

emen

t: 3.

8%●

Wen

del a

nd K

KR

did

an

LB

O o

n L

egra

nd in

200

2●

Priv

ate

equi

ty fu

nd (M

onta

gu):

4.8%

● I

nstit

utio

nal i

nves

tor (

Gol

dman

Sac

hs

Cap

ital P

artn

ers)

: 4.1

%

Schn

eide

r E

lect

ric

(1)

2002

Mar

chT

acit

rene

wal

of

the

cont

ract

ev

ery

year

Bet

wee

n ba

nks

(BN

P-Pa

ribas

) and

in

sura

nce

com

pani

es

(AX

A, A

GF

) – 3

.5%

Low

● 2

ow

ners

with

mor

e th

an 3

% e

quity

(t

otal

: 7.4

%)

● A

uto-

cont

rol &

em

ploy

ees (

9.9%

)

● I

nstit

utio

nal i

nves

tors

(CD

C):

3.9%

● I

nsur

ance

com

pany

(AX

A):

3.5%

Schn

eide

r E

lect

ric

(2)

2006

Sept

em-

ber

Tac

it re

new

al

of th

e co

ntra

ct

ever

y ye

ar

Bet

wee

n Sc

hnei

der

and

Axa

Low

● 3

ow

ners

with

mor

e th

an 3

% e

quity

(t

otal

: 12.

7%)

● A

uto-

cont

rol a

nd

empl

oyee

s (9.

9%)

● I

nstit

utio

nal i

nves

tors

(cap

ital g

roup

): 5%

● I

nstit

utio

nal i

nves

tors

(CD

C):

4.4%

● I

nsur

ance

com

pany

(AX

A):

3.5%

Not

es:

1 O

wne

rshi

p co

ncen

trat

ion

at t

he t

ime

of t

he c

ontr

act.

Cri

teri

a: I

nter

med

iate

con

cent

rati

on: s

ome

larg

e ow

ners

(>

3%

equ

ity)

, yet

no

maj

orit

y ow

ner;

Hig

h co

ncen

trat

ion:

exi

sten

ce o

f a

maj

orit

y ow

ner;

Low

con

cent

rati

on: s

ome

larg

e ow

ners

but

the

sum

of

larg

e ow

ners

is lo

wer

tha

n 10

%2

16.7

% o

f vo

ting

rig

hts.

3 E

lisab

eth

Bad

inte

r, da

ught

er o

f th

e fo

unde

r, M

arce

l Ble

uste

in-B

lanc

het.

4

Equ

ally

ow

ned

by W

ende

l and

KK

R.


Proposition 2 Shareholder agreements are more likely to be found in companies in which shareholders have a long-term interest (for example, families, corporations).

The heterogeneity of goals of shareholders is well established in the management literature (for example, Bushee, 1998 and 2001; Verstegen Ryan and Schneider, 2002, 2003, for institutional investors; Thomsen and Pedersen, 2000, for a large range of owner identities). Prior research suggests that owners have diff erent objectives and strategic preferences according to their identities (Bethel and Liebeskind, 1993; Hoskisson et al., 2002: Tihanyi et al., 2003; Gaspar et al., 2005). Our data indicate that shareholder agreements are more likely to be signed between owners who have non-fi nancial objectives. By non-fi nancial objectives, we mean objec-tives that are not strictly linked to the value maximization of the share-holders’ equity. They span personal and strategic objectives. For example founders’ families such as Pernod or Badinter (the founders’ children) may want to keep the culture and values of the founders alive in the fi rm. Corporations such as Kirin and Dentsu appear to have strategic reasons to surrender their external control rights.

Kirin is the largest spirits company in Japan with a long tradition of joint ventures to access markets and technology. In the 1970s, it signed a joint venture with Seagram to distribute its whisky brands in Japan. Shortly after Seagram’s spirits division was sold to Pernod Ricard and Diageo in 2002, Kirin signed an agreement with the two companies to retain and acquire sales rights in Japan for the former Seagram portfolio’s brands. Thus Kirin’s equity share (3 percent) in Pernod Ricard seems to serve other purposes than fi nancial returns, such as expanding in the non-beer busi-ness. Figuratively speaking, Pernod Ricard became part of the ‘keiretsu’.

Similarly, it is hard to understand why Dentsu would assent to sign an agreement that strongly limits its controlling power over Publicis (when it has formally 18 percent of the equity) for purely fi nancial reasons.4 Secondary data suggest that at the time of the agreement (2002) Dentsu was facing growth imperatives – just after its introduction on the stock market – and was struggling with a downward advertisement market in Japan. Two years before, Dentsu had made some diversifi cation invest-ments abroad, particularly in Bcom 3, a large US agency network. In 2002, it sold its dominant share in Bcom 3 to Publicis – offi cially to free up some cash – and signed what was described by the top management of the fi rms as a ‘strategic alliance’ or ‘business tie up’ with Publicis.5 Dentsu was prob-ably agreeing to trade off control over Publicis for a business alliance that would possibly off set the slow growth of its Japanese activity and enhance its global off er of communication services.


A slightly diff erent logic applies to the Club Med agreement. Why would Accor set up a contract with several new institutional investors at the time when it exited from Club Med? Secondary data, notably press articles, suggest that Accor may have been pressured by the French government to fi nd a ‘substitute’ owner that would replace Accor (as the reference owner) and thus protect Club Med from a takeover by foreign investors. Club Med is commonly considered one of the ‘national corporate jewels’ of France, one that needs to be protected against hostile bids.

Another important non-fi nancial objective – which we have not seen documented in previous research – is what we could call ‘preventive owner-ship’ or ‘preventive control’, which consists in preventing third parties from acquiring control. To be sure, preventive control is not equal to protection against takeovers. It is a way for corporate owners to prevent their direct competitors from expanding their footprint. For example, in the Club Med case, it is likely that Accor did not want a foreign competitor like Hilton to establish a strong position in the leisure business by taking over Club Med. Similarly, in the Publicis case, it was probably important for Dentsu to prevent Publicis from being acquired by competitors like WPP or Grey. In this case, control over management is less important than preemption over competitors. Hence, we make the following proposition:

Proposition 3 Shareholder agreements are more likely to be found in companies in which owners have non-fi nancial objectives (for example, strategic alliance interests for corporate owners, continuity for family ownership, ‘national protectionism’ for government owners, preemption – in contrast to investors who prefer to avoid the loss of fl exibility).

Our cases also suggest that the leading shareholder of the agreement, that is, the one initiating the agreement and managing the negotiation process, has a strong historical link with the fi rm and seeks to maintain this link. For example, in Publicis and Pernod Ricard, the leading negotiating shareholders are the founders’ families. We argue that they tried to main-tain their historical control over the fi rm. As both fi rms grew over the years, they needed additional fi nancial back-up and went public, with a strong share of the fi rm remaining in the founders’ hands (in 2000, the Badinter/Bleustein-Blanchet family still had 40 percent of the fi rm’s equity; in 2001, 28 percent). Both shareholder agreements followed just after major exter-nal acquisitions (BCom 3 for Publicis in 2002; Allied Domecq for Pernod Ricard in 2006) suggesting that the founders’ families used the contracts to protect themselves against a loss of power in the new entity.

The Legrand contract also indicates that the incumbent shareholders (the two private funds) sought to maintain their control of Legrand after


the fi rm was introduced on the stock market. Initially, in 2002, at the time of the LBO, KKR and Wendel had signed a very detailed agreement on Legrand. Once they fl oated part of the equity, they were required to reveal their agreement, and thus turned to a simplifi ed version. One of our inter-viewees told us very strongly that ‘[they] use shareholder agreements to maintain [their] power in the fi rm’.

These cases suggest that when incumbent owners seek to maintain their power in the fi rm, the costs of contracting may be off set by the benefi ts of control. Thus:

Proposition 4 Shareholder agreements are more likely to be found in companies in which an incumbent shareholder (for example, family, private equity) seeks to maintain dominant control.

4.2 Nature of the Industry

Prior studies show that fi rms’ ownership structures vary across industries (Demsetz and Lehn, 1985; Pedersen and Thomsen, 1997; Villalonga, 2005). For example, family ownership is more prevalent in industries such as food manufacturing and media while government ownership – to take extreme examples – is more frequent in the weapon and aircraft industries. Similarly we would expect that shareholder agreements are not equally dis-tributed across all types of industries. There are two main reasons for this. First, as previously shown, there are ownership conditions (related to own-ership concentration and identity of owners) under which the benefi ts of shareholder agreements exceed the costs, and these conditions may change across industries. Second, shareholder agreements generate costs of ‘lock-in’ because they tie up the signing parties for several years without much fl exibility. Lock-in is less costly in relatively stable and certain industries than in dynamic industries where short-term changes may be required. Our cases show that all of the fi rms in our sample operated in mature industries (wine and spirits, electrical equipment, advertising). Findings are reported in Table 11.2. We formally state this proposition as follows:

Proposition 5 Shareholder agreements are more likely to be found in companies that operate in relatively stable businesses in which the costs of lock-in for a couple of years are small.

This fi nding departs from previous research on interfi rm ownership and strategic alliances which suggests that equity arrangements among fi rms are more frequent in knowledge intensive industries with high R&D intensity (Allen and Phillips, 2000) such as biotechnologies and electronics.


Here, the main role of interfi rm shareholdings is to facilitate knowledge fl ows between fi rms (Mowery et al., 1996, Gomes-Casseres et al., 2006). In contrast, in our cases, knowledge sharing does not appear to be a dominant motive in transactions between shareholders.

4.3 Nature of the Contract Items

The benefi ts of shareholder contracts will be particularly high (and exceed the costs) when no other mechanism eff ectively addresses the issues included in the contract. One of our interviewees told us:

We do not use shareholder agreements on a standalone basis but in conjunction with other elements such as charters and commercial law. We see shareholder agreements as one of the tools of a larger toolbox which help protect our inter-ests in the fi rm. Yet, these contracts are useful because we have a lot of fl exibil-ity in the content, and we may address issues that are not explicitly taken into account by charters or by commercial law.

What are these issues? Our cases point to diff erent categories, related to equity rights, control rights and non-control and equity issues. Table 11.3 provides a summary of the issues while the details of the contracts are in Appendix 11.2. Three categories of issues are distinguished: equity rights, control rights, and other (non-equity and control) issues.

Equity rights relate to the control of the equity by the signing sharehold-ers. They encompass preemption clauses, tag-along and drag-along rights (respectively the right of joint exit and the obligation of joint exit) and rights of approval (clauses d’agrément) which allow the current shareholders to avoid ‘undesirable’ new shareholders entering the fi rm. As an illustration, the Publicis (2002) shareholder agreement states that Dentsu will not be able to transfer or sell its equity shares in Publicis until July 2012; after July 2012, Mrs Badinter has a preemptive right to buy Dentsu’s shares.

Table 11.2 Shareholder agreements and the nature of industry

Firm Industry Industry development

Pernod Ricard Wine and spirits MaturePublicis Communication and

advertisingMature

Club Med Leisure and holiday village operator

Mature

Legrand Electrical equipment MatureSchneider Electric Electrical distribution Mature


Control rights pertain to the allocation of board seats, the nomination process of the directors and the committees, the allocation of voting rights and the obligation to ‘act in concert’ with respect to strategic decisions. Legrand provides a good illustration. The shareholder contract specifi es that the board will comprise 11 members: three representatives of KKR, three representative of Wendel, two independent board members and three top managers. It also stipulates that the strategic committee will be

Table 11.3 Shareholder agreements and the nature of the contract items

Firm Control rights Equity rights Non-equity and control-related issues

Board structure

Votingrights

Rights to sell and acquire

shares

Cash fl ow rights

Strategic alliances/

business deals

Pernod Ricard

X X X

Publicis X X X XClub Med X X X XLegrand X X X XSchneider X

Notes:Board structure refers to the rules attached to the composition of the board of directors and its sub-committees. Shareholders’ agreements encompass the following board-related items: number of board seats granted to the signing shareholders, total number of members on the board, including total number of independent board members; nomination process for board members and chairman; rights to propose nominees and obligation to accept the appointments made by some of the signing parties. Voting rights refer to (1) the rights given to the contract parties relative to the strategic decisions and (2) the total voting rights granted to each of the parties. In the fi rst case, the shareholders’ agreement will specify the type of strategic decisions for which the signing parties agree to vote in concert and those for which they may express an individual opinion. In the second case, it will specify the maximum number of votes given to each of the contract parties. Rights to sell and acquire shares refer to the rules attached to the sale and purchase of equity shares. Shareholders’ agreements specify the minimum holding period of shares, the preemptive rights given to the signing shareholders, and the process that needs to be followed in case of sell-out. They also defi ne the rights and obligations in case of takeover attempts. Includes clauses for joint exit, rights of approval, tag-along and drag-along rights. Cash fl ow rights refer to the allocation of cash fl ows between the signing parties, in particular when the fi rm is sold to a third party or goes public. Strategic alliances refer to agreements between the signing parties about some joint activity, which may range from common distribution channels to joint production and knowledge transfer. Business deals refer to more ad-hoc transactions between the fi rm and one of the signing investors.


chaired by a KKR representative while the compensation committee will be chaired by a Wendel representative. As for Club Med, the contract explicitly states that all signing shareholders will support the strategy of the current management.

Non-equity and control issues relate to business relationships either among the shareholders or between the shareholders and the fi rm. For example, in the Club Med contract, Fipar, a real estate institutional inves-tor and one of the signing parties, is granted the right to strike a deal with Club Med for its real estate assets.

Our fi ndings suggest that the equity and control rights issues included in the shareholder agreements cannot easily be addressed by standard open market transactions like hedging or fi nancial options, partly because they are conditional on what other parties contract do, and partly because of the relatively long time horizon which they cover. Thus the following proposition:

Proposition 6 Shareholder agreements are more likely to be found in companies in which shareholders need to address complex and conditional issues characterized by an intermediate level of information asymmetry (which can benefi t from third-party arbitration in case of disagreement).

Previous research suggests that market mechanisms do not fully address expropriation issues between shareholders (Shleifer and Vishny, 1997; Johnson et al., 2000; Djankov et al., 2008) such as tunneling and self-dealing. Two types of expropriation issues (also called principal–principal issues) need to be distinguished: expropriation of minority shareholders by large shareholders, and expropriation among the large shareholders.

La Porta et al. (1999) indicate that legal protection of minority inves-tors is scarce in countries with French civil law origin. Although French law based investor protection has been more favorably regarded in recent research (Djankov et al., 2008), we fi nd it interesting to examine whether French shareholder agreements are substitutes for limitations in legal investor protection, using mechanisms such as the granting of board seats and additional voting power to minority investors. But our cases do not provide evidence of this. The rights of the minority investors signing the contracts tend to be linked to the interests of the dominant owner, with no rebalancing taking place. For example, Kirin, a minority investor, clearly agrees to be on the backseat, leaving the Ricard family with the full deci-sion and control power.

However, our cases point to the relevance of shareholder agreements to mitigate expropriation among large shareholders. As an illustration, the shareholder contract between KKR and Wendel (which have equal


stakes in Legrand) stipulates how the two owners will share the control and profi ts of the fi rm after the IPO. It forbids one of the parties to pursue an opportunistic behavior at the expense of the other. A revealing quote from our interviewee:

When we sign a shareholder agreement, we act as complete paranoiacs. We investigate all possible scenarios under which we could lose control and money because of the other party’s opportunism. We make sure all the scenarios are included in the contract, with clear resolution processes. Never trust another shareholder!

Under these circumstances, shareholder agreements may represent an effi -cient instrument to moderate potential confl ict of interests between large shareholders. Hence, we propose:

Proposition 7 Shareholder agreements are more likely to be initiated by large investors seeking to protect their bargaining power than by small shareholders seeking substitute mechanisms for weak legal protection.

Prior studies have shown that fi rms are more likely to adopt anti-takeover measures when managers have high discretion and low owner control (Brickley et al., 1988). We found that shareholder agreements are commonly used by large insiders to protect themselves against the entry of ‘undesirable’ shareholders. Schneider Electric and Axa for example have cross-ownership. In the event of a hostile takeover of Schneider, AXA has the right to purchase all AXA shares still owned by Schneider; and con-versely for Schneider. As for the Club Med shareholder agreement, it stipu-lates that, in case of takeover, investors are allowed to sell their shares only if Club Med’s board of directors has given its agreement on the takeover. Another good example of the use of agreement against takeover is Pernod Ricard. The shareholder agreement with Kirin reinforced the family’s equity share by increasing its equity block; yet it did not seem safe enough to secure family control. The following year (2007), a new blockholder, Albert Frère, who is reported to be a ‘40-year-old friend of the family’, entered the fi rm. Albert Frère had been on Pernod Ricard’s board of directors from 1991 to 1995. In addition, in line with the new fi nance law ‘Breton’, Pernod Ricard adopted a measure whereby convertible bonds can be given for free to existing shareholders (maximum 50 percent capital) in case of a hostile takeover, consequently increasing the cost of the takeover. Thus:

Proposition 8 Shareholder agreements are more likely to be found in companies in which there is a takeover risk (for example, companies with large free cash fl ows).


4.4 Nature of the Network Ties

A stream of research has examined the ties between directors and top man-agers and has revealed that the nominations of directors are not random but linked to social and professional networks (Davis et al., 2003; Conyon and Muldoo, 2006; Kirchmaier and Kollo, 2007). In France in particular, there are strong ‘small world’ eff ects mostly related to top management’s membership of the ‘elite’ schools (Nguyen-Dang, 2006). In such a context, it is likely that the choice of partners for shareholder agreements follows some network rules. Our cases show that this is often so.

For example, in the Club Med agreement, all signing shareholders have somewhat tight connections. Accor had close links with the govern-ment, both through its founders (notably Mr Pelisson, who also became the mayor of Fontainebleau, a ‘posh’ city on the outskirts of Paris), and through Mr Espalioux, the CEO, who studied at ENA, the top administra-tive school in France. Accor sold part of its equity in Club Med to a consor-tium of investors who were related to the French government. Icade is the real estate arm of CDC, a ‘hybrid’ institutional investor strongly connected with the government.6 Fipar Holding is the CDC equivalent for Morocco. Air France Finance is the fi nance arm of Air France, the former national French airline company, still partly owned by the French government and with a strong connection to Accor (for example, they have a common payment card and share part of their loyalty programs). While ex ante the network ties probably facilitated the signature of the contract, ex post they also increased the enforcement of the contract because of potential social sanctions in case the contract is breached.

A close look at the Schneider Electric agreement also reveals many informal relationships among the signing shareholders (Axa, Schneider Electric, BNP-Paribas, and AGF for the 2002 agreement), in particular, board ties. Over the period 2000–07, the board of AXA (one of France’s top insurance companies) included Michel Pebereau (CEO of Bank BNP-Paribas, one of France’s top banks) and Henri Lachman (CEO and subsequently chairman of Schneider Electric) as members. Until 2002, Claude Bebear (CEO of AXA) was a direc-tor on the board of Schneider Electric. Directors and managers have mutual board ties which possibly give them more scope for ‘gentlemen’s agreements’ not to interfere in the strategic decisions of socially connected parties for fear of retaliation. The Schneider Electric agreement seems to refl ect a social arrangement between managers of top French corporations, banks and insur-ance companies. From these various cases, we propose the following:

Proposition 9 Shareholder agreements are more likely to be found in companies whose leading offi cers and directors have social ties, such as


belonging to the ‘elite’ network (where formal contacts can be backed up by social sanctions).

In the case of Kirin–Ricard and Dentsu–Publicis it is relatively clear that the foreign fi rms did not have the same strong social ties. But recent research has indicated that alliances may be a means to achieve social ties and legitimacy (Koka and Prescott, 2002; Dacin et al., 2007). Since minor ownership shares have been used to express a commitment to future business relations among Japanese fi rms (in the so-called keiretsu system), the contractual relations with the French fi rms could be seen as an international extension of a traditional Japanese practice. In addition, a closer look at the Dentsu–Publicis deal reveals the existence of signifi cant informal ties – although not at the CEO level – between the two companies. Secondary sources indicate that the founder of Publicis (Marcel Bleustein-Blanchet) and the founder of Dentsu (Hideo Yoshida) knew each other from the 1960s. The CEO of Publicis, Maurice Levy, made a revealing comment when announcing the arrangement with Dentsu: ‘Friendly ties were established in the sixties between Marcel Bleustein-Blanchet, our founder, and Mr Hideo Yoshida. I am glad that this partnership off ers us new opportunities to build on this tradition.’

5. IMPACT OF SHAREHOLDER AGREEMENTS

Shareholder agreements bias the formal ownership structure by intro-ducing idiosyncratic arrangements between selected shareholders. This phenomenon of ‘contracting around ownership’ may constitute an impedi-ment to the effi ciency of the markets because it introduces some complexity in the allocation of control and cash fl ows among shareholders. Yet, the confl icting perspectives of the fi nance and strategic management streams of literature suggest that we still have no clear view on the eff ects of share-holder agreements.

On the one hand, the fi nance literature indicates that protections against takeovers such as poison pills (included in company charters) destroy value (Gompers et al., 2003). Takeovers and particularly hostile takeovers are demonstrated to be an important source of value creation for target fi rm shareholders (for example, Schwert, 2000). Accordingly, we would expect a similar negative eff ect to apply to shareholder agree-ments, given that they restrict hostile takeovers and deter entry of new shareholders.

On the other hand, concentrated control is believed to create value under some circumstances (Thomsen and Pedersen, 2000) and the


strategic management literature suggests that equity arrangements between shareholders may be positive for the fi rm. They can promote strategic alliances – complementary or additive – between fi rms (Allen and Phillips, 2000), ultimately creating fi rm value. In addition, long-term shareholders, such as banks, corporations and government, may be in a better position to expand the fi rm’s resource base and access critical resources for the fi rm (for example, bank loans, lobbying of governmen-tal bodies, R&D funding) (Pfeff er and Salancik, 1978). Accordingly, we would expect some positive reaction of the markets over shareholder agreements.

Our cases suggest that the impact of shareholder agreements is highly contingent upon their content. Findings are reported in Table 11.4. Market reactions to the announcement of the Club Med and Schneider Electric (2006) deals were clearly negative, while they appear to be positive for the Pernod Ricard and Legrand agreements.

As explained in the previous section, the Club Med and Schneider Electric contracts function as defense mechanisms against takeovers (see also details of contracts in Appendix 11.2). The fi rst one points to the

Table 11.4 Impact of shareholder agreements

Firm Announcement date of the shareholder agreement

+/−1 month +/−2 months

Change in company

stock price (%)

Change in the CAC 40 index (%)

Change in company

stock price (%)

Change in the CAC 40 index (%)

Pernod Ricard

27 March 2006

+4.9 +1.9 +3.9 +2.4

Publicis 24 May 2002 −9 −7.7 −22 −15Club Med 21 June 2006 −12.7 +0.7 −16.6 −1.2Legrand 27 April 2006 4.8 −1.1 N.A. (IPO

occurring on 7 April 2006)

Schneider Electric

15 March 2002 (renewal)

−3.2 −2.8 −2.6 −2.9

Schneider Electric

19 May 2006 −11.6 −6.8 −10.9 −6.4

Notes:Variations of the fi rm’s stock price +/−1 month or +/−2 months around the announcement date of the agreement.The CAC 40 index refl ects the stock price variation of the 40 largest companies in France.The source for the stock price data is Datastream.


creation of a coalition that replaces the ‘reference’ owner and protects Club Med against takeovers. The later one (2006 version) is largely focused on protection against takeovers, with mutual preemption rights between Axa and Schneider in case of hostile bids.

In comparison, Pernod Ricard and Publicis’ agreements combine both strategic and fi nancial components. Kirin seems to have traded its investor power for some potential business agreement with Pernod Ricard, thus expanding out of its traditional (slow growth) Japanese beer business. Dentsu went for a global alliance with Publicis following the sell-out of Bcom 3. It thus gave up its role as an external shareholder for a larger fi nancial and strategic agreement. As offi cially announced by Dentsu:

Dentsu Inc. has reached today a basic agreement to form a strategic global alliance with a new company created through the merger of Bcom3, a U.S. based communications group and Publicis Groupe . . . In addition, Dentsu and Publicis will discuss working together on specifi c projects on a global basis. (Dentsu’s website)

We may argue that the Publicis agreement was not positively greeted because the positive perspective of a strategic alliance was probably already included in the announcement of the deal between Publicis, Dentsu and Bcom 3.

Finally, the Legrand agreement seems to follow a diff erent logic. KKR and Wendel were the company’s two main shareholders at the time of the IPO (initial public off ering). The renewed commitment in the fi rm and the willingness to ‘bind their fate’ over a somewhat longer period of time can be a reassuring signal for the markets, which may be especially important in IPOs. Private equity funds are known to be focused on value maximi-zation and to exert strong monitoring over the fi rm management. Thus their agreement communicates to the fi nancial markets that agency issues related to adverse selection will be minimal. In addition, long-term inves-tors and particularly those with a strong fi nancial objective in mind bring additional value through some continuity in strategies.

Overall, our cases indicate that the markets are ‘smart’ enough to distin-guish between various types of contracts and to value those that promise a positive long-term eff ect on the fi rm. Thus, we propose the following:

Proposition 10 Shareholder agreements are more likely to be negatively perceived when they are primarily defense mechanisms against takeovers.

Proposition 11 Shareholder agreements are more likely to be positively perceived when they off er additional strategic benefi ts.


Proposition 12 Shareholder agreements are more likely to be positively perceived when they signal an increased commitment to value creation objectives.

We propose that these hypotheses are empirically testable in event studies which distinguish between alternative types of shareholder agree-ments based on industry and fi rm characteristics. For example, companies characterized by substantial free cash fl ow, low debt and low valuation ratios and companies in newly deregulated industries with restructuring potential are more likely to be takeover targets. In such fi rms we expect news of shareholder agreements to be associated with negative abnormal returns. In contrast we expect that shareholder agreements with a business case – which extend the technological or sales capacity of the companies involved – will tend to generate positive abnormal returns. Moreover, shareholder agreements among owners with clear value-maximizing objec-tives – such as private equity funds – are also more likely to be associated with positive abnormal returns.

6. DISCUSSION AND CONCLUSION

This chapter explores the determinants and consequences of shareholder agreements among large listed fi rms in France. Shareholder agreements are contractual instruments that organize the relationships between sharehold-ers ‘backstage’, that is, behind the formal ownership structure. Because of their confi dentiality, they are often hard to study. Yet, without access to these ‘backstage’ agreements, it is diffi cult to understand what is going on in the fi rm. The formal ownership structure may provide an inadequate picture of the allocation of power between shareholders and between shareholders and managers.

Using a sample of shareholder agreements among listed fi rms in France, we look at the cost/benefi t trade-off of these contracts versus other governance instruments such as increased ownership and market discipline. We defi ne the costs of shareholder agreements as the transac-tion costs (negotiating, renegotiating and enforcing the contracts) and the costs of lock-in. The benefi ts of shareholder agreements vary according to the issues faced by the shareholders. We propose that shareholder contracts are particularly eff ective instruments for fi rms with specifi c ownership patterns (intermediate concentration of ownership, long time horizons and non-fi nancial goals) and in certain (mature) industries. In addition, we conjecture that shareholder agreements are more likely to be considered when there are expropriation risks between large shareholders


(principal–principal issues) and takeover risks. Finally, we propose that shareholder agreements will be more eff ective and frequent among members of the same ‘elite’ network, as the enforcement of the contract can be backed by social sanctions.

With respect to the economic performance, we propose that the value of shareholder agreements is highly contingent upon the content of the contract. Agreements emphasizing protection against hostile takeovers will tend to destroy value. In contrast, agreements including both strategic and fi nancial components and those signaling long-term commitment by value-maximizing owners will tend to increase shareholder value. We conjecture that shareholder agreements are likely to create value when stability of ownership is required to give the fi rm a secure strategic direction. Despite remarkable changes in stock ownership – the average holding period for common stock has fallen signifi cantly over the past decade – we have seen very little research on the importance of the stability of ownership to the continuity of strategy and fi nancial results. We would argue that stability of ownership may under certain circumstances provide the right background for fi rms to progress, and that shareholder contracts can be an instrument in this regard. Moreover, we fi nd reasons to believe that fi rms can occasionally benefi t from having the shareholders assign ownership (control) rights to a competent owner who can then exercise authority to the benefi t of the share-holders as a group. Shareholder contracts provide a fl exible framework for this, because they have limited duration and need to be renewed at regular intervals.

We regard this chapter as a fi rst step towards improving our understand-ing of the relatively unexplored domain of shareholder agreements in listed fi rms. Obviously, there are limitations to this study that need to be made explicit.

First, we chose to focus on a very limited number of ‘diff erent’ cases. Expanding our sample would improve the generalizability of the fi ndings.

Second, because of the availability of the agreements, we focus on the French institutional context. Research looking at cross-country compari-sons of ownership structures and corporate governance practices under-lines the importance of the institutional context and more particularly of company law (La Porta et al., 1999; Gedajlovick and Shapiro, 1998; Pedersen and Thomsen, 1997). Thus, we should test our propositions in other contexts, for example in both countries with similar French civil law origin and in countries with common law origin. The protection of minor-ity investors, which is held paramount in common law countries, may limit the scope for agreements between blockholders. For example, Roe (1991) demonstrates how US law has historically blocked cooperation between minority investors because of a fear of insider trading and cornering the


market. Countries with other legal traditions – like France – may be more amenable to contracts between shareholders.

Nevertheless, we think this chapter contributes to the current debate on ownership from both the fi nancial contracting literature and the govern-ance literature perspectives.

The emerging theory of fi nancial contracting has understandably been preoccupied with fundamental issues like diff erences in generic types of debt and equity and what these imply for fi rm behavior and perform-ance. But the fact is that the real-world fi nancial contracts are often more complicated because they combine generic liability types with specifi c contractual provisions which infl uence the allocation of decision, control and cash fl ow rights. This chapter demonstrates that shareholder contracts tailor-make the relationships between shareholders. Thus we highlight the need for a systematic analysis of these contracts for fear that an important dimension of the fi nancial structure should be missed. In addition, we provide indications of the conditions under which shareholder agreements are more likely to be effi ciently used.

The governance stream of literature has been relatively silent on share-holder agreements. In particular, cross-country studies have often over-looked this dimension when studying ownership structures in French civil law origin countries (for example, Gedajlovic and Shapiro, 1998; La Porta et al., 1999; Faccio and Lang, 2002). This chapter points to the importance of an in-depth understanding of less visible contextual variables such as shareholder agreements when studying governance systems as well as individual fi rms. As it turns out, theories emphasizing relational capitalism seem quite right; some of the relationships can be documented by studying interfi rm contracts.

We see several possible avenues for future research. First, one could test the propositions on a larger sample of fi rms within France (circa 300 detailed contracts available). Another analysis would be to measure the rel-ative infl uence of the formal and real controls on fi rm outcomes. Thirdly, it would be interesting to test the propositions in other institutional contexts. As explained earlier, we expect shareholder agreements to be particularly frequent in countries characterized by intermediate degrees of ownership concentration such as France, Germany and Belgium, but less common in countries like the USA and the UK (fi rms with low ownership concentra-tion) or Italy (fi rms with high ownership concentration).

Finally, these fi ndings have implications for corporate governance guidelines. On the premise that all materially relevant information (that is, with the potential to infl uence stock prices) should be disclosed quickly in a well-functioning stock market, shareholder agreements should be made public information, since they can infl uence corporate strategy and stock


prices. In this respect, we believe that the French approach (a public reg-ister) is a good solution.

NOTES

1. This chapter originated as a paper prepared for the Workshop on ‘The Economics of the Modern Firm’, Jönköping, Sweden , 21–22 September 2007. It has benefi ted from com-ments by Benito Arunada and an anonymous referee.

2. In 2005. Source: AMF. Note that some companies have been delisted while others have gone public over the period 1997–2007.

3. These two disclosure requirements tie in with a third requirement on ownership thresh-olds (‘declaration de franchissement de seuils et declarations d’intentions’ – article I-233-7 of commercial law) which asks shareholders to notify the market authority when their shares in a fi rm go above the 5, 10 and 15 percent voting or equity thresholds.

4. The contract between the Badinter family and Dentsu states that Dentsu will commit to nominate ‘all management team members proposed by E. Badinter’ as well as ‘all supervisory board members who have been chosen by E. Badinter’. In addition, it will ‘vote in favor of E. Badinter’s decisions in the cases of change of Publicis’ charters, M&A, distribution of dividends, capital off erings and share repurchase’.

5. ‘Dentsu Inc. announced today that it has reached an agreement to form a strategic global alliance with a new company created through the merger of Bcom3 Group, Inc., a U.S. based communications group and Publicis Groupe S.A., a major European com-munication group headquarted in France’ (published in Dentsu’s website – Investors’ relations – 7 March 2002. Emphasis added). Yukata Narita, Dentsu’s president, further commented: ‘We aim to provide the very best marketing communication services cover-ing every domain in the world market by integrating the power of the three corporations. By doing so, we believe we can win the confi dence of clients and establish the very best global network’ (emphasis added).

6. CDC invests funds collected through the ‘Livret A’ (typical non-risk placement for French households) and acts as a long-term owner in a large number of French fi rms. The CDC director is nominated by the French president.

REFERENCES

Adams R. and D. Ferreira (2007), ‘One share, one vote: the empirical evi-dence’, ECGI Working Paper Series in Finance, Finance Working Paper No. 177/2007.

Aghion, P. and P. Bolton (1992), ‘An incomplete contracts approach to fi nancial contracting’, Review of Economic Studies, 59 (3), 473–94.

Allen, J. and G. Phillips (2000), ‘Corporate equity ownership, strategic alliances and product market relationships’, The Journal of Finance, 55 (6), 2791–816.

Amihud, Y. and B. Lev (1981), ‘Risk reduction as a managerial motive for con-glomerate mergers’, Bell Journal of Economics, 12, 605–17.

Anderson, R. and D. Reeb (2003), ‘Founding-family ownership and fi rm perform-ance: evidence from the S & P 500’, Journal of Finance, 58 (3), 1301–28.

Baysinger, B., R. Kosnik and T. Turk (1991), ‘Eff ects of board and ownership structure on corporate R&D strategy’, Academy of Management Journal, 34 (1), 205–14.


Bethel, J. and J. Liebeskind (1993), ‘The eff ects of ownership structure on corpo-rate restructuring’, Strategic Management Journal, Summer Special Issue, 14 (2), 237–53.

Brickley, J., J. Lease and C. Smith (1988), ‘Ownership structure and voting on anti-takeover amendments’, Journal of Financial Economics, 20, 267–91.

Burkart, M. and S. Lee (2007), ‘The one share–one vote debate: a theoretical per-spective’, ECGI Working Paper Series in Finance, Finance Working Paper No. 176/2007.

Bushee, B. (1998), ‘The infl uence of institutional investors on myopic R and D investment Behavior’, The Accounting Review, 73 (3), 305–33.

Bushee, B. (2001), ‘So institutional investors prefer near-term earnings over long run value?’, Contemporary Accounting Research, 18 (2), 207–46.

Carson, S., A. Madhok and T. Wu (2006), ‘Uncertainty, opportunism, and govern-ance: the eff ects of volatility and ambiguity on formal and relational contract-ing’, Academy of Management Journal, 49 (5), 1058–77.

Chemla, G., M. Habib and A. Ljungqvist (2007), ‘An Analysis of shareholder agreements’, Journal of the European Economic Association, 5 (1), 93–121.

Conyon, M. and M. Muldoo (2006), ‘The small world of corporate boards’, Journal of Business Finance and Accounting, 33 (9–10), 1321–43. Available at SSRN.

Dacin, T., C. Oliver and J.-P. Roy (2007), ‘The legitimacy of strategic alliances: An Institutional Perspective’, Strategic Management Journal, 28 (2), 169–87.

Danielson, M. and J. Karpoff (1998), ‘On the uses of corporate governance provi-sions’, Journal of Corporate Finance, Contracting, Governance and Organization, 4, 347–71.

David, P., M. Hitt and J. Gimeno (2001), ‘The role of institutional investors in infl uencing R&D’, Academy of Management Journal, 44, 144–57.

Davis, G., M. Yoo and W. Baker (2003), ‘The small world of the American corpo-rate elite, 1982–2001’, Strategic Organization, 1 (3), 301–26.

De Miguel, A., J. Pindado and C. De la Torre (2004), ‘Ownership structure and fi rm value: new Evidence from Spain’, Strategic Management Journal, 25, 1199–207.

Demsetz, H. and K. Lehn (1985), ‘The structure of corporate ownership: causes and consequences’, Journal of Political Economy, 93 (6), 1155–77.

Denis, D.J., D.K. Denis and S. Atulya (1997), ‘Agency problems, equity ownership and corporate Diversifi cation’, Journal of Finance, 52 (1), 135–160.

Denis, D.J., D.K. Denis and S. Atulya (1999), ‘Agency theory and the infl uence of equity ownership structure on corporate diversifi cation’, Strategic Management Journal, 20 (11), 1071–6.

Desai, M.A., C.F. Foley and J.R. Hines, Jr. (2004), ‘Economic eff ects of regional tax havens’, NBER Working Paper No. W10806, available at http://srn.com/abstract5601108.

Djankov, S., R. La Porta, F. Lopez de Silanes and A. Shleifer (2008), ‘The law and economics of self-dealing’, Journal of Financial Economics, 88 (3), 430–65.

Doz, Y. and G. Hamel (1998), Alliance Advantage, Boston, MA: Harvard Business School Press.

Eisenhardt, K.M. and M.E. Graebner (2007), ‘Theory building from cases: oppor-tunities and challenges’, Academy of Management Journal, 50 (1), 25–32.

Ellickson, R. (1991), Order without Law, Cambridge, MA: Harvard University Press.

Faccio, M. and L. Lang (2002), ‘The ultimate ownership of Western European corporations’, Journal of Financial Economics, 65, 365–95.


Gaspar, J.M., M. Massa and P. Matos (2005), ‘Shareholder investment horizons and the market for corporate control’, Journal of Financial Economics, 76 (1), 135–65.

Gedajlovic, E. and D. Shapiro (1998), ‘Management and ownership eff ects: evi-dence from 5 countries’, Strategic Management Journal, 19 (6), 533–55.

Gedajlovic, E. and D. Shapiro (2002), ‘Ownership structure and fi rm profi tability in Japan’, Academy of Management Journal, 45 (3), 565–75.

Glaser, B. and A. Strauss (1967), The Discovery of Grounded Theory: Strategies for Qualitative Research, Chicago: Aldine.

Gomes-Casseres, B., J. Hagedoorn and A. Jaff e (2006), ‘Do alliances promote knowledge fl ows?’, Journal of Financial Economics, 80 (1), 5–33.

Gompers, P., J. Ishii and A. Metrick (2003), ‘Corporate governance and equity prices’, Quarterly Journal of Economics, 118 (1), 107–55.

Graves, S. (1988), ‘Institutional ownership and corporate R&D in the computer industry’, Academy of Management Journal, 31 (2), 417–28.

Hart, O. (2001), ‘Financial contracting’, Journal of Economic Literature, 39 (4), 1079.

Hill, C. and S. Snell (1988), ‘External control, corporate strategy and fi rm per-formance in research-intensive industries’, Strategic Management Journal, 9, 577–90.

Hill, C. and S. Snell (1989), ‘Eff ects of ownership structure and control on corpo-rate productivity’, Academy of Management Journal, 32 (1), 25–46.

Holderness, C. and D. Sheehan (1988), ‘The role of majority shareholders in publicly-held corporations’, Journal of Financial Economics, 20 (1), 317–46.

Hoskisson, R., M. Hitt, R. Johnson and W. Grossman (2002), ‘Confl icting voices: the eff ects of institutional ownership heterogeneity and internal governance on corporate innovation strategies’, Academy of Management Journal, 45 (4), 697–716.

Johnson, S., R., La Porta, F. Lopez-de-Silanes and A. Shleifer (2000), ‘Tunneling’, American Economic Review, 90 (2), 22–7.

Kaplan, S. and P. Strömberg (2003), ‘Financial contracting theory meets the real world: an empirical analysis of venture capital contracts’, Review of Economic Studies, 70 (243), 281–315.

Kirchmaier, T. and M. Kollo (2007), ‘Who wants sirs and lords on their boards? The importance of prestige and social networks on UK boards’, paper presented at EURAM, Paris.

Kogut, B. (1988), ‘Joint ventures: theoretical and empirical perspectives’, Strategic Management Journal, 9, 319–32.

Koka, B. and J. Prescott (2002), ‘Strategic alliances as social capital: a multidimen-sional view’, Strategic Management Journal, 23 (9), 795–817.

La Porta, R., F. Lopez-de-Silanes and A. Shleifer (1999), ‘Corporate ownership around the world’, Journal of Finance, 54 (2), 471–517.

La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishny (2000), ‘Investor protection and corporate governance’, Journal of Financial Economics, 58 (1–2), 3–27.

Lane P., A. Cannella and M. Lutbatkin (1998), ‘Agency problems as antecedents to unrelated mergers and diversifi cation: Amihud and Lev reconsidered’, Strategic Management Journal, 19 (6), 555–78.

Lee, P. and H. O’Neill (2003), ‘Ownership structures and R&D investments of


US and Japanese fi rms: agency and stewardship perspectives’, Academy of Management Journal, 46 (2), 212–25.

Lerner, J. and R. Rajan (2006), ‘NBER Conference on corporate alliances’, Journal of Financial Economics, 80 (1), (April), 1–3.

Macaulay, S. (1963), ‘Non-contractual relations in business: a preliminary study’, American Sociological Review, 28, 55–70.

Mathews, R.D. (2006), ‘Strategic alliances, equity stakes, and entry deference’, Journal of Financial Economics, 80, 35–79.

McConnell, J. and H. Servaes (1990), ‘Additional evidence on equity ownership and corporate value’, Journal of Financial Economics, 27, 595–612.

Mikkelson, W.H. and M. Megan Partch (1997), ‘The decline of takeovers and dis-ciplinary managerial turnover’, Journal of Financial Economics, 44, 205–28.

Mnookin, R. and L. Komhauser (1979), ‘Bargaining in the shadow of the law: the case of divorce’, 88 Yale L. J. 950.

Mowery, D., J. Oxley, and B. Silverman (1996), ‘Strategic alliances and interfi rm knowledge transfer’, Strategic Management Journal, 17, 77–91.

Nguyen-Dang, Bang (2006), ‘Does the rolodex matter? Corporate elite’s small world and the eff ectiveness of boards of directors’, EFA 2006 Zurich Meetings Paper. Available at SSRN.

Pedersen, T. and S. Thomsen (1997), ‘European patterns of corporate ownership: a twelve-country study’, Journal of International Business Studies, 28 (4), 759–78.

Peirce C.S. (1931–5), Collected Papers, Cambridge, MA: Harvard University Press.

Pfeff er, J. and G. Salancik (1978), The External Control of Organization: a Resource Dependence View, New York: Harper and Row.

Poppo, L. and T. Zenger (2002), ‘Do formal contracts and relational governance function as substitutes or complements?’, Strategic Management Journal, 23 (8), 707–26.

Roe, M. (1991), ‘A political theory of corporate fi nance’, Columbia Law Review, 91 (1), 10–67.

Scharfstein, D. (1988), ‘The disciplinary role of takeovers’, Review of Economic Studies, 55, 185–99.

Schwert, G.W. (2000), ‘Hostility in takeovers: in the eyes of the beholder?’, Journal of Finance, 55, 2599–640.

Shavell, S. (2004), Foundations of Economic Analysis of Law, Belknap Press of Harvard University Press.

Shleifer, A. and R. Vishny (1986), ‘Large shareholders and corporate control’, Journal of Political Economy, 94, 461–88.

Shleifer, A. and R. Vishny (1997), ‘A survey of corporate governance’, Journal of Finance, 52 (2), 737–83.

Thomsen, S. and T. Pedersen (2000), ‘Ownership structure and economic perform-ance in the largest European companies’, The Strategic Management Journal, 21 (6), 689–705.

Tihanyi, L., R. Johnson, R. Hoskisson and M. Hitt (2003), ‘Institutional ownership diff erences and international diversifi cation: the eff ects of boards of directors and technological opportunity’, Academy of Management Journal, 46 (2), 195–213.

Verstegen Ryan, L. and M. Schneider (2002), ‘The antecedents of institutional investor activism’, The Academy of Management Review, 27 (4), 554–73.

Verstegen Ryan, L. and M. Schneider (2003), ‘Institutional investor power and heterogeneity’, Business and Society, 42 (4), 398–430.


Villalonga, B. (2005), ‘Family-controlled industries’, paper presented at the Bentley College Finance seminar, September.

Villalonga, B., and R. Amit (2006), ‘How do family ownership, control, and man-agement aff ect fi rm value?’, Journal of Financial Economics, 80 (2), 385–417.

Williamson, O. (1979), ‘Transaction-cost economics: the governance of contractual relations’, Journal of Law & Economics, 22 (2), 233–261.

Williamson, O. (1985), The Economic Institutions of Capitalism, New York: Free Press.

Williamson, O. (2005), ‘The economics of governance’, American Economic Review, 95(2), 1–18.

Yin, R. (1984), Case Study Research, Design and Methods, Beverly Hills: Sage Publications.

Zaheer, A. and N. Venkatraman (1995), ‘Relational governance as an interorgani-zational strategy: an empirical test of the role of trust in economic exchange’, Strategic Management Journal, 16 (5), 373–92.


APPENDIX 11.1 SHAREHOLDER AGREEMENTS IN FRENCH LISTED FIRMS

Table A11.1 Number of shareholder agreements

Year Number of (new) shareholder agreements1

1997 161998 261999 342000 562001 372002 662003 452004 Not available2005 162006 582007 25

Note: 1 Including changes in shareholder contracts (avenants) and changes in equity shares held by shareholders included in shareholder agreement. Excluding breach and end of contracts (résiliation, declaration de fi n de concert, fi n de clauses, caducité d’une convention) as mentionned in the ‘comment’ section of the database.

Source: amf-france.org


APPENDIX 11.2 DETAILED CONTENT OF SHAREHOLDER AGREEMENTS

Table A11.2

Firm Agreement

Pernod Ricard

Agreement on the voting patterns/voting rights● Both investors agree to vote in concert● In case of disagreement between parties, Kirin commits to vote

in favor of all resolutions proposed by the board of directors of Pernod Ricard and to equally vote against resolutions that were not accepted by the board on issues related to:

● Nomination and compensation of directors ● Modifi cation of the fi rm’s charters ● M&A ● Extraordinary dividends ● Measures against takeoversAgreement on the sale and purchase of shares● Kirin commits not to sell its shares before the end of the agreement

(31 Dec. 2007)● After Dec. 2007, Pernod Ricard has a preemptive right to buy Kirin’s

shares at the following price: the average between (1) the average weighted stock price over the 30-day period before Kirin announced its willingness to sell out and (2) the average weighted stock price over the 30-day period before Kirin eff ectively sells its shares.

Publicis Agreement on the board structure● Dentsu will be granted 2 seats on the supervisory board (as long

as it owns at least 10% of the equity); in case the total number of directors increases, Dentsu will be granted additional seats in proportion to its voting rights.

● Dentsu commits to nominate or maintain all supervisory board members who have been chosen by E. Badinter

● Dentsu commits to nominate E. Badinter or any representative (proposed by her) as the chairman of the supervisory board

● Dentsy commits to nominate all management team members proposed by E. Badinter

● A strategic committee (named ‘special committee’ will be formed. Members will be nominated by E. Badinter and Dentsu (with E. Badinter having the discretion to nominate the majority of members)

Agreement on the voting patterns/voting rights● Dentsu will not be able to own more than 15% of voting rights

(33.5% for E. Badinter)● Dentsu commits to vote in favor of E. Badinter’s decisions in the

following cases:


Table A11.2 (continued)

Firm Agreement

● Change of Publicis’ charters ● M&A ● Distribution of dividends ● Capital off erings ● Share repurchases● Dentsu may freely vote (after consultation with E. Badinter) on

the related topics ● Transfer of assets ● Granting of subscription rights ● ‘Reserved’ capital off erings ● Transaction involving E. Badinter, Dentsu or a subsidiary of

Publicis● Dentsu commits to vote in favor of the certifi ed accounts, after

Dentsu’s comments have been taken into account by the fi nancial auditors

Agreement on the sale and purchase of shares● In case of seasoned off ers (that is, share issues by companies

who have already listed shares), Dentsu will be granted an anti-dilution right. Yet it will not be able to participate in the off er through preferred subscription rights

● Dentsu will not be able to transfer or sell its equity shares in Publicis until July 2012

● After July 2012, E. Badinter has a preemptive right to buy Dentsu’s shares

● Dentsu commits not to make any special arrangements with Publicis’ management without prior notice of E. Badinter. Conversely for E. Badinter

Club Med Agreement on the board structure● Fipar is granted the right to propose the nomination of one

director (as long as it owns 4% equity)● All investors commit to vote in favor of this nominee, and ‘fi re’

the directors that would be requested by Fipar● Accor will keep one representative on the board of directorsAgreement on the voting patterns/voting rights● All investors confi rm to support current management’s strategyAgreement on the sale and purchase of shares● Investors agree not to sell any of their shares without informing

the other investors for a period of 2 years● Investors agree not to increase their ownership level (on an

individual or collective basis) until either the agreement or the date when the group of investors will own less than 20% of Club Med’s equity or voting rights



Firm Agreement

● After two years, investors have the preemptive right on the purchase of shares sold by other investors

● The parties may unanimously decide to lift the ban on additional share purchases so as to increase their shares in Club Med’s equity

● Agreement on takeovers: in case of takeover, investors are allowed to sell their shares only if Club Med’s board of directors has given its agreement on the takeover. If one of the parties wishes to make a competitive off er, it may terminate the agreement

Non-equity and control issues● Icade (real estate arm of institutional investor CDC) has joined

the agreement under the specifi c condition that it will conclude a contract with Club Med related to real estate issues

Legrand Agreement on the board structure● The board will include 11 members● Until the initial period (2 years and 3 months after the IPO date),

the parties agree that the board will be composed of: ● 3 representatives of each signing party ● 2 independent board members ● 3 top managers● After the initial period, Wendel and KKR commit that the board

will be constituted by a majority of board members nominated by both parties

● In addition, Wendel and KKR will be granted seats in proportion to their respective voting rights

● The governance structure will include ● A strategic committee, chaired by a KKR representative ● A compensation committee, chaired by a Wendel

representative ● An audit committee, chaired by an independent board memberAgreement on voting patterns/voting rights● Wendel and KKR forbid to vote in favor of granting dual rights

to shareholders holding Legrand’s shares for more than two years● The chairman of the board will be granted signifi cant discretion

with respect to the daily management of the fi rm, except on decisions relative to

● Share off er and buy back ● Subscription of new debt or early pay back ● Acquisition of equity shares in other fi rms, acquisition of other

businesses and JV for deals above €50m



Firm Agreement

● Sell-out of businesses asset or participation above €50m ● Agreement or modifi cation of Legrand’s 3-year strategic plan

and annual budget ● Firing or nominating auditors ● Any projects that would entail the full or partial transfer of

Legrand’s assets ● Any deal that would result in equity increase of equity reduction,

including convertible debt or preferred shares ● The cancelling out of double voting rights or any decision that

would modify the voting rights attached to Legrand’s shares ● Any modifi cation of the governance rules, such as the

composition of the board ● The introduction of Legrand’s shares in a stock market other

than Euronext ● A voluntary liquidation of the fi rm or any decision that would

generate a collective procedure against Legrand ● Any modifi cation of Legrand’s charters that would favor one of

the parties ● Any transaction or treaty if amounts at stake exceed €50m● The parties commit that Lumina White (Lumina Participation) will

vote in accordance with KKR and Wendel. In case of disagreement between the parties, Lumina White will conform to instructions given its owners in relation to their respective shares

Agreement on the sale and purchase of shares● Wendel and KKR both commit not to sell their equity shares

before the end of the ‘restrictive’ period (the minimum between (1) 18 months after the expiration of the lock-up period for syndicate loans and (2) date when the parties have jointly agreed they could sell a portion of their stocks)

● Some transactions will however be allowed: ● Cessions in favor of entities which are fully owned by either

Wender or KKR (sociétés apparentées) ● Cessions that do not exceed €10m, in so far as the other party

has been informed at least the day before the transaction ● Cessions of shares in favor of a board member of Legrand, to the

extent that it does not exceed what is written in the charters● After the restriction period, the sell-out of Legrand’s shares will be

unrestrained as long as it is consistent with ● The right of preemptive off er ● The ban on block sell-out and joint sell-out (‘tag along’)

applicables to blocks ● Cessions that are forbidden by the investors’ agreement contract



Firm Agreement

● The stipulations of ‘off ering rights agreement’ and ‘tag along agreement’

● Preemptive right: each party commits to inform the other party when it wishes to sell its shares. The remaining party has the right to make a preemptive off er (at a price which equals or is superior to the one off ered by the selling party)

● Block sell-out: when one of the parties wishes to sell out its shares in ‘blocks’, it is required to inform the other party through a letter. The recipient has 5 days to also inform the seller that it equally wishes to sell its shares. The seller commits to inform the other parties in advance of the conditions of the block sell-out

● Under the ‘tag along agreement’, if the informed party does not wish to sell its shares, the seller will be authorized to sell all of its shares. If the second party also wants to sell its shares, a specifi c rule of ‘share allocation’ will be enforced. Each party will be authorized to sell only a specifi c portion of their shares.

● Each party has agreed not to sell its blocks of shares to an industrial fi rm above a value of €100m

● The above conditions do not hold for: ● Cessions of shares authorized along prior conditions ● Cessions sold within a seasoned off er led by a banking syndicate

(following a guarantee contract) ● Swaps by one of the parties between Legrand securities and

Legrand stocks or other fi nancial instruments ● Cessions in the context of a takeover ● All cessions ruled by the ‘tag along agreement’ (agreement relative

to the joint cession of Legrand’s shares after the IPO)● Agreement on takeovers by one of the parties ● Each party commits to get the written consent of the other party

before its proceeds to a takeover off er. The informed party has three days to give its answer. Beyond this period, agreement is assumed. In case of disagreement, the ‘takeover’ party is expected to incur all costs related to the off er

● If after the takeover off er, one of the parties becomes a majority owner and the other one a minority owner, a new investors’ agreement will be concluded. In any case, this new agreement will give the minority investor a veto right on all strategic decisions on Legrand as long the minority investor holds 20% of the voting rights

● In addition, a joint ‘exit right’ will be implemented if the majority owner wants to sell its block equity



Firm Agreement

Schneider Electric

2002. Agreement on the sale and purchase of shares● Second modifi cation (avenant) of the investors’ agreement

contract signed in 1993 between AXA, BNP-Paribas & AGF. Updates the respective equity shares included in the agreement, i.e. 3%, 1.4% and 0.4% of capital

2006. Agreement on the sale and purchase of shares● AXA commits to keep at least 2.5m equity shares in Schneider● Schneider commits to keep at least 8.8m equity shares in AXA● In case of a hostile takeover of Schneider, AXA has the right to

purchase all AXA shares still owned by Schneider; conversely for Schneider

292

12. Board governance of family fi rms and business groups with a unique regional datasetLluís Bru and Rafel Crespí

1. INTRODUCTION

This study is a detailed description with methodological contribution to the measurement of a set of family business groups in the Balearics region in Spain that belong to either the Balearic Family Business Association (ABEF) or the nationwide association, the Spanish Family Business Institute (IEF).

Before we discuss the object of our study, we will examine the main aspects that characterize the family business as an economic organization. Next, we will display their economic activity and relevance in the region’s economy.

Subsequently, we will take a closer look at the economic behaviour and organization of the 556 companies that belong to the 50 family busi-ness groups. In our description we will examine these companies on two levels. First, each individual company will be contemplated as a separate economic entity. Second, we will off er a specifi c description of each family business group.

For this, we have used the data we have on the family companies and their boards of directors as the essential basic information for our study. The main methodological innovation of our study is that the Spanish ‘two-surnames’ system allows us to analyse in detail the family ties among administrators at both the fi rm and the business group level.

2. OUR DEFINITION OF THE FAMILY BUSINESS

What is understood by the term ‘family business’? We might defi ne the family business as a company that fulfi ls two basic requirements: persistent belonging to individuals within a single family circle, and being governed by one or more of the members of that family.

Board governance of family fi rms and business groups 293

When applying this defi nition to specifi c companies, to avoid ambiguity it is important to be more precise about what we exactly mean by a family company. Three factors are taken into consideration in our attempt to defi ne a family business.1

First, the ownership of the company by a family is essential for such company to be defi ned as a family business. Typically we use the term ‘family fi rm’ when the majority of the capital, with the corresponding voting rights, is owned by individuals from a single family circle, in such a manner that we can be sure that the family do govern the fate and future of the company. Whilst the control of the company can occasionally be attained with a minority of shares, it is most common to see most of the share capital in the hands of the family. To possess the majority of the voting rights means having the power to take all sorts of strategic and operative decisions within the company. Of course, the greater the per-centage of ownership, the stronger the family’s infl uence on the fate of the company.

The second defi ning feature of the family business is that the manage-ment of the company is controlled by the family members, who are also the primary decision makers. However, when analysing any fi rm, we know it is important to distinguish between the management of the company and its control. In the family company, when it is said that one or more members of the family take part in the management, this could mean that such members undertake the control and management activities simul-taneously – this is frequently the case in fi rst-generation companies and small businesses – or it may imply that they only undertake the control of the company, while the business is managed by professional managers who do not belong to the family. This latter case is more typical of com-panies in which the ownership is distributed among many members of the third and subsequent generations of a family and is also commonly seen in the case of large companies. Although the management activities of the family company are frequently conceived of as remaining in the hands of the family members, it is no less true that, if the family actively controls the company, this would be enough of a determining factor in the com-pany’s decision-making process, and thus in the path that the company will follow.

The third defi ning characteristic of the family company is the family’s continuous involvement over time, through successive generations of the same family. It makes no sense to speak of a family business if the company does not continue to be a long time under the control of the family circle.2 This aspect greatly limits the scope of enterprises that we refer to when we speak of a family company and defi nes certain aspects of business organization as distinguishing features of the family business.


Thus, the transmission of the ownership, the requirements established to enable family members to become a part of the company, the leadership in the successive generations and the necessary attraction of professionals who are not family members are relevant issues in any study of the family business.

3. DATA SOURCES AND METHODOLOGY

The quantitative information used in this study comes from the Spanish section of the Amadeus database, created by Bureau Van Djick, which basically compiles data that Spanish companies are required to record with the Companies Registry. Because this database is computerized, its contents can be processed, as we will see below. The large number of com-panies included in the database (virtually all of the Spanish companies) has enabled us to cover information on large, medium and small compa-nies throughout Spain over several years. In fact, the number of Spanish companies included in the last database update was 830,000, and for the autonomous region of the Balearic Islands this fi gure was 26,747.

There are essentially four types of information in the database that are relevant to our study: (i) the fi nancial statements, (ii) information on the activities that the companies are engaged in, (iii) the list of administrators of the companies and the positions they hold, and fi nally (iv) the ownership listings, including both company shareholders and companies partially owned by other companies.

The fi nancial statements, including the balance sheets and operating statements, give us information on the size of the companies analysed and off er us an approach to the structure of their share capital. Certain proportions of business debt and earnings can also be compared among companies.

The information on the business line of activity of each company is primarily based on that company’s assigned NACE (economic business activity) code. This classifi cation makes it possible to assign each company a highly specifi c economic activity code, up to four digits, gradually adding in three-, two- and one-digit codes, to progressively specify the economic activity, while at the same time enabling companies with similar economic activities to be grouped together.

The list of company administrators is essential to our purpose, as it allows us to measure the extent to which family businesses entrust the seats on their boards of directors (and therefore the fi rms’ governance and management powers) to family members. Here we can make use of the information that the Spanish ‘two-surnames’ incorporate. This surname


system is very suitable for genealogical purposes, because it has two features that help to establish kinships. First, married women usually do not change their name; and secondly, every newborn has two surnames or family names (apellidos in Spanish): the fi rst is the father’s fi rst surname, and the second is the mother’s fi rst surname.3 Having the names and surnames of the admin-istrators makes it possible to process family ties on the computer, using fi rst two surnames (enabling us to infer whether or not two administrators are siblings) and then one surname, which allows us to trace the generational family ties among administrators of diff erent generations (parents and chil-dren, grandparents and grandchildren, uncles/aunts and nieces/nephews, and so on) as well as among those of the same generation (cousins).

By way of example, consider the very simple board of directors of Barcelo Corporacion Empresarial SA, a fi rm pertaining to the Barcelo family included in our sample:

Board of directors of Barcelo Corporacion Empresarial SABarcelo Vadell, Simon Pedro

Barcelo Tous, GuillermoBarcelo Vadell, FranciscaGonzalez Rodríguez, Raul

The names of the board’s members are stated in the following order: father’s surname, mother’s surname, and fi nally Christian name (possibly two, as in Simon Pedro). The coincidence of two surnames in exactly the same order (Barcelo fi rst, then Vadell) allows us to infer that Simon Pedro and Francisca are siblings; whereas the fact that there is only one surname that coincides between them and Guillermo Barcelo Tous allows us to infer that they either pertain to a diff erent generation or are cousins (the order also allows us to discard some family ties; for instance, Guillermo cannot be Francisca’s son, since Barcelo would then appear in second place). Finally, the fourth member of the board is not identifi ed as a member of the family, since there is no coincidence whatsoever of surnames.

Finally, the lists of company shareholders show the proportions of share capital held by the individual shareholders of the fi rms, establishing who the last shareholder is. This information is not available for all of the companies included in the database. Another relevant type of information is the structure of the companies that control other subsidiary companies, with information on the percentage of their interest in their affi liates. This information is particularly useful to ascertain business groups and shed light on the relations between the companies and the ties between their administrators.


Based on the personal information of the members of the ABEF and the Balearic families that belong to the IEF, we have used the Amadeus data-base to outline the diff erent business groups in the Balearics. Our starting point was the notion that a company belongs to the family group whenever it is controlled by the family. To establish the business groups, we applied the following steps: (i) initially the family group included the companies in which the corporate partner who was a member of the above-mentioned associations fi gured as a company board member with signifi cant owner-ship of the companies; (ii) subsequently, the group came to include com-panies in which its direct family members also fi gured as board members; (iii) this extended to the companies that were partially owned by the above companies, with the condition that such ownership consisted of at least one-fourth of the share capital; (iv) once fi rms were identifi ed, we checked with representatives of the family fi rms associations the identifi cation of fi rms within family business groups, correcting case by case any possible error which stemmed from the automated processes described in steps (i) to (iii).

Finally, diff erent controls were applied to the data to ensure that the companies analysed were indeed engaged in an actual economic activity. We eliminated from our sample any companies in the process of liquida-tion, those with no existing income and asset volume data for 2004 fi nan-cial year, which was the last year fully published in the database, as well as the companies in which either of these two measures did not reach €60 000. Following the application of these selection criteria, the resulting number of sample companies was 556.

Below we will describe the group of companies analysed. Prior to a detailed description of the family businesses, the object of our study, we off er some aggregate fi gures on these companies that show the signifi cant amount of economic activity that they generate, and thus the importance of studying them. Subsequently, we explore these companies in some detail by applying two diff erent procedures: fi rst, we will analyse the com-panies individually; next, we group the companies according to the family that controls them, which gives us a group of companies for each of the 50 families of the Balearic Islands that form part of the ABEF or IEF.

3.1 The Relevance of Associated Family Companies in the Region

To address the relative importance of the member companies of the family business association in the Balearics, we can compare the fi gures of their economic activity with the economic activity volume generated in the autonomous region at large, namely the regional gross domestic product (GDP), or we can consider the number of workers employed by these


companies compared with the number of active workers in the Balearic Islands.4

In Table 12.1 we see that the 556 companies in our sample had a total asset volume of more than 14 200 million euros in 2004. If we add up the income volume generated by each of the 556 sample companies, our total is 10 060 million euros. By way of reference, for the year 2004 the value of the gross domestic product for the Balearic Islands was approximately 20 900 million euros. Obviously these fi gures are not comparable in the sense that the asset is a stock measure and the GDP a fl ow measure. To attain a fi gure that can be compared with the regional GDP we must refer to the added value generated by the companies in our sample. For this item, the value was 2867 million euros.

As to the employee volume, the aggregate fi gure for 372 of the 556 com-panies that we have information on comes to more than 70 000 workers. Once again and to attain a point of reference, according to the statistics of the INE (Spanish Institute of Statistics), the employment volume for the year 2004 in the Balearic Islands was 455 000.5

Table 12.1 describes some aggregated data of interest on our sample of family businesses, including the value of their equity, which comes to nearly 6600 million euros; the earnings before taxes, which are approxi-mately 486 million euros; and the total corporation tax, which comes to 139.26 million euros.

Another aspect worth bearing in mind is the organization and control of the companies. All in all, the 556 companies have structured their gov-erning bodies with a total of 2244 board members, and an average of 4.04 members, where 77.6 per cent of them are men, in 4.5 per cent of the cases another company fi gures as a board member in the offi cial register, and nearly 18 per cent of board members are women. This fi rst approach to associated family companies in the Balearic Islands gives us an idea – albeit

Table 12.1 Aggregate values of the main magnitudes of the sample family companies

Total value in thousands of €

Total assets 14 200 000 Revenue 10 600 000 Employees (number) 70 269 Added value 2 867 307 Equity 6 599 452 Earnings before taxes 486 969 Corporation tax 139 261


a rather imprecise one – of their relative importance, and thus reveals their most salient magnitudes. A more in-depth study will enable us to charac-terize them according to what we might refer to as the typical or average company.

4. COMPANIES UNDER THE CONTROL OF ASSOCIATED FAMILIES

4.1 The Representative Company

Characterizing the typical business based on our sample of 556 compa-nies is by no means easy, given their vastly diverse sizes and the diff erent sectors of activity that they belong to. Nevertheless, we see that the average company has assets of around 25.5 million euros, a revenue volume of 19.6 million euros and equity of approximately 5.19 million euros for the year 2004, as shown in Table 12.2. However these average values repre-sent a great deal of dispersion among the 556 companies in the sample. The standard deviations are more than fi ve times higher than the average values, which points to the bias introduced by several extreme observa-tions, due to their very high values, for any of the fi nancial magnitudes that we are contemplating.

Hence, the median of the 556 sample companies can better characterize the typical company. We can now see that the observation of the company holding the central position in our sample is smaller than what the average of the observations could lead us to believe: it has an asset value of 2.6 million euros and a revenue volume of 1.27 million euros, it generates an added value of 543 million euros, and its equity is approximately 1 million euros.

Panel B of Table 12.2 informs us about the boards of directors of these fi rms. The representative company has an average of four board members, one of which is a woman and the rest of which are men, with the exception of several seats assigned to legal entities (companies). The average business age of the 556 sample companies is 17.5 years. This is not excessively far from the median, which is 14 years.

Panel C of Table 12.2 displays data on the average number of employ-ees, which represents a great deal of dispersion, making the interpretation of this item quite diffi cult. Nevertheless, by examining the average data we can see that a third of the income from operations is allocated to staff salaries, and that the average return for shareholders is 10.4 per cent. This value goes down to 2.4 per cent when placed in relation to the return on assets. Finally, we can give information on the solvency ratio, which takes


in the proportion of the company’s assets in relation to its net worth. In this case, the average value is 43.7 per cent, which is very similar to the median for the same item. This review of the fi nancial magnitudes and administrative bodies enables us to make an initial approach. However, the high variability on the observed magnitudes suggests that a more detailed study is in order, distinguishing the characteristics of the compa-nies according to their size or their sector of activity, as we shall see in the sections below.

4.2 Diff erences According to Size

In this section we have established the parameters for company compari-sons by grouping them according to their size. To do so, we have divided

Table 12.2 Descriptive statistics of average values of the main magnitudes of the sample family companies (2004)

Values in thousands of €

Panel A average median standard deviation

Total assets 25 519 2 683 142 167 Revenue 19 683 1 278 100 152 Added value 5 194 543 30 035 Equity 11 870 1 001 69 350 Earnings before taxes 889 23 5 414 Corporation tax 255 4 1 456

Panel B

Size of board of directors 4.04 4 2.8Men on board of directors 3.13 3 2.28Women on board of directors 0.721 0 1.06Companies on the board of directors

0.182 0 0.647

Panel C

Employees (number) 189 27 1 033 Company age (years) 17.5 14.0 12.9Staff costs/income from operations (%)

33.2 22.3 53.2

Solvency ratio (%) 43.7 42.2 40.2Return on shareholders’ investments (%)

10.4 5.2 84.5

Return on assets (%) 2.4 1.3 18.8


the sample into three groups, corresponding to what we shall refer to as small, medium and large enterprises, according to their asset volume. Each tertile of the 556 sample companies is made up of 185 companies.6 This procedure aims to show more homogeneity than the analysis in the above section, amid the companies of each tertile, and in turn enables us to see any diff erences that may emerge among the companies that belong to dif-ferent size groups.

The values of Panel A in Table 12.3 illustrate the signifi cant diff erences between small, medium and large enterprises. Thus, the average small company has an asset volume of 484 000 euros and a revenue volume of 711 000 euros, generates an added value of 238 000 euros, and its average earnings before taxes are 7000 euros. As a group, the 185 small companies only generate 126.6 million euros, for a total asset volume of 89.5 million euros. Their overall earnings before taxes are around 1.27 million euros.

At the opposite end of the spectrum, large companies together account for more than 90 per cent of the revenue volume, number of employees and asset volume for the sum total of all the companies in our sample. On average, the large companies have 33.8 million euros in equity, and the earnings before taxes of the average company within this group come to nearly 2.5 million euros.

The typical medium company generates a revenue volume of 2.7 million euros, with assets of 3.14 million euros, and equity for a value of 1.46 million euros. As a group, these 186 medium companies account for approximately 500 million euros in assets and revenue, with aggregate earnings before taxes of 24.67 million euros.

Beyond the descriptive fi gures mentioned above it is interesting to observe the behaviour of the magnitudes associated with the corporate governing bodies according to their size. In Panel B in Table 12.3 we can see that, for the same number of companies in each size group, the total number of board members is 901 for large enterprises, 543 for small enter-prises, and 803 for medium companies. Indeed, it is plain to see that the average size of the board of directors goes up as the company grows larger in size. The average values are 2.94 board members for the small compa-nies, 4.34 for the medium companies, and 4.84 for the large companies. This comes as no surprise, if we consider that larger companies may require a greater participation in their government. For example, the corporate governance report for Spanish listed companies, Corporate Governance Report of the Companies with Values Listed in Stock Exchanges for the 2004 year, published by the regulator of the Spanish Stock Exchange, the Comision Nacional del Mercado de Valores (CNMV), reveals that the number of board members also rises with the size of the companies listed on the Spanish Stock Exchange, although the average number of members

301

Tab

le 1

2.3

Ave

rage

and

tota

l val

ues o

f the

mai

n m

agni

tude

s of t

he sa

mpl

e fa

mily

com

pani

es fo

r thr

ee si

ze le

vels

, bas

ed

on a

sset

s

smal

lm

ediu

mla

rge

Pane

l A (t

hous

ands

€)

aver

age

tota

lav

erag

eto

tal

aver

age

tota

l

Tot

al a

sset

s48

4 89

587

3 140

58

0 901

72

678

13 50

0 000

R

even

ue71

1 12

6 629

2 726

49

3 350

55

101

10 00

0 000

E

mpl

oyee

s (nu

mbe

r)16

1 4

50

28

3 772

44

0 65

047

Add

ed v

alue

238

43 61

3 83

6 15

3 870

14

431

2 669

824

Equ

ity21

0 38

792

1 460

27

0 012

33

821

6 290

648

Ear

ning

s bef

ore

taxe

s7

1 273

13

5 24

647

2 492

46

1 049

C

orpo

ratio

n ta

x7

1 185

53

9 7

46

694

128 3

30

Pane

l B (n

umbe

r of p

erso

ns)

Size

of b

oard

of

dire

ctor

s2.

94

543

4.34

80

3 4.

84

901

Men

on

boar

d of

di

rect

ors

2.40

44

4 3.

38

625

3.62

67

3

Wom

en o

n bo

ard

of

dire

ctor

s0.

41

76

0.74

13

6 1.

02

189

Com

pani

es o

n th

e bo

ard

of d

irect

ors

0.12

23

0.

22

41

0.20

37

Not

e:

Dat

a co

rres

pond

to th

e 20

04 y

ear,

with

185

com

pani

es fo

r the

smal

l and

larg

e te

rtile

s and

186

for t

he m

ediu

m c

ompa

ny te

rtile

.


is not directly comparable with the companies in our sample, as the average number of board members for the listed companies is 9.7.

While there are diff erences in the number of board members according to the size of the family companies, diff erences can also be seen in their composition. The proportion of women on the boards of directors of the small companies is 14 per cent, becoming 16.9 per cent in the medium companies and reaching 21 per cent in the large companies.

Finally, the size-based classifi cation enables us to observe some propor-tions regarding the cost, profi tability and structure of the family busi-nesses. Table 12.4 shows no signifi cant diff erence in the average age of the medium and large enterprises, standing at approximately 19 years for both groups. However, the average age of the small companies is markedly lower: 13.9 years. Moreover, it seems that the small companies tend to be more labour intensive, as they allocate a larger proportion of their revenue to staff salaries (38.2 per cent, in relation to the 31 per cent allocated by medium and large companies). The solvency ratio oscillates between 41 per cent and 46 per cent, with no apparent pattern related to the average size of the companies.

The profi tability ratios display the greatest diff erences in the average values according to company size. The returns for shareholders are higher for small companies than for large ones, whilst, on the other hand, the return on assets is greater for the large and medium companies. The medium companies, in this comparison, present the lowest shareholders’ returns of the three groups and the highest total returns on assets.7 In any case, we must proceed with caution when considering these average values, for two reasons. The fi rst is the heterogeneity of the companies included in each

Table 12.4 Average values of the proportions of the main magnitudes of the sample family companies for three size levels, based on assets

small medium large

Age (years) 13.9 19.9 18.6Staff cost/income from operations (%)

38.2 31.2 31.4

Solvency ratio (%) 43.6 46.0 41.4Return on shareholders’ investments (%)

22.1 −1.4 10.7

Return on assets (%) 0.6 3.9 2.7

Note: Data correspond to the 2004 year, with 185 companies for the small and large company tertiles and 186 companies for the medium company tertile.


group, as these average values are not weighted, meaning that they place the same relative importance on the smallest company within the tertile as they do on the largest. The second reason lies in the database, which conveys the information provided by the companies to the Trade Registry. A signifi cant number of the very small companies present unaudited accounts, as the regulations in force do not require them to be audited.

4.3 Sector of Activity Diversity

Indeed, the vast disparities in the sizes of the companies allow us to observe diff erences in some of their behaviour variables such as profi tability and the composition of their boards of directors. All the same, there are operational aspects of the companies that could potentially aff ect their profi tability, and which are tied to their specifi c business activity. In this section we present diff erent magnitudes, assessing the companies together according to their sector of activity at the NACE one-digit level.

Table 12.5 illustrates the main measurement magnitudes of the com-panies by sector of activity, and shows the distribution of the sample companies among the eight major sectors into which the sample has been divided. Most of the economic activity, according to the aggregate values of asset and revenue volume, resides in three sectors of activity: hotel and catering, transport and communications, and real estate and busi-ness services. These three sectors embrace the majority of the companies associated with tourism, which is characteristic of the Balearic economy. Along these lines, Table 12.5 suggests that the distribution of economic activity for family fi rms follows a similar pattern to that of the Balearic GDP. Moreover, the distribution of the companies in the sample, which was calculated with diff erent variables such as the number of companies in the sample, asset value and revenue level, is similar to the relative weight of each of the sectors in the Balearic economy, as shown in Table 12.5. The distribution of the number of companies in the sample is to a large degree in line with the data on the regional GDP composition, with the exception of the grouping of other activities, which takes in such diverse activities as education, healthcare and veterinary activities, social services, personal services and fi nancial intermediation. The fi gures of the companies in hotel and catering and in real estate and business services are representative of the considerable weight of the tourist industry in the Balearic Islands, as it includes lodging in hotel establishments and other forms of rentals, and also what is known as the complementary supply.

By asset volume, the average size of the hotel companies is signifi cantly greater than those of agriculture, industry, construction and trade, despite the fact that its number is equal to the sum of those that form these other


sectors. The high average revenue volume of the companies in transport and communications is due to the fact that this sector includes the Balearic Islands’ major travel agencies,8 which are well-established nationally and internationally and have relatively very high revenues and assets.

To assess profi tability, labour expenses and fi nancial structure, Table 12.6 displays the corresponding proportions after weighting these values according to the size of the companies, which in turn was calculated accord-ing to the asset volume of each company. This prevents us from attributing to a small company the same weight that we attribute to the large sample companies in the same sector. Particularly worthy of note in the column on staff expenses over income from operations is the high intensity of labour in the primary sector and the ‘other activities’ group, which is essentially made up of personal services. As to the solvency ratios, the low values for the construction companies are salient, which is now typical of this sector and of the transport and communications sector, given the high debt rates in relation to their equity. As regards the returns on assets, the values oscillate within margins that display little dispersion among the sectors. The return on stockholders’ equity, however, exhibits greater dispersion, although

Table 12.5 Relative relevance of the economic activity sectors in the 556 sample companies

Companies in sample

(%)

Contribution of the sector to regional GDP (%)

Companies’ sample

assets (%)

Companies’ sample

revenues (%)

Agriculture, livestock, fi shing and power

1.26 3.64 0.12 0.05

Industry 6.29 5.75 1.59 2.41Construction 6.29 10.43 1.75 1.90Trade and repairs 9.53 9.93 2.70 7.45Hotel and catering 21.76 25.22 43.74 22.53Transport and communications

12.59 8.96 8.03 41.55

Real estate and business services

35.61 18.16 40.98 23.59

Other activities 6.65 17.92 1.09 0.52

Note: Comparison of relative weights of the magnitudes of asset size and revenue volume, as well as the number of sample companies, with industry distribution reported by the National Institute of Statistics (INE) in the regional accounts. The industry groupings are based on the NACE 1-digit level.


it must be viewed with caution, as some of the companies do not present audited fi nancial statements, as we have mentioned, and the absence of certain observations could distort the average values of the sample.

5. BUSINESS GROUPS UNDER THE CONTROL OF ASSOCIATED FAMILIES

Although we have explored the family companies in the section above, we have to consider the fact that each family can control more than one company. Naturally a family will not take a certain business decision for each company separately, but rather will consider the group of companies under its control as a whole.

This calls for a description of business groups under the control of each family in the sample. The key to making this possible resides in the ability to ‘build’ the group of companies to be assigned to each of the families, so as to assess the decisions of the associated corporate families in the Balearic Islands.9 The fi rst issue that we can address by studying the family groups is

Table 12.6 Proportions of profi t, fi nancial structure and expenses by activity sector

No. of companies

Return on shareholders’ investments

(%)

Return on assets

(%)

Solvency ratio (%)

Staff cost / income

from operations

(%)

Agriculture, livestock, fi shing and power

7 25.6 4.0 51.6 43.1

Industry 35 2.4 3.6 38.6 24.3Construction 35 26.3 5.0 23.5 15.7Trade and repairs 53 16.7 8.5 46.0 8.5Hotel and catering 121 7.0 3.6 45.0 33.9Transport and communications

70 39.5 4.2 18.3 10.9

Real estate and business services

198 8.0 2.7 60.3 38.5

Other activities 37 −5.6 3.0 54.2 71.5Total 556 10.5 3.4 48.8 32.6

Note: The proportions were weighted according to the size of each company, bearing in mind each company’s total assets.


the concentration or dispersion of sizes among the diff erent family groups, and, within them, the diff erences we observe among the companies within each group. By individually studying the companies in the above section, we have seen vast diff erences in size, sector and even governing body com-position. By attaching these companies to their respective family groups, the questions regarding heterogeneity continue to be of application among companies of a single-family group.

We also explore the families’ corporate diversifi cation strategy, distin-guishing the degree of sector diversifi cation, according to whether diversi-fi cation revolves around a main business activity (related diversifi cation) or whether sector diversifi cation is more intense, meaning that it is not neces-sarily linked to any initial core business. Finally, we delve more deeply into the administrative and management bodies of the business groups and the degree of involvement of the controlling family.

5.1 Family Group Heterogeneity

Among the family groups studied, the typical one has an average of 11 companies, with a total asset value of 283 million euros and an average revenue nearing 213 million euros, and provides employment to a total of 1464 employees. These averages represent a great deal of dispersion, however. Another way of characterizing the ‘usual’ family group would be to consider the median of the values mentioned above: 7 companies per family group, with a total revenue volume of around 29 million euros, and total assets of nearly 42 million euros. These are family group companies with an age of around 18 years and are run and controlled by the second or third generation of the family.

To address the structure of the companies that form the family groups, we can present two opposite structures. On one hand, there are groups made up of several companies (on average seven), all of which are equally important as regards the volume of activity or the asset volume. On the other hand, there may be business groups that are formed by many com-panies but there is one of them that generates nearly all of their aggregate activity, while the other companies are virtually insignifi cant. What is the most common structure among the family groups in our sample? What we see is that they tend more towards the second situation than the fi rst, although with signifi cant diff erences.

Indeed, Table 12.7 shows that the largest company of each of the 50 family groups of our study generates 53 per cent of the total asset volume and 44 per cent of the total revenue of the group.10 The degree to which the activity is concentrated in a reduced number of companies can also be seen in the fact that the three largest companies in each group represent three


quarters of the family group activity, both in terms of asset volume and in terms of revenues. For 36 of the 50 family groups that have formed fi ve or more companies, the total activity of those fi ve largest companies accounts for nearly 85 per cent of the group total, while the fi fth most important company represents only 3.6 per cent of the group’s activity.

As regards their legal format, 270 of the 556 are limited companies, 282 are public limited companies and 4 are other types of organizations. Among the largest companies in the 50 family groups, 34 are public limited companies and 16 are limited companies.

Thus, we can confi rm that the business structure of the associated family companies in the Balearics is complex, though it is built upon a small number of companies. In many cases, hidden beneath the number of companies that sustain the family business structure there is a diversifi ca-tion strategy. Indeed diversifi cation into activities that may or may not be similar to the core activities of the family business is important from the perspective of the diversifi cation of risks. In the next section we apply the appropriate method to explore their diversifi cation strategies.

5.2 Measurement of Family Group Diversifi cation

When we speak of business diversifi cation we refer to the entry of a company, business group or shareholder into a number of diff erent eco-nomic activities. Alternatively, a non-diversifi cation strategy means focus-ing on a single activity or line of business.

Moreover, it is common in the business and the economics literature to distinguish between related and unrelated diversifi cation. Related diversi-fi cation implies involvement in several diff erent lines of activity that share a group of corporate resources and the same organizational abilities or skills. The sharing of technology, sales forces or distribution activities might be considered an example of related diversifi cation. When these

Table 12.7 Distribution of the importance of the largest and succeeding companies in each group in the study sample

Importance of the company in the group

Proportion of total activity (%)

Assets Accumulated Revenue Accumulated

largest 53.0 53.0 44.1 44.1second 17.7 70.7 19.0 63.1third 6.6 77.3 13.0 76.1fourth 4.6 81.9 3.8 79.9fi fth 2.9 84.9 2.0 81.9


technological, commercial or skill-related connections do not exist among diff erent business units, diversifi cation is understood as being unrelated.11

Undoubtedly, any abstract classifi cation ranging between the extremes of ‘non-diversifi cation’ and ‘maximum unrelated diversifi cation’ enables the practice of many diff erent levels or degrees of diversifi cation. The actual positioning of each company or business group on this scale between the extremes can be empirically measured thanks to the existence of universally accepted sectorial classifi cation standards. We will use below the European sectorial classifi cation (NACE) system implemented by Eurostat, which uses four-digit codes to classify the diff erent types of activities. There are also three-digit groupings, and two-digit divisions, which can be broken down into one-digit divisions for large-scale sectorial grouping. Moreover the database we use assigns each company a main activity code and another secondary activity code, which allows us to measure the distance between the sectorial activities, bringing us closer to the concepts of related and unrelated diversifi cation.

In the case of family groups, the relevant unit with which to study the level of diversifi cation is the family group, not the individual company.12 A fi rst simple way to measure the family business groups’ sectorial diversifi cation consists of simply counting the number of diff erent activity codes within a group. For the three- or four-digit NACE codes, a high number of diff erent codes will mean a high degree of related diversifi cation. A high number of diff erent NACE one-digit codes, on the other hand, would be indicative of the group’s unrelated diversifi cation. Obviously, if all of the activity revolves around a single NACE code, regardless of the number of companies that make up the group, this tells us that the family group has chosen to focus its economic activity on very few activities. In other words, it has decided not to diversify.

Table 12.8 illustrates how the median number of activities of the family business groups is amid three diff erent codes, if measured within the NACE one-digit code for large sectors. When we further specify the breakdown of the codes, the averages for two- and three-digit codes are 3 and 4 respec-tively. The company groups with an average of 6 companies are devoted to 4 diff erent activities if calculated with the NACE 3-digit code, or to 3, if the NACE calculation is within the one-digit sphere.

For the values of the average number of activities of the 50 family busi-ness groups taken together, the number of diff erent sectors increases with the level of precision (number of digits considered) from 2.76 for a one-digit NACE to 4.9 for a three-digit NACE. In other words, the 11.2 companies of the average are devoted to only 2.76 diff erent NACE one-digit activities, whilst the number of diff erent activities of these companies with greater sectorial precision is 4.9. However, these levels of diversifi cation provide little information, as it is diffi cult to establish points of comparison. Nor


can such points of comparison be established with non-family-business groups, as it becomes diffi cult to defi ne where these other groups begin and end. Moreover, there are no references for other studies on family business groups in other geographic regions, thus making the assessment of these levels of sectorial dispersion more diffi cult.

All the same, Table 12.8 displays some interesting results: if we divide our sample of 50 family business groups into three sub-groups according to their size, we will see how the small groups form an average of around 3.25 companies per group, the medium family groups have 6.8 companies and the large groups show an average of 22.7. The overall median value of these large family groups is 16 companies per group. In other words, we see that the greater organizational complexity of large family groups by virtue of their numerous companies is actually based on businesses within the same activity sector. At the other extreme, for the smaller family groups each branch of activity appears to be proportionally closer to a diff erent company, particularly for the three-digit sectorial classifi cation.

This fi rst approach to the diversifi cation strategies of the family business groups leads us to the conclusion that large groups tend to diversify more in absolute terms, which is probably explained by the fact that their larger size entails a large number of companies, in relation to the medium and small family business groups. We can also conclude that when entering new economic activity sectors, which may or may not be interrelated, the large family groups do so with a larger number of companies, whereas the small family groups tend more towards the structure of companies that are each associated with a diff erent activity, if they opt for several companies.

Table 12.8 Family groups and sectorial diversifi cation according to number of diff erent NACE codes and group size

Panel A: median values Number of diff erent

NACE codes Number of companies per

groupFamily group size 1 digit 2 digits 3 digits

Small 2 2 3 3Medium 3 3 4 6Large 3 6 6 16Overall 3 3 4 6.5

Panel B: average values

Small 2.00 2.31 2.69 3.25Medium 2.59 3.12 3.76 6.88Large 3.65 5.82 8.12 22.76Overall 2.76 3.78 4.90 11.12


A descriptive study based on calculating the number of diff erent activi-ties that a family business group has become involved in is highly sim-plistic. Such approach assumes that each and every one of the activities bears the same weight within the business group, thus skewing the results towards potentially fi ctitious diversifi cation rates.

Hence, we next proceed to a more sophisticated analysis that takes into account the relative importance of each of the companies within the group. The literature on business diversifi cation often makes use of the entropy measure to off er a weighted assessment of the degree of sectorial diversifi -cation, enabling the distinction between related and unrelated diversifi ca-tion. Our entropy measure uses a weighted average of the activities within the business group at a NACE three-digit level in relation to diversifi cation at a one-digit level.13 This is a typical continuum measurement that assigns higher values to high diversifi cation levels and lower values to low diver-sifi cation levels.

By way of example, consider two business groups with construction, hotel and food activities. Group A has 90 per cent of its activity in hotels, 5 per cent in construction and another 5 per cent in food. Group B has one-third of its activity in each of the aforementioned sectors. If we only take into account the diff erent NACE numbers in each case, there are 3 in both groups, leading to the misleading conclusion that Group A is just as diversifi ed as Group B. The entropy index instead gives us a value of 0.39 for Group A, whereas for Group B the index is 1.09. The entropy index is thus a more precise depiction of the concept of business diversifi cation.

As occurs with the NACE number diversifi cation measure, the simple observation of the entropy index values off ers little information on the inten-sity of diversifi cation. Once again, because we have no non-family business groups as a control group, our study must be limited to the diff erences in behaviour according to the size of the diff erent family business groups.

Panel A of Table 12.9, which shows the entropy index diversifi cation median values for each of the three family group sizes, does not diff er con-siderably from Panel B, which displays the average of each of these group-ings. We must point out that related diversifi cation is more important than unrelated diversifi cation for any family group size. In fact, in Panel B, the average of the entropy coeffi cient for all of the business groups together is 0.93 while the unrelated diversifi cation is 0.56. Hence related diversifi -cation accounts for 62 per cent of the total diversifi cation, and unrelated diversifi cation explains the remaining 38 per cent. These values show how diversifi cation particularly revolves around the group’s core business areas, though with some dispersion in areas that are not directly related to such central business activities.

This behaviour is more pronounced in the small family groups, which


often hinge two-thirds of their diversifi cation on a main activity. Even though the entropy measure plots the relative importance of an economic activity sector within the family group, the diversifi cation value is greater in large groups than in small groups, for both related and unrelated diver-sifi cation. As a result, we see that the larger family groups diversify more than smaller groups, even when the measure of diversifi cation weights the relative importance of each activity. This eff ect of diversifi cation is based more on related diversifi cation for the small groups than for the larger groups.

The concept of entropy takes in the organizations’ natural tendency to become more complex through time. In our case, diversifi cation comes to take on this tendency in connection with the growth and age of the organi-zations, and particularly for family businesses in which successive genera-tions are admitted into the company. In the case that this assertion is true, the level of diversifi cation will grow with the overall age of the companies that make up the family business group.

In fact, as can be seen in Figure 12.1, the business groups that amass greater experience, with more consolidated and more numerous compa-nies, adopt higher levels of diversifi cation. For well-established family groups with several generations of experience, the evolution through time that accompanies the years of business experience leads to the decision to diversify to a greater extent than the diversifi cation generally undertaken by younger family groups.

This study of diversifi cation gives us a reference of the behavioural business patterns of the family groups to diversify risks by investing in

Table 12.9 Family groups and entropy coeffi cient for related and unrelated diversifi cation according to the size of the business group

Panel A: median values Diversifi cation

Family group size Related Unrelated Total

Small 0.71 0.55 1.38Medium 0.79 0.60 1.51Large 0.90 0.68 1.75Overall 0.81 0.6 1.48




diff erent sectors of activity. It is not only the reduction of fi nancial risks that explains the tendency to diversify. Economic effi ciency reasons, such as the attainment of economies of scope, capitalizing on specifi c skills or aptitudes within the company, or the mere need to allocate the funds generated by the business activity can explain this tendency to diversify.

On the other hand, diversifi cation also brings along potential expenses, given the need to incorporate new skills into unrelated activities, diffi cul-ties in implementing systems of control when the activities are diverse and even ineffi ciencies in the allocation of resources outside the discipline of the capital markets.

The advantages and disadvantages of business diversifi cation are proc-essed by the controlling shareholders, which by defi nition in the case of family groups are the family members. All the same, diversifi cation can require the incorporation of new skills or simply of new fi nancial partners, which can lead to the possible dilution of family control. These aspects are further explored in the next section.

5.3 Business Group Directors and the Family

The control of the companies within a family group can be organized into many diff erent possible models between two extremes: at one end of the spectrum, the governing bodies can revolve around a small number of people who amass offi ces as board members in each of the companies;

0

500

1000

0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5

Small groupsA

ccum

ulat

ed n

umbe

r of

yea

rsof

com

pani

es o

f the

gro

up

Medium groups Large groups

Total diversification valuesAdjustment

Figure 12.1 Level of sectorial diversifi cation (entropy) and accumulated age of the companies within the family group


and, at the other end, the administrative bodies can be entrusted to a large number of individuals that barely repeat from one company to another.

If we form a hypothetical board of directors of the business group, by adding up the members of the boards of directors within the family groups, we can shed some light on these matters. The representative family group of our sample, made up of 11 companies and 4 board members per company, could be described, at the lowermost limit, as a family group entrusted to a number of people that equals the size of the largest board in the group.14 On the other hand, the maximum possible number of board members could be applied, in which case it would be made up of 44 diff erent people. Table 12.10 illustrates the average and median values for the 50 family groups in our sample. Panel A shows how the median number of companies per family group is 6 and the number of board member posts to be covered is 23. On average, the family groups entrust these 23 board positions to 10 diff erent individuals. Panels C and D of Table 12.10 contain the proportions of each type of relative in relation to the size of the board. The median (Panel C) for the dispersion of individuals is 43 per cent, and the average for such items is 41 per cent, as can be seen in the same column in Panel D.

As we have seen in the previous section, there are signifi cant diff erences in the median values for the business groups according to their size per asset volume. Thus, the small groups with 3 companies use 4 diff erent indi-viduals to cover the 7 positions. Abounding in this sub-sample of smaller family groups are companies with sole administrators. This circumstance makes the proportions to which they become open to diff erent individuals higher than in larger size business groups.

For the large groups, as illustrated in Panel B, the number of diff erent individuals is higher, simply due to the larger size of their boards of directors. However, the proportion of diff erent individuals is the lowest of the three business group sizes, with a proportion of 38 per cent, as shown in Panel D.

As a result, we can conclude that the associated family groups in the Balearics rely on a common nucleus of 6 people for every 10 board posi-tions to be covered. This proportion is smaller in the small family groups than in the larger ones, which in absolute terms place their posts in the hands of a larger number of diff erent individuals.

These measures of diversity or dispersion of individuals do not necessarily comply with the dictates of the family centre of control. Thus, for companies that have outside members or bring in experts in certain lines of business for technological reasons, the dispersion rates can vary considerably.

Yet, from the perspective of family groups, the salient question is how many of the individuals that form part of the boards of directors actually share family ties. The data available in the public registries, which in this


case come from the Trade Registry (information recorded in the SABI database we use), give us the identities of the members of the diff erent boards of directors. One way to examine family control and take advan-tage of the name and two-surname structure in force in Spain is to look at the number of family ties among board members. Thus, the individuals who share both fi rst and second surnames will be listed as siblings, and those who share only one surname will be listed as ‘cousins’. The sibling kinship listings will seldom lead to an error when considering family ties,

Table 12.10 Opening of family groups to non-family board members, for diff erent degrees of kinship, according to the size of the business group

Panel A: median values

Dispersion of board members (number)

Family group size

Number of companies per group

Board members

Diff erent individuals

Diff erent ‘siblings’

Diff erent ‘cousins’

Small 3 7 4 3.5 2.75Medium 6 22 9 7 4.5Large 16 59 34 29 19Overall 6 23 10 8 6


Dispersion of board members (number)

Small 3.25 10.69 7.13 4.81 3.88Medium 6.88 24.18 11.12 8.65 6.59Large 22.76 93.59 35.29 30.29 19.53Overall 11.12 44.44 18.06 14.78 10.12

Panel C: median proportions

Dispersion of board members (proportion)


Panel D: average proportions

Dispersion of board members (proportion)



with the exception of situations that stem from second marriages or other special cases. Undoubtedly however, our cousin listing covers many family relations beyond those known as cousins in daily language, as it includes, in addition to actual cousins, parent–child and grandparent–grandchild relations, and uncle/aunt–niece/nephew relations.

The concurrence of surnames on a board of directors of a family group is a reasonable indication of the degree of kinship among its members, and its consideration thus enables us to quantitatively evaluate the extent to which families keep the control of the business groups in the hands of family members or the degree to which they are open to the inclusion of non-family members on their boards of directors.

Among the measures of dispersion of board members, Table 12.10 also off ers the median and average values for the number of ‘diff erent siblings’ in the Balearic family groups. For a typical business group, taking the median values (Panel A), with 6 companies and 23 board members, the administrative bodies are made up of 10 diff erent individuals. Eight of them have diff erent pairs of surnames, suggesting that 2 of the 10 are sib-lings. Moreover, 6 of the 10 members show no surname concurrence with any of the other members, leading us to believe that there are 2 individuals of the remaining 8 who have a family relationship that we have generically listed as ‘cousins’.

These median measures of dispersion for the family group are not pro-portionally very diff erent from the values presented by the average, which can be seen in Panel D of Table 12.10. In fact, every 10 seats on the boards of directors are covered by a little more than 4 individuals (average 41 per cent proportion). More than 3 of those individuals are not siblings (average 33 per cent proportion), and nearly 2 of them have no family relationship that can be inferred from the surname (average 23 per cent proportion).

Panel D of Table 12.10 also shows us the existing diff erences between the small family groups and the larger ones, as regards the proportion of board members that share family ties. The small family groups incorporate greater proportions of individuals without sibling or ‘cousin’ family ties onto their boards of directors. These proportions of outsiders to the family are considerably lower in the large family groups, despite the fact that the latter tend to incorporate a larger number of non-family members on their boards, by virtue of the larger numbers of board seats.

5.4 Sector Diversifi cation and Family Control

There is one last aspect that allows us to fi nd relations between two of the points that we have discussed thus far: on one hand, diversifi cation as a


relevant aspect of the family group policy, and on the other, the degree to which the boards of directors are open to include board members from outside the family.

Here, the relevant question is whether the family groups that opt for greater sectorial diversifi cation do so through tighter family control over the governing bodies or whether they rather tend to be more open to the inclusion of outsiders on their boards of directors. There are numerous theory-based arguments in favour of and against family control of diver-sifi ed companies. Family expansion in new business areas may require the incorporation of new members, whether fi nancial (who contribute monetary resources) or technological (who bring in new know-how or organizational skills in order to carry out new activities). For a given size of the family that controls the business group, embarking on new business ventures can depend on the family members’ management skills or knowl-edge of the new sector. If these requirements are fulfi lled, one could expect greater degrees of diversifi cation to be accompanied by a larger number of members from outside the family, and the proportion of family members on the boards of the companies within the group would be lower than on those of the groups that do not diversify.

On the other hand, the expansion into new business areas can also mean the potential loss of family control over the activities of the group of com-panies. The incorporation of new executive directors or managers of new business areas may require greater board supervision and control, which will be exercised by appointing family members as board members. In these situations, the mechanisms of trust are what justify a strong family pres-ence to prevent the loss of control over the important decisions made in the group’s companies. If this eff ect is prevalent, we might anticipate that the higher the degree of diversifi cation, the higher the rates of family members on the boards of directors in the family business groups.

Table 12.11 shows the average values and proportions of the groups’ openness to non-family members on the boards of directors, and divides the business groups according to their overall diversifi cation level (meas-ured by means of the entropy index). The results of Panel A are clear: greater diversifi cation goes with a larger number of diff erent individuals as well as a larger number of sibling board members. The cousin-relation trend follows a similar pattern: the greater the sectorial diversifi cation, the larger the number of individuals with no surname concurrence. This trend is explained by the size of the business group: greater sectorial diversifi ca-tion also occurs with a higher number of group companies. This necessary opening of the company to incorporate new talent or new partners into the diversifi ed activities is explained by how they complement the family administrators.


Panel B of Table 12.11, nevertheless, shows that as the business groups diversify more, they become less open to non-family members. Indeed, the proportion of non-sibling board members in relation to the total number of board members for the group is 45 per cent for the family groups with more limited sectorial diversifi cation, whereas for more diversifi ed groups the rate is only 34 per cent. A similar result is seen when the type of family relation taken into consideration is limited to one surname. In such cases, the less diversifi ed groups display a 36 per cent proportion of ‘non-cousin’ individuals, whereas the more diversifi ed groups only have a 21 per cent dispersion rate, suggesting that the remaining 79 per cent of the members share a surname.

This study can lead to the conclusion that the more diversifi ed business groups incorporate a larger number of new talents into their boards of directors, and that such new individuals do not necessarily share any family relations with already-existing members. Nevertheless, the control eff ect is also prevalent, as the proportion of non-family members brought into the group is increasingly lower in relation to the members that share some sort of family tie.

Table 12.11 Opening of family groups to non-family board members, for diff erent degrees of kinship, according to the degree of sectorial diversifi cation

Panel A: average values Dispersion of board members (number)

Sectoral diversifi cation

Number of companies per group

Board members




Low 3.25 10.69 7.13 4.81 3.88Medium 6.88 24.18 11.12 8.65 6.59High 22.76 93.59 35.29 30.29 19.53Overall 11.12 43.46 18.06 14.78 10.12

Panel B: average proportions Dispersion of board members (proportion)

Sectoral diversifi cation




Low 0.67 0.45 0.36Medium 0.46 0.36 0.27High 0.38 0.32 0.21Overall 0.42 0.34 0.23


6. CONCLUSIONS

This chapter presents a general description of the business groups of fami-lies that belong to the ABEF and IEF associations in the Balearic Islands. The fi rst thing to be verifi ed, their economic relevance in the general level of economic activity of the Balearic Islands, has become evident when we evaluate their weight through aggregate measures such as total regional employment and GDP.

We have presented a description of the organizational structure of the companies that belong to the associated families, which are analysed indi-vidually. Organizational aspects such as the size of the boards of directors and the most relevant economic information from their annual accounts reveal large disparities in activity sectors and company size. The relatively high presence of women in the governing bodies of the companies (by Spanish standards) is also worthy of note.

The 566 sample companies are not intended to be a representative sample of the economic activity of the Balearics. However the distribution of their activities in the diff erent sectors, and particularly those directly or indirectly associated with tourism, are not signifi cantly diff erent from the information of the aggregate offi cial statistics.

We also have analysed the characteristics of the business groups that have been formed around the 50 member families studied. Once again, the diversity of the sizes has allowed us to undertake a homogeneous comparison after grouping them according to their asset volume. These groups of companies are formed around one or two central companies, which generate nearly 70 per cent of their activity, despite the fact that the average number of companies per group is eleven. The largest family business groups present higher degrees of sectorial diversifi cation, although it is the smallest groups that base a larger proportion of their diversifi cation on activities that are not related to their main core busi-ness activity.

The study of the governing bodies of the family groups reinforces the idea of family control in the sense that these groups rely on a small group of people to serve as board members in their companies, and the types of family relations among them are diverse. This phenomenon is more pro-nounced in the largest family groups: by having larger boards of directors, the number of diff erent people that they rely on is also larger. However, these larger groups also show a larger proportion of board members that share family ties than the smallest family groups do.

The trust eff ect inherent in the inclusion of offi cers that share family ties is predominant over the entry of non-family member offi cers for the com-panies with the highest rates of sectorial diversifi cation, demonstrating that


diversifi cation strategies are possible while keeping the family structure of the business groups intact.

Future studies could enhance the description of these family groups by following two basic approaches. The fi rst would be to gradually incorpo-rate new sources of data that would allow us to describe in greater detail the elements discussed here and explore new aspects of the organization of the family groups. For example, it would be interesting to look into the requirements for the degree of participation of women on the boards of directors of these companies. Could such participation be contingent on the founding generation’s continued control of the business group or on successive generations being incorporated into it? We have taken an aggre-gate approach to family relations, using public information. A detailed study of the specifi c posts held by the family members and outsiders might clarify the concept of family control, while enabling us to ascertain the true role of the women in this mechanism of control. Secondly, here we have off ered a static description of these family business groups. It would be interesting to follow their evolution through time, which would allow us to examine the determining factors of the groups’ business organization deci-sions in greater detail, their diversifi cation strategies, their organizational structure and even their profi tability.

NOTES

1. See Shanker and Astrachan (1996) and Sharma (2003) for a discussion and diff erent defi nitions of family fi rms, based on the degree of family involvement in the fi rm. Another literature develops a typology of family fi rms along the potential combinations of three axes: the ownership of the fi rm, the family structure and the characteristic of the company (see Gersick et al., 1997; Neubauer and Lank, 1998).

2. Indeed, it has been argued that one main non-pecuniary benefi t for family members to own and control a fi rm is the satisfaction of transferring the fi rm to the descendants (see Casson, 1999).

3. The law has recently been modifi ed in Spain, so that the order of surnames can be changed: fi rst the mother’s surname, and then the father’s surname. This change can be done by mutual agreement of both parents, or by the choice of the concerned individual when he/she reaches the age of majority (18 years). Until now this modifi cation has had almost no practical impact on the structure of surnames in Spain.

4. In any case, the reader will have to bear in mind that the data on the companies studied do not exclusively refer to the economic activity or employment generated in the Balearic Islands. Indeed, when a hotel chain of an ABEF member, for example, has establishments in Mexico, and its earnings are calculated within a Spanish company that is included in the database that we have used, such economic activity is under the control of the member families of the ABEF and is therefore taken into account in our study of the economic activity of the family businesses under control of ABEF-member families, although it is not an economic activity in the Balearic Islands. Similarly, our study calculates the activity of the Banca March, where it is a well-known fact that this insti-tution has a large number of offi ces outside of the Balearics that generate employment


and economic activity outside of the Islands. By our defi nition, this is a new economic activity under the control of a Balearic family business.

5. Let us recall again that we must proceed with caution when drawing these comparisons, as the sample includes companies under the control of the families of family business associations in the Balearic Islands, regardless of where their activity is carried out, and in some cases a signifi cant part of such activity is undertaken outside of the Islands.

6. Actually, the middle tertile is made up of 186 companies. 7. The univariant study presented does not allow us to reach any conclusions or explana-

tions for the non-monotonic behaviour of a variable, although we could make some speculations on this signifi cant fact.

8. These also include those of the group Globalia. 9. Section 3 above explains the procedure we follow to assign the companies to Balearic

ABEF- and IEF-member families.10. If, rather than measuring the importance in relation to the total activity volume, the

average relative importance of the largest company is measured and expressed in the percentage of each family group, such percentage goes up to 56 per cent.

11. There is a large literature that tries to evaluate the possible benefi ts of diversifi cation for fi rms; see Campa and Kedia (2002), Grant et al. (1988), Hadlock et al. (2001) and Villalonga (2004). For the particular case of family fi rms, see Anderson and Leeb (2003).

12. Anyway, let us mention that, for the family fi rm, Anderson and Reeb (2003) fi nd that family fi rms diversify less than non-family fi rms.

13. The use of the entropy index for measuring diversifi cation is based on work by Jacquemin and Berry (1979). For further details, see Appendix 12.1, which displays the breakdown of the entropy index.

14. For a group of 11 companies with 4 board members each, the minimum number of people to whom the administration would be entrusted in this case would be 4.

REFERENCES

Anderson, R.C. and D.M. Reeb (2003), ‘Founding-family ownership, corporate diversifi cation, and fi rm leverage’, The Journal of Law and Economics, 46, 653–84.

Campa, J.M. and S. Kedia (2002), ‘Explaining the diversifi cation discount’, The Journal of Finance, 57 (4), 173–62.

Casson, M. (1999), ‘The economics of the family fi rm’, Scandinavian Economic History Review, 47, 10–23.

Gersick, K.E., J.A. Davis, M. Hampton and I. Lansberg (1997), Generation to Generation: Life Cycles of the Family Business, Boston: Harvard Business School Press.

Grant, R.M., A.P. Jammine and H. Thomas (1988), ‘Diversity, diversifi cation, and profi tability among British manufacturing companies’, Academy of Management Journal, 31, 771–801.

Hadlock, C., M. Ryngaert and S. Thomas (2001), ‘Corporate structure and equity off erings: are there benefi ts to diversifi cation?’, Journal of Business, 74 (4) 613–35.

Jacquemin, A.P. and C.H. Berry (1979), ‘Entropy measure of diversifi cation and corporate growth’, Journal of Industrial Economics, 27, 359–69.

Neubauer, F. and A.G. Lank (1998), The Family Business: Its Governance for Sustainability, London: Macmillan.

Shanker, M.C. and J.H. Astrachan (1996), ‘Myths and realities: family businesses’


contribution to the US economy: a framework for assessing family business statistics’, Family Business Review, 9 (2), 107–23.

Sharma, P. (2003), ‘Stakeholder mapping technique: toward the development of a family fi rm typology’, mimeo, Laurier Business & Economics.

Villalonga, B. (2004), ‘Diversifi cation discount or premium? New evidence from the business information tracking series’, Journal of Finance, 59 (2), 475–502.


APPENDIX 12.1 THE ENTROPY INDEX

This index is used as a measurement of diversity. In our case, it allows us to assess the extent to which a family group diversifi es its activity in diff erent economic sectors.

If Pij is the proportion of the assets of a given business group in activity i (NACE 3-digit code) in industry j (NACE one-digit code), the entropy index for the total diversifi cation of this business group is calculated as follows:

Total diversification 5 DT 5 am

j51a

n

i51[Pij 3 ln(1/Pij) ] for Pij 2 0

Jacquemin and Berry (1979) propose the following breakdown of the total diversifi cation into related and unrelated diversifi cation:

Related diversification 5 DR 5 am

j51Pja

n

i51c aPij

Pjb 3 lnaPij

Pjb d

Non-related diversification 5 DNR 5 am

j51Pj 3 lnaPij

Pjb

It must be noted that these two measurements and the measurement of total diversifi cation are consistent in the sense that the measurement of the entropy of the total diversifi cation is the sum of the related and unrelated diversifi cation, DT 5 DR 1 DNR.

Reference

Jacquemin, A.P. and C.H. Berry (1979), ‘Entropy measure of diversifi cation and corporate growth’, Journal of Industrial Economics, 27, 359–69.

323

13. Better fi rm performance with employees on the board?R. Øystein Strøm*

1. INTRODUCTION

This chapter deals with the impact of co-determination1 upon fi rm per-formance. Two confl icting views on the benefi ts of co-determination exist. One says that co-determination increases fi rm performance, either because employee directors supply outside directors with information they would otherwise not have access to (Freeman and Reed, 1983; Blair, 1995), or because co-determination is a safeguard against dismissal, inducing employees to invest in fi rm-specifi c human capital (Zingales, 2000; Becht et al., 2003). The other view is that owners’ and employees’ interests are not aligned, and therefore allowing employees into the boardroom means that confl icting goals are pursued. When decision makers with diff erent objectives share in the board’s decisions, its focus may become unclear (Tirole, 2001), its decision time longer (Mueller, 2003), and its decision quality inferior.2 The prediction is that fi rm performance will be lower than it could otherwise be.

Even though co-determination is important in many European coun-tries,3 few fi rm-level studies have been made of its fi rm performance impact. This chapter is an attempt to bring more academic research to the still under-researched (Goergen, 2007) comparison of fi rm perform-ance in shareholder determined companies and co-determined companies. Earlier studies give mixed results, showing a negative impact in German fi rms (Fitzroy and Kraft, 1993; Schmid and Seger, 1998; and Gorton and Schmid, 2000, 2004), Canadian (Falaye et al., 2006), and Norwegian (Bøhren and Strøm, 2008), but a positive impact in a later German study (Fauver and Fuerst, 2006).

Compared with former literature, the simultaneous equation estimation of the relationship between fi rm performance and explanatory variables is the distinctive feature of this chapter. The need for simultaneous modelling arises from the fact that the presence of employee directors may induce shareholders to adjust other governance mechanisms, notably board


composition and leverage, in order to neutralize the co-determination eff ects (Buchanan and Tullock, 1962). Employee directors may have a direct impact upon fi rm performance, but also an indirect eff ect. This also means that board composition is at least partly determined by employee directors. Thus, the chapter necessarily also relates to the board endogene-ity issue (Hermalin and Weisbach, 2003). And since the data cover several periods, it is possible to test the reverse causation hypothesis that fi rm performance determines board composition (Hermalin and Weisbach, 1998). The simultaneous equation setup allows not only the discovery of endogeneity in governance mechanisms, but also a quantifi cation of its importance compared with direct eff ects. I am unaware of former literature containing a measure of the endogeneity eff ects.

The chapter’s results come from a panel data set of non-fi nancial fi rms spanning the 14 years from 1989 to 2002, containing fi nancial information, data on ownership, and board composition data. Employee representation was mandated by law in 1972 in Norway, and regulations have remained almost unchanged since (Aarbakke et al., 1999). The data on employee directors seem to be superior to those pulled from German and Canadian institutional frameworks. While the employee director in a German board may be elected from the national labour union, and the Canadian evidence is from fi rms where employees have considerable shareholdings, in Norway the employee director must be employed in the company. Furthermore, because the mandatory employee director rules only apply to certain fi rms, some fi rms have employee directors, others have none. Thus, the study avoids the Dow (2003, p. 87) objection that empirical investigations on the eff ects of employee directors suff er from a lack of control group. Thus, unlike previous studies, the Norwegian institutional framework allows comparison between similar fi rms with and without employee directors. This setting allows for sharper estimates of the co-determination eff ects.

The chapter has relevance for the emerging regulation literature on boards (Hermalin, 2005). Because the sample includes both the co- determined (by regulation) and the shareholder determined kind, I can study the eff ects of governance regulation by comparing the two sub-samples.

Compared with the related Bøhren and Strøm (2008) study, I introduce a number of new features. I construct a board structure index that cap-tures many standard board characteristics in the same manner as Bertrand and Mullainathan (2001), I add fi nancial leverage and average wage as new explanatory variables, I perform a system estimation rather than a single-equation estimation, and I make separate regressions for various sub-samples, for instance employee director fi rms only. These steps should yield better estimations of the employee director impact than the partial regressions in Bøhren and Strøm (2008), and should also subject the

Better fi rm performance with employees on the board? 325

co-determination hypothesis to more severe robustness tests. Furthermore, I confi rm their results when using individual board characteristics instead of the board index.

In order to fully utilize the information in the panel data, I use the fi xed eff ects model (Woolridge, 2002), employing a three-stage least squares (3SLS) methodology in system estimations. With the fi xed eff ects method, I am able to remove fi rm heterogeneity, as did Palia (2001). Therefore, few (if any) control variables are needed.

Using Tobin’s Q as the measure of fi rm performance, the results confi rm the employee directors’ negative relationship to fi rm performance in earlier studies, but also show a positive indirect eff ect on the board index and leverage. This refl ects endogeneity, but the economic signifi cance of the indirect eff ects turn out to be much smaller than the direct. The reverse cau-sation hypothesis also fi nds confi rmation, since lagged fi rm performance is signifi cant regarding both the board index and leverage. But again, the indirect eff ects of the lagged fi rm performance are low compared with the direct. I fi nd clear diff erences in the various board characteristics’ impact upon fi rm performance in sub-samples of co-determined and shareholder-determined fi rms. This means that regulations have costs, both in relation to fi rm performance and in the remaking of boards. The results stand up to a number of robustness tests, including alternative performance meas-ures (stock return and accounting return on assets), and also to dividends replacing leverage.

The chapter proceeds as follows. In the next section, a brief review of the literature is given. Then, in Section 3 testable implications are spelled out. Section 4 contains data sources and institutional background, while Section 5 discusses estimation methodology. Then Section 6 shows results, in Section 7 robustness checks are undertaken, and Section 8 concludes.

2. LITERATURE REVIEW

Few empirical studies of co-determination have been undertaken. Evidence in Fitzroy and Kraft (1993), Schmid and Seger (1998), and Gorton and Schmid (2000, 2004) shows that co-determination has a negative economic eff ect upon fi rms in Germany, where employees have the right to equal representation in the Aufsichtsrat with shareholders. Recently, Fauver and Fuerst (2006) fi nd a positive relationship to performance in a 2003 sample of German companies in information intensive industries. In the regressions with all industries, however, the relationship is not signifi cant. The German data often contain two kinds of employee directors, some elected from among the employees in the company and others, national


union representatives. In contrast, the Norwegian system is such that only persons employed in the company may be elected. Thus only a company and not a national union representative may sit on a Norwegian board. Presumably, company employed persons are more authentic stakeholders than their national union representatives.

Using Canadian data, Falaye et al. (2006) fi nd that fi rms giving employ-ees a greater voice in corporate governance spend less on new capital, take fewer risks, grow more slowly, create fewer new jobs, deviate more from value maximization, show greater cash fl ow problems, and exhibit lower labour and total factor productivity. This chapter is set in a diff erent institutional environment. The Canadian employee directors are elected in their capacity as owners of company shares. The infl uence of these direc-tors on fi rm performance thus picks up two eff ects, one as a supplier of labour services, the other as owner. None of these studies use panel data or simultaneous data estimations.

Using Norwegian data, Bøhren and Strøm (2008) show that the employee director variable has a negative impact upon Tobin’s Q. They also fi nd evidence of interdependencies among board characteristics. However, they do not explore the indirect eff ects of co-determination, nor do they carry their analyses into sub-samples of co-determined and share-holder determined fi rms. In this chapter, the analysis is extended to include eff ects upon average wage, leverage is a new governance variable, and the board index is defi ned; I employ simultaneous equations modelling, and perform regressions in sharply defi ned sub-samples. The robustness tests are also more extensive, as I use return on assets and stock return as new dependent variables, and also vary the defi nition of the board index.

3. THEORY AND HYPOTHESES

3.1 Stakeholder or Interest Group?

Board decisions include the formulation and control of strategy, larger investments and disinvestments, and the determination of the company’s organization. Employee directors’ infl uence upon these decisions may have long-time impact upon fi rm performance. Therefore, an analysis over a long period of time is needed to detect the eff ects. The impact upon fi rm performance could be positive, non-signifi cant, or negative.

One possibility is that the fi rm performance and employee link is posi-tive. Blair and Stout (1999) view stakeholders as members in a team pro-duction. Since stakeholders make fi rm-specifi c investments, it is in their interest to co-operate. The fi rm-specifi c human capital investments make


the employees residual claimants to much the same extent as shareholders (Zingales, 2000; Becht et al., 2003). The upshot is that employees should be represented on the board, and that this co-determination will lead to improved fi rm performance. The conclusion rests on the argument that the stakeholders’, including the shareholders’, interests are aligned.

How could this be manifested in the board? Employee directors could have a dual informational role in bringing inside information to the board (Blair, 1995, p. 16) but also relating board information to the employees (Freeman and Lazear, 1995). Since employees are in the middle of the day-to-day running of the company, they may bring valuable operational knowledge to the board. The information may expand on or contrast with information from the CEO. Thus, the information set available to the outside directors is enlarged. This comes close to viewing the employee director in the same role as the insider in the Raheja (2005) theory, although in this model the insider is willing to furnish the outside directors with information only if this furthers his own career interests. Secondly, the role as messengers of board information to the employees at large could be of particular value in the case of personnel reductions or plant closures, when the board may want to instil an understanding for the need for drastic measures among employees. The dual informational role of employee directors should lead to better fi rm performance (Fauver and Fuerst, 2006).

Another possibility is that co-determination has no signifi cance for fi rm performance. This may come about through co-optation (Pfeff er, 1981, pp. 166–73). In the board employee directors are exposed to fi duciary duties and conformity pressures to accept the shareholder value logic. Also, since the employee director is made co-responsible for decisions with adverse outcomes for employees, the decisions carry higher legitimacy among employees. If co-optation is the case, interests are again aligned, but this time because employee directors have taken on the views of shareholders. The eff ect upon fi rm performance should be non-signifi cant.

The third possibility is that the co-determination impact upon fi rm performance is negative. It may be hard to accept the premise that stake-holder interests are aligned. If this were so, co-determination would be an effi cient economic organizational mode, and fi rms would adopt this mode voluntarily (Jensen and Meckling, 1979; Hansmann, 1996). But while shareholders seek to maximize residual income, employees want to maximize pay and the protection of fi rm-specifi c human capital,4 that is, a part of the residual income. The inconsistency of these two objectives makes the board decision process longer and more diffi cult (Mueller, 2003). The fi rm’s objectives may become unfocused, and the CEO may develop capabilities as a compromise maker rather than a shaper of the fi rm under a


clear objective (Tirole, 2001, 2002). The implied consensual decision model in co-determination means that the fi rm pursues stability and predictability instead of bold new moves (Siebert, 2005). If employee directors are suc-cessful, they should infl uence the average wage positively. The unfocused decision structure should result in weaker fi rm performance. I call this the interest confl ict model for reference, and hypotheses stemming from the model are set forth in the next section.

When objectives diverge, shareholders and employees may game against each other so as to further their own interests. Employees may furnish information strategically to further their own interests (Pistor, 1999; Hopt, 1998), and they may use moral arguments in parallel. Information strategizing could take the form of economizing on the supply of internal information to the board. For instance, employee directors may not inform of low productivity units in the organization. Another form could be infor-mation leakage from the board.5 Employee directors will hardly inform their fellow workers only on matters that owners and management fi nd in their interests to inform about. Stakeholder theorists seem to assume only benefi cial information dissemination through employee directors. Furthermore, moral arguments against, for instance, plant closures or high management pay may be put forward, too. The shareholder elected direc-tors may have trouble withstanding such arguments, since they may experi-ence large personal costs and small personal gains from making decisions that aff ect employees adversely (Baker et al., 1988). Taking the issue to the public attention could make the decision even harder for the shareholder elected directors. Thus, even though the employees are in a minority posi-tion in the board, they may infl uence board decisions to their advantage. Their access to board information seems to be vital in this respect.

But the presence of employee directors may have indirect eff ects upon the use of other governance mechanisms as well. Shareholders may adjust governance mechanisms in order to neutralize the co-determination impact. This is analogous to the situation Buchanan and Tullock (1962) point out, that when an exogenous regulation is imposed upon a (political) committee, it will try to compensate for the regulatory eff ect by placing a heavier weight on the unregulated. These previously unexplored indirect eff ects make a simultaneous equations approach necessary. In the remainder of this section governance mechanisms and hypotheses about interactions are explained.

3.2 Simultaneity and Endogeneity

In a simultaneous equations system some variables are endogenous, others exogenous. In the present setup, the exogenous variables are the fraction of employee directors, the lagged fi rm performance, the fi rm size, and fi rm


risk. Since employee directors are imposed from outside the fi rm, they must constitute an exogenous variable. These variables determine fi rm perform-ance and average wage, but also the intervening governance variables, the board characteristics and leverage. Thus, the intervening governance vari-ables and the average wage are at least partly determined by the employee directors and lagged fi rm performance. The simultaneous setup gives the researcher the opportunity to recognize the governance variables’ endo-geneity, but at the same time also to measure the magnitude of the eff ect relative to their direct eff ects upon fi rm performance.

Specifi cally, the co-determination hypothesis says that the mechanism of employee directors has a negative relationship to fi rm performance, but a positive one to average wage, the board characteristics, and leverage. The reverse causation hypothesis says that lagged fi rm performance is associated with governance variables and average wage, but that signs are uncertain. The remainder of this section concerns explanations of variables and their relationships.

In this chapter, shareholders may adjust the board characteristics and the leverage. In order to achieve a reliable measure, and in the interest of economy, I build an index by including board characteristics that have proven to be important in board studies. The board index BI is:6

BI 5 DH 1 BN 2 BS 2 G (1)

DH is directors’ holdings, BN is the board network, BS is board size, and G is gender. The board index construction follows the Bertrand and Mullainathan (2001) procedure, as each index variable in equation (1) is standardized to have average zero and standard deviation 1 before sum-mation. The sum is then standardized. This gives a continuous variable, in contrast to the Gompers et al. (2003) type of index. Their governance index is based upon a subjective allocation of categorical points for reasons that restrict shareholder rights, and then summed over all characteristics. Since all variables in equation (1) are continuous, the resulting index is continu-ous as well, and this is an advantage in estimations. Another advantage is that the index is likely to be more stable in sub-samples than the individual variables. The interpretation is that the higher the board index, the better is the board structure. It should be positive towards fi rm performance and negative towards average wage. If it is complementary to leverage, a posi-tive sign will appear.

The choice of variables in the index refl ects important board character-istics that are decision variables for shareholders. Directors’ ownership represents the need for the board to be aligned with shareholders, the network variable the need for the board to be informed, the board size


and gender diversity the need for the board to be decisive. The signs in are common fi ndings in the literature. The ownership literature (Morck et al., 1988; McConnell and Servaes, 1990) confi rms the positive sign on directors’ ownership share, and so do studies taking other board character-istics into account, for example, Bøhren and Strøm (2008).7 The network variable is little used in studies of boards, but Bøhren and Strøm (2008) fi nd a positive sign.8 It comprises direct and indirect connections to other listed non-fi nancial fi rms stemming from directors’ multiple board seats. A variety of studies, such as Yermack (1996) and Eisenberg et al. (1998), document that performance decreases with increasing board size. The rela-tionship between gender and fi rm performance may be more controversial, as Shrader et al. (1997), Smith et al. (2006), and Bøhren and Strøm (2008) report a negative relationship, whereas Carter et al. (2003) fi nd the oppo-site. I perform robustness tests with other defi nitions, described in Section 4, to test the choice of index.

Next, I include leverage. A higher leverage will decrease the fi rm’s free cash fl ow, and will, therefore, limit the potential for agency costs (Easterbrook, 1984; Jensen, 1986). Perotti and Spier (1993) model how the lower free cash fl ow may be used as a bargaining tool against employees, implying better fi rm performance and lower average wage. Both eff ects should point to higher fi rm performance from higher leverage. However, the complexity of leverage leads to an indeterminate prediction. On the one hand, given the presence of employee directors, owners may fear higher debt may bring even higher decision costs. If, as Easterbrook (1984) supposes, higher leverage brings the lender into closer oversight of the fi rm, the fi rm may end up with three decision makers with potentially divergent interests. Furthermore, if the leverage is also used to signal investment prospects (Myers, 1977), a high leverage used to discipline employees can be taken to signal weak investment opportunities in the fi rm. Another aspect is that, as Tirole (2006, pp. 51–3) points out, higher leverage may cause costs related to illiquidity and bankruptcy. This complexity of leverage means that the sign is uncertain. It could be the case that shareholders in co-determined fi rms adjust the leverage in an eff ort to neutralize employee directors to a greater extent than they do in shareholder determined fi rms. In a simulta-neous equations setup, Brick et al. (2005) fi nd a negative relationship.

Thus, I expect employee directors to be associated with better board composition and higher leverage. If these are successful from the share-holder point of view, a positive indirect eff ect may compensate for the negative direct employee director eff ect upon fi rm performance. In the stakeholder theory, the employee director should be a welcome addition to the board, and thus carry a positive sign to fi rm performance, while the indirect eff ects should not appear.


In addition to the endogeneity induced by employee directors, the reverse causation hypothesis says governance mechanisms may be at least partly determined by past performance (Hermalin and Weisbach, 1998, 2003). The signs on the board index and the leverage may be diffi cult to set out. In the Hermalin and Weisbach (1998) bargaining model the CEO bargains over pay and monitoring intensity. Good past fi rm performance gives the CEO a better bargaining position, which he will use to reduce monitoring. This means that the association between past fi rm perform-ance and governance mechanisms should be negative. However, it may well be that governance mechanisms are improved after a good perform-ance, for instance, since the fi rm learns good practices. Since shareholders may adjust either board composition or leverage, or both, leverage and board composition may be either complements or substitutes (Agrawal and Knoeber, 1996). Thus, the sign is ambiguous.

I study the direct and indirect eff ects of employee directors in a simul-taneous setup. Since the lagged fi rm performance is included, the system is dynamic. Taken together, and with constants suppressed, this results in the system of equations

(2)

where FP is fi rm performance, and FPt21 indicates one period lag; W stands for the average wage, BI is the board index, DE is the leverage (debt to equity), ED is employee director, FS is fi rm size, FR is fi rm risk, and uit is the error term. The main hypotheses are summarized below the coeffi cients. Thus, the co-determination hypothesis is set out for the ED variable.

4. DATA AND INSTITUTIONAL BACKGROUND

The sample comprises all non-fi nancial fi rms listed on the Oslo Stock Exchange (OSE) at year-end at least once during the period 1989 to 2002.9


Board data are collected from the handbook Kierulfs Håndbok for the fi rst years, and from the national electronic register at Brønnøysund from 1995. The register provides information on name, date of birth, and director status (chairman, vice-chairman, ordinary member, and employee direc-tor). The CEO’s name and date of birth are recorded as well. The CEO or director name gives gender information. Data on board and CEO owner-ship, as well as outside ownership concentration, are pulled from the public securities register, while share price and accounting data come from OSE’s data provider (Oslo Børs Informasjon). The ownership structure data cover every equity holding by every investor in each sample fi rm. By interna-tional standards, the size and quality of the data are considerable.

The data for this chapter span the period from 1989 to 2002. During this period, the law regulating the governance of the companies is from 1972, with amendments in 1987 (‘Aksjeloven’), and a new law in 1997 (‘Allmennaksjeloven’). The regulations for representation have been unchanged since 1987. In this respect, there is no before-and-after situation, as with the Cadbury Committee (1992) report in the UK, in the sample period.

As a general rule, fi rms with more than 200 employees must have at least two employee directors, or at least one-third of the board.10 In the size bracket 31 to 200 employees, the fi rm must have labour board seats if a majority of the employees vote in favour, fi rst with one representative in the 31 to 50 bracket, then two in the 51 to 200. The employee director must be employed in the company. A number of important Norwegian indus-tries are exempted from these rules, that is, the employees have no rights of representation in these industries. These include newspapers, news agen-cies, shipping, oil and gas extraction and fi nancial fi rms. The characteris-tics of employee board representation mean that some fi rms have employee directors, others do not, and also that co-determined fi rms have diff erent fractions of employee directors. Thus, an implication of the regulations is that comparisons of two sets of diff erently governed but otherwise similar fi rms can be made, and that further analyses can be carried out in sub-samples of, say, co-determined fi rms with more than 200 employees. This data property answers the Dow (2003, p. 87) objection that the study of co-determined fi rms lacks a proper control group. I defi ne the employee director variable as the fraction of employee directors, unlike most former studies, which only use employee directors as a dummy variable.

This institutional framework off ers advantages over the German and Canadian studies referred to in Section 2, since the Norwegian employee directors represent an authentic stakeholder group. The German regula-tions are such that one-third of the employee representatives on German boards need to be labour union offi cials (Siebert, 2005). Presumably, the


union offi cials are supposed to look after the interests of workers in general, not only those in the fi rm. No such minimum is required in Norway, and the employee directors need to be employed in the fi rm. The Canadian co-determination comes about when workers are also shareholders in the company. This might cause a confl ict of interests, when the optimal policy from the shareholder point of view is detrimental to the optimal policy for workers. In Norway, employee directors are elected in their capacity as workers in the fi rm, not through their shareholdings.

The initiative for employee representation came from a joint committee of the Labour Party and the major employee union (LO) in the early 1960s. However, concurrent with this initiative, LO and the employer associa-tion (NAF) ran a ‘co-operative project’ together with researchers to study co-determination in selected companies. This was in the consensus and co-operation spirit that arose from common war-time experience. The ques-tion was not only about co-determination, but also about new production methods. Later, the need for co-determination in order to improve pro-ductivity was the guiding principle of the offi cial document NOU (1985:1), whose recommendations were unanimous, as opposed to the original 1971 report. The insider information argument was behind the codifi cation of employee board representation in Norway. Thus, it seems as if the lawmak-ers were familiar with stakeholder theory. Bråthen (1982, p. 14) interprets the law on co-determination to imply that profi t maximization is no longer the single objective of the company. Employees’ interests now become one of several objectives the fi rm has to consider. Thus, a harmony of interests model is behind the regulations on co-determination in Norway.

Next, I report some descriptive statistics on employee directors. Table 13.1 shows the number of employee directors in fi rms according to employment size. The table shows the percentage of fi rm-year observa-tions of employee directors in various employment sizes. It turns out that in fi rms where employees may demand representation, few do so. In the 101–200 employees category, 61.5 per cent do not have employee directors. Furthermore, in the highest category, where representation is compulsory if the industry is not exempted, employees have no board seats in about one-third of the companies. Among the fi rms that do have employee direc-tors, the legal minimum, two representatives, is found in the majority of cases. Very few have four employee board seats. Thus, the Jensen and Meckling (1979) conjecture that co-determination requires law backing seems to be supported in our Norwegian data.

Next, Table 13.2 shows the distribution of employee directors according to industry, and also the percentage of fi rms with no employees on the board in each year. Exempted industries such as Energy and Transport (including shipping) have no employee directors to a higher degree than average. The


low representation in Hotels, restaurants and entertainment is perhaps due to high labour turnover. The two industries Health care equipment and sup-plies and Software and supplies also have a lower than average representa-tion. These are industries where the human capital element should be above average, and co-determination of extra value, according to stakeholder theory. Yet, obviously, employees do not demand board seats to a great extent. The time trend is that fi rms with no employee directors increase in relative importance. Thus, nothing in the overall descriptive statistics shows that co-determination is a preferred organizational mode. Firms seem to avoid it if they can, and keep it to a minimum if they cannot.

Variable defi nitions are shown in Table 13.3, which also presents the main characteristics of variables in the analysis in the two main sub- samples of co-determined and shareholder determined fi rms. The table shows that a large number of variables are distributed diff erently in the two sub-samples. The fi rm performance variables Tobin’s Q and stock return are not signifi cantly diff erent, while the ROA in co-determined fi rms is signifi cantly higher than in shareholder determined fi rms. Apart from directors’ holdings, all other variables are signifi cantly diff erent at the 5.0 per cent level or better. Obviously, the two types of fi rms are diff erent.

The table shows that the fraction of employee directors is 0.301, or slightly below the minimum requirement for the 2001 employee size group.

Table 13.1 The percentage of fi rms with zero or more employee directors by employment size

Employees Employee directorsN 0 1 2 3 4

0–30 98.4 0.5 0.5 0.5 0.0 19031–50 95.8 2.1 2.1 0.0 0.0 4851–100 73.5 5.3 18.6 2.7 0.0 113 101–200 61.5 4.5 24.4 9.6 0.0 156200+ 33.5 8.1 30.0 27.1 1.3 1006 Total 49.5 6.3 24.0 19.3 0.9N 749 96 363 292 13 1513

Note: The table shows the percentage of fi rms having employee directors, according to employment categories. N is the number of fi rms in the employee directors or the number of employees category. The number of employees category refl ects the regulations on co-determination (Aarbakke et al., 1999). With more than 200 employees co-determination is compulsory. In the 31 to 200 bracket co-determination is realized if an employee majority demands it, with a larger proportion of representation with a larger workforce. In all categories, including the above 200 employees, fi rms in some industries are exempted from the rules.

335

Tab

le 1

3.2

The

per

cent

age

of fi

rms w

ith z

ero

or m

ore

empl

oyee

dire

ctor

s by

indu

stry

and

the

perc

enta

ge w

ith z

ero

by y

ear

Indu

stry

Em

ploy

ee d

irect

ors

% o

fto

tal

N

Yea

r %

no

empd

ir N

0

12

34

Ene

rgy

77.7

1.1

8.2

12.7

0.3

16.

0 3

5419

8950

.5

95

Mat

eria

ls17

.8 8

.539

.534

.1 0

.0

5.8

129

1990

49.5

99

C

apita

l goo

ds34

.5 2

.835

.327

.0 0

.4 1

1.4

252

1991

48.4

93C

omm

erci

al se

rvic

es49

.4 8

.925

.316

.5 0

.0

3.6

79

1992

45.3

95T

rans

port

77.1

5.8

5.6

11.6

0.0

18.

8 4

1419

9348

.4

91

Aut

os a

nd c

ompo

nent

s 0

.0 4

.369

.626

.1 0

.0

1.0

23

1994

52.4

1

03C

onsu

mer

art

icle

s, cl

othe

s24

.018

.048

.010

.0 0

.0

2.3

50

1995

61.3

1

86H

otel

s, re

st.,

ente

rtai

nmen

t90

.9 0

.0 9

.1 0

.0 0

.0

2.5

55

1996

60.9

1

92M

edia

24.3

8.1

35.1

21.6

10.8

3.

4

7419

9762

.8

215

Ret

ailin

g46

.2 6

.224

.623

.1 0

.0

2.9

65

1998

59.0

2

17

Foo

d/st

aple

s Ret

ailin

g50

.0 0

.050

.0 0

.0 0

.0

0.4

8

1999

57.3

2

13B

ever

ages

36.4

0.0

36.4

27.3

0.0

3.

5

7720

0058

.4

209

Hea

lth c

are

equi

p./s

uppl

ies

75.0

0.0

5.0

20.0

0.0

0.

9

2020

0160

.9

202

Phar

mac

eutic

als b

iote

ch.

55.2

3.4

24.1

13.8

3.4

1.3

29

2002

61.8

1

99R

eal e

stat

e88

.5 3

.1 8

.5 0

.0 0

.0

5.9

130

So

ftw

are/

supp

lies

71.4

5.8

15.3

6.9

0.5

8.

6 1

89

Har

dwar

e/eq

uipm

ent

40.2

14.

523

.220

.7 1

.2 1

0.9

241

Tel

ecom

. 15

.8 5

.331

.647

.4 0

.0

0.9

19

T

otal

57.4

5.7

20.0

16.3

0.7

100.

022

0857

.4

2209

Not

e:

The

tabl

e sh

ows t

he d

istrib

utio

n of

em

ploy

ee d

irect

ors a

cros

s ind

ustr

ies.

The

Glo

bal I

ndus

try

Cla

ssifi

catio

n St

anda

rd (G

ICS)

is u

sed.

The

w

hole

or p

arts

of t

he in

dust

ry m

ay b

e ex

empt

ed, f

or in

stan

ce th

e E

nerg

y (h

ydro

pow

er a

nd p

etro

leum

) sec

tor.

Tra

nspo

rt c

onta

ins t

he im

port

ant

ship

ping

segm

ent.

Med

ia is

exe

mpt

ed a

s wel

l, bu

t in

som

e fi r

ms c

o-de

term

inat

ion

com

es a

bout

thro

ugh

unio

n ne

gotia

tions

. ‘E

mpd

ir’ is

shor

t-ha

nd

for e

mpl

oyee

dire

ctor

s.

336

Tab

le 1

3.3

Defi

niti

ons o

f var

ious

boa

rd m

easu

res a

nd th

eir m

ain

stat

istic

al p

rope

rtie

s

Shar

ehol

der d

eter

min

edC

o-de

term

ined

F si

gnM

ean

Med

ian

Std

N

Mea

nM

edia

nSt

d N

Tob

in’s

Q1.

461

1.10

51.

156

867

1.50

11.

162

1.06

477

30.

459

Stoc

k re

turn

16.1

09−

1.70

012

1.51

5 7

7417

.666

2.52

078

.204

724

0.77

0 R

OA

3.27

26.

220

18.8

40 8

386.

531

8.21

013

.883

771

0.00

0 A

vera

ge w

age

558.

442

340.

878

1516

.360

677

355.

909

316.

306

222.

129

762

0.00

0 D

irect

ors’

hold

ings

0.06

50.

000

0.18

9 9

660.

063

0.00

00.

187

825

0.82

8 N

etw

ork

0.18

00.

198

0.11

512

640.

191

0.20

80.

075

942

0.01

5 Si

ze1

4.83

45.

000

1.33

012

675.

341

5.00

01.

271

942

0.00

0 G

ende

r10.

024

0.00

00.

078

1267

0.04

50.

000

0.10

194

20.

000

Boa

rd in

dex

0.19

20.

233

1.87

7 9

65−

0.27

1−

0.07

81.

956

825

0.00

0 L

ever

age

2.38

71.

165

5.95

5 8

571.

903

1.04

43.

216

761

0.04

6 D

iv. p

ayou

t rat

e0.

197

0.00

00.

747

960

0.26

10.

085

0.56

482

20.

042

Em

pdir

0.00

00.

000

0.00

012

672.

282

2.00

00.

707

942

0.00

0 E

mpd

irfra

c 0

.000

0.00

00.

000

1267

0.30

10.

300

0.08

294

20.

000

Firm

size

5.42

75.

462

0.78

8 9

056.

071

6.02

10.

725

801

0.00

0 Sy

stem

atic

risk

0.82

80.

724

0.74

9 8

880.

707

0.69

00.

535

794

0.00

0 V

olat

ility

0.91

80.

646

1.20

0 8

850.

738

0.58

40.

597

788

0.00

0

337

Not

es:

Tob

in’s

Q is

mar

ket v

alue

div

ided

by

book

val

ue o

f ass

ets;

Stoc

k re

turn

is th

e ra

w st

ock

retu

rn c

orre

cted

for d

ivid

end

and

stoc

k sp

lit; R

OA

is

acco

untin

g pr

ofi ts

on

book

val

ue o

f ass

ets;

Ave

rage

wag

e is

the

loga

rithm

of t

otal

wag

es d

ivid

ed b

y th

e nu

mbe

r of e

mpl

oyee

s; D

irect

ors’

hold

ings

is

the

perc

enta

ge o

f dire

ctor

s’ ow

ners

hip;

Net

wor

k is

a su

mm

ary

mea

sure

of t

he b

oard

’s di

rect

and

indi

rect

rela

tions

to o

ther

fi rm

s thr

ough

m

ultip

le d

irect

orsh

ips (

see

endn

ote

8); S

ize1

is th

e bo

ard

size

of sh

areh

olde

r ele

cted

dire

ctor

s; G

ende

r1 is

the

frac

tion

of w

omen

of t

he sh

areh

olde

r el

ecte

d di

rect

ors;

Boa

rd in

dex

is a

sum

mar

y m

easu

re o

f the

abo

ve b

oard

var

iabl

es; L

ever

age

is th

e bo

ok v

alue

of d

ebt o

n bo

ok v

alue

of e

quity

; D

ivid

end

payo

ut ra

te is

div

iden

ds o

n ne

t inc

ome;

Em

pdir

is th

e nu

mbe

r of e

mpl

oyee

dire

ctor

s div

ided

by

the

num

ber o

f dire

ctor

s; E

mpd

irfra

c is

the

frac

tion

of e

mpl

oyee

dire

ctor

s in

the

tota

l boa

rd; F

irm si

ze is

the

natu

ral l

ogar

ithm

of a

ccou

ntin

g in

com

e; S

yste

mat

ic ri

sk is

the

com

pany

’s ex

posu

re to

mar

ket c

hang

es (e

quity

bet

a); V

olat

ility

is th

e fi r

m’s

tota

l risk

mea

sure

d as

its y

early

stan

dard

dev

iatio

n.

The

‘F si

gn’ s

how

s the

sign

ifi ca

nce

of th

e te

st o

f the

nul

l hyp

othe

sis th

at th

e tw

o gr

oup

mea

ns a

re e

qual

, est

imat

ed fr

om a

n an

alys

is of

var

ianc

e (A

NO

VA

). L

ow v

alue

s ind

icat

e re

ject

ion

of th

e nu

ll hy

poth

esis.

The

F v

alue

is fo

und

by d

ivid

ing

the

Bet

wee

n G

roup

s Mea

n Sq

uare

by

the

Err

or

Mea

n Sq

uare

(Joh

nson

and

Wic

hern

, 198

8, p

. 235

).


Like the fi ndings in Table 13.2, this is evidence that the fi rms attempt to minimize the employee director importance.

The two fi rm groups diff er in background variables, notably fi rm size. The co-determined fi rms are larger on average. This warrants paying par-ticular attention to the largest fi rm size groups in regressions, in order to control for fi rm size biases.

5. ESTIMATION AND METHOD

I estimate the relationships in equation (2) with simultaneous equations regressions on the full samples as well as sub-samples. The equations spell out behavioural relationships between variables. Since the equilibrium model of governance is not known, reduced form estimation is not pos-sible (Greene, 2003, section 15.2). The equations are behavioural, but not structural in the sense of belonging to an equilibrium model.

The fi xed eff ects method (Woolridge, 2002) is common to all regressions. Fixed eff ects estimation amounts to removing the individual heterogeneity of fi rms contained in the fi xed eff ect ci.11 Remember the error term in the system equation (2) is uit, which contains the fi xed eff ect ci and a idiosyn-cratic eff ect vit, which varies over time and companies; i refers to the fi rm number, and t is the time period. When demeaning the variables, the fi xed eff ect element disappears. So does the constant term.

I use the three-stage least squares (3SLS) methodology in estimations. The 3SLS is an instrumental variables estimation method where the instru-ments are the predicted values of the dependent variable in a regression on all the explanatory variables in the system (Greene, 2003, p. 398). The predicted values are found from GLS regressions, and iterations are taken until convergence is achieved. Meaningful overall measures, such as R2 in OLS regressions, are not available. Instead, I include a Wald test (Greene, 2003, p. 107) to study whether all coeffi cients in a given equation are zero.

The danger in simultaneous equation estimation lies in the model speci-fi cation (Greene, 2003). If, for instance, a misspecifi cation has occurred in the fi rst equation, the mistake may contaminate all other equations as well. To investigate if this propagation of misspecifi cation is a serious problem, I perform several robustness tests.

I perform estimations in the full sample and for sub-samples. First, the model in equation (2) is estimated on the full sample with Tobin’s Q as fi rm performance, and then on sub-samples of co-determined and shareholder determined fi rms. The sub-sample tests will reveal whether results from the overall sample really apply to co-determined fi rms alone, or whether the employee director eff ect is merely due to diff erence in sampling. I further


partition the sample to include only fi rms with more than 200 employees, when co-determination is compulsory. This will remove fi rm size eff ects.

In robustness tests, I perform an estimation with all index variables included individually (the right hand side of equation (1) on p. 329), as well as an estimation of a wider defi nition of the board index,12 this time including non-signifi cant eff ects in Bøhren and Strøm (2008) as well. Further robustness tests include replacing Tobin’s Q with ROA and stock return as a dependent variable, and replacing leverage with the dividend payout rate. Also, I remove the lagged fi rm performance in order to inves-tigate whether parameter estimates remain stable. The last robustness test is a test of the Fauver and Fuerst (2006) information hypothesis, which I interpret to mean that in information intensive industries fi rm performance is improved with co-determination. This regression should show whether their positive employee director result is also the case in Norway.

The explanatory variables are assumed to be simultaneous with fi rm performance. Since board members are predominantly elected in the late spring, the new board should also have had some time to make a noticeable impact upon fi rm performance, measured at year-end. This assumption is reasonable, given some market effi ciency.

6. ECONOMETRIC EVIDENCE

Do employee directors improve fi rm performance, and are governance mechanisms at least partly endogenously determined? This section reports simultaneous regression results of the model in equation (2). I estimate for the whole sample, and then turn to sub-samples of co- determined and shareholder determined companies, and for fi rms with more than 200 employees. All regressions are done with standardized values. This means that comparisons of economic importance can be read off from coeffi cient values.

I start with estimations of the model in equation (2) for the entire sample. Table 13.4 shows the estimation results. The Wald tests show that no equa-tion supports the null hypothesis that all coeffi cients are zero. Comparing the two sections of the table, signs and coeffi cient values are very much the same. Thus, I restrict comments to the case of systematic risk in the upper section.

The co-determination hypothesis says that the employee director vari-able is negative to fi rm performance, and positive to the board index and leverage. Table 13.4 confi rms this except for the leverage, where only the sign is as predicted. Furthermore, the co-determination hypothesis implies a positive impact on average wage. Here too, only the sign is confi rmed.


Table 13.4 Is co-determination associated with negative fi rm performance and positive governance mechanisms? Full sample (N51135) estimations using systematic and fi rm specifi c risk

IndependentVariable

Dependent variable

Tobin’sQ

Averagewage

Boardindex

Leverage

Tobin’s Q lagged 0.106** 0.028 0.065* −0.094** Average wage −0.035 −0.038 0.204** Board index 0.122** −0.030 −0.191** Leverage −0.041** 0.145** −0.171** Employee

directors−0.119** 0.072 0.314** 0.044

Firm size −0.141** −0.027 −0.060 0.129** Systematic risk 0.004 0.030 0.010 −0.035 Wald c2 test 79.516 39.396 82.612 87.617 p-value 0.000 0.000 0.000 0.000 Tobin’s Q lagged 0.105** 0.032 0.062* −0.082** Average wage −0.031 −0.036 0.181** Board index 0.119** -0.029 −0.182** Leverage −0.034 0.131** −0.167** Employee

directors−0.123** 0.066 0.311** 0.042

Firm size −0.136** −0.019 −0.066 0.145** Volatility −0.045** 0.028 −0.003 0.116**Wald c2 test 85.137 36.095 79.369 103.055 p-value 0.000 0.000 0.000 0.000

Notes:The table reports the simultaneous equation estimation of the system of equations in (2) with systematic risk (upper part) and fi rm-specifi c risk (lower part). The dependent variable is Tobin’s Q, which we measure as the market value of the fi rm over its book value. Variables are defi ned in Table 13.3. Each variable is time demeaned in the regressions. For each fi rm and each variable, I time demean by subtracting a given year’s observation from the fi rm’s overall mean. The table shows the estimates based on the standardized variables, which we construct by deducting each observation from its mean value and dividing by its standard deviation. Fixed eff ects estimation in 3SLS framework with standardized variables. All non-fi nancial fi rms on Oslo Stock Exchange 1989 to 2002. The Wald test (Greene, 2003, p. 107) is here a test of the null hypothesis that the coeffi cients in the given equation are all zero. A low value indicates null hypothesis rejection. If R is the q 3 K matrix of q restrictions and K coeffi cients, g the K vector of coeffi cients, and r the vector of the q restrictions, the Wald c2 (q) statistic is c2 (q) 5 (r 2 Rg) r [RSXR r ]21 (r 2 Rg) , where SX is the estimated covariance matrix of coeffi cients. The test results show that a hypothesis that all coeffi cients are zero must be rejected in all relations at the 1% level. Signifi cant results at the 5% (10%) level are marked with ** (*).


This weaker result may be due to pay being determined by external market conditions. Thus, the direct and indirect eff ects of co-determination are partly confi rmed. Consequently, employee directors carry a negative asso-ciation with fi rm performance, and shareholders tend to take compensa-tory actions to alleviate the infl uence of employee directors. The board index is at least partly endogenously determined.

Are the board index and the leverage positively related to fi rm perform-ance and negatively to average wage? For the board index, this is confi rmed for fi rm performance, but the only sign is as expected for average wage. Thus, a better composed board will improve fi rm performance. On the other hand, leverage is against the free cash fl ow hypothesis expectations in both fi rm performance and average wage. A higher leverage indicates a lower fi rm performance and higher average wage. In conclusion, the gov-ernance hypothesis is not fully confi rmed.

The negative association between leverage and fi rm performance con-fi rms fi ndings in empirical studies (Barclay et al., 1995; Rajan and Zingales, 1995; Brick et al., 2005). I off er two alternative explanations to the free cash fl ow hypothesis: the fear of higher decision costs in a situation with three decision makers, that is, shareholders, employees, and banks; and the negative signalling eff ect of a high leverage (Myers, 1977).

Also note the complementarity between the board index and leverage (Agrawal and Knoeber, 1996). The sign is negative and signifi cant. Thus, the two governance mechanisms are substitutes rather than complements.

The Hermalin and Weisbach (1998) reverse causation hypothesis is only partly confi rmed, as the board index is positive and leverage is negative. Lower leverage should bring lower monitoring intensity. The results are signifi cant, indicating that good performance leads to a better board index and to an easier debt burden. In all, endogeneity is confi rmed, as both the board index and the leverage are at least partly determined from the pres-ence of employee directors and from past performance.

Are shareholders able to neutralize the employee director by adjustments in the board index and the leverage, taking the employee director relation-ship to average wage into consideration as well? Since the variables are standardized to have average zero and standard deviation 1.0 in regressions, coeffi cients can be compared. They show that the direct eff ect is stronger than the indirect eff ect on the board index. For the negative direct employee director eff ect is now 0.119, while the indirect eff ect upon the board index is positive and 0.314. Since the board index is now 0.122 to fi rm perform-ance, the positive, indirect impact of employee directors through the board index is only 0.038 (5 0.122 3 0.314), or 31.1 per cent of the direct board index eff ect. The shareholders are able to compensate 31.9 per cent of the negative direct eff ect of employee directors through adjustments to board


characteristics. Furthermore, the employee director also impacts positively upon average wage, which is negatively related to fi rm performance. Even though the average wage is not signifi cant in the overall sample, it is for co-determined fi rms, as I shortly report. The same applies to leverage. Likewise, the economic signifi cance of the indirect eff ects from the lagged fi rm per-formance is very low, being 0.01 for both the board index and the leverage.

Thus, the economic magnitude of the indirect eff ects from employee directors or past fi rm performance upon fi rm performance is small com-pared to the direct eff ect of the board index and the leverage. Endogeneity matters, but not very much.

The volatility measure in the lower section of Table 13.4 gives two inter-esting relationships in the board index and the leverage equations. It turns out that only leverage has the expected positive and signifi cant sign. The Raheja (2005) theory of board composition implies that the board index is positively related to fi rm risk. For volatility the opposite sign obtains.

Next, the model is studied in sub-samples. If regulation plays a role, a less than optimal board composition is likely to follow. Therefore, we should observe stronger and more signifi cant coeffi cients in the co- determined fi rms than in the shareholder determined. Table 13.5 is a report on the two sub-samples of fi rms.

Note that the Wald test shows rejection of the null hypothesis that all variables have zero signifi cance. Furthermore, a Chow dummy variable test rejects the hypothesis that the coeffi cients of the sub-samples are equal to those in the overall sample. Thus, there is a diff erence between co-determined and shareholder determined fi rms.

In the co-determination sub-sample the employee director eff ects are even more pronounced than in the overall sample. The negative employee director impact upon fi rm performance is about 45 per cent higher than in the overall sample and the indirect eff ect on the board index increases even more. Now, the employee director variable is signifi cant in relation to leverage and to average wage. Thus, the co-determination hypothesis is even more strongly confi rmed in the sub-sample of only co-determined fi rms than in the overall sample.

The board index and the free cash fl ow hypotheses come out more in line with expectations in the co-determined fi rms too. Now a signifi cant result for the board index towards average wage appears. Leverage turns out to be negative and signifi cant towards average wage, while positive in the overall sample. In shareholder determined fi rms, signifi cant results are fewer and of diff erent sign. Leverage is positively correlated with average wage, in contrast to the co-determined fi rms.

In both sub-samples the board index and leverage are negatively related. Thus, the substitution result from the overall sample is confi rmed in the


Table 13.5 Is fi rm performance (Tobin’s Q) diff erently related to governance mechanisms in co-determined and in shareholder determined fi rms?

IndependentVariable

Dependent Variable

Tobin’sQ

AverageWage

BoardIndex

Leverage

Co-determined fi rms N5639Tobin’s Q lagged 0.303** 0.011 0.059 −0.069* Average wage −0.089 −0.520** −0.156**Board index 0.118** −0.175** −0.274** Leverage −0.103** −0.080** −0.414** Employee

directors−0.173** 0.186** 0.484** 0.140**

Firm size 0.077 −0.001 −0.210** −0.083 Systematic risk 0.034 −0.010 0.027 −0.035Wald c2 test 112.123 83.056 212.330 89.279 p-value 0.000 0.000 0.000 0.000 Shareholder determined fi rms N5496Tobin’s Q lagged −0.072** 0.030 0.041 −0.101 Average wage −0.023 0.032 0.272**Board index 0.121** 0.035 −0.158**Leverage −0.018 0.208** −0.109** Firm size −0.296** −0.077 −0.021 0.237**Systematic risk −0.048 0.060 −0.006 −0.028Wald c2 test 57.543 30.281 9.986 45.030 p-value 0.000 0.000 0.076 0.000 Chow dummy

variable test c2 (7): 62.160 p-value 0.000

Notes:The table reports the simultaneous equation estimation of the system of equations in (2) with co-determined fi rms in the upper part and shareholder determined fi rms in the lower part. The dependent variable is Tobin’s Q , which we measure as the market value of the fi rm over its book value. Variables are defi ned in Table 13.3. Each variable is time demeaned in the regressions. For each fi rm and each variable, I time demean by subtracting a given year’s observation from the fi rm’s overall mean. The table shows the estimates based on the standardized variables, which we construct by deducting each observation from its mean value and dividing by its standard deviation. Fixed eff ects estimation in 3SLS framework with standardized variables. All non-fi nancial fi rms on Oslo Stock Exchange 1989 to 2002. The Wald test is explained in Table 13.4. The test results show that a hypothesis that all coeffi cients are zero must be rejected in all relations at the 1% level, except one where a 7.7% level is required. The Chow (Greene, 2003, Ch. 7) dummy variable test is an exclusion test for the null hypothesis that variables formed by a co-determination dummy variable interacted with each of the explanatory variables are all zero. Low value indicates hypothesis rejection. The test result shows that the hypothesis that the two sub-samples have equal coeffi cients must be rejected. Signifi cant results at the 5% (10%) level are marked with ** (*).


sub-samples. We also see that the past fi rm performance endogeneity hypothe-sis gains less support in the sub-samples than in the overall sample. In fact, only the negative leverage result in the co-determined sub-sample is signifi cant.

Another diff erence exists for fi rm size. Firm size is negative and sig-nifi cant in the fi rm performance equation in shareholder determined fi rms, while positive in co-determined ones. Also, in the leverage equation the signs are reversed, and signifi cant in shareholder determined fi rms only. The latter confi rms ‘stylized facts’ about the positive relationship between fi rm size and leverage (Harris and Raviv, 1991).

An interpretation of the diff erence in sub-samples is that in shareholder determined fi rms the board composition is closer to the optimal, and therefore exogenous characteristics such as fi rm size play a larger role. The large diff erences between samples confi rm the Buchanan and Tullock (1962) theory.

Are the results arrived at so far driven by a fi rm size eff ect? Table 13.6 shows regressions for all fi rms with more than 200 employees in the upper part, while the lower part is limited to the largest co-determined fi rms. The 2001 employee sample shows results very similar to those in the entire sample in Table 13.4 in the upper part, and for the co-determined fi rms in Table 13.5 in the lower part. Thus, the former results are not due to some fi rm size eff ect. In fact, even among fi rms where co-determination is com-pulsory, the main co-determination hypothesis is confi rmed.

Looking back, the co-determination and governance hypotheses are con-fi rmed. Tests in sub-samples do not overturn these conclusions; on the con-trary, they add to their strength. For instance, while the employee director eff ect is negative for leverage in the overall sample, it is positive in the co-determined sub-sample, as the hypothesis predicts. Thus, having representa-tives of one stakeholder group, the employees, in addition to shareholders on the board does not improve fi rm performance, as a stakeholder (Freeman and Reed, 1983; and Blair, 1995) or a new economy position (Zingales, 2000; Becht et al., 2003) implies. Instead, the results point to confl ict of interests among the stakeholders. Furthermore, evidence of substitution between the board index and leverage is present in all regressions. I also fi nd evidence of endogeneity (or reverse causation) from past fi rm performance, but with opposite signs to those predicted in Hermalin and Weisbach (1998). However, the indirect eff ects of employee directors and past fi rm perform-ance upon fi rm performance through the board index and leverage are small compared with the direct eff ects from the board index and leverage. Endogeneity counts, but has low economic signifi cance.

The negative relation between employee directors and fi rm performance is in agreement with Fitzroy and Kraft (1993), Schmid and Seger (1998), Gorton and Schmid (2000, 2004), Falaye et al. (2006), and Bøhren and


Table 13.6 Are the employee director direct and indirect (endogenous) eff ects upheld in all fi rms with more than 200 employees and in co-determined fi rms with more than 200 employees?

IndependentVariable

Dependent Variable

Tobin’sQ

AverageWage

Boardindex

Leverage ratio

200+ employee fi rms, N5814Tobin’s Q lagged 0.168** 0.008 0.139** −0.090*Average wage −0.012 −0.101* 0.075 Board index 0.094** −0.041* −0.145** Leverage −0.065** 0.041 −0.197** Employee directors −0.107** 0.049 0.420** −0.006 Firm size −0.083 0.103* −0.093 0.172** Systematic risk 0.013 0.045 0.025 -0.101** Wald c2 test 73.709 13.037 85.068 46.726 p-value 0.000 0.042 0.000 0.000200+ employees co-determined, N5565Tobin’s Q lagged 0.358** 0.025 0.091* −0.060 Average wage −0.111* −0.478** −0.359**Board index 0.047 −0.148** −0.256**Leverage −0.126** −0.180** −0.413** Employee directors −0.148** 0.093** 0.522** 0.099**Firm size 0.008 0.031 −0.248** −0.028 Systematic risk 0.017 −0.027 0.008 −0.041Wald c2 test 110.682 73.992 190.267 99.414 p-value 0.000 0.000 0.000 0.000

Notes:The table reports the simultaneous equation estimation of the system of equations in (2) with all fi rms larger than 200 employees in the upper part and all co-determined fi rms larger than 200 employees in the lower part. The dependent variable is Tobin’s Q , which we measure as the market value of the fi rm over its book value. Variables are defi ned in Table 13.3. Each variable is time demeaned in the regressions. For each fi rm and each variable, I time demean by subtracting a given year’s observation from the fi rm’s overall mean. The table shows the estimates based on the standardized variables, which we construct by deducting each observation from its mean value and dividing by its standard deviation. Fixed eff ects estimation in 3SLS framework with standardized variables. All non-fi nancial fi rms on Oslo Stock Exchange 1989 to 2002. The Wald test is explained in Table 13.4. The test results show that a hypothesis that all coeffi cients are zero must be rejected in all relations at the 1% level, except one, where a 4.3% level is required. Signifi cant results at the 5% (10%) level are marked with ** (*).


Strøm (2008), but at odds with Fauver and Fuerst (2006). None of these studies contain simultaneous equations models, and only the Bøhren and Strøm (2008) paper investigates the endogeneity of board mechanisms. I will return to the Fauver and Fuerst (2006) and Bøhren and Strøm (2008) articles in the following robustness section.

7. ROBUSTNESS CHECKS

I perform robustness checks on the defi nitions of the board index, fi rm performance, and leverage. In addition, I check for the absence of serial dependence of the fi rm performance, that is, whether lagged fi rm perform-ance is zero. Finally, I check the Fauver and Fuerst (2006) results in two sub-samples of information industries and other industries. With simulta-neous equations, changes in one place are likely to propagate throughout the system. Thus, diff erent coeffi cient values and signifi cance from the original formulation are quite likely to appear. Fortunately, the results largely confi rm those in Section 6.

Do the co-determination results survive when the individual board mechanisms are used in place of the board index? Table 13.7 shows simul-taneous regressions results when all four board characteristics making up the board index enter the regressions individually. Former results for co- determination largely apply. The employee director variable is negative to Tobin’s Q, and positive to average wage and leverage. For the board charac-teristics, only the relation to board size is signifi cant. On the other hand, the hypotheses on governance variables are upheld for all board characteristics but the gender variable. It turns out to be non-signifi cant in the Tobin’s Q relation. The other variables are as expected, and their coeffi cients are close to those Bøhren and Strøm (2008) fi nd in partial GMM estimations. These authors also discuss endogeneity. Even though the estimations are not directly comparable, none of the signifi cant results in Table 13.7 confl ict with the endogeneity results in Bøhren and Strøm (2008). The second endo-geneity eff ect from lagged fi rm performance is signifi cant in the leverage but not in any of the board variables. However, the signs on the individual board variables conform to the positive sign of the board index in earlier tables.

Besides these main points, Table 13.7 contains many new details, which it is beyond this chapter to explore. For instance, the substitution eff ect between the board index and leverage in former tables now turns out to concern network, while leverage is a complement to board size and gender. Thus, overall the results are well in line with former fi ndings, except for the lagged fi rm performance relationship to governance variables.

In Table (13.8) I have modifi ed the board index to include all board

347

Tab

le 1

3.7

The

em

ploy

ee d

irec

tor d

irec

t and

indi

rect

(en

doge

nous

) eff

ect

s upo

n fi r

m p

erfo

rman

ce w

hen

the

indi

vidu

al

boar

d va

riab

les a

re u

sed

inst

ead

of th

e bo

ard

inde

x (N

5 1

135)

Var

iabl

eT

obin

’sQ

Ave

rage

wag

eD

irect

ors’

hold

ings

Net

wor

kB

oard

size

Gen

der

Lev

erag

e

Tob

in’s

Q la

gged

0.10

6**

0.02

6−

0.02

20.

071

−0.

030

−0.

027

−0.

091*

*A

vera

ge w

age

−0.

039

−0.

049*

0.12

7**

0.04

30.

097*

*0.

199*

*D

irect

ors’

hold

ings

0.05

1*

−0.

057*

0.04

60.

131*

*0.

024

−0.

038

Net

wor

k0.

091*

* 0.

061*

*0.

019

0.07

0**

−0.

067*

*−

0.09

0**

Boa

rd si

ze−

0.06

2**

0.05

20.

136*

*0.

175*

*0.

122*

*0.

212*

*G

ende

r −

0.02

50.

124*

*0.

026

−0.

177*

*0.

129*

*0.

082*

*L

ever

age

−0.

041*

* 0.

140*

*−

0.02

3−

0.13

1**

0.12

4**

0.04

5**

E

mpl

oyee

dire

ctor

s-0

.114

**

0.10

2**

0.07

40.

043

−0.

565*

*−

0.00

50.

108*

Firm

size

−0.

144*

* −

0.04

30.

082

0.04

40.

258

0.00

60.

089

Syst

emat

ic ri

sk0.

001

0.02

2−

0.10

4**

0.09

3**

0.03

4−

0.04

1−

0.03

2W

ald

c2 tes

t86

.300

65.3

9545

.170

57.5

2030

1.55

158

.640

98.4

43

p-va

lue

0.00

00.

000

0.00

00.

000

0.00

00.

000

0.00

0

Not

es:

The

tabl

e re

port

s the

sim

ulta

neou

s equ

atio

n es

timat

ion

of th

e sy

stem

of e

quat

ions

in (2

) whe

n th

e in

divi

dual

var

iabl

es m

akin

g up

the

boar

d in

dex

repl

ace

the

boar

d in

dex.

The

boa

rd in

dex

cons

ists o

f dire

ctor

s’ ho

ldin

gs, n

etw

ork,

boa

rd si

ze, a

nd g

ende

r. T

he d

efi n

ition

of d

irect

ors’

hold

ings

is

the

frac

tion

of o

wne

rshi

p fo

r the

boa

rd a

s a w

hole

; net

wor

k is

info

rmat

ion

cent

ralit

y (W

asse

rman

and

Fau

st, 1

994)

, see

not

e 9;

the

boar

d siz

e is

the

num

ber o

f sha

reho

lder

ele

cted

dire

ctor

s; an

d ge

nder

is d

efi n

ed a

s the

num

ber o

f sha

reho

lder

ele

cted

wom

en o

ver b

oard

size

.

The

dep

ende

nt v

aria

ble

is T

obin

’s Q

, w

hich

we

mea

sure

as t

he m

arke

t val

ue o

f the

fi rm

ove

r its

boo

k va

lue.

Var

iabl

es a

re d

efi n

ed in

Tab

le 1

3.3.

E

ach

varia

ble

is tim

e de

mea

ned

in th

e re

gres

sions

. For

eac

h fi r

m a

nd e

ach

varia

ble,

I tim

e de

mea

n by

subt

ract

ing

a gi

ven

year

’s ob

serv

atio

n fr

om

the

fi rm

’s ov

eral

l mea

n. T

he ta

ble

show

s the

est

imat

es b

ased

on

the

stan

dard

ized

var

iabl

es, w

hich

we

cons

truc

t by

dedu

ctin

g ea

ch o

bser

vatio

n fr

om

its m

ean

valu

e an

d di

vidi

ng b

y its

stan

dard

dev

iatio

n.

Fix

ed e

ff ect

s est

imat

ion

in 3

SLS

fram

ewor

k w

ith st

anda

rdiz

ed v

aria

bles

. All

non-

fi nan

cial

fi rm

s on

Oslo

Sto

ck E

xcha

nge

1989

to 2

002.

The

Wal

d te

st is

exp

lain

ed in

Tab

le 1

3.4.

The

test

resu

lts sh

ow th

at a

hyp

othe

sis th

at a

ll co

effi c

ient

s are

zer

o m

ust b

e re

ject

ed in

all

rela

tions

at

the

1% le

vel.

Si

gnifi

cant

resu

lts a

t the

5%

(10%

) lev

el a

re m

arke

d w

ith *

* (*

).


variables used in Bøhren and Strøm (2008) as specifi ed in note 12 to check whether the board index is sensitive to the selection of board character-istics. The overall Wald tests are strong and the signifi cance of the coef-fi cients are almost similar to what earlier full sample results in Table 13.4 show. We note that the impact of the employee director variable is less in the new board index, and is now signifi cant in its positive relationship to average wage. Thus, the co-determination hypothesis is supported with

Table 13.8 Does a wide defi nition of the board index change the relationship between fi rm performance, employee directors and governance mechanisms? (N51135)

IndependentVariable

Dependent Variable

Tobin’sQ

AverageWage

BoardIndex 2

Leverage

Tobin’s Q lagged 0.114 ^** 0.027 0.009 −0.109**Average wage −0.031 −0.136 ^** 0.204**Board index 2 0.091 ^** −0.102 ^** −0.083**Leverage −0.048 ^** 0.143 ^** −0.077 ^** Employee directors −0.094 ^** 0.077 ^* 0.152 ^** −0.004 Firm size −0.129 ^** −0.048 −0.218 ^** 0.126*Systematic risk 0.005 0.030 0.005 −0.037Wald c2 test 65.962 54.197 42.933 56.620 p-value 0.000 0.000 0.000 0.000

Notes:The table reports the simultaneous equation estimation of the system of equations in (2) when all individual variables enter the board index, and not just directors’ holdings, network, board size, and gender. The added variables are outside owner concentration, independence, CEO director, exported and imported directors, and board age dispersion. Outside owner concentration is the sum of squared equity fractions across all the fi rm’s outside owners; independence is the board tenure of the non-employee directors minus the tenure of the CEO; CEO director equals 1 if the CEO is a member of his company’s board and zero otherwise; exported CEO is the number of outside directorships held by the fi rm’s CEO; imported CEO is the proportion of CEOs from other companies on the board; board age dispersion is the standard deviation of board age. The dependent variable is Tobin’s Q, which we measure as the market value of the fi rm over its book value. Variables are defi ned in Table 13.3. Each variable is time demeaned in the regressions. For each fi rm and each variable, I time demean by subtracting a given year’s observation from the fi rm’s overall mean. The table shows the estimates based on the standardized variables, which we construct by deducting each observation from its mean value and dividing by its standard deviation. Fixed eff ects estimation in 3SLS framework with standardized variables. All non-fi nancial fi rms on Oslo Stock Exchange 1989 to 2002. The Wald test is explained in Table 13.4. The test results show that a hypothesis that all coeffi cients are zero must be rejected in all relations at the 1% level. Signifi cant results at the 5% (10%) level are marked with ** (*).


this new board index, although with lower coeffi cient values. The endog-eneity eff ect of a lagged fi rm performance loses signifi cance in the board index relation. The same happens when individual board characteristics replace the board index, and the eff ect also disappears in the shareholder determined sub-sample. Thus, a preliminary conclusion is that the reverse causation in the board index relation seems to be sensitive to the specifi ca-tion of the index and in sub-samples.

The conclusion from the discussion of the two previous tables is that the results are upheld; in particular, the co-determination hypothesis is confi rmed.

Now I turn to variations on fi rm performance, using the stock return and ROA instead of Tobin’s Q. The stock return and ROA may be seen as two extremes in performance measurement, the one only market based, the other only accounting based. Bhagat and Jeff eris (2002) argue in favour of accounting measures, noting that market measures may contain an antici-pation bias, since accounting numbers may be manipulated during a given year. Since our data span 14 years, this accounting manipulation should be a minor concern. These two measures of fi rm performance should together provide an adequate framework for robustness tests.

The results for the full sample are given in Table 13.9. Since the results in the sub-samples largely parallel those found for the full sample, the sub-sample results are not reported. The results in Table 13.9 largely replicate those already found for Tobin’s Q in Table 13.4. The co-determination and the governance hypotheses show the same confi rmations. As before, leverage is negative in the fi rm performance equation. Again, the board index and leverage are substitutes. Endogeneity (or reverse causation) is evident in both fi rm performance specifi cations, although at diff erent variables. For the stock return the lagged stock return is signifi cant in fi rm performance and leverage, as before. One would expect this to happen with accounting numbers due to earnings management or conservative accounting practices (Watts, 2003), which would induce serial correlation. However, lagged performance is signifi cant for only the board index for the accounting measure ROA. Overall, Table 13.9 supports earlier fi ndings.

The upshot is that alternative performance measures do not upset con-clusions reached with Tobin’s Q. Therefore, further robustness tests may well proceed with Tobin’s Q as the dependent variable.

Next, Table 13.10 shows results when the dividend payout rate replaces leverage, and Tobin’s Q is the fi rm performance in the upper part, while in the lower part the lagged fi rm performance is removed. Dividend payout rate is gauged as the annual dividend as a fraction of the earnings before interest, taxes, depreciation, and accruals (EBITDA). During the period of study, share buybacks were illegal in Norway.


Table 13.9 The employee director direct and indirect (endogenous) eff ects when the stock return and the return on assets (ROA) defi ne fi rm performance

IndependentVariable

Dependent Variable

Tobin’sQ

AverageWage

BoardIndex 2

Leverage

Stock return, N51019Stock return lagged −0.242** −0.046** −0.008 −0.072**Average wage −0.056 −0.057* 0.077**Board index 0.132** −0.055* −0.223**Leverage −0.232** 0.056** −0.169** Employee directors −0.165** 0.052 0.302** 0.039 Firm size −0.112 0.012 −0.026 0.117*Systematic risk −0.138** −0.003 −0.010 −0.043Wald c2 test 123.539 15.975 78.542 61.258 p-value 0.000 0.014 0.000 0.000

ROA N51135ROA lagged −0.008 −0.043 0.063** −0.011 Average wage −0.129** −0.028 0.106**Board index 0.066** −0.022 −0.195**Leverage −0.170** 0.077** −0.183** Employee directors −0.125** 0.062 0.322** 0.060 Firm size 0.033 −0.010 −0.051 0.103 Systematic risk −0.024 0.014 0.025 −0.048Wald c2 test 68.694 15.739 86.037 57.564 p-value 0.000 0.015 0.000 0.000

Notes:The table reports the simultaneous equation estimation of the system of equations in (2) when the stock return replaces Tobin’s Q in the upper part and the return on assets replaces Tobin’s Q in the lower part. The dependent variable is the stock return, defi ned as the raw stock return adjusted for dividend and stock splits; alternatively, as the return on assets, gauged as the accounting profi ts on book value of assets. Variables are defi ned in Table 13.3. Each variable is time demeaned in the regressions. For each fi rm and each variable, I time demean by subtracting a given year’s observation from the fi rm’s overall mean. The table shows the estimates based on the standardized variables, which we construct by deducting each observation from its mean value and dividing by its standard deviation. Fixed eff ects estimation in 3SLS framework with standardized variables. All non-fi nancial fi rms on Oslo Stock Exchange 1989 to 2002. The Wald test is explained in Table 13.4. The test results show that a hypothesis that all coeffi cients are zero must be rejected in all relations at the 1% level, except for the average wage, where at least a 1.6% level is needed. Signifi cant results at the 5% (10%) level are marked with ** (*).


Table 13.10 The relationships between fi rm performance, employee directors and governance mechanisms when dividend payout rate replaces leverage ratio and when lagged fi rm performance is removed

IndependentVariable

Dependent Variable

Tobin’sQ

AverageWage

BoardIndex 2

Dividend Payout

Dividend payout rate, N51150Tobin’s Q lagged 0.106** 0.014 0.088** 0.035 Average wage −0.046* −0.075** −0.025 Board index 0.125** −0.059** −0.021 Dividend payout rate 0.005 −0.011 −0.012 Employee directors −0.117** 0.082* 0.317** 0.092 Firm size −0.178** 0.004 −0.106* 0.022 Systematic risk 0.002 0.028 0.020 −0.066Wald c2 test 79.275 8.236 48.154 4.785 p-value 0.000 0.221 0.000 0.572

Tobin’s Q

AverageWage

BoardIndex

Leverage

No lag N51333Average wage −0.062** −0.018 0.201**Board index 0.131** −0.014 −0.169**Leverage −0.058** 0.152** −0.161** Employee directors −0.117** 0.058 0.306** 0.029 Firm size −0.151** 0.021 −0.066 0.100**Systematic risk 0.027 0.015 0.017 −0.031Wald c2 test 71.943 45.253 85.976 84.992 p-value 0.000 0.000 0.000 0.000

Notes:The table reports the simultaneous equation estimation of the system of equations in (2) when the dividend payout rate replaces leverage in the upper part and the lagged fi rm performance is removed in the lower part. The dependent variable is Tobin’s Q, which we measure as the market value of the fi rm over its book value. Each variable is time demeaned in the regressions. For each fi rm and each variable, I time demean by subtracting a given year’s observation from the fi rm’s overall mean. The table shows the estimates based on the standardized variables, which we construct by deducting each observation from its mean value and dividing by its standard deviation. I use fi xed eff ects estimation in 3SLS framework with standardized variables. The sample comprises all non-fi nancial fi rms on Oslo Stock Exchange 1989 to 2002. The Wald test is explained in Table 13.4. The test results show that a hypothesis that all coeffi cients are zero must be rejected in all relations at the 1% level, except for the average wage and the dividend payout relations in the upper part, where I cannot reject the hypothesis. Signifi cant results at the 5% (10%) level are marked with ** (*).


The striking results are fi rst that the dividend payout rate is nowhere signifi -cant as an independent variable, and second, as a dependent variable no vari-able in the system is related in a signifi cant way. In fact the Wald test cannot reject the hypothesis that all coeffi cients in the dividend payout rate equation are zero. An exclusion test (not reported) for the dividend payout rate cannot confi rm that the variable coeffi cient is diff erent from zero. Thus, the dividend payout rate is an inferior substitute for leverage. Moreover, the results for the other variables are not aff ected, even though changes in one part of a simultaneous system may bring about new values in other parts. Therefore, the results in Table 13.10 increase the confi dence in the original model.

The lower part of Table 13.10 shows results when the lagged fi rm performance is left out. The reason for the removal is that lagged fi rm performance induces bias (Hsiao, 2003, pp. 71–2), since the errors are no longer independent of the regressors. The smaller the bias, the larger is the number of periods in the panel and the closer to zero is the auto-correlation coeffi cient on lagged fi rm performance. Furthermore, if the explanatory variables apart from the lagged fi rm performance have very persistent elements, the bias will not disappear. This persistence can be a concern in governance studies. For instance, the fi rm’s board size is likely to be fairly stable. To test for the seriousness of this bias, I include static system regres-sions, that is, with no lagged performance.

Comparing the results from the no lagged fi rm performance regres-sion with the original estimates in Table 13.4, we see that practically all signs are maintained, and also that coeffi cient values are quite similar. The co-determination hypothesis is confi rmed. For average wage on fi rm performance, the variable is signifi cant in the static specifi cation but not in the dynamic. But overall the results from the dynamic estimations are upheld. Apparently, the low auto-correlation coeffi cient, the rather long time period and the small persistence in the explanatory variables warrant the use of the dynamic specifi cation in Table 13.4.

I also run a regression (not reported) with all explanatory variables lagged one period for the entire sample. This regression shows far fewer signifi cant results, and, although the signs are the same as before, this specifi cation is far inferior to the main regression in Table 13.4. Again, this points to a contemporaneity in governance mechanisms.

Finally, I run a test for the Fauver and Fuerst (2006) information hypoth-esis in the sub-samples. The authors assume information signifi cance to trade, transportation, and manufacturing industries. Using the same GICS industry classifi cation as in Table 13.2, I allocate Capital goods, Transport, Consumer articles, Retailing, Food and staples retailing, Health care equip-ment and supplies, and Telecommunications to the information intensive industries, while the rest are in other industries. Co-determined fi rms are


distributed in the two sub-samples almost as in the total population, with 61.1 per cent without employee directors in the Other industries category against 57.4 in the full sample. A test for the Fauver and Fuerst (2006) information hypothesis is that the employee director variable is positive in the information intensive industries. Table 13.11 shows the results.

The main interest is in the employee director, that is, the co-determination hypothesis. Both sub-samples show a negative and signifi cant coeffi cient on the employee director variable. The Chow test shows that the two sub-samples are diff erent, but the main Fauver and Fuerst (2006) hypothesis is not supported.

Overall, the results for the robustness tests do not invalidate the results found in Table 13.4.

8. CONCLUSION

In this chapter I pose the question whether board representation of one group of non-owner, the employees, improves fi rm performance. I con-clude it does not. The conclusion runs counter to claims from stakeholder theorists (Freeman and Reed, 1983; Blair, 1995) and some fi nancial econo-mists (Zingales, 2000; Becht et al., 2003) that co-determination improves fi rm performance. Instead the results support most former fi ndings in the empirical literature (Fitzroy and Kraft, 1993; Schmid and Seger, 1998; Gorton and Schmid, 2000, 2004; Falaye et al., 2006; Bøhren and Strøm, 2008) that employee board representation reduces fi rm performance.

The Norwegian regulations on co-determination provide the institu-tional framework. Co-determination is required by law for fi rms with more than 200 employees, and is an option if an employee majority demands so in fi rms having between 30 and 200 employees. A number of industries are exempted, and in all industries employees exercise their option. Thus, testing can take place using sub-samples, for instance in co-determined and shareholder determined sub-samples. For the whole sample, nearly 60 per cent do not have employee directors. The percentage has been rising during our period from 1989 to 2002. For fi rms with more than 200 employees, two-thirds have employee directors. The resultant data set is of a panel nature.

I estimate a system of simultaneous equations where employee direc-tors, fi rm size (sales), fi rm systematic risk, and one period lagged Tobin’s Q are the exogenous variables, and Tobin’s Q, average wage, board index, and leverage are the endogenous. The board index is constructed from important board characteristics, that is, directors’ holdings, the board’s network, board size and the female fraction. The free cash fl ow hypothesis (Easterbrook, 1984; Jensen, 1986) warrants the use of leverage.


Table 13.11 Firm performance, employee directors and governance mechanisms in sub-samples of information intensive industries and other industries

IndependentVariable

Dependent Variable

Tobin’sQ

AverageWage

BoardIndex

Leverage

Information industries N5533Firm performance lag 0.277** 0.210** 0.127 −0.240**Average wage −0.041 −0.150** 0.440**Board index 0.015 −0.093** −0.040 Leverage −0.024 0.217** −0.032 Employee directors −0.081** 0.182** 0.197** −0.044 Firm size 0.007 0.167* −0.060 0.141 Systematic risk 0.003 0.055 0.062 0.009Wald c2 test 60.354 88.323 17.790 61.376 p-value 0.000 0.000 0.007 0.000 Other industries N5601Firm performance lag 0.069* -0.040 −0.036 −0.088**Average wage −0.050 −0.113** 0.047 Board index 0.153** −0.131^ ** −0.040 Leverage −0.081** 0.044 −0.033 Employee directors −0.117* −0.033 0.157** −0.016 Firm size −0.215** −0.129^ * −0.145** 0.131*Systematic risk 0.062 0.013 −0.165** −0.133**Wald c2 test 41.646 14.737 35.541 15.583 p-value 0.000 0.022 0.000 0.016 Chow dummy variable test c2 (7): 28.362 p-value 0.000

Notes:The table reports the simultaneous equation estimation of the system of equations in (2) when the full sample is sub-divided into informationally intensive industries in the upper part and other industries in the lower. Using the same GICS industry classifi cation as in Table 13.2, informationally intensive industries are Capital goods, Transport, Consumer articles, Retailing, Food and staples retailing, Health care equipment and supplies, and Telecommunications, while the rest are in other industries. The dependent variable is Tobin’s Q, which we measure as the market value of the fi rm over its book value. Each variable is time demeaned in the regressions. For each fi rm and each variable, I time demean by subtracting a given year’s observation from the fi rm’s overall mean. The table shows the estimates based on the standardized variables, which we construct by deducting each observation from its mean value and dividing by its standard deviation. I use fi xed eff ects estimation in 3SLS framework with standardized variables. The sample comprises all non-fi nancial fi rms on Oslo Stock Exchange 1989 to 2002. The Wald test is explained in Table 13.4. The test results show that a hypothesis that all coeffi cients are zero must be rejected in all relations at the 1% level, except for a 2.3% level in the average wage relation in the Other industries estimation. The Chow dummy variable test is explained in Table 13.5. The test result indicates that coeffi cient values are diff erent in the two sub-samples. Signifi cant results at the 5% (10%) level are marked with ** (*).


The setup allows the testing of direct and indirect employee director eff ects upon fi rm performance. The indirect eff ects constitute a test of endo-geneity (Hermalin and Weisbach, 2003). The lagged fi rm performance gives a test of the reverse causation hypothesis (Hermalin and Weisbach, 1998) that past fi rm performance determines current governance. Furthermore, it allows testing of complementarity between the two governance variables board index and leverage (Agrawal and Knoeber, 1996).

Regressions are performed on the whole sample, the sub-samples of co-determined and shareholder determined fi rms, and then the sub-samples of fi rms with more than 200 employees. I use a fi xed eff ects model imple-mented in a three-stage least squares (3SLS) estimation.

In all regressions, the estimated coeffi cient for employee directors is signifi cantly negative. Moreover, the economic importance becomes larger as regressions proceed from the overall sample to the sub-sample of co-determined fi rms, and then to co-determined fi rms with 200 employees or more. The result is at odds with Fauver and Fuerst (2006), who fi nd a positive relationship when a dummy employee director variable is inter-acted with information intensive industries. In sub-samples of information intensive and other industries I confi rm the negative employee director correlation to Tobin’s Q. Overall, the results support agency theory and reject stakeholder theory.

The indirect eff ects are also present. Employee directors are positively associated with average wage, the board index, and, in co-determined samples, leverage. For the board index, this means that shareholders improve board composition so as to neutralize the negative employee direc-tor eff ect, as Buchanan and Tullock (1962) predict. However, this neutral-izing eff ect falls far short of the negative direct employee director eff ect. The lagged fi rm performance is signifi cantly positively related to the board index and negatively to leverage. This result runs counter to the Hermalin and Weisbach (1998) reverse causation theory that earlier fi rm performance determines board composition. Thus, the results show endogeneity eff ects, but the economic signifi cance falls far below the importance of the direct eff ect. The negative direct eff ect of employee directors is only partially com-pensated for by a better board. Endogeneity matters, but not very much.

Furthermore, leverage turns out to be negatively related to fi rm perform-ance, contrary to the free cash fl ow hypothesis (Easterbrook, 1984; and Jensen, 1986). The negative association with fi rm performance confi rms fi ndings in empirical studies (Barclay et al., 1995; Rajan and Zingales, 1995; Brick et al., 2005).

Jensen and Meckling (1979) argue that co-determination can only survive if supported by law. The long-term data set employed here sup-ports this view. Evidently, owners have good economic reasons for not


choosing the co-determination form of organization if they can. This also implies that there are costs to maintaining co-determination required by law. First, I document the negative impact of employee representation upon fi rm performance. Second, shareholders try to work around the regulations by strengthening aspects of board characteristics that are left unregulated. Thus, co-determination, supported by law, has costs. Therefore, these results are relevant for the emerging literature on board regulation (Hermalin, 2005).

NOTES

* Acknowledgements: I have benefi ted from comments made by Øyvind Bøhren, Ole Gjølberg, Roswitha King, Gudbrand Lien, participants at the 7th workshop on Corporate Governance and Investment, Jönköping, 2006, and at the 2nd International Business Economics Workshop, Majorca, 13–14 September 2007. Pål Rydland and Bernt Arne Ødegaard have guided me to data.

1. Co-determination is defi ned as employee board representation (Jensen and Meckling, 1979; Furubotn, 1988).

2. Tirole (2002, p. 118) argues that these ‘[c]onfl icts of interest among the board generate endless haggling, vote-trading and log-rolling. They also focus managerial attention on the delicate search for compromises that are acceptable to everyone; managers thereby lose a clear sense of mission and become political virtuosos.’ In a similar vein, Hansmann (1996, p. 44) states that ‘because the participants [that is, stakeholders] are likely to have radically diverging interests, making everybody an owner threatens to increase the costs of collective decision making enormously.’

3. According to the EIRO (1998) full employee representation is found in Austria, the Nordic countries, and Germany, while the Netherlands and France have systems closer to a consultative function for employee representatives.

4. In a recent booklet, the long-time employee director Svein Stugu (2006) says that the main objective is to prevent plant closures. Mergers, takeovers, and outsourcing must also be prevented.

5. Stugu (2006, p. 63) says that opposition to plant closures was organized in co-operation with representatives of the local community, but that this could only be done eff ectively if labour representatives had access to internal information.

6. The variables are defi ned as follows. Directors’ holdings is defi ned as the fraction of equity owned by the board of directors; board network is the information centrality, constructed from network theory (Wasserman and Faust, 1994), see note 8 below; board size is the number of shareholder elected directors; gender is the proportion of shareholder elected female directors.

7. I keep only a linear specifi cation in the board ownership relation, despite evidence in Morck et al. (1988) and McConnell and Servaes (1990) pointing towards a concave relationship. The Bøhren and Strøm (2008) study fi nds no signifi cance in the squared term, maybe due to the inclusion of other board characteristics.

8. Network theory uses concepts such as nodes and lines. In our setting, a node is a fi rm, and a line between two fi rms represents a joint director in the two fi rms. We defi ne geodesic gjk as the shortest path between two nodes j and k, and G as the total number of nodes. The node i is designated as ni. Using Wasserman and Faust (1994, pp. 192–7), our information centrality measure is constructed in the following way. Form the G 3 G matrix A with diagonal elements aii equal to 1 plus the sum of values for all lines incident to ni and off -diagonal elements aij, such that aij 5 0 if nodes ni and nj are not adjacent, and aij 5 1 2 xij if nodes ni and nj are adjacent. xij is the value of the link from fi rm ni


to nj, that is, 0 or 1. The inverse of A, which is C 5 A21, has elements {cij}, where we defi ne T 5 gG

i51cii and R 5 gGj51cij. The information centrality index for fi rm ni is:

Ci(ni) 51

cii 1 (T 2 2R) /G The index measures the information content in the paths that originate and end at a

specifi c fi rm. 9. The OSE had an aggregate market capitalization of 68 billion USD equivalents by year-

end 2002, ranking the OSE 16 among the 22 European stock exchanges for which compa-rable data are available. During the sample period from 1989 to 2002, the number of fi rms listed increased from 129 to 203, market capitalization grew by 8 per cent per annum, and market liquidity, measured as transaction value over market value, increased from 52 per cent in 1989 to 72 per cent in 2002 (sources: www.ose.no and www.fi bv.com).

10. The main sources are Bråthen (1982) and Aarbakke et al. (1999) and a government report (NOU 1985:1). In order to maintain readability, specifi c references have been dropped in tables and text.

11. For every individual fi rm, an overall average is constructed. Then, from each company observation the overall individual average is subtracted.

12. In addition to the variables in (2), I include outside owner concentration, independence, CEO director, exported and imported directors, and board age dispersion. Outside owner concentration is the sum of squared equity fractions across all the fi rm’s outside owners; independence is the board tenure of the non-employee directors minus the tenure of the CEO; CEO director equals 1 if the CEO is a member of his company’s board and zero otherwise; exported CEO is the number of outside directorships held by the fi rm’s CEO; imported CEO is the proportion of CEOs from other companies on the board; board age dispersion is the standard deviation of board age.

REFERENCES

Aarbakke, M., J. Skåre, G. Knudsen, T. Ofstad and A. Aarbakke (1999), Aksjeloven og Allmennaksjeloven (Company Law), Oslo: Tano Aschehoug.

Agrawal, A. and C.R. Knoeber (1996), ‘Firm performance and mechanisms to control agency problems between managers and shareholders’, Journal of Financial and Quantitative Analysis, 31 (3), 377–97.

Baker, G.P., M.C. Jensen and K.J. Murphy (1988), ‘Compensation and incentives: Practices vs. theory’, Journal of Finance, 43 (3), 593–616.

Barclay, M.J., C.W. Smith and R.L. Watts (1995), ‘The determinants of corporate leverage and dividend policies’, Journal of Applied Corporate Finance, 7 (4), 4–19.

Becht, M., P. Bolton and A. Röell (2003), ‘Corporate governance and control’, in G. Constantinides, M. Harris and R. Stulz (eds), Handbook of the Economics of Finance, Volume 1A, Amsterdam: North-Holland, pp. 1–109.

Bertrand, M. and S. Mullainathan (2001), ‘Are CEOs rewarded for luck? The ones without principals are’, Quarterly Journal of Economics, 116 (3), 901–32.

Bhagat, S. and R.H. Jeff eris (2002), The Econometrics of Corporate Governance Studies, Cambridge, MA: The MIT Press.

Blair, M.M. (1995), Ownership and Control: Rethinking Corporate Governance for the Twenty-fi rst Century, Washington DC: The Brookings Institution.

Blair, M.M. and L.A. Stout (1999), ‘A team production theory of corporate law’, Virginia Law Review, 85 (2), 247–328.


Bøhren, Ø. and R.Ø. Strøm (2008), ‘Aligned, informed, and decisive: characteris-tics of value-creating boards’, in R.Ø. Strøm, Three Essays on Corporate Boards, Dr. Oecon. thesis, BI Norwegian School of Management, Oslo, Norway, pp. 21–60.

Bråthen, T. (1982), Bedriftsforsamlingen i aksjeselskaper (The Supervisory Board in Shareholder Owned Companies), Oslo: Tanum Nordli.

Brick, I.E., D. Palia and C.-J. Wang (2005), ‘Simultaneous estimation of CEO compensation, leverage, and board characteristics on fi rm value’.

Buchanan, J.M. and G. Tullock (1962), The Calculus of Consent: Logical Foundations of Constitutional Democracy, Ann Arbor: The University of Michigan Press.

Cadbury Committee (1992), Report of the Committee on the Financial Aspects of Corporate Governance, London: Gee and Co. Ltd.

Carter, D.A., B.J. Simkins and W.G. Simpson (2003), ‘Corporate governance, board diversity and fi rm value’, Financial Review, 38, 33–53.

Dow, G.K. (2003), Governing the Firm: Workers’ Control in Theory and Practice, Cambridge, UK: Cambridge University Press.

Easterbrook, F.H. (1984), ‘Two agency-cost explanations of dividends’, American Economic Review, 74 (4), 650–59.

Eisenberg, T., S. Sundgren and M.T. Wells (1998), ‘Larger board size and decreas-ing fi rm value in small fi rms’, Journal of Financial Economics 48, 35–54.

European Industrial Relations Observatory On-line (EIRO) (1998), ‘Board-level employee representation in Europe’, September.

Falaye, O., V. Mehrotra and R.Morck (2006), ‘When labor has a voice in corporate governance’, Journal of Financial and Quantitative Analysis, 41 (3), 489–510.

Fauver, L. and M.E. Fuerst (2006), ‘Does good corporate governance include employee representation? Evidence from German corporate boards’, Journal of Financial Economics, 82 (3), 673–710.

FitzRoy, F.R. and K. Kraft (1993), ‘Economic eff ects of codetermination’, Scandinavian Journal of Economics, 95 (3), 365–75.

Freeman, R.B. and E.P. Lazear (1995), ‘An economic analysis of works councils’, in J. Rogers and W. Streeck (eds), Works Councils: Consultation, Representation, and Cooperation in Industrial Relations, Chicago: The University of Chicago Press, pp. 27–50.

Freeman, R. E. and D. L. Reed (1983), ‘Stockholders and stakeholders: a new per-spective on corporate governance’, California Management Review, 25, 88–106.

Furubotn, E.G. (1988), ‘Codetermination and the modern theory of the fi rm: a property-rights analysis, Journal of Business, 61 (2), 165–81.

Goergen, M. (2007), ‘What do we know about diff erent systems of corporate gov-ernance?’, ECGI Finance Working Paper No. 163/2007.

Gompers, P., J. Ishii and A. Metrick (2003), ‘Corporate governance and equity prices’, Quarterly Journal of Economics, 118 (1), 107–55.

Gorton, G. and F.A. Schmid (2000), ‘Universal banking and the performance of German fi rms’, Journal of Political Economy, 58 (1–2), 29–80.

Gorton, G. and F.A. Schmid (2004), ‘Capital, labor, and the fi rm: a study of German codetermination’, Journal of the European Economic Association, 2 (5), 863–905.

Greene, W.H. (2003), Econometric Analysis, 5th edn, New York: Prentice Hall.Hansmann, H. (1996), The Ownership of Enterprise, Cambridge, MA: Harvard

University Press.Harris, M. and A. Raviv (1991), ‘The theory of capital structure’, Journal of

Finance 46 (1), 297–355.


Hermalin, B.E. (2005), ‘Trends in corporate governance’, Journal of Finance, 60 (5), 2351–84.

Hermalin, B.E. and M.S. Weisbach (1998), ‘Endogenously chosen boards of directors and their monitoring of the CEO’, American Economic Review, 88 (1), 96–118.

Hermalin, B.E. and M.S. Weisbach (2003), ‘Boards of directors as an endogenously determined institution: a survey of the economic literature’, Economic Policy Review, 9 (1), 7–26.

Hopt, K.J. (1998), ‘The German two-tier board: experience, theories, reform’, in K.J. Hopt, H. Kanda, M.J. Roe, E. Wymeersch, and S. Prigge (eds), Comparative Corporate Governance: The State of the Art and Emerging Research, Oxford: Oxford University Press, pp. 227–58.

Hsiao, C. (2003), Analysis of Panel Data, 2nd edn, Cambridge: Cambridge University Press.

Jensen, M.C. (1986), ‘Agency cost of free cash fl ow, corporate fi nance and take-overs’, American Economic Review, 76, 323–39.

Jensen, M.C. and W. Meckling (1979), ‘Rights and production functions: an application to labor-managed fi rms and codetermination’, Journal of Business, 52 (4), 469–506.

Johnson, R.A. and D.W. Wichern (1988), Applied Multivariate Statistical Analysis, 2nd edn, Englewood Cliff s, NJ: Prentice-Hall.

McConnell, J. and H. Servaes (1990), ‘Additional evidence on equity ownership and corporate value’, Journal of Financial Economics 27, 595–612.

Morck, R., A. Shleifer and R. Vishny (1988), ‘Management ownership and market valuation: an empirical analysis’, Journal of Financial Economics, 20, 293–315.

Mueller, D.C. (2003), Public Choice III, Cambridge, UK: Cambridge University Press.

Myers, S.C. (1977), ‘Determinants of corporate borrowing’, Journal of Financial Economics, 5, 147–75.

NOU (1985), Videreutviklingen av bedriftsdemokratiet (Further development of co-determination), Oslo: Statens forvaltningstjeneste.

Palia, D. (2001), ‘The endogeneity of managerial compensation in fi rm valuation: a solution’, Review of Financial Studies, 14 (3), 735–64.

Perotti, E. and K.E. Spier (1993), ‘Capital structure as a bargaining tool: the role of leverage in contract renegotiation’, American Economic Review, 83 (5), 1131–41.

Pfeff er, J. (1981), Power in Organizations, Cambridge, MA: Ballinger Publishing Company.

Pistor, K. (1999), ‘Codetermination: a socio-political model with governance externalities’, in M.M. Blair and M.J. Roe (eds), Employees and Corporate Governance, Washington, DC: Brookings Institution Press, pp. 163–93.

Raheja, C.G. (2005), ‘Determinants of board size and composition: a theory of cor-porate boards’, Journal of Financial and Quantitative Analysis, 40 (2), 283–306.

Rajan, R.G. and L. Zingales (1995), ‘What do we know about capital structure? Some evidence from international data’, Journal of Finance, 50 (5), 1421–60.

Schmid, F.A. and F. Seger (1998), ‘Arbeitnehmermitbestimmung, Allokation von Entscheidungsrechten und Shareholder Value’ (Codetermination, decision right allocation, and shareholder value), Zeitschrift für Betriebswirtschaft, 68 (5), 453–73.

Shrader, C.B., V.B. Blackburn and P. Iles (1997), ‘Women in management and


fi rm fi nancial performance: an exploratory study’, Journal of Managerial Issues, 9 (3), 355–72.

Siebert, H. (2005), The German Economy: Beyond the Social Market, Oxford: Oxford University Press.

Smith, N., V. Smith and M. Verner (2006), ‘Do women in top management aff ect fi rm performance? A panel study of 2300 Danish fi rms’, International Journal of Productivity and Performance Management, 55 (7), 569–93.

Stugu, S. (2006), Orkla: Bedriftskultur og bedriftsdemokrati 1991–2004 (Orkla: Company culture and democracy 1991–2004), Oslo: DeFacto.

Tirole, J. (2001), ‘Corporate governance’, Econometrica, 69 (1), 1–36.Tirole, J. (2002), Financial Crises, Liquidity, and the International Monetary System,

Princeton: Princeton University Press.Tirole, J. (2006), The Theory of Corporate Finance, Princeton: Princeton University

Press.Wasserman, S. and K. Faust (1994), Social Network Analysis: Methods and

Applications, Cambridge, UK: Cambridge University Press.Watts, R.L. (2003), ‘Conservatism in accounting: Part I: Explanations and implica-

tions’, Accounting Horizons, 17 (3), 207–21.Woolridge, J.M. (2002), Econometric Analysis of Cross Section and Panel Data,

Cambridge, MA: The MIT Press.Yermack, D. (1996), ‘Higher market valuation of companies with a small board of

directors’, Journal of Financial Economics, 40, 185–212.Zingales, L. (2000), ‘In search of new foundations’, Journal of Finance, 55 (4),

1623–53.

361

14. The determinants of German corporate governance ratingsWolfgang Drobetz, Klaus Gugler and Simone Hirschvogl

1. INTRODUCTION

In recent years many countries have introduced ‘corporate governance codes’. These codes represent unmistakable improvements in minority shareholder right protection as well as transparency, and they generally entail a movement towards Anglo-Saxon institutions. Many of the rules in these codes are only recommendations, however, and there is much scepticism that best-practice recommendations and/or principles-based approaches are eff ective substitutes for more rule-based approaches, such as the US Sarbanes-Oxley Act. This is all the more the case since there is the widespread perception that markets do not function well in punish-ing deviant behaviour of managers, particularly in Continental Europe, where regulators tend to rely heavily on principles-based approaches in their attempts to reform corporate governance. There are many reasons to believe that markets are less of a constraint on managerial discretion in Continental Europe than in the US or the UK, in particular. For example, ownership and voting right concentration is tremendous, liquidity of shares is low, and there is frequently a separation between cash fl ow and voting rights.1 In general, therefore, the ‘exit option’ is less of a threat to fi rms’ management, and the ‘voice’ of institutional investors, in particular, ought to be strengthened.

The existing literature on codes is scant at best, and if it exists it is on the eff ects of corporate governance codes on performance.2 Most recently, Drobetz et al. (2004) construct a corporate governance rating for 91 German fi rms and fi nd that the rating of a fi rm positively aff ects market value and the returns to shareholders. Their empirical analysis reveals that for the median fi rm a one standard deviation change in the governance rating results in about a 24 per cent increase in the value of Tobin’s Q.


This evidence notwithstanding, there are problems with the link between corporate governance code ratings and fi rm performance. First, there is the well-known endogeneity problem already mentioned above: if only ‘good’ fi rms adopt the code (for example, because the costs are low, since they fulfi l the code anyway), one must expect a positive code–performance relationship.3 Second, a high rating on the code only indirectly aff ects performance: a high rating must correlate with the true ‘spirit’ of good governance, which only then can aff ect performance.4 In fact, very little is known about the underlying mechanism that relates corporate governance practices and fi rm performance (for example, see Shleifer and Wolfenzon, 2002).

A more cautious approach of analysing corporate governance codes is adopted in this chapter. We take one step back and do not try to assess the impact of code fulfi lment on the performance of companies (which is, of course, ultimately the most interesting question). Instead, we use the corporate governance rating constructed by Drobetz et al. (2004) for publicly listed German fi rms and analyse the determinants of this rating. This approach has the advantage that we do not run into the same endo-geneity problems with the determinants of code fulfi lment that we would encounter by trying to assess the eff ects on fi rm performance. For example, one cannot sensibly argue that a high or low governance rating aff ects the voting rights of the largest shareholder or the size/composition of the supervisory board. It must be the case that the decision making process is determined or at least monitored by the largest shareholder and/or the board, and their decisions naturally aff ect compliance with the code. The decision to improve corporate governance practices and attitudes should be made in awareness of its consequences and obligations (for example, see Demsetz and Lehn, 1985). However, we only have a cross-section of data at hand, which may also limit our analysis.

Our results show that there is a non-linear relationship between owner-ship concentration and the corporate governance rating. Moreover, fi rms with larger boards have lower ratings, but fi rms that apply US-GAAP or IAS rules or use an option-based remuneration plan have higher ratings.

The remainder of this chapter is structured as follows. Section 2 devel-ops our hypotheses, which are subject to empirical testing. Because our corporate governance rating mainly refers to the rules and recommenda-tions of the German Corporate Governance Code, we give a brief and general comparative analysis of the governance codes in place throughout the European Union in Section 3. Section 4 describes the data, Section 5 presents our empirical results, and Section 6 concludes.

The determinants of German corporate governance ratings 363

2. HYPOTHESES

Germany is the prototype of an insider system of fi nance and control, and thus our hypotheses as to the determinants of ratings must refl ect its insti-tutional background. The most striking fact of even large, listed fi rms in Germany is that ownership and voting right concentration is tremendous. While the median largest ultimate voting block in US or UK listed fi rms is well below 10 per cent, it is above 50 per cent in Germany, Italy and Austria (see Becht and Röell, 1999). Therefore, presumably the owner of this block has ultimate control over the company and can decide which stance to adopt with regard to the code of good corporate governance. Accordingly, we hypothesize that the voting power of the largest shareholder aff ects the code rating. We also develop the notion that the size of the board of direc-tors, a fi rm’s accounting principles, and its method of executive remunera-tion impact the code ratings.

2.1 Ownership Concentration

In the literature two main eff ects of large shareholders have been disentan-gled (for example, see Claessens et al., 2002; Gugler et al., 2003a). First, with increasing cash fl ow rights of the largest shareholder, there is a posi-tive incentive eff ect. A good code rating – provided it is awarded by the capital market – increases the value of the fi rm and, hence, the value of the ownership stake of the largest shareholder. S/he should therefore have an incentive to comply with the code. However, there is a second, nega-tive entrenchment eff ect. The larger the voting rights of the largest share-holder, the more entrenched s/he is and the more s/he can infl uence the decision making process. A high code rating achieved by making it easier for small shareholders to cast their votes in general assemblies, increasing transparency by disclosing information on individual compensation of management and the supervisory board, or agreeing to strict incompat-ibility regulation, to give a few examples, is not necessarily in the largest shareholder’s interest. We summarize the discussion as follows:

Hypothesis 1 Ownership concentration is non-linearly related to the cor-porate governance rating. At low to intermediate holdings of the largest shareholder the entrenchment eff ect outweighs the incentive eff ect and we expect a negative relation between ownership concentration and the corporate governance rating. At high levels of ownership concentra-tion the incentive eff ect outweighs the entrenchment eff ect and, hence, we expect a positive relation between ownership concentration and the corporate governance rating.


2.2 Board Size

Our second determinant of code compliance is the size of the supervisory board. The decision making process in the supervisory board is likely to be aff ected by its size for at least two reasons. First, coordination problems are larger on a large board than on a small board. Jensen (1993) and Lipton and Lorsch (1992) suggest that large boards can be less eff ective than small boards, presuming that the emphasis on politeness and courtesy in board-rooms is at the expense of truth and frankness. Specifi cally, when boards become too big, agency problems (such as director free-riding) increase and the board becomes more symbolic and neglects its monitoring and control duties. Moreover, large boards may refl ect an inadequate percep-tion of the true executive function, particularly in fi rms with public involve-ment. Supporting this rather ad hoc proposition, Yermack (1996) was the fi rst to report empirical evidence for a negative relationship between board size and fi rm valuation (see also Eisenberg et al., 1998; Beiner et al., 2004). Second, on a large board it is likely that more confl icting groups of stakeholders, such as representatives of large shareholders, employees, and creditors, are represented than on smaller boards. Third, many companies do have a (and if so, at most one) representative of small shareholders. However, the larger the board the less weight this representative has at a ballot. All of these arguments lead us to:

Hypothesis 2 Larger boards tend to be reluctant to adopt ‘good’ corpo-rate governance practices and, hence, board size is negatively related to the corporate governance rating.

2.3 Accounting Principles

There are several papers that fi nd signifi cant eff ects of accounting practices on the performance of companies as well as on the distribution of profi ts among stakeholders, such as dividends or interest payments on debt (see, for example, La Porta et al., 1997, 1998, 2000; Gugler et al., 2003b, 2004). In Germany there are three possibilities as to how fi rms are allowed to account: US-GAAP (US Generally Accepted Accounting Principles), IAS (International Accounting Standards) and HGB (‘Handelsgesetzbuch’). US-GAAP and IAS contain much stricter rules on accounting practices than HGB, which is the national law standard for accounting, particu-larly with respect to transparency and details of information. Due to its conservative approach (for example, historical cost accounting), HGB accounting appears to favour debtholders and large shareholders versus minority shareholders.


Accounting according to international standards and compliance with the code can be viewed as complements for a number of reasons. First, many of the requirements of code compliance are antedated by the decision to account according to international principles. Thus the marginal costs of code compliance are smaller for these fi rms than for fi rms using HGB. Second, fi rms that account with US-GAAP or IAS may want to signal their good investment opportunities, and code compliance is one way to achieve this goal. Finally, although we explicitly account for fi rm size (total assets) in the determinants regressions below, part of the co-variation in account-ing principles and code rating may be attributable to fi rm size (for example, due to measurement errors of true fi rm size), which is a main determinant of international accounting. Accordingly, we formulate:

Hypothesis 3 Firms accounting according to US-GAAP or IAS have higher corporate governance ratings than fi rms accounting according to HGB.

2.4 Executive Remuneration

Our fi nal variable aff ecting code rating is whether or not the fi rm has adopted an option-based remuneration plan. Diamond and Verrechia (1982) and Holmström and Tirole (1993) developed models that are based on the interaction of capital markets and contingent compensation. Giving managers an equity stake in the fi rm is a solution to ensure that managers pursue the interests of shareholders without necessarily increasing manage-rial entrenchment. Provided that a high governance rating is awarded by the capital market, management of fi rms using option-based remuneration has an incentive to comply with the code. Therefore, we formulate:

Hypothesis 4 Firms that use an option-based remuneration plan have higher corporate governance ratings than other fi rms.

3. CODES OF GOOD CORPORATE GOVERNANCE

3.1 European Corporate Governance Codes

Recently, all EU member states have adopted at least one governance code document.5 It is generally acknowledged that the legal framework for corpo-rate governance is most eff ective if it aims at ensuring: (i) fair and equitable treatment of all shareholders, (ii) managerial and supervisory body account-ability, (iii) transparency as to corporate performance, ownership structure and governance, and (iv) corporate responsibility. While the codes originate


from countries with very diverse cultures, fi nancing traditions, ownership structures, and legal origins, they are remarkably similar in their general notion of ‘best practice’ corporate governance rules. In fact, codes appear to serve as a converging force in corporate governance practices.

Nevertheless, two observations are noteworthy. First, the coverage of the codes diff ers substantially due to diff erences in legal origins and frameworks. While some codes address general principles and practices of corporate governance, other nations establish these in company laws and securities regulation. Second, while some codes strongly emphasize the supervisory body holding managers accountable to a broad base of rela-tively dispersed shareholders (for example, in the UK), other codes focus on the protection of minority shareholders to ensure equal treatment to a dominant shareholder (for example, in Germany).

The codes have three stated objectives: (i) stakeholder and/or share-holder interests, (ii) the work of supervisory and managerial bodies, and (iii) disclosure requirements. The majority of codes recognize that corpo-rate success, shareholder profi t, employee security and well-being, and the interests of other stakeholders are strongly interrelated. They generally call for shareholders to be treated equitably, disproportional voting rights to be avoided or at least fully disclosed to all shareholders, and removal of bar-riers to shareholder participation in general meetings, whether in person or by proxy.6 Despite structural diff erences between two-tier and unitary board systems, they all stress that supervisory responsibilities are distinct from management responsibilities. Many suggest practices designed to enhance the distinction between the roles of the supervisory and manage-rial bodies, including supervisory body independence, separation of the chairman and CEO roles, and reliance on board committees (such as the nominating committee).7 Finally, all codes contain various disclosure requirements. An issue that has received specifi c public attention is the greater voluntary transparency as to executive and director compensa-tion.8 In addition, the codes also support the increasing public interest in disclosure as regards director independence (in both one-tier board and two-tier board systems), share ownership, and, in many instances, issues of broader social concerns.

With regards to code enforcement, the prescriptions supplement and complement the mandatory prescriptions provided by company and secu-rities laws and listing rules. However, they are non-imperative and lack mandatory compliance authority. The vast majority of codes merely require companies to provide greater voluntary disclosure of governance practices, including disclosure about the extent of compliance with a par-ticular code recommendation. Listed companies are required to disclose whether they comply with the specifi ed code and explain any deviations


(‘comply or explain’). Even though compliance with code provisions is wholly voluntary, reputational market forces can result in signifi cant com-pliance pressures. Finally, codes are increasingly used by investors and market analysts, rating agencies, shareholder monitoring groups and com-mentators to benchmark supervisory and management bodies.

3.2 The German Corporate Governance Code

After a few private interest groups began establishing best practices of corporate governance in the late 1990s, in June 2000 the German federal government appointed a commission with the goal to formulate proposals for modernizing German corporate law. This report prepared the ground for the development of a national code for improving the management and control functions of publicly quoted companies. The results were elabo-rated into the code of conduct by a second, follow-up commission. The German Corporate Governance Code was fi nally published on 26 February 2002, and the Transparency and Disclosure Act (TransPuG), which took eff ect on 26 July 2002, obliges publicly quoted companies to apply the code recommendations. The code is an example of self- commitment by the corporate sector and requires disclosure on the ‘comply or explain’ rule described in Section 3.1.

The stated goal of the code is to ‘promote the trust of international and national investors, customers, employees and the general public in the management and supervision of listed German stock corporations’.9 This is in contrast to the Anglo-Saxon view of corporate governance, where there is little room for the general public. Nevertheless, the code constitutes a regime shift in the German corporate governance system by taking a sur-prisingly pragmatic view on the ‘fundamental’ diff erences in stakeholder and shareholder interests, an issue that has been fi ercely debated in particu-lar in the German literature (for example, see Albach, 2003).

4. DATA DESCRIPTION

4.1 A German Corporate Governance Rating

The corporate governance rating applied in this chapter is from Drobetz et al. (2003, 2004). They construct a broad, multifactor corporate governance rating, which is based on responses to a survey sent out to a broad sample of German publicly listed fi rms. To qualify for inclusion in the corporate governance rating, each practice and attitude (i) had to refer to a governance element that is not (yet) legally required and (ii) needed to be considered as


international market practice from an investor’s perspective. Most proxies included in the rating represent recommendations and suggestions of the German Corporate Governance Code. Note that while the former work according to the comply-or-explain principle, the latter are wholly volun-tary. A few other governance proxies originate from the DVFA German Corporate Governance Scorecard,10 from CalPERS German Market Principles, and from the Deminor Corporate Governance Checklist.

In total, the rating contains 30 governance proxies divided into fi ve categories: (1) corporate governance commitment, (2) shareholder rights, (3) transparency, (4) management and supervisory board matters, and (5) auditing. A representative question from each category is listed below:

Are there fi rm-specifi c corporate governance guidelines set out in ●

writing?Are there measures in place to facilitate the personal exercising of ●

shareholder voting rights (for example, via internet) and to assist the shareholders in the use of proxies?Are the fi xed and variable remuneration elements as well as share ●

ownership (including existing option rights) of members of the management and supervisory board published separately and in individualized form in the notes to the fi nancial statements?Are there supervisory board committees to deal with complex ●

matters (such as audit, compensation, strategy)?Are there fi rm-specifi c rules to ensure that the auditor does not ●

perform other services for the fi rm (such as consulting work)?

A questionnaire with all 30 governance proxies was sent out to all fi rms in the four principal market segments of the German stock exchange: DAX 30 (blue-chip stocks), MDAX (mid-cap stocks), NEMAX 50 (index of growth fi rms), and SDAX (small-cap stocks), comprising a total of 253 fi rms. Data collection was completed at the end of March 2002. Overall, the survey had a response ratio of 36 per cent, which results in a sample of 91 German fi rms.

The construction principles of the aggregate governance rating are kept simple. Twenty-fi ve basis points are added for each acceptance level of the respective proxy in a fi ve-scale answering range. For each fi rm the aggregate rating is an unweighted sum of the basis points across all proxies, ranging from 0 (minimum) to 30 (maximum).11 Hence, the dependent variables in our empirical analysis are: OVERALL (aggregate corporate govern-ance rating), CG_UNT (governance commitment), CG_AKT (shareholder rights), CG_TRA (transparency), CG_ENT (management and supervisory board matters), and CG_ABS (auditing).


The histogram in Figure 14.1 shows that the rating over the 91 fi rms in our sample is slightly skewed to the right. More than 40 per cent of the fi rms have a rating between 20 and 23. Nevertheless, governance proxies display a suffi ciently wide distribution to mitigate a possible sample selec-tion bias in the survey.

Panel A in Table 14.1 presents summary statistics of the dependent vari-ables. Due to data limitations for the independent variables, the sample in our empirical analysis is reduced to 80 fi rms. The average rating is 19.51, with fi rm ratings ranging from 9.75 to 27.25. The sub-indices with the highest ratings are CG_ENT (management and supervisory matters) and CG_TRA (transparency), which can be explained by the fact that these areas are strongly accompanied by laws and regulation.

4.2 Explanatory Variables

The data for ownership structure/voting rights are based on the CD-ROM ‘Wer gehört zu Wem?’ (‘Who owns whom?’, 30 April 2002) or taken

0

2

4

6

8

10

12

0 2 4 8 126 10 14 16 18 20 22 24 26 28 30

Num

ber

of fi

rms

Corporate governance rating

Note: This fi gure shows the distribution of the survey-based corporate governance rating (CGR) for 91 German public fi rms from Drobetz et al. (2004). The survey was sent out in February 2002, and the data collection was completed by the end of March 2002. The rating represents an unweighted sum of the basis points (on a fi ve-scale answering range) for all governance proxies in fi ve broad categories: (1) corporate governance commitment, (2) shareholder rights, (3) transparency, (4) management and supervisory board matters, and (5) auditing. The corporate governance rating ranges from 0 (minimum) to 30 (maximum). The ratings in the fi gure are rounded to the nearest integer.

Figure 14.1 Distribution of the German corporate governance rating


Table 14.1 Summary statistics

Variables Mean Median Minimum Maximum Obs.

Panel A: Aggregate rating and componentsOVERALL 19.51 19.75 9.75 27.25 85CG_UNT 2.27 2.00 0.00 5.00 85CG_AKT 3.07 3.00 0.00 5.00 85CG_TRA 4.55 4.75 2.00 5.00 85CG_ENT 5.98 6.25 1.25 9.50 85CG_ABS 3.63 3.75 1.00 5.00 85Panel B: Aggregate rating by ownership concentrationVR1<25% 21.42 3325%≤VR1.50% 18.67 2150%≤VR1.75% 17.30 21VR1≥75% 19.47 8Panel C: Independent variablesVR1 37.13 31.85 4.60 100.00 80VR1^2 2112.94 1014.45 21.16 10,000.00 80VR1_25 19.80 25.00 4.60 25.00 85VR1_25to50 12.01 9.10 0.00 25.00 85VR1_50 6.45 0.00 0.00 50.00 80BOARDSIZE 10.29 8.00 3.00 21.00 85GAAP 0.26 0.00 0.00 1.00 85IAS 0.32 0.00 0.00 1.00 85OPTION 0.60 1.00 0.00 1.00 85TA 13.78 13.22 8.26 20.64 85TQ 1.63 1.17 0.46 8.02 85

Notes:The variables relating to the corporate governance score are drawn from the study by Drobetz et al. (2004), estimates on the ownership structure are based on the CD-ROM ‘Wer gehört zu Wem’ (‘Who owns whom?’, 30 April 2002) and from ‘BaFin (Bundesanstalt für Finanzdienstleistungsaufsicht)’ March 2002; all the rest of the calculations are based on annual reports as of end 2001. The independent variables are: OVERALL (corporate governance rating), CG_UNT (governance commitment), CG_AKT (shareholder rights), CG_TRA (transparency), CG_ENT (management and supervisory board matters), CG_ABS (auditing). The corporate governance variables are: VR1 denotes the voting rights of the largest ultimate shareholder, and VR1^2 is the squared value of VR1. VR1_25 equals the voting rights of the largest shareholder if the voting rights are below 25%; in any other case it is set to 25%. VR1_25to50 is 0 if the voting rights of the largest investor are below 25%; if the voting rights are beyond or equal to 25% and below 50%, this variable is calculated as voting rights –25%. VR1_50 is set to 0 if the voting rights are below 50%; in any other case it is computed as voting rights –50%. BOARDSIZE is the number of directors on the company’s supervisory board. GAAP is a dummy variable and equals 1 if US-GAAP are used as accounting standards in the annual reports and equals 0 otherwise. IAS is a dummy variable and is set to 1 if IAS are used as accounting standards in the annual reports and equals 0 otherwise. OPTION is a dummy variable and equals 1 if the fi rm uses an option-based remuneration plan and 0 otherwise. TA is defi ned as the natural logarithm of the book value of total assets. TQ equals the ratio of market value of equity plus liabilities divided by book value of total assets.


from ‘BaFin’ (Bundesanstalt für Finanzdienstleistungsaufsicht, March 2002). All other variables are based on annual reports as of end 2001. To appropriately capture the distribution of control rights and decision power among shareholders, we use voting concentration as a proxy for ownership concentration. Following the hypothesis formulated in Section 2, we con-struct several variables related to voting concentration. VR1 denotes the voting rights of the largest ultimate shareholder. To account for a possibly non-linear relationship between ownership concentration and the corpo-rate governance rating, VR1^2 is the squared value of VR1. Alternatively, we follow Morck et al. (1988) and use the following variables to estimate piecewise linear regressions:

VR1_25 5 voting rights of the largest shareholder if voting rights , 25%, 5 25% if voting right of the largest shareholder $ 25%;VR1_25to50 5 0 if the voting rights of the largest shareholder , 25%, 5 voting rights of the largest shareholder minus 25% if 25% ≤ voting rights , 50%, 5 25% if voting rights of the largest shareholder $ 50%;VR1_50 5 0 if voting rights of the largest shareholder , 50%, 5 voting rights minus 50% if voting rights $ 50%.

Panel B in Table 14.1 presents a breakdown of the aggregate corporate governance rating by four breaking points of ownership concentration. The average rating is higher than 21 points if the largest shareholder holds less than 25 per cent in voting rights, but it is lower than 19 (18) points if VR1 is larger than 25 per cent (50 per cent) but smaller than 50 per cent (75 per cent). The average rating again increases above 19 points if the fi rm is in super-majority control (VR1.75%). This hints at possible non-linear eff ects of the largest shareholder on fi rms’ ratings.

BOARDSIZE denotes the number of directors on the company’s supervisory board. The data are taken from the fi rms’ annual reports as of year-end 2001. GAAP is a dummy variable and equals 1 if a fi rm uses US-GAAP as the accounting standards in its annual report, and equals 0 otherwise. Similarly, IAS is a dummy variable and is set to 1 if IAS are used as accounting standards, and equals 0 otherwise. OPTION is a dummy variable that equals 1 if the fi rm uses an option-based remuneration plan, and 0 otherwise. Finally, we use two additional control variables: (i) SIZE is defi ned as the natural logarithm of the book value of total assets, and (ii) TQ refers to the Tobin’s Q, approximated as the ratio of market value of equity plus liabilities divided by book value of total assets. Summary statis-tics of the independent variables are presented in Panel C of Table 14.1.


5. EMPIRICAL RESULTS

5.1 Main Empirical Results

Table 14.2 presents our main results. In equation (1), VR1 and SIZE are the only independent variables. The corresponding coeffi cient on VR1 is negative and statistically signifi cant at the 5 per cent level. Controlling for size, we observe that larger voting rights are associated with lower governance ratings, indicating that the entrenchment eff ect, on average, dominates the alignment eff ect. We will explore this relation in more detail below. As expected, the coeffi cient on SIZE is positive and statistically sig-nifi cant. The explanatory power (adjusted R2) for this simplest regression specifi cation is almost 20 per cent.

In equation (2), BOARDSIZE is included as an additional explanatory variable. Confi rming our second hypothesis, the relationship between board size and the governance rating is signifi cantly negative. The analy-sis again controls for fi rm size, taking into account that larger fi rms also possess larger boards. This result confi rms the hypothesis by Jensen (1993) and Lipton and Lorsch (1992), suggesting that larger boards are hampered by coordination and communication problems. In addition, the decision fi nding process may be complicated by more confl icting groups of stake-holders on larger boards.

Equation (3) contains the full set of explanatory variables, where the possibly non-linear relationship between the corporate governance rating and the voting rights by the largest shareholders is captured using the three variables related to the breakpoints described in Section 4.2. We fi nd supporting evidence for all four hypotheses. First, there is some evi-dence that the relationship between the corporate governance rating and ownership concentration is non-linear. At intermediate holdings of the largest shareholder the entrenchment eff ect dominates the incentive eff ect, as indicated by the negative and signifi cant coeffi cient on the VR25_50 variable. However, with ownership concentration above 50 per cent, the incentive alignment eff ect begins to dominate, as refl ected by the positive (albeit insignifi cant) coeffi cient on the VR1_50 variable. Together, these results imply a U-shaped relationship between the corporate governance rating and ownership concentration. In addition, confi rming our second hypothesis, board size is signifi cantly negatively related to the corporate governance rating even when we include all our explanatory variables. Our empirical results further support the third hypothesis, that fi rms accounting according to US-GAAP or IAS have higher governance ratings than fi rms accounting according to HGB. This is indicated by the signifi cant positive coeffi cients on both the GAAP and IAS dummy variables. Finally, we fi nd

373

Tab

le 1

4.2

Mai

n eq

uatio

ns

Eq.

VR

1V

R1^

2V

R1_

25V

R1_

25to

50V

R1_

50B

OA

RD

SIZ

EG

AA

PIA

SO

PTIO

NT

AC

onst

.O

bs.

Adj

. R2

(1)

−0.

0300

0.54

5713

.073

780

0.19

65(−

2.18

)**

(3.7

2)**

*(5

.92)

***

(2)

−0.

0296

−0.

3429

1.25

846.

7290

800.

2793

(−2.

26)*

*(−

3.14

)***

(4.7

3)**

*(2

.31)

**(3

)0.

0061

−0.

0969

0.04

68−

0.25

363.

2071

1.66

691.

2759

0.95

557.

4942

800.

4555

(0.1

0)(−

2.09

)**

(1.5

5)(−

2.54

)**

(3.5

9)**

*(2

.31)

**(1

.81)

*(3

.87)

***

(2.7

6)**

*(4

)−

0.08

340.

0007

−0.

2447

3.09

421.

6288

1.20

020.

9627

8.28

9680

0.44

70(−

2.08

)**

(1.7

5)*

(−2.

44)*

*(3

.48)

***

(2.2

4)**

(1.6

9)*

(3.8

8)**

*(3

.06)

***

Not

es:

The

est

imat

ing

sam

ple

cont

ains

85

Ger

man

fi rm

s; va

riatio

ns a

re d

ue to

dat

a lim

itatio

ns. T

ime

perio

d: 2

002.

The

tabl

e sh

ows t

he re

sults

from

O

LS

regr

essio

ns o

f the

cor

pora

te g

over

nanc

e ra

ting

as d

epen

dent

var

iabl

e on

the

mai

n co

rpor

ate

gove

rnan

ce m

echa

nism

s alo

ng w

ith th

e co

ntro

l va

riabl

e. E

q. (1

) is a

par

tial m

odel

with

ow

ners

hip

conc

entr

atio

n ex

plai

ning

the

corp

orat

e go

vern

ance

scor

e. E

q. (2

) int

rodu

ces a

noth

er c

orpo

rate

go

vern

ance

mec

hani

sm, t

he fi

rm’s

boar

d, in

to th

e eq

uatio

n. E

q. (3

) inc

lude

s all

corp

orat

e go

vern

ance

mec

hani

sms a

nd is

sim

ilar t

o th

e pi

ece-

wise

lin

ear r

egre

ssio

n es

timat

ed b

y M

orck

et a

l. (1

988)

; how

ever

, oth

er tu

rnin

g po

ints

are

use

d. E

q. (4

) allo

ws f

or n

on-li

near

ities

by

incl

udin

g a

squa

red

term

. The

dep

ende

nt v

aria

ble

refe

rs to

the

corp

orat

e go

vern

ance

ratin

g as

cal

cula

ted

by D

robe

tz e

t al.

(200

4). T

he c

orpo

rate

gov

erna

nce

varia

bles

ar

e: V

R1

deno

tes t

he v

otin

g rig

hts o

f the

larg

est u

ltim

ate

shar

ehol

der,

and

VR

1^2

is th

e sq

uare

d va

lue

of V

R1.

VR

1_25

equ

als t

he v

otin

g rig

hts

of th

e la

rges

t sha

reho

lder

if th

e vo

ting

right

s are

bel

ow 2

5%; i

n an

y ot

her c

ase

it is

set t

o 25

%. V

R1_

25to

50 is

0 if

the

votin

g rig

hts o

f the

larg

est

inve

stor

are

bel

ow 2

5%; i

f the

vot

ing

right

s are

bey

ond

or e

qual

to 2

5% a

nd b

elow

50%

, thi

s var

iabl

e is

calc

ulat

ed a

s vot

ing

right

s –25

%. V

R1_

50 is

se

t to

0 if

the

votin

g rig

hts a

re b

elow

50%

; in

any

othe

r cas

e it

is co

mpu

ted

as v

otin

g rig

hts –

50%

. BO

AR

DSI

ZE

refe

rs to

the

num

ber o

f dire

ctor

s on

the

com

pany

’s su

perv

isory

boa

rd. G

AA

P is

a du

mm

y va

riabl

e an

d eq

uals

1 if

US-

GA

AP

are

used

as a

ccou

ntin

g st

anda

rds i

n th

e an

nual

re

port

s and

equ

als 0

oth

erw

ise. I

AS

is a

dum

my

varia

ble

and

is se

t to

1 if

IAS

are

used

as a

ccou

ntin

g st

anda

rds i

n th

e an

nual

repo

rts a

nd e

qual

s 0

othe

rwise

. OPT

ION

is a

dum

my

varia

ble

and

equa

ls 1

if th

e fi r

m u

ses a

n op

tion-

base

d re

mun

erat

ion

plan

and

0 o

ther

wise

. The

con

trol

var

iabl

e is

TA

, the

nat

ural

loga

rithm

of t

he b

ook

valu

e of

tota

l ass

ets.

***/

**/*

den

otes

sign

ifi ca

nce

at th

e 0.

01/0

.05/

0.10

err

or le

vel,

resp

ectiv

ely.


supporting evidence for our fourth hypothesis, that fi rms with option-based remuneration plans have higher governance ratings than other fi rms. The coeffi cient on OPTION is (marginally) signifi cantly positive. Again, the regression controls for fi rm size, as measured by TA, confi rming that larger fi rms exhibit a higher governance rating. The explanatory power is reasonably high, with an adjusted R-square of 45.5 per cent.

In equation (4) we use VR1 and VR1^2, in addition to all other explana-tory variables, to measure the non-linear relationship between the govern-ance rating and ownership concentration. The results confi rm our previous fi ndings. The coeffi cients on VR1 and VR1^2 are signifi cantly negative (at the 5 per cent level) and positive (at the 10 per cent level), respectively, again indicating a U-shaped relationship between the governance rating and ownership concentration. All other coeffi cient estimates are as before.

5.2 Robustness Tests

In order to determine the reliability of our results, we conduct two robust-ness tests for equation (4) in Table 14.2. First, we test whether industry eff ects drive the results and estimate a fi xed-eff ects model. Using the Dow Jones STOXX classifi cation scheme, the model incorporates intercepts for 18 industries. The estimation results are shown in Table 14.3. Compared with the previous results in Table 14.2, VR1 and IAS are now signifi cant only at the 10 per cent level, and the squared term VR^2 as well as BOARDSIZE and OPTION turn insignifi cant. The notion that industry is a determinant of board size, compensation packages and accounting standards should come as no surprise. For example, supervisory boards of traditional indus-tries tend to be larger, while boards of ‘New Economy’ fi rms are smaller. Furthermore, the optimal compensation package is likely to be infl uenced by the presence of asymmetric information between principal and agent, by the riskiness of the fi rm’s environment and by its ‘asset specifi city’ (for example, see Demsetz and Lehn, 1985). All of these fi rm characteristics are likely to be infl uenced by the industry a fi rm operates in.

As a second robustness test we include one (so far) possibly omitted variable, namely the performance of a fi rm. Firms with better performance and higher valuations could be more inclined to choose better corporate governance instruments, since it may be cheaper for them as they fulfi l most of the recommendations anyway. We apply Tobin’s Q as a measure of fi rm valuation and use this variable as an additional explanatory vari-able for the governance rating. To account for endogeneity, we estimate a two-stage least-squares regression. The fi rst-stage regression involves a regression of Tobin’s Q on all exogenous variables and the following instrument variables: industry dummies (18 industry dummies according


to the Dow Jones EURO-STOXX classifi cation), a fi rm’s beta value calcu-lated from monthly stock returns over the period from 1998 to 2001, and the natural logarithm of the age of the fi rm. The second-stage regression applies all governance mechanisms and the fi tted value of Tobin’s Q as the explanatory variables. As shown in Table 14.3, Tobin’s Q is insignifi cant, a Hausman-test accepts exogeneity, and all the results for the diff erent corporate governance mechanisms remain essentially the same, both in

Table 14.3 Robustness tests

Industry fi xed eff ects Endogeneity (2SLS)

Coeff . t-value Coeff . t-value

VR1 −0.0880 (−1.79)* −0.0737 (−1.71)*VR1^2 0.0007 (1.37) 0.0007 (1.52)BOARDSIZE −0.1000 (−0.72) −0.2776 (−2.58)**GAAP 3.5001 (2.99)*** 2.5774 (2.40)**IAS 1.5394 (1.84)* 1.6891 (2.24)**OPTION 1.1015 (1.13) 1.1029 (1.40)TA 0.6701 (1.91)* 1.0926 (3.85)***TQ 0.3518 (0.75)Const. 11.0366 (2.86)*** 6.2065 (1.71)*Obs. 80 77Adj. R2 0.3685 0.4458Hausman-test:c2(7) 2.55p-value 0.9232

Notes:The estimating sample contains 85 German fi rms; variations are due to data limitations. Time period: 2002. The table shows robustness test for equation (4) from Table II. The fi rst specifi cation applies an industry fi xed-eff ect model to equation (4). The second specifi cation uses two-stage least squares whereby in the fi rst stage TQ (Tobin’s Q, ratio of market value of equity plus liabilities divided by total book value of assets) is regressed on the following instruments: industry (refers to 18 industries from Dow Jones EURO-STOXX classifi cation), beta value calculated from monthly stock returns over the period from 1998 to 2001, and the natural logarithm of the age of the fi rm. A Hausman-test accepts exogeneity. The dependent variable refers to the corporate governance score as calculated by Drobetz et al. (2004). The corporate governance variables are: VR1 denotes the voting rights of the largest ultimate shareholder, and VR1^2 is the squared value of VR1. BOARDSIZE is the number of directors on the company’s supervisory board. GAAP is a dummy variable and equals 1 if US-GAAP are used as accounting standards in the annual reports and equals 0 otherwise. IAS is a dummy variable and is set to 1 if IAS are used as accounting standards in the annual reports and equals 0 otherwise. OPTION is a dummy variable and equals 1 if the fi rm uses an option-based remuneration plan and 0 otherwise. The control variable is TA, the natural logarithm of the book value of total assets. ***/**/* denotes signifi cance at the 0.01/0.05/0.10 error level, respectively.


the magnitude of the coeffi cients and their level of signifi cance. Overall, these results indicate that our previous results for the baseline regressions in Table 14.2 are not affl icted by the inclusion/exclusion of Tobin’s Q.

5.3 Results for the Components of the Governance Rating

In this section we split the aggregate rating into its fi ve components: (1) shareholder rights, (2) management and supervisory board matters, (3) transparency, (4) governance commitment, and (5) auditing. The results of the regressions using the respective sub-indices as dependent variables are shown in Table 14.4.

Shareholder rights (equation (1) in Table 14.4) encompass criteria such as the one-share-one-vote principle, subscription rights for capital increases, and modern communication (that is, internet) used for the general meeting and/or the voting process. As can be seen from equation (1) in Table 14.4, there is no positive part in the relation between this sub-index and VR1. Regressing VR1 linearly on the shareholder rights rating, VR1 is estimated signifi cantly (at the 5 per cent level) negative. This indicates that the largest shareholder is particularly wary of code recommendations that increase the control rights of minority shareholders.

A signifi cantly negative/positive relationship between VR1 and a sub-index are obtained for the categories management and supervisory board matters (equation 2) and auditing (equation 5). Management and super-visory board matters encompass dimensions such as remuneration and performance criteria of board members; disclosure of individual board members’ variable and fi xed pay components in the annual reports; selection process of directors; separate committees within the board; and the number of board members’ directorships. We therefore argue that this category (besides shareholder rights) is the most relevant with respect to corporate governance improvement. Given that board size also has a signifi cantly neg-ative infl uence, our main results are confi rmed strongest for this sub-index.

None of our corporate governance variables are signifi cant in the regression for the category transparency (equation 3). Together with the fact that the average rating is extremely high (4.55 out of a maximum of 5), this refl ects the general understanding in Germany as well as in other Continental European countries that transparency is vital for good cor-porate governance, and not even large shareholders can oppose this. A major improvement in transparency legislation was achieved when the European Union’s Transparency Directive (88/627/EEC) was transposed into German law and became eff ective at the beginning of 1995.

The component related to governance commitment (equation 4) investi-gates whether there are corporate governance guidelines set out in writing,

377

Tab

le 1

4.4

Com

pone

nts o

f the

cor

pora

te g

over

nanc

e ra

ting

Eq.

VR

1V

R1^

2B

OA

RD

SIZ

EG

AA

PIA

SO

PTIO

NT

AC

onst

.O

bs.

Adj

. R2

(1)

−0.

0046

−0.

0000

3−

0.02

560.

1152

0.31

890.

4539

0.22

180.

0721

800.

313

(−0.

38)

(−0.

23)

(−0.

83)

(0.4

2)(1

.44)

(2.0

9)**

(2.9

3)**

*(0

.09)

(2)

−0.

0590

0.00

06−

0.13

751.

2683

0.50

070.

4978

0.40

012.

0826

800.

313

(−2.

68)*

**(2

.52)

**(−

2.49

)**

(2.6

0)**

*(1

.25)

(1.2

7)(2

.93)

***

(1.4

0)(3

)0.

0019

−0.

0000

01−

0.01

48−

0.05

30−

0.08

920.

1962

0.08

643.

3753

800.

043

(0.2

6)(−

0.01

)(−

0.79

)(−

0.32

)(−

0.65

)(1

.47)

(1.8

5)*

(6.6

1)**

*(4

)0.

0080

−0.

0001

−0.

0494

0.92

390.

1061

−0.

1883

0.19

17−

0.00

2180

0.06

5(0

.41)

(−0.

69)

(−1.

01)

(2.1

4)**

(0.3

0)(−

0.55

)(1

.59)

(−0.

00)

(5)

−0.

0297

0.00

03−

0.01

740.

8397

0.79

220.

2406

0.06

282.

7616

800.

276

(−2.

53)*

*(2

.61)

***

(−0.

59)

(3.2

2)**

*(3

.72)

***

(1.1

5)(0

.86)

(3.4

7)**

*

Not

es:

The

est

imat

ing

sam

ple

cont

ains

85

Ger

man

fi rm

s; va

riatio

ns a

re d

ue to

dat

a lim

itatio

ns. T

ime

perio

d: 2

002.

The

tabl

e sh

ows t

he re

sults

from

O

LS

regr

essio

ns o

f the

com

pone

nts o

f the

cor

pora

te g

over

nanc

e ra

ting

as d

epen

dent

var

iabl

es o

n th

e m

ain

corp

orat

e go

vern

ance

mec

hani

sms

alon

g w

ith th

e co

ntro

l var

iabl

e. T

he d

epen

dent

var

iabl

es in

the

diff e

rent

equ

atio

ns a

re: E

q. (1

) sha

reho

lder

righ

ts (‘

Akt

ionä

rsre

chte

’). E

q. (2

) m

anag

emen

t and

supe

rviso

ry b

oard

mat

ters

(‘E

ntsc

heid

ungs

- u. K

ontr

ollg

rem

ien’

). E

q. (3

) tra

nspa

renc

y (‘T

rans

pare

nz’).

Eq.

(4) g

over

nanc

e co

mm

itmen

t (‘U

nter

nehm

ensa

usric

htun

g un

d C

orpo

rate

Gov

erna

nce’

). E

q. (5

) aud

iting

(‘A

bsch

luss

prüf

ung’

). T

he c

orpo

rate

gov

erna

nce

varia

bles

ar

e: V

R1

deno

tes t

he v

otin

g rig

hts o

f the

larg

est u

ltim

ate

shar

ehol

der,

and

VR

1^2

is th

e sq

uare

d va

lue

of V

R1.

BO

AR

DSI

ZE

is th

e nu

mbe

r of

dire

ctor

s on

the

com

pany

’s su

perv

isory

boa

rd. G

AA

P is

a du

mm

y va

riabl

e an

d eq

uals

1 if

US-

GA

AP

are

used

as a

ccou

ntin

g st

anda

rds i

n th

e an

nual

repo

rts a

nd e

qual

s 0 o

ther

wise

. IA

S is

a du

mm

y va

riabl

e an

d is

set t

o 1

if IA

S ar

e us

ed a

s acc

ount

ing

stan

dard

s in

the

annu

al re

port

s and

eq

uals

0 ot

herw

ise. O

PTIO

N is

a d

umm

y va

riabl

e an

d eq

uals

1 if

the

fi rm

use

s an

optio

n-ba

sed

rem

uner

atio

n pl

an a

nd 0

oth

erw

ise. T

he c

ontr

ol

varia

ble

is T

A, t

he n

atur

al lo

garit

hm o

f the

boo

k va

lue

of to

tal a

sset

s. **

*/**

/* d

enot

es si

gnifi

canc

e at

the

0.01

/0.0

5/0.

10 e

rror

leve

l, re

spec

tivel

y.


or whether there is a corporate governance representative reporting on corporate governance issues to the supervisory board. The only signifi cant governance variable is GAAP. This could be explained by the fact that fi rms employing US-GAAP are those which strive for a listing in the US, where corporate governance is organized more formally, and where it is more common to structure the corporation according to corporate govern-ance guidelines.

Finally, the sub-index referring to auditing (equation 5) is based on the following questions. Do quarterly reports contain segment reporting? Are there fi rm-specifi c rules to ensure that the auditor does not perform other services for the fi rm? Does the annual report contain information about the risk-management system of the corporation? Besides the signifi cant nega-tive/positive VR1 infl uence, international accounting standards (IAS and GAAP) exert a positive and signifi cant infl uence on auditing. International accounting standards, which are supposed to reveal more information than national accounting standards, also raise the quality of auditing.

6. CONCLUSION

There is mounting empirical evidence that there is a relationship between the quality of fi rm-level corporate governance and fi rm valuation. Ultimately, this is the only reason why corporate governance issues should be of inter-est for fi nancial economists at all. Unfortunately, all empirical studies are inherently plagued with endogeneity problems, as causality could well run from performance to governance. This chapter tries to circumvent the problem of causality by taking one step back and investigating the determinants of good corporate governance as measured by the corporate governance rating of Drobetz et al. (2004). It is ultimately the owners who decide (or at least monitor the decision) on whether or not to adopt better governance practices. Ownership structure may thus be regarded as exogenous and even more so in Continental European countries, where signifi cant ownership concentration is the rule rather than an exception. Similarly, the structure of the supervisory authorities can be expected to aff ect the governance rating of a fi rm. The board of directors ultimately takes the decisions with respect to all governance issues (and, hence, has to assume responsibility for all corporate governance malfunctions).

While our research question is clearly a lot more modest than directly exploring the link between corporate governance and fi rm valuation, we still uncover several interesting interrelationships within fi rms. We confi rm the non-linear relationship between ownership concentration and the quality of fi rm-level governance familiar from previous governance/


performance studies. We interpret it as being caused by two opposing infl u-ences, incentive alignment and entrenchment, and document a signifi cant entrenchment eff ect at intermediate holdings of the largest shareholder (between 25 and 50 per cent). With increasing holdings of the largest shareholders (more than 50 per cent), there are positive wealth eff ects and, hence, incentive eff ects start to dominate. Our results hold up strongest when analysing the sub-index relating to management and supervisory board matters. In addition, fi rms with larger board size have lower govern-ance ratings, but fi rms that apply US-GAAP or IAS rules and/or use an option-based remuneration plan have higher governance ratings.

It is worth putting our results into perspective. First, there is a positive and reassuring message. Corporate governance codes potentially improve the governance and decision making processes of companies. Otherwise, if pro-visions were not binding anyway, large shareholders or large boards had no need to oppose (some of) them (for example, transparency of executive pay). Second, however, there is a more negative and cautious conclusion that follows from our results. Large shareholders still have a tight grip on com-panies in Continental European countries and veto recommendations that might lead to a loss of their control and power, such as recommendations for one-share-one-vote or disclosure of individual board members’ pay.

NOTES

1. See Barca and Becht (2001), Becht and Röell (1999), Gugler (2001), and La Porta et al. (1998) for analyses of ownership and voting right concentration. See Becht (1999) on liquidity, and see Claessens et al. (2002) and Gugler and Yurtoglu (2003a) on the separa-tion of voting and cash fl ow rights. Pagano et al. (2002) show that European markets having the highest trading costs, lowest accounting standards and poorest shareholder protection fare worst in attracting and retaining cross-border listings.

2. There is a much more developed literature on the eff ects of corporate governance mecha-nisms on performance, though. Using fi rm-level data from 27 developed countries, La Porta et al. (2002) fi nd that better shareholder protection is associated with higher valua-tion of corporate assets. Gompers et al. (2003) report for a broad sample of US fi rms that fi rms with stronger shareholder rights receive higher valuations and have higher profi ts, higher sales growth, and lower capital expenditures. Klapper and Love (2003) use fi rm-level data from 14 emerging stock markets and also report that better corporate govern-ance is highly correlated with better operating performance and higher market valuation.

3. On the endogeneity issue and suggestions for cure, see Börsch-Supan and Köke (2002) and Gugler and Yurtoglu (2003b).

4. For example, Cuervo (2002) argues that especially in civil law countries such as Germany the codes of good governance can be applied formally, following the letter but not the spirit of the law, since they cannot be legally enforced.

5. Specifi cally, a variety of organizations have issued governance codes, including govern-mental entities, committees and commissions organized or appointed by governments, stock exchange related bodies as well as business, industry and academic associations. In addition to national codes, several pan-European and international governance


codes have emerged (such as the OECD Principles of Corporate Governance). For the codes of almost 40 countries, see http://www.ecgi.org/codes/all_codes.htm. For an extensive comparative analysis we refer to http://europa.eu.int/comm/internal_market/en/company/company/news/corp-gov-codes-rpt_en.htm.

6. Baums and Fraune (1994) report that only 58 per cent, on average, of all voting rights are represented at the annual meeting of a German publicly listed fi rm.

7. For empirical analysis see Loderer and Peyer (2002). 8. Because two-thirds of the German fi rms included in the DAX blue-chip index opted

out and do not report the salaries of each director separately, the public discussion has intensifi ed only recently. There are even suggestions by major political parties to legally force disclosure of individual compensation (see ‘Keeping stumm’, Economist, 21–27 August 2004).

9. See the German Corporate Governance Code (2002), www.corporate-governance-code.de.

10. DVFA is the German Society of Investment Analysis and Asset Management.11. More in-depth analysis in Drobetz et al. (2004) shows that an equal-weighting scheme

is not a restrictive assumption.

REFERENCES

Albach, H. (2003), ‘Corporate governance – large and small corporations, agents, principals, competitors’, in Fandel, G., U. Backes-Gellner, M. Schlüter and J.E., Staufenbiel (eds), Managing Enterprises of the New Economy by Modern Concepts of the Theory of the Firm, Proceedingsband zur Tagung der Erich-Gutenberg-Arbeitsgemeinschaft.

Barca, Fabrizio and Marco Becht (eds) (2001), The Control of Corporate Europe, Oxford: Oxford University Press.

Baums, T., and C. Fraune (1994), ‘Institutionelle Anleger und Publikumsgesellschaft: Eine empirische Untersuchung’, working paper, University of Frankfurt.

Becht, M. (1999), ‘European corporate governance: trading off liquidity against control’, European Economic Review, 43 (4–6), 1071–83.

Becht, M. and A. Röell (1999), ‘Blockholdings in Europe: an international com-parison’, European Economic Review 43 (4–6), 1049–56.

Beiner, S., W. Drobetz, F. Schmid and H. Zimmermann (2004), ‘Is board size an independent corporate governance mechanism?’, Kyklos, 57 (3), 327–58.

Börsch-Supan, A. and J. Köke (2002), ‘An applied econometricians’ view of empir-ical corporate governance studies’, German Economic Review, 3 (3), 295–326.

Claessens, S., S. Djankov, J.P.H. Fan and L.H.P Lang (2002), ‘Disentangling the incentive and entrenchment eff ects of large shareholders’, Journal of Finance, 57, 2741–71.

Cuervo, A. (2002), ‘Corporate governance mechanisms: a plea for less code of good governance and more market control’, Corporate Governance: An International Review, 10 (2), 84–93.

Demsetz, H. and K. Lehn (1985), ‘The structure of corporate ownership: causes and consequences’, Journal of Political Economy, 93 (6), 1155–77.

Diamond, D. and R. Verrechia (1982), ‘Optimal managerial contracts and equilib-rium security prices’, Journal of Finance, 37, 275–87.

Drobetz W., A. Schillhofer and H. Zimmermann (2003), ‘Ein Corporate Governance Rating für deutsche Publikumsgesellschaften’, Zeitschrift für Betriebswirtschaft, 74 (1), 1–21.


Drobetz W., A. Schillhofer and H. Zimmermann (2004), ‘Corporate govern-ance and expected stock returns: evidence from Germany’, European Financial Management, 10 (2), 267–93.

Eisenberg, T., S. Sundgren and M. Wells (1998), ‘Larger board size and decreasing fi rm value in small fi rms’, Journal of Financial Economics, 48, 35–54.

Gompers, P., L. Ishii and A. Metrick (2003), ‘Corporate governance and equity prices’, Quarterly Journal of Economics, 107–55.

Gugler, K. (2001), Corporate Governance and Economic Performance, Oxford: Oxford University Press.

Gugler, K. and B.B. Yurtoglu (2003a), ‘Corporate governance and dividend pay-out policy in Germany’, European Economic Review, 47 (4), 731–58.

Gugler, K. and B.B. Yurtoglu (2003b), ‘Average Q, marginal Q, and the relation between ownership and performance’, Economics Letters, 78 (3), 379–84.

Gugler, K., D.C. Mueller and B.B. Yurtoglu (2003a), ‘Separating the wealth and entrenchment eff ects of insider ownership on investment performance’, Center for Business Studies, Department of Economics, University of Vienna, Working Paper Series.

Gugler, K., D.C. Mueller and B.B. Yurtoglu (2003b), ‘The impact of corporate governance on investment returns in developed and developing countries’, The Economic Journal, 113, 511–39.

Gugler, K., Mueller, D.C. and Yurtoglu, B.B. (2004), ‘Corporate governance, capital market discipline and the returns on investment’, Journal of Law and Economics, 47 (2), 589–633.

Holmström, B. and J. Tirole (1993), ‘Market liquidity and performance monitor-ing’, Journal of Political Economy, 101, 678–709.

Jensen, M.C. (1993), ‘The modern industrial revolution, exit, and the failure of internal control systems’, The Journal of Finance, XLVIII (3), 831–80.

Klapper, L.F. and I. Love (2003), ‘Corporate governance, investor protection, and performance in emerging markets’, Journal of Corporate Finance, 195, 1–26.

La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishny (1997), ‘Legal deter-minants of external fi nance’, Journal of Finance, 52 (3), 1131–50.

La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishny (1998), ‘Law and fi nance’, Journal of Political Economy, 106 (6), 1113–55.

La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishny (2000), ‘Agency problems and dividend policies around the world’, Journal of Finance, 55, 1–33.

La Porta, R., F. Lopez-de-Silanes, A. Shleifer and R. Vishny (2002), ‘Investor pro-tection and corporate valuation’, Journal of Finance, 57, 1147–70.

Lipton, M. and J. Lorsch (1992), ‘A modest proposal for improved corporate gov-ernance’, Business Lawyer, 48, 59–77.

Loderer, C. and U. Peyer (2002), ‘Board overlap, seat accumulation and share prices’, European Financial Management, 8, 165–92.

Morck, R., A. Shleifer and R. Vishny (1998), ‘Management ownership and market valuation: an empirical analysis’, Journal of Financial Economics, 20, 293–315.

Pagano, M., A. Roell and J. Zechner (2002), ‘The geography of equity listing: why do companies list abroad?’, Journal of Finance, 57, 2651–94.

Shleifer, A. and Wolfenzon, D. (2002), ‘Investor protection and equity markets’, Journal of Financial Economics, 66, 3–27.

Yermack, D. (1996), ‘Higher market valuation of companies with a small board of directors’, Journal of Financial Economics, 40, 185–211.

382

15. Top management, education and networkingMogens Dilling-Hansen, Erik Strøjer Madsen and Valdemar Smith*

1. INTRODUCTION

The corporate governance literature has primarily focused on agent prob-lems in management and, consequently, on the misallocation of resources as a result of bad decision making by managers as they do not have an incentive to behave in the best interests of all share- and debt-holders. Another important theme in the literature is misallocation of resources as a result of cash fl ow expropriation of major shareholders at the expense of minority shareholders; see, for example Tirole (2006). However, good deci-sion making is not only a question of the right long-run target or incentives of the management; it also depends on their knowledge and the level of sig-nifi cant information about the competitive and technological environment of the fi rm. In that respect personal or professional networks may play an important role for the fi rm managers.

The chapter uses a social network approach to analyse the importance in respect to fi rm performance of this external network between manag-ers and board members of diff erent fi rms by using a data base of the largest Danish companies. The corporate networking activity is divided into two types: networking between fi rms with a common owner struc-ture and networking between independent fi rms. We use the Bonacich centrality approach to measure the networking activity; see Bonacich (1987). Moreover the primary goal of the paper is to analyse whether fi rm performance is aff ected by the strength of the professional networks with other fi rms held by top management, that is, CEOs and members of the supervisory board of the fi rm. Recognizing that the strength of network and fi rm performance are potentially endogenous, we set up a system of simultaneous equations in order to explain fi rm performance and network strength between fi rms.

Top management, education and networking 383

2. NETWORKING AND FIRM PERFORMANCE

The traditional discussion of the eff ect of networking is linked to the rela-tion between ownership structure and performance: a positive eff ect of concentrated ownership is found if alignment of incentives is dominant and a negative eff ect is found if the entrenchment eff ect dominates. Although a substantial part of all small fi rms is controlled by a family with a majority share in the company, a lack of formal relations between independent fi rms is unlikely. It is often seen that a manager of one fi rm is associated with another fi rm as a board member. Furthermore, as verifi ed by La Porta et al. (1999), even the majority (70 per cent) of the top 20 fi rms in 27 developed countries are controlled by wealthy families or foundations through control pyramids and other structures. Only in the United States and the United Kingdom are widely held corporations predominant among large fi rms.

Moreover, in many countries the ownership structure has developed into networks by cross-holding of shares between companies, but even without common ownership structures between two fi rms, persons employed in the top management of the fi rms create formal relations when they are members of an outside board. If the networking connections between top managers are clustered in relatively small but concentrated groups a small world is identifi ed. Conyon and Muldoon (2006) fi nd that there may be signs of small world characteristics in USA, Germany and UK, but the concentration found could also have been the eff ect of a random process. However, Kogut and Walker (2001) studied the ownership ties in large German corporations and found that the ownership networks constitute a small world and that this ‘Rhineland capitalism’ is robust against glo-balization with increasing foreign direct investment. Using a sample of Danish fi rms, Thomsen and Sinani (2005) fi nd the same stability over time in the corporate network. This study does not control for the existence of ownership ties as opposed to ties between independent board members or CEOs.

Ownership and performance have been analysed intensively. It has been documented that fi rms with dispersed ownership perform worse than closely held fi rms due to the agency problem; see, for example, the survey by Shleifer and Vishny (1997). And in their survey of the empirical evi-dence for family controlled fi rms, Morck et al. (2005) conclude that ‘large block-shareholdings by a family, and family involvement in management, need not destroy value, and may even add value for public shareholders’. However, the fi nding that the average family fi rm is doing as well or better than other fi rms does not rule out cases where families mismanage their fi rms.

Burkart et al. (2003) examined whether the founder of a fi rm wants to


surrender control of the fi rm to professional managers. On the one hand, the family has potential benefi t from owner control; on the other hand, by self-control it misses the opportunity to hire the best qualifi ed manager of the fi rm. When the benefi t of owner control is smaller than the foregone benefi t of not having a professional management, the family may choose to hand over control of the fi rm. However, this may not be an easy decision for many families/founders and there is some evidence that family owned fi rms that appoint an insider family CEO perform signifi cantly worse than family owned fi rms that appoint an outside professional CEO; see Bennedsen et al. (2005).

Very few empirical studies have examined the social networking of board members and the board of directors (including the CEO) and its infl uence on fi rm decision and performance. As an exception, Booth and Deli (1996) report a negative correlation between fi rm performance and social networking using the simple number of outside directorships. The argument for this fi nding is high opportunity cost of spending time at another fi rm, that is, external networking may benefi t the individual but the hiring fi rm does not get a positive spin-off from this activity.

3. NETWORKING ACTIVITY

A social network perspective and social network analysis are an effi cient way to analyse the eff ect of informal knowledge transfer between fi rms. The information fl ow is created by persons interacting even though they are employed in legally independent fi rms, and the overall expectation is that there is a positive relation between interaction and performance of the fi rms involved. A simple example is the process of separation of ownership and control: handing over the CEO position to an external manager is a drastic decision and a more gentle way to create a structure with professional corporate governance is probably to invite an external manager to take a seat in the boardroom. Of course, this would enhance the professional decision of the board and leave the family control over the fi rm unaltered.

In this chapter, we examine the external ties to other fi rms of board members and the management team and calculate a measure of the fi rm’s integration in management networks between fi rms. Our measure covers both the situation where an external professional is appointed to the board of a fi rm and the situation where the CEO of the fi rm or some of its board members are appointed to the board of an external fi rm. The latter case may also signal that the internal CEO or board member has the profes-sional qualifi cation to be appointed to an external board.


The expected positive eff ect on performance of this fi rst step on the line to professional corporate governance is based on several arguments. Firstly, fi rms who invite other professionals into their boardroom may benefi t from the professional managers’ advice and at the same time retain the documented benefi t of owner control, which they could partly lose by handing over control to an external CEO, who may expropriate fi rm value and increase the monitoring cost for the family. Secondly, in fi rms with minority shareholders the external board members or the CEOs may reduce the majority shareholders’ possibility of appropriating the cash fl ow at the expense of the minority shareholders and cause the fi rm’s value to increase. Next, board members’ and managers’ external links to other fi rms may bring in valuable information and knowledge concerning the competitive environment, best practice and threats from innovation or other developments of importance for the management of the fi rm. Information sharing in the professional networks may increase fi rm value also in cases where members of the controlling family are invited to be part of the management system in another fi rm. Finally, information sharing could bring about an understanding between managers within the market that potentially could promote collusion in the market and bring in more value to the fi rm at the expense of the consumers.

The value of information and knowledge sharing may depend on the educational level of CEOs and board members. Thus the absorp-tion capacity concerning the exploitation of information is expected to depend on the formal education of the managers, meaning that educa-tion improves the quality of their decisions. This argument is in line with standard human capital theory. We therefore expect that a higher level of education of CEOs and board members will increase the performance of the fi rm and for managers with many external links the eff ect may even be stronger.

Even though the arguments for a positive relation are prevalent there are also arguments for a negative relation. Booth and Deli (1996) argue that the negative eff ect is caused by the opportunity cost of spending time at another fi rm: the networking person may benefi t from the networking but the parent fi rm experiences no spin-off from the board membership (it may even lose if the person with social relations uses confi dential information from the parent fi rm). Finally, there may be an ambiguous eff ect of using networking as a control mechanism for a parent fi rm; in line with the clas-sical entrenchment argument (see Morck et al., 1988), board members of a subsidiary will pursue the goals of the parent fi rm and without full informa-tion of the subsidiary’s market situation these goals may aff ect subsidiary performance negatively.


4. IMPLEMENTATION OF THE SOCIAL NETWORK ANALYSIS

4.1 Firm Interaction

Firms interact with other fi rms in several ways and naturally the most important interaction is a consequence of transactions between fi rms in the value chain. Even though some of these transactions are automated using digital technologies, there are still a signifi cant number of transactions involving human interaction between people in diff erent fi rms but also within the fi rm. Focusing on the fi rm’s executive level, that is, the board of managers and the board of executives, the top management, implementa-tion of strategies is expected to result in communication across various groups of individuals. Thus, communication between the top management and other staff within the same fi rm is part of the management process, but the fact that there is business-based communication and relations between separate fi rms necessitates that it must be explained by other factors such as exercising control, information seeking, knowledge sharing. Depending on the reasons for external networking diff erent eff ects on fi rm perform-ance may be expected.

Table 15.1 illustrates the direction of human interaction between fi rms in a simple two-fi rm model. Basically the focus of this chapter is networking between top managers, that is, the top left corner of the table. In general, the motive behind human interaction between two fi rms may be purely business orientated or personal, but sometimes it is the contact between the top management in the two fi rms that is interesting when analysing fi rms’ long-run performance. Of course, if the two fi rms are formally related, there is an ownership structure that may control both fi rms and, in that case, the fi rms can be seen as a joint unit. At the other extreme, there are no joint owners at all and, in this situation, the interaction between the fi rms is defi ned as purely business related.

4.2 Measuring Power using Social Network Analysis

Measuring the interaction between top managers using a social network approach is based on data that diff er from data used in traditional quan-titative analysis. Network data is defi ned by actors (also called nodes) and by relations (also called links or ties) and in this chapter, we analyse the eff ect of relations (ties) created by persons in top management. In a social network context, fi rms are the nodes in the network; persons create the rela-tions or ties between these nodes and the ties are symmetric if the relation goes in both directions.


A fi rm without ties to other fi rms will obviously have no ‘social power’, and using a full network approach with all fi rms in the sample (in contrast to an ego network) these fi rms will perform worse. A fi rm with many ties will, on the other hand, perform better because information will fl ow more easily between relevant fi rms; that is, there will always be a path between fi rms with many ties associated. The ‘social power’ of a fi rm is therefore directly related to the number of ties connected to the fi rm and the rela-tive importance of the nearest neighbours – a fi rm with many relations to other (important) fi rms will benefi t from these relations (centrality in social network analysis).

Other measures, mainly based on distance, connectivity, reachability and number of paths, are also focused on the assumption that the power of an actor placed in the ‘centre’ of the network will be greater than the power of an actor placed on the periphery; see Bonacich (1987), Conyon and Muldoon (2006), Freeman (1979) and Wasserman and Faust (1994).

Mizruchi and Bunting (1981) fi nd that centrality based network analysis is superior to examining corporate control. The Bonacich power measure of centrality (see Bonacich, 1987) is a general measure of power based on cen-trality and it is used to measure the ‘social power’ of a fi rm in this chapter. Actors with more connections are more likely to be powerful because they can directly contact other actors and, moreover, the power is dependent on the connections the actors in the neighbourhood have. The Bonacich measure for node i in a network is a function of two parameters, a and b:

Bonacichi 5 ci(a, b) 5 SAij(a 1 bcj) (1)

The Bonacich measure of centrality is a standard measure of centrality using the sum of connections (links) weighted with centralities (Aij is the

Table 15.1 Formal interaction between person employed in a fi rm and business partners

Firm B

Top management Other staff

Firm A Top management Joint fi rms Joint fi rmsBusiness related fi rms Business related fi rms

Other staff Joint fi rms Joint fi rmsBusiness related fi rms Business related fi rms

Notes: Firm A and fi rm B are defi ned as ‘joint fi rms’ if they are legally connected, e.g. by common owners. Firm A and fi rm B are defi ned as ‘business related fi rms’ if there is no legal tie between the fi rms.


adjacency matrix describing the tie between nodes and cj is the centrality of node i). The fi rst parameter, a, is a normalizing parameter, while b refl ects the individual’s status compared with that of the neighbours connected to the fi rst node.

If b is positive then the power of each node is a positive function of the status of the connected neighbours. If the power is interpreted as bargain-ing power then we experience increased power when neighbours are power-less and this situation is covered by a negative value of b.

The value of a is selected automatically, but choosing the right value of b has been discussed; see Bonacich (1987) and Bonacich and Lloyd (2004). Still, the standard application of the Bonacich measure is based on a positive value of b whose absolute value is less than the absolute value of the reciprocal of the largest eigenvalue of the adjacency matrix, Aij; see Borgatti et al. (2002). If b50, formula (1) will no longer take any second-ary eff ects into account and the ‘social power’ of the fi rm will simply be the number of ties connected to the fi rm. Both approaches (b50 and optimal b) are applied in the empirical part of this chapter.

Almost all social network analysis is based on the assumption that power is closely related to centrality and the discussion about ‘betweenness’, ‘nearness’ and ‘degree’ results in arguments in favour of using Bonacich’s centrality measure c(a, b); see Bonacich (1987) and Freeman (1979). Firms (nodes in network analysis) are related to each other through persons in the top management and the ties are constructed using information on ownership structure and type of relation (‘betweenness’ and ‘nearness’) and strength (‘degree’).

We use the full network approach in contrast to a special network using snowballing methods to select fi rms; see Hanneman and Riddle (2005). Using the full network approach, attention is paid to the network structure between ‘all’ large fi rms in Denmark and the ties between fi rms are estab-lished via persons in the top management. This means that ties between the large fi rms in the sample and smaller fi rms outside are neglected; however we identify the network structure regardless of the number and strength of the ties.

4.3 Measuring Ties between Firms using Information on Ownership Structure

A tie is identifi ed if a person in the top management of one fi rm has an equivalent role in another fi rm; that is, we only defi ne a tie if a person has a formal position in the other fi rm. The ownership structure is used to identify two types of ties, internal and external ties, and a fi rm may have both types of ties.


An internal tie or an owner tie is identifi ed if a fi rm is related by owner-ship to another fi rm. The formal link can have diff erent appearances, but existence of a pyramid structure or information on an owner share larger than 50 per cent is the most frequent way of identifying an owner tie. This defi nition focuses on the actual control over a fi rm and, even though the controlling fi rm has the majority of control, it is not certain that the net-working person actually owns the majority shares.

Other links between fi rms without a formal ownership structure are defi ned as external ties. Persons in the top management are either manage-ment board members or board members:

1. chief executive offi cer2. other members of the board of management3. chairman of the board4. members of the board

and a person with a tie can have any of the four roles in the other fi rm. If a person has several involvements with other fi rms, we identify a tie for each and we use a symmetric approach, because it is not possible to identify the primary activity. Using the symmetric approach we do not distinguish between the eff ect of a tie between fi rm 1 and fi rm 2 and the correspond-ing tie from fi rm 2 to fi rm 1; that is, we implement the same network eff ect whether one or two ties are found. Finally, we use binary weights because there is no objective method to determine the strength of the relations.

5. DATA DESCRIPTION

The empirical analyses are based on a sample of the 1000 largest Danish fi rms supplying information on economic performance available. The indi-vidual relations between persons in the fi rms are defi ned by the relations between persons belonging to the top management. Thus we focus on the external relations between top managers and members of the supervisory board (board of directors). Top CEOs can be members of the supervisory boards of other fi rms and board members can be members of boards or CEOs of another fi rm and so on.

Information on economic performance is based on the offi cial annual report of the fi rm, which includes turnover and standard measures of profi tability. The information on the relations between the persons in top management is based on a private on-line company supplying information on ownership structure and persons in the top management (CEOs and board members).1 Only relations between persons in legally independent


fi rms are recorded and relations between profi t-oriented fi rms and non-profi t organizations are not included.

5.1 Descriptive Statistics

In the sample of large fi rms used in the network analysis, we end up with 999 fi rms with valid information for the year 2004 (descriptive statistics are presented in Table 15.2). The general picture of industry structure and fi rm performance is that there is no signifi cant diff erence between the relatively large fi rms (more than 500 employees) and smaller ones. The minimum effi ciency scale (MES) is unchanged, but the average market share is of course a bit higher for large fi rms. We fi nd a relatively stable average profi t

Table 15.2 Descriptive statistics for small and large fi rms

Small and medium sized fi rms

(< 500 employed)

Large fi rms(≥ 500

employed)

All fi rms

Number of fi rms 747 252 999(74.8%) (25.2%) (100.0%)

Industry structureMarket share 0.149 0.233 0.170MES (log (lower

quartile of turnover))8.47 8.55 8.49

PerformanceProfi t 0.099 0.103 0.100Labour productivity

(mill DKK/empl)3.43 2.51 3.20

Governance structureNumber of persons in

management1.4

(0.8)2.2

(1.3)1.6

(1.0)Number of persons on

the board5.7

(2.5)6.7

(3.1)5.9

(2.7)Links from board of

management1.9

(3.5)4.5

(6.1)2.5

(4.4)– of which external

links0.4

(1.2)1.2

(2.9)0.6

(1.8)Links from the board 10.0

(12.5)13.5

(14.4)10.9

(13.0)

Note: An external link is defi ned as a link from the management board to another fi rm not owned partially or entirely by the fi rst fi rm. Standard deviation is shown in brackets. Information about fi rm performance and other fi nancial information is based on 999 fi rms with valid data. Results on external links between the fi rm and other fi rms are based on valid information from 854 fi rms.


rate (ROE, defi ned as net result compared with net capital) even though this measure is very unstable with a standard deviation about 2½ times higher than the average.

Information about number of persons on the management board and the board is based on valid information from 854 fi rms and there is a clear diff erence between relatively large and small fi rms in respect to the number of persons on the board. We fi nd a relatively small management board, and the overall average of 1.6 persons on the board of managers is the result of a right-skewed distribution with one person as mode value (in about 50 per cent of all fi rms, an owner is on the management board) and a few fi rms with 6 persons on the management board.

As expected, larger fi rms tend to have a higher number of persons on the management board and on average 60 per cent of the managers are repre-sented in another external fi rm (an external fi rm is defi ned as a fi rm with no obvious owner dependency of the fi rst fi rm). The number of persons on the board is around 6 and these persons are very active in respect to repre-sentation on other boards and on average we fi nd 11 links from the board to another external board. However, these data do not allow us to identify the relative importance of the other fi rm, so the networking activity is in general somewhat overrated.

5.2 Educational Background

The educational background of the CEO and the chairman of the board is based on the same source, GreensOnline. The information is reported by the persons themselves and we can therefore expect a relative overrepresen-tation of persons without any formal educational background as a result of no information. The results are presented in Figure 15.1 and we do fi nd a high number of persons in top management without formal education. Only about 30 per cent of the CEOs or chairmen of the board have a higher education and about half of the chairmen of the board have a business related education. The CEOs on the other hand are more oriented toward a business related education if the person has a higher education.

The relation between the relatively low educational background and a tendency towards professional CEOs is illustrated in Table 15.3. The table underlines a number of interesting results. Firstly, we fi nd that a relatively low level of formal education among managers is prevalent and less than one-third have a master’s degree. Secondly, there is a signifi cant depend-ency between the educational background of the chairman of the board and the CEO. If the chairman has a master’s degree then the probability of a CEO with a master’s degree is about 50 per cent higher than expected. If the chairman does not have a higher education we fi nd the opposite


relation with an almost 20 per cent lower probability of fi nding a CEO with a higher education. In other words, we fi nd a signifi cant overrepresentation of chairmen and CEOs with the same educational background. These fi nd-ings are especially signifi cant for the group of large fi rms with more than 500 employees (these results are not shown in Table 15.3).

6. EMPIRICAL MODEL AND RESULTS

The estimation results of the estimated models are listed in Tables 15.5a–c. A basic productivity model is presented in Table 15.5a and we fi nd the general translog specifi cation superior to the simple Cobb-Douglas model. No social networking activity is included and there is no support for a sig-nifi cant eff ect of educational level of the top management. Education and networking activity may be seen as complementary activities and, if that is the case, we can still fi nd a positive eff ect from networking without any eff ect from education. The relation between educational level of the top management and social networking is shown in Table 15.4. The correlation between internal and external networking is relatively low (below 30 per

0%

10%

20%

30%

40%

50%

60%

70%

None listed Commercialbackground

Undergraduatebusinesseducation

Highereducation

Higherbusinesseducation

Education

Rel

ativ

e di

stri

butio

n in

per

cen

tChief executiveChairman of the board

Figure 15.1 Educational background of the top management


cent) and, on average, there is an external and an internal link in 30 per cent of the fi rms. The average propensity to participate in social network-ing activities is doubled if only top managers with a higher education are considered – even though formal education has an eff ect on performance, social networking activity complements education.

Concerning fi rm performance (Table 15.5a), which is measured by the productivity eff ect in an augmented production function, 65 per cent of all variation in fi rm turnover is explained by the simple model. Of course, the coeffi cients for labour and capital are correctly signed and highly signifi cant and refl ect approximately constant returns to scale in the pro-duction. As we have no data for the fi rms’ use of raw and other materials, we include a manufacturing dummy to correct for the fact that the retail sector especially has a high level of other intermediate inputs. As expected, the coeffi cient is negative and highly signifi cant. The negative size eff ect

Table 15.4 Average number of links (social networking) and education

High educational level All fi rms

Internal networking 0.66 (1.04) 0.31External networking 0.71 (0.88) 0.31

Notes: High educational level is defi ned as at least a master’s degree. The averages reported are the number of links to other external and internal fi rms, and the numbers in brackets are the average numbers for fi rms with high education for both the CEO and the chairman of the board.

Table 15.3 Formal education of executives in large Danish fi rms: relation between chairman and executive manager

Chairman of the board

Low education High education

Executive manager

Low education 504(477)

136(163)

640(75%)

High education 132(159)

82(55)

214(25%)

Total 636(74%)

218(26%)

854(100%)

Notes: The table shows the highest formal education of 854 executives in the largest Danish fi rms. A high education is defi ned as a university based education (master’s degree). Numbers in brackets are expected number given independence between the two characteristics. Row and column totals are presented together with percentages. c2-test of independence rejects H0; c2 (1)524.6; P(c2 (1).24.6) < 0.0001


could also refl ect a higher degree of in-sourcing for larger fi rms which have enough scale in production to pursue even some of the more specialized tasks in the production.

The number of persons in the management board relative to fi rm size does not aff ect performance directly but the interaction between fi rm size and man-agement size has a positive and signifi cant eff ect on productivity. This result

Table 15.5a Single equation (OLS) models: fi rm productivity

Dependent variable Cobb-Douglas Translog specifi cation

Estimation method OLS OLS OLS

Intercept 6.120(0.186)

8.482(1.169)

8.620(1.277)

Labour 0.670**(0.033)

0.864**(0.303)

0.698**(0.333)

Capital 0.265**(0.015)

−0.224*(0.134)

−0.153(0.133)

Labour*Labour −0.071**(0.032)

−0.060*(0.034)

Capital*capital 0.002(0.005)

−0.002(0.005)

Labour*capital 0.063**(0.021)

0.068**(0.021)

Size (dummy) −0.311**(0.142)

Management size/employment

5.270(6.963)

2.862(6.909)

Size* (manag. size/employment)

110.48**(37.95)

Manufacturing (dummy) −0.272**(0.049)

Market share 0.262**(0.104)

Education of the CEO 0.004(0.061)

Education of the chairman −0.033(0.059)

Adj. R2 0.602 0.634 0.649Number of observations 989 850 850

Notes: Numbers in brackets are standard errors of the estimated parameters. * indicates that the estimated parameter diff ers signifi cantly from zero at the 10% level of signifi cance and ** at the 5% level.


supports the human capital argument: relatively small fi rms can be controlled effi ciently by one manager but large fi rms take advantage of a relatively large management board with complementary skills. Market share aff ects fi rm performance positively and signifi cantly and this is in line with Demsetz’s effi ciency hypothesis, whereby the most effi cient fi rms gain market shares, and with Porter’s home based hypothesis, whereby multinational fi rms need a strong and competitive home market to be innovative and competitive.

The standard productivity model is enhanced with the external social networking variables in Table 15.5b. The fi rst column (b50) uses the simple number of ties as a proxy for social networking power while column two uses the ‘optimal b’, and it is clear that there is no signifi cant eff ect on fi rm performance from social networking.

External networking activity and fi rm performance using a SURE specifi cation allowing for interdependence between fi rm performance and networking is reported in the last part of Table 15.5b and their eff ect on performance from networking is the same. In the equation for the social networking activity, the coeffi cients of educational level are positive and signifi cant as expected.

The general picture is an insignifi cant coeffi cient to all tested variants of the Bonacich power measure of centrality when we look at the external net-working. This leads to the conclusion, in line with Booth and Deli (1996), that persons involved in networking activities may benefi t from the activity, but fi rms representing these persons are not sharing any of this surplus.

Table 15.5c presents the model with internal networking activities. The correlation between internal and external networking is relatively low and therefore there are no signifi cant changes in the results if external network-ing activity is reported in the same model. The basic model reports the same results and, while the eff ect of external networking is insignifi cant and with positive and negative signs, the overall picture is that internal networking activities do increase productivity of the fi rms involved. Using the optimal version of the Bonacich power measure (including secondary values of a network), we fi nd a signifi cant positive eff ect on performance. However, the overall picture is not convincing and the signifi cance disappears when the models are estimated using SURE methodology (last columns in Table 15.5c).

7. CONCLUSION

Networking among persons in top management positions may serve as a means for sharing knowledge and, furthermore, the educational level might increase the capacity of gaining advantages from new information.


Table 15.5b External networking eff ects in models for fi rm ties and productivities

Dependent variable Productivity SURE (produc., network)

Model b50 ‘optimal b’ Productivity Network


8.49(1.30)

8.63(1.29)

0.028(0.045)

Labour 0.698**(0.334)

0.718**(0.336)

0.697**(0.334)

Capital −0.154(0.136)

−0.141(0.135)

−0.154(0.136)

Labour*labour −0.060*(0.034)

−0.060*(0.034)

−0.060*(0.034)

Capital*capital −0.002(0.005)

−0.002(0.005)

−0.002(0.005)


0.066**(0.021)

0.068**(0.021)


−0.272**(0.049)

−0.273**(0.049)

0.029(0.057)


0.264**(0.104)

0.262**(0.104)

Size (dummy for emp. ≥ 500)

−0.312**(0.143)

−0.309**(0.142)

−0.313**(0.143)

0.493**(0.066)


2.87(6.92)

2.89(6.91)

2.85(6.91)

Size* (man. size/empl) 110.60**(38.22)

110.17**(37.97)

110.60**(38.22)

Networking (Bonacich power)

−0.001(0.028)

0.001(0.003)

0.001(0.028)

Education of the CEO 0.343**(0.066)

Education of the chairman

0.387**(0.065)

Minimum effi ciency scale, MES

0.741**(0.057)

Adj. R2 0.648 0.648 r(produc.,netw.)50.002Number of observations 850 850 850

Notes: Numbers in brackets are standard errors of the estimated parameters. * indicates that the estimated parameter diff ers signifi cantly from zero at the 5% level of signifi cance and ** at the 1% level. Models for productivity and networking activity are estimated simultaneously using a SURE (seemingly unrelated regressions) procedure.


Table 15.5c Internal networking eff ects in models for fi rm ties and productivities

Dependent variable Productivity SURE (produc., network)

Model b50 ‘optimal b’ Productivity Network


8.61(1.27)

8.51(1.28)

0.140(0.050)

Labour 0.725**(0.334)

0.675**(0.333)

0.724**(0.334)

Capital −0.145(0.133)

−0.133(0.132)

−0.145(0.133)

Labour*labour −0.062*(0.033)

−0.056*(0.034)

−0.062*(0.034)

Capital*capital −0.002(0.005)

−0.002(0.005)

−0.002(0.005)


0.065**(0.021)

0.067**(0.021)


−0.265**(0.049)

−0.269**(0.049)

−0.110*(0.063)


0.263**(0.104)

0.264**(0.104)

Size (dummy for emp. ≥ 500)

−0.308**(0.142)

−0.297**(0.142)

−0.309**(0.142)

0.452**(0.073)


2.52(6.91)

2.01(6.90)

2.51(6.91)

Size* (man. size/empl) 109.80**(37.95)

108.64**(37.85)

109.79**(37.95)

Networking (Bonacich power)

0.028(0.025)

0.002**(0.001)

0.030(0.025)

Education of the CEO 0.269**(0.073)

Education of the chairman

0.344**(0.072)

Minimum effi ciency scale, MES

0.002(0.279)

Adj. R2 0.649 0.651 r(produc.,netw.)5−0.002Number of observations 850 850 850

Notes: Numbers in brackets are standard errors of the estimated parameters. * indicates that the estimated parameter diff ers signifi cantly from zero at the 5% level of signifi cance and ** at the 1% level. Models for productivity and networking activity are estimated simultaneously using a SURE (seemingly unrelated regressions) procedure.


In addition, networks among CEOs could facilitate and lead to collusion among the competitors in the market. Therefore, the expected infl uence on fi rm performance of this networking and education is positive.

The empirical analysis is based on a sample of large Danish fi rms and it shows no eff ect of the educational level of the top management on perform-ance. On the other hand, we fi nd interdependency between the educational level of the CEO and the chairman of the board. Education also aff ects the overall attitude towards networking positively.

The eff ect of networking between top managers on their fi rms’ perform-ance is in general not found. However, we fi nd a weak signifi cant positive eff ect on fi rm performance regarding internal network activities, which means that social networking activities may be interpreted as control of subsidiaries, but we fi nd no support for a conjecture of a positive eff ect on fi rm performance from the ‘old boy networks’.

NOTES

* We would like to thank two anonymous referees and the participants of the 7th Workshop on Corporate Governance and Investments in Jönköping, April 2006, for helpful comments.

1. Online access: www. GreensOnline.dk

REFERENCES

Bennedsen, M., K. Nielsen, F. Pérez-González and D. Wolfenzon (2005), ‘Inside the Family Firm: The Role of Families in Succession Decisions and Performance’, WP 2005-13, Centre for Industrial Economics, University of Copenhagen.

Bonacich, P. (1987), ‘Power and Centrality: A Family of Measures’, The American Journal of Sociology, 92 (5), 1170–82.

Bonacich, P. and P. Lloyd (2004), ‘Calculating Status with Negative Relations’, Social Networks, 26, pp. 331–338.

Booth, J. and D. Deli (1996), ‘Factors Aff ecting the Number of Outside Director-ships Held by CEOs’, Journal of Financial Economics, 40, 81–104.

Borgatti, S.P., M.G. Everett and L.C. Freeman (2002), Ucinet for Windows: Software for Social Network Analysis, Harvard, MA: Analytic Technologies.

Burkart, M., F. Panunzi and A. Shleifer (2003), ‘Family Firms’, Journal of Finance, 58, 2167–201.

Conyon, M. and M. Muldoon (2006), ‘The Small World of Corporate Boards’, Journal of Business Finance & Accounting, 33, 1321–43.

Freeman, L.C. (1979), ‘Centrality in Social Networks Conceptual Classifi cation’, Social Networks, 215–239.

Hanneman, Robert A. and Mark Riddle (2005), Introduction to Social Network Methods, Riverside, CA: University of California, Riverside (published in digital form at http://faculty.ucr.edu/~hanneman/)


Kogut, B. and G. Walker (2001), ‘The Small World of Germany and the Durability of National Networks’, American Sociological Review, 66, 317–5.

La Porta, R., F. López-de-Silans and A. Shleifer (1999), ‘Corporate Ownership Around the World’, Journal of Finance, 54, 471–517.

Mizruchi, M. and D. Bunting (1981), ‘Infl uence in Corporate Networks: An Examination of Four Measures’, Administrative Science Quarterly, 26, 475–89.

Morck, R., A. Shleifer and R.W. Vishny (1988), ‘Management Ownership and Market Valuation: An Empirical Analysis’, Journal of Financial Economics, 20, 293–315.

Morck, R., D. Wolfenzon and B. Yeung (2005), ‘Corporate Governance, Economic Entrenchment and Growth’, Journal of Economic Literature, 43, 657–722.

Shleifer, A. and R.W. Vishny (1997), ‘A Survey of Corporate Governance’, Journal of Finance, 52, 737–83.

Tirole, J. (2006), The Theory of Corporate Finance, Princeton, NJ: Princeton University Press.

Thomsen, Steen and Evis Sinani (2005), ‘A Small World in a Small Country: Ownership and Board Ties in Denmark 1991–2001’, Working Paper, Copenhagen Business School.

Wasserman, S. and K. Faust (1994), Social Network Analysis: Methods and Applications, Cambridge, MA: Cambridge University Press.

401

accounting principles, German corporate governance ratings 364–5

Adams, R. 142administered price hypothesis 48agency

costs 140, 141, 163, 210and institutional ownership 229,

230principal–agent contracts 4, 52,

53–4, 86, 87principal–agent model 202, 204and property rights theory 86theory 86, 99, 188

Alchian, A. 31, 86, 87allocation and economic growth

124–7all-or-none trading rule 24Amadeus database 294antitrust 4–5

crisis in (1970) 11–12general application to 18–19inhospitality tradition in 22objections to exchanges 26–7

appropriable quasi-rent 68, 79arbitrage pricing theory (APT) 171,

180, 183‘The Architecture of Complexity’

(Simon) 31Areeda, Philip 23, 24Arrow, K.J. 105, 130asset specifi city 14, 19–20associated families, Balearic region see

family businesses, Balearic regionauction rule, US 192, 197authority

bureaucratic 88defi nitions 83–4, 90delegation of 89formal and real 88–90, 91managerial see managerial authority

automotive industry 67–8

backward integration 21Bain, J. 20Balearics region (Spain), family fi rms

in see family businesses, Balearic region

bargaining power, and authority 85Barnard, Chester 14, 83Baumol, William 49bauxite ore, raw materials

procurement 21–2Bebchuk, L. 142behavioral attributes 14Bennedsen, M.B. 143Berle, A.A.

The Modern Corporation and Private Property 2, 45, 46

on ownership concentration 140, 141, 142, 162

and takeover regulation 190Bertrand, M. 329bilateral dependency, asset specifi city

14Bjuggren, P.-O. 148, 151, 152, 155Black, B. 211black box of fi rm, opening

antitrust analysis 19exchange agreements 26fi rm size, limits to 30and market organization 1, 11, 33and neoclassical model 64

Blair, M.M. 326Blake, H.M. 33board composition 5–7board neutrality rule, Thirteenth

Company Law Directive 209board size, German corporate

governance ratings 364Bøhren, O. 324, 326, 330Bonacich, P. 382, 387–8Booth, J. 385boundary of fi rm issue, scaling up

31

Index

402 Index

bounded rationality (contractual incompleteness) 14, 79

Brick, I.E. 330Burkart, M. 142, 383–4business judgment rule, US 192business opportunities space (universal

state space) 104, 107, 108and competence bloc theory 111–12

CAPM (capital asset pricing model) 4

conventional 180institutional risk and uncertainty

170–71, 175, 178, 182multi-beta 171

cargo shipping 71Carlsson, Bo 113Carter, D.A. 330cash fl ows

cash fl ow rights and control rights 142, 143, 163

cash fl ow rights and performance 142, 158

and dual-class shares 163and neoclassical model 50rates of return on 49role in investment 56–7

Chandler, Alfred 33Cheung, Stephen 86, 87Chicago School, antitrust scholars

from 18civil law

shareholder value and legal conception of fi rm 186–90

takeover bidsEurope 201–4Japan 204–7

Claessens, S. 142Club Med, shareholder agreements

261, 267, 271, 273, 275, 276Coase, Ronald H. 15, 63, 85, 87, 108

‘The Nature of the Firm’ 84co-determination, impact upon fi rm

performance 6, 323–54board employee directors 326, 330,

331negative employee director effect

327–8, 355data and institutional background

331–8, 335–7

econometric evidence 339–46, 340, 343, 345

employee directors, effects 6estimation and method 338–9fi xed effects estimations 325, 338lagged fi rm performance 328, 352leverage 330, 342, 355literature review 325–6methodology 338, 339reverse causation hypothesis, and

co-dermination hypothesis 329robustness checks 339, 346–53, 347,

350–51simultaneity and endogeneity 323–4,

328–31stakeholder or interest group 326–8theory and hypotheses 326–31three-stage least squares (3SLS)

methodology 6, 325, 338Wald test 342, 352see also employees

cognition, transaction cost economics 14

cognitive competence 13commercialization competence 114company interest, concept 190Company Law Directive (13th),

takeover bids 194, 196, 201, 203, 204, 209

Company Law Review Steering Committee/Group 187, 188

competence 13bloc theory see competence bloc

theorybusiness, nature of 110–13commercialization 114horizontal diversity 116–17receiver 113and social capital 121specifi cation of fi rm in EOE 107–9venture capital 115

competence bloc theory 107actors in 3, 110, 111, 113–17decision structure of bloc 111defi nitions 126hierarchies, limits 112industrial spillover generator, bloc

as 117institutions and incentives/

competition 120–21

Index 403

nature of business competence and efficiency of project selection 110–13

theory of fi rm 117–20vertical completeness of bloc 115–16

competitionendogenous growth through 121–4and institutions 120–21product market 51

Compustat Global database 148concentration of ownership

in Anglo-Saxon countries 144and descriptive statistics 235–9Germany, corporate governance in

7, 363intermediate levels of 263, 264–5and investment performance 141–2non-linear effects on performance

141–2, 156, 162–3, 226and ownership structure 142in Scandinavia 144, 145

see also Scandinavian countriesconglomerates 33contract of affreightment (COA) 71, 72contracts

contract as framework/contract as legal rules 17

employment 66, 84–5, 100fi rm as nexus of 64, 65–8, 66freight 72hazards, contractual 20insecure, risk of 170long-term 68in maritime transport 66, 75–8and markets and fi rms 68–70principal–agent 52, 53–4, 86, 87spot 76–7wide spectrum of 86, 87see also contractual perspective of

fi rm; freedom of contractscontractual perspective of fi rm 2,

64–70fi rm as nexus of contracts 64, 65–8,

66fl exibility 68maritime industry 2, 74–8mutual dependency 2, 69–70, 78specialization and institutions 65water tightness 68see also contracts

contractual specifi cities 76control enhancing mechanisms (CEM)

142control rights, shareholder agreements

270–71Conyon, M. 383coordination 93–8

problem 85–6, 94, 96corporate control

economists’ view of market 209–13and investment, in Scandinavia

140–44market for 52, 53, 213–14stock market prices and market

213–16corporate governance

Anglo-Saxon model 212, 225Codes 365–7French model 189in Germany 361–79governance structures 14–15and ownership 227–32principal–agent model 202, 204in Scandinavia 139–40, 143–4,

145corporate return, in Scandinavia 149,

151–6cash-fl ow rights and performance

158concentration of control, voting

rights and performance 159dual-class shares 160, 161and ownership structure 156–7,

160–62corporate value, concept of 206, 207corporation

Anglo-Saxon version 43centralized/de-centralized 33and early economists 43–6managerial discretion 51–4, 56managerialist challenge 48–51‘marginalist’ controversies 46–8M-form structure 33modern see modern corporationrecent developments 54–5

Cosh, A. 216creative destruction process,

Schumpeterian 106, 121, 124credible commitments 25–8Cronqvist, H. 143

404 Index

cross-holdings 142, 143, 203, 204Cyert, R.M. 49

de Beers 24, 25de la Torre, C. 240De Soto, Hernando 168–9Debreu, G. 130dedicated assets 69, 80Delaware courts, US 189, 192delegation

of authority 96–8of discretion 92, 97

Deli, D. 385Demsetz, H. 35

on experimentally organized economy 105

on managerial authority, knowledge economy 86, 87

on ownership concentration 141, 144

on scaling up 31Denmark

dominant fi rm 152dual-class shares 143, 144networking in see networking: in

Denmarkdescriptive statistics 235–9, 390–91diffusion, technological 125–6disclosure requirements, shareholder

agreements 258discount rate 169, 170discretion

defi ned 88–9delegation of 89, 97exercise by employees 89managerial 51–4, 56

discriminating alignment hypothesis 15distribution, vertical market relations

22dividends, and institutional ownership

see institutional ownership and dividends

Drobetz, W. 361, 362, 378dual-class shares 3

and cash-fl ow rights/control rights 163

Denmark 143, 144effects 142Finland 143, 144Norway 3, 143

ownership and performance 160, 161Sweden 3, 5, 143, 144

early economists 43–6Easterbrook, F.H. 330economic mistakes, informational

assumptions 105educational background, networking

391–2efficient capital market 2, 51, 52–3Eisenberg, T. 330Eliasson, G. 108, 109, 119Elster, Jon 34employees

on board see co-determination, impact upon fi rm performance

costs of bargaining with 91discretion, exercise of 89, 100and hostile bids 198, 199, 200

employment contracts 66, 84–5, 100endogenous growth

competition, through 121–4micro-to-macro model, Swedish

122Salter curves 122–3

entrenchment effect 141, 142, 163, 383, 385

entrepreneurs 114, 115, 168EOE (experimentally organized

economy) see experimentally organized economy (EOE)

equity rights, shareholder agreements 269

Ericsson 152Europe

Codes of Corporate Governance 365–7

takeover bids 201–4Eurostat 149exchange agreements 25–8

Canadian Study 26, 28entry fees 26–7growth and supplementary supply

restraints 27–8objections to exchanges 26–7petroleum exchanges 26

executive remuneration, German corporate governance ratings 365

experimentally organized economy (EOE) 104–27

Index 405

allocation and economic growth 124–7

business opportunities space 104, 107, 108, 111–12

competence bloc theory see competence bloc theory

competence specifi cation of fi rm in 107–9

critical mass 117dominant selection problem 109, 110efficient selection in 111endogenous growth 121–4fl exibility 117informational assumptions 104–10macro dynamics, experimental

selection 109–10MOSES model 127, 128opportunities space assumption 109property rights 120static equilibrium 128tacit dimension 108

Faccio, M. 142Falaye, O. 326Fama, Eugene 51, 118family businesses, Balearic region

292–319business group directors and family

312–15business groups under control of

associated families 305–17companies under control of

associated families 298–305data sources and methodology

294–8defi nition of family business 292–4family group heterogeneity 306–7measurement of family group

diversifi cation 307–12relevance of associated family

companies in region 296–8representative company 298–9sector diversifi cation and family

control 315–17sector of activity diversity 303–5size, differences according to 299–

300, 301, 302–3‘two-surnames’ system 6

Faure, M. 167Fauver, L. 325, 352, 353

Federal Trade Commission, US 26, 35Ferreira, D. 142ferry/cruise market, maritime industry

71fi nancial contracting 255, 279Finland

dominant fi rm 152, 153dual-class shares 143, 144

fi rm interaction, social network analysis 386

fl ip-in/fl ip-over, anti-takeover defences 193

Ford Motor Company, raw materials procurement 21

forward contractscontractual specifi cities 76tramp shipping 70

Frank, Robert 50freedom of contracts 168, 262–74freight market 70–71Fuchs, Victor 11Fudenberg, Drew 14, 35Fuerst, M.E. 325, 352, 353functional efficiency of capital markets

146, 162fundamental valuation efficiency

(FVE) 213

General Motors 54Georgescu-Roegen, N. 112Germany, corporate governance in

361–79accounting principles 364–5board size 364Code of Corporate Governance

367‘comply or explain’ kind 6 see also corporate governance:

Codesconcentration of ownership 7, 363data description 367–71empirical results 372, 373, 374executive remuneration 365explanatory variables 369, 370, 371HGB rules 364, 365hypotheses 363–5IAS rules 362, 364, 365ownership concentration 363ratings 367–9

components of 376, 377, 378

406 Index

robustness tests 374–6US-GAAP rules 362, 364, 365, 378

Gierke, Otto von 190Gilson, R. 211Gompers, P. 329governance structures 14–15Grabowski, H. 49Gugler, K. 148, 155–6

Hall, R. L. 47, 48Hannah, Leslie 200, 201Hart, Oliver 86Hayek, Friedrich 14‘Heisenbergian fl ux’, economy in 105Heritage Foundation 167, 173, 177Hermalin, B.E. 331hierarchies, governance structures 15High Level Group of Experts, on

takeover bids 201, 202, 203, 204Hitch, C.J. 47, 48Holmstrom, Bengt 14, 92horizontal diversity, competence

116–17horizontal mergers 32Hughes, A. 216human capital (production factor) 65Hume, David 168, 180hybrid contracting, governance

structures 15

IBM 111, 118Imperial–Shell exchange agreement

27–8incentives

business opportunities space (universal state space) 107

concentration of ownership 141and institutions 120–21in knowledge economy 92Särimner effect 107use of authority from perspective of

90–93Industry and Trade (Marshall) 45information arbitrage efficiency (IAE)

213informational assumptions

allocation and economic growth 126–7

competence specifi cation of fi rm in EOE 107–9

economic mistakes 105grossly ignorant actor 105industrial development theory

104–10Särimner effect 106–7

innovations, market for 114in-or-out trading rule 24, 25institutional ownership and dividends

5agency arguments 229, 230Breusch-Pagan/Cook-Weisberg test

241concentration, and descriptive

statistics 235–9earnings trend model, modifi ed

232–4empirical results and analysis

239–45, 246FGLS estimations 241, 242, 243fi xed effects estimations 244full and partial adjustment models

232Hausman test 243hypotheses 230–32institutional shareholdings, positive

effect on dividend changes 231non-linear relationship 231, 241OLS regression 241, 245research methodology 232–4signalling arguments 229–30taxation arguments 228–9variables 234, 235vote-differentiated shares 231–2,

242, 244, 245Waud model 232, 233see also ownership

institutional risk and uncertaintyadaptations to 85estimations of risk and return 171–2,

180fi rst-pass regression 173, 177freedom of contracts 168insecure property rights and

contracts, risk of 170models and results 177–80, 179net present value 169political risk 168portfolio theory and investment

170–71, 180property rights 4, 168–9

Index 407

regression specifi cation error test (RESET) 180

research data 172–3, 174–5, 176risk-free rate plus risk premiums

181–2second-pass regression 177, 178transaction attributes 14world market portfolio 177

Institutional Shareholders Committee, guidelines 217–18

institutionsand incentives/competition 120–21institutional environment,

importance 3–5and specialization 65

intermediate product market transaction (paradigm transaction) 15–18

Node A (unassisted market) 16Node B (unrelieved hazard) 16–17,

23Node C (credible commitment) 17,

23Node D (integration) 17, 18simple contractual schema 16

International Country Risk Guide (ICRG) 173, 177, 178

International Stock Exchange, Pre-emption Group 217

investor rights protection (IRP) 173

Japan, takeover bids 204–7Jensen, Michael 30–31, 54–5, 92,

140–41, 143joint stock companies 43Jorgenson, Dale 50, 211

Kaplan, A.D.H. 47Kaplan, S. 255Kasper, W. 168Kay, John 214Kelly, Marjorie 208Kennedy, Allen 208Kenney, Roy 24Keynes, John Maynard/Keynesian

economics 46, 113Kindahl, James 48kinked-demand schedule hypothesis

47, 48Kirzner, I.M. 105

Klein, Benjamin 17, 24, 68Knight, Frank 29knowledge economy 3

centralized decision making 95diminishing use of authority in 3,

92, 98–9information dispersal 91–2informational assumptions 104investment in assets 91managerial authority in 82–99

Kogut, B. 383Koopmans, Tjalling 11Kuth, E. 57

La Porta, R. 144, 226, 383Lang, L.H.P. 142lateral integration 20–21Lee, S. 142legal conception of fi rm, and

shareholder value, in common and civil law 186–90

Legrand, shareholder agreements 261, 267–8, 276

Lehn, K. 144lens of contract/governance 12, 13, 34Lester, Richard 47Lewis, Tracy 29life-cycle hypothesis 49, 50liner market, maritime industry 71Lintner, J. 232Llewellyn, Karl 17

Magirou, E. 70managerial authority

and bargaining power 85centralized 94–6change in relative use of 90–98

diminishing of use, in knowledge economy 3, 92, 98–9

from incentive perspective 90–93from production coordination

perspective 93–8coordination problem 85–6delegating, setting 96–8in fi rms and markets 84–90incentive perspective, use from

90–93managerial, in knowledge economy

82–99measurement costs 87

408 Index

orders 82–3, 87, 99production coordination

perspective, use from 93–8and property rights 82, 85–6relations between employer/

employees 2–3subordinate’s acceptance of 84see also authority

managerial discretionconstraints on 51–2

strength of constraints 52–4end to 56

managersmanagerialist challenge 48–51market for 51, 53of private-sector companies 186see also managerial authority;

managerial discretionmandatory bid rule, United Kingdom

194, 209Manne, Henry 52March, J.G. 49marginal q, use of 3, 4, 55

agency hypothesis 148cumulative distribution 150–51defi nitions 145–6measurements 146, 162ratios 146in Scandinavia 3, 149, 150–51, 152see also Tobin, J./Tobin’s q

marginalist pricing models 46–8maritime industry 2

bulk shipping, contracting practices 77

carrier and shipper, link between 78

characteristics of maritime transport 70–74, 77–8

contractual perspective 74–8contracts in maritime transport

66, 72, 75–8see also contractual perspective

of fi rmeconomic organization 74–8freight loading, and scaling-up 31freight market 70–71shipping company 72–4structure of shipping services

in relation to cars/car manufacturers 69

third-party ship management 66, 74–5, 79

tramp shipping 70, 71vessel as fungible asset 67

mark-up pricing model 47market

corporate control 209–16see also corporate control

economists’ view of, for corporate control 209–13

for innovations 114for managers 51, 53‘pure vanilla’ type 1and stock market prices 213–16

market organizationantitrust see antitrustcredible commitments 25–8intermediate product market

transaction (paradigm transaction) 15–18

lens of contract/governance 12, 13, 34microanalytics 13–15and opening black box of fi rm 1,

11, 33price theoretic issues 22–5vertical market see vertical market

relationssee also corporation; modern

corporationmarket-for-corporate control 52Marris, Robin 49, 52Marshall, Alfred 2, 45, 46Masten, S. 75Matthews, R.C.O. 35Maury, B. 143McConnell, J.J. 141Means, G.C. 46–7, 48

The Modern Corporation and Private Property 2, 45

on ownership concentration 140, 141, 142, 162

Meckling, William 30–31, 92, 140–41, 143

mergers 32, 208Merrick Dodd, E. 190METI (Japanese economics ministry)

206Meyer, J.R. 57micro-to-macro model, Swedish 122,

124, 125

Index 409

Miguel, A. 245Milgrom, Paul 14, 92Mill, John Stuart 2, 44, 45, 46Mitroussi, K. 74, 75modern corporation

conglomerates 33fi rm size, limits to 29–30horizontal mergers 32scaling up 30–32

modifi ed earnings trend model, institutional ownership 232–4

Modigliani and Miller cost of capital 50

The Modern Corporation and Private Property (Berle and Means) 2, 45, 46

Moller-Maersk 152monopolies, oligopoly–monopoly

comparisons 32moral hazard 91, 92Morck, R. 139, 141, 227Morgan Stanley world market index

171, 172, 173MOSES (Model of the Swedish

Economic System) 127, 128Mueller, D.C. 145, 148, 149Muldoon, M. 383Mullainathan, S. 329Muris, Timothy 19, 35mutual dependency, contractual

perspective of fi rm 2, 69–70, 78

NACE (economic business activity) codes 294, 309

one-digit level 303two- and three- digit levels 308, 310

‘The Nature of the Firm’ (Coase) 84NBS (Nippon Broadcasting System)

205neoclassical model 2, 45

assumptions 106‘black box’ theory 64corporate control and investment

140investment 50limitations of 50and marginalist controversies 46, 47

net present value (NPV) 169networking

activities 384–5

Bonacich approach 382, 387–8data description 389–92in Denmark 7, 382–98descriptive statistics 390–91educational background 391–2empirical model/results 392–5and fi rm performance 383–4network ties

French shareholder agreements 273–4

internal and external 7, 389, 392measuring, using ownership

structure information 388–9nodes and lines 356social network analysis 386–9translog specifi cation 392

new growth theory 114New York Stock Exchange index 171nexus of contracts, fi rm as 64, 65–8Nielsen, K.M. 143Nilsson, M. 143Nippon Broadcasting System (NBS)

205‘Nirvana fallacy’ 105, 128Node A (unassisted market)

contracting 16Node B (unrelieved hazard) 16–17, 23,

27, 35Node C (credible commitment) 17,

23, 35Node D (integration) contracting 17,

18Nokia 152Norsk Hydro 152Nortel 214North, D.C. 168Norway

dominant fi rm 152, 153, 154dual-class shares 3, 143marginal q, use of 4proportionality principle 144

oligopolies 32, 47one-share-one vote principle 142, 144opportunism 14organization of economic activities 2–3over-searching 24–5ownership

and capital 65–6categories 235

410 Index

concentration see concentration of ownership

and corporate governance 227–32institutional see institutional

ownership and dividendsnature of, and shareholder

agreements 262–8, 264–5nominal 142pyramid 142, 143single owners 143structure see ownership structure

ownership structureand board composition/fi rm

performance 5–7and concentration of ownership 142and corporate return, Scandinavia

156–7, 158–9, 160–62measuring network ties between

fi rms using information on 388–9

Pajuste, A. 143Panel on Mergers and Takeovers, UK

194paradigm transaction 15–18Pernod Richard, shareholder

agreements 259–60, 266, 267, 272, 276

Perotti, E. 330physical capital (production factor) 65Pindado, J. 240Pirrong, Stephen 64, 75–6poison pills, anti-takeover defence 193political risk 167, 168Political Risk Group 167Porter, Michael 212portfolio theory and investment

170–71, 180Pound, J. 229, 230predatory pricing 23–4present value (PV) 169price discrimination 22–3price rigidity 47, 48price theoretic issues

marginal cost pricing test 24mark-up pricing model 47output test 24over-searching 24–5predatory pricing 23–4price discrimination 22–3

repositioning 24Robinson-Patman Act 23trading rules 24, 25

principal–agent contracts 4, 52, 53–4principal–agent corporate governance

model 202, 204private equity market, competence

bloc theory 110product market competition 51, 52product variation 112production coordination, use of

authority from perspective of 93–8profi t disgorgement 193profi t maximization 47, 48, 141‘proper purposes’ doctrine 196property rights

and agency 86experimentally organized economy

120insecure, risk of 170institutional risk and uncertainty

168–9and managerial authority 82, 85–6nexus of contracts, fi rm as 65–6

property rights protection (PRP) index 173

proportionality principle, Norway 144

Publicis, shareholder agreements 260, 266, 267, 276

pyramid ownership 142, 143

Raheja, C. G. 327Rathenau, Walther 190raw materials procurement, vertical

market relations 21–2Reardon, Elizabeth 54, 145, 148, 149regression specifi cation error test

(RESET) 180remuneration, German corporate

governance ratings 365repositioning, predatory pricing 24RESET (regression specifi cation error

test) 180Robinson-Patman Act (Anti-Price

Discrimination Act) 1936 23robustness tests

co-determination, impact upon fi rm performance 339, 346–53, 347, 350–51

Index 411

German corporate governance 374–6

Roe, M. 278Roll, R. 171, 177Ruback, Richard 54rules, formal and informal 168,

170

Särimner effect, informational assumptions 106–7

scaling up 30–32Scandinavian countries 3, 139–63

cash-fl ow rights and performance 158

corporate control and investment 140–44

corporate governance in 139–40, 143–4, 145

corporate return in 149, 151–6and ownership structure 156–7,

158–9, 160–62dual-class shares 142, 160, 161homogeneity of 140, 144hostile bids rare in 143largest countries in 153–4marginal q, use of 3, 149, 150–51,

152micro-to-macro model, Swedish 122,

124, 125over-investment 4research methodology 145–9vote-differentiated shares in 142,

143, 157, 162Schneider Electric, shareholder

agreements 261, 272, 273, 275Schumpeter, J. 55, 130

creative destruction process 106, 121, 124

on innovator and entrepreneur 129

self-interest, transaction cost economics 14

Servaes, H. 141shareholder agreements 5, 253–80

antecedents of 262–74background 255–6cases

Club Med 261, 267, 271, 273, 275, 276

Legrand 261, 267–8, 276

Pernod Richard 259–60, 266, 267, 272, 276

Publicis 260, 266, 267, 276Schneider Electric 261, 272, 273,

275control rights 270–71defi nition of 255–6disclosure requirements 258duration of 257empirical setting and methods

256–61equity rights 269in France 256–8impact 274–7listed fi rms, used by 256–7minority investors, protection 271more likely to be found, where

in companies with intermediate levels of ownership concentration 263, 264–5

incumbent shareholders seeking to maintain dominant control 268

large investors seeking to protect bargaining power 272

long-term interest of shareholders 266–7

non-fi nancial objectives of owners 267–8

shareholders addressing complex and conditional issues with intermediate information asymmetry 271–2

social ties of leading officers and directors 273–4

stable businesses with small lock-in costs 268–9

takeover risk 272nature of contract items 269–72,

270nature of industry 268–9nature of ownership 262–8, 264–5negatively perceived where 276network ties 273–4non-equity and control issues 271positively perceived where 276–7related literature 256sources and methods 258–9typical contracts 257written contracts 255

412 Index

shareholder value and legal conception of fi rm, in common and civil law 186–90

shareholdersfree-riding by 197, 230issues addressed by, with

intermediate information asymmetry 271–2

long-term interest of 266–8as owners of fi rm 67primacy, notion of 187, 208see also shareholder agreements

shark repellents, anti-takeover defence 192–3

ship-management companies 75shipping company, and marine

industry 72–4see also maritime industry

Shleifer, A. 142Short, H. 232‘short-termism’ 212, 247Shrader, C.B. 330Siebert, Calvin 50Simon, Herbert 13

on authority 83–4, 85, 90on knowledge economy 91, 97on managers 49on scaling up 31

Sinani, E. 383Singh, A. 212, 216size of fi rm

and associated families 299–300, 301, 302–3

limits to 29–30Skogh, G. 167Smith, Adam 2, 43–4, 45, 46, 63

on property rights 168, 180The Wealth of Nations 44, 65

Smith, N. 330social capital, and competence 121social network analysis

fi rm interaction 386measuring power using 386–8measuring ties between fi rms

388–9Solow, R. 30specialization 65, 67Spier, K.E. 330spot markets/contracts 15, 70, 76–7Standard and Poor’s 500 index 171

The State of Competition in the Canadian Petroleum Industry 26

static equilibrium, EOE 128Stigler, George 47, 48stock market

corporate control, market for 214–15

as evolutionary mechanism 213–14

in developed countries 175efficiency types 146, 162, 213mispricing of shares 214–15over- or under- estimation of 160pricing process 213in Scandinavia 148–9swings, sensitivity to 151technology boom (1995–2000) 214see also takeover bids

Stout, L.A. 326strategic acquisitions market,

competence bloc theory 110Streit, M.E. 168Strine, Leo 189Strøm, R.O. 324, 326, 330Strömberg, P. 255Stuckey, John 21Summers, Larry 212Sweden

dividends 228, 230dominant fi rm 154dual-class shares 3, 5, 143, 144economic tradition 115innovation capacity 119micro-to-macro model 122, 124,

125MOSES model 127, 128mutual funds 227shipping industry 64, 71taxation system 228–9

Sweezy, P.M. 47, 48

takeover bidsanti-takeover defences 192–3, 196,

207British model 194–201City Code 194, 197, 202, 203civil law model, mainland Europe

201–4Company Law Directive (13th) 194,

196, 201, 203, 204, 209

Index 413

High Level Group of Experts on 201, 202, 203, 204

hostile takeovers 4, 55and employees 198, 199, 200rare, in Scandinavia 143rise of 191

‘just say no’ defence 191, 192legal regulation of 190–209origins of takeover regulation

190–92regulation in emerging and

transition systems 208–9stakeholder statutes 193tender offers 55US model 192–4

taxation, and institutional ownership 228–9

teamwork 31, 86technological core 20, 31technological diffusion 125–6temporal specifi cities 75, 76tender offers 55theory of fi rm 31, 117–20thermal economies 20third-party ship management 66, 74–5,

79Thomsen, S. 383three-stage least squares (3SLS)

methodology 6, 325, 338time charter, freight contract 72time contracts, contractual specifi cities

76time series analysis 171Tirole, J. 330Tobin, J./Tobin’s q 143, 146, 162

co-determination, impact upon fi rm performance 325, 326, 343

defi nition of Tobin’s q 145Germany, corporate governance in

375marginal q distinguished 148

see also marginal q, use ofmeasuring 142relationship between ownership and

Tobin’s q 141Tokyo Stock Exchange (TSE) 207total market value of fi rm, defi ned 148trading rules 24, 25tramp shipping 70, 71transaction cost economics

and authority 3cognition and self-interest 14costly nature of transactions 64intermediate product market

transaction (paradigm transaction) 17

lateral integration 20lens of contract/governance 13and managerial authority 82predatory pricing 23Robinson-Patman Act 23scaling up 31shareholder agreements 5

transactions, attributes of 14transparency, and takeover regulation

192Turner, Donald 23, 24

United KingdomCity Code on Takeovers and

Mergers 194, 197, 202, 203Companies Act 2006 188Company Law Review Steering

Committee/Group 187, 188deregulation policy 190–91mandatory bid rule 194, 209networking and fi rm performance

383privatization policy 190takeover bids 194–201, 203

United Statescorporate law 189, 192Delaware courts 189, 192deregulation policy 190–91economy, during 1990s 43Federal Trade Commission 26, 35marginal q, use of 55networking and fi rm performance

383privatization policy 190Sarbanes-Oxley Act 361takeover bids 192–4

utility maximization 141

value added chains 67–8Veblen, Thorstein 141venture capital market 110, 115vertical integration 36, 67, 69, 70vertical market relations 19–22

distribution 22

414 Index

lateral integration 20–21raw materials procurement 21–2

vertical market restrictions 12, 22Vishny, R. W. 142vote-differentiated shares

and dividend changes 231–2institutional ownership and

dividends 231–2, 242, 244, 245in Scandinavia 142, 143, 157, 162

voyager charter, freight contract 72

Walker, G. 383Walras, L. 130Walras–Arrow–Debreu (WAD) model

106, 127, 129

The Wealth of Nations (Smith) 44

Weber, Max 88Weisbach, M. S. 331Weiss, Leonard 48Wibert, D. K. 151Wicksell, Knut 105Wieberg, D. 152, 155Williams Act, US 192, 193Williamson, Oliver 49, 63, 68

Yermack, D. 330Yurtoglu, B. B. 148

Zeckhauser, R. 229, 230

the modern firm, corporate governancedocshare01.docshare.tips/files/6637/66375734.pdf · 2016. 6....

Documents