Conglomeration: Good, Bad, or Unavoidable?¤

Martin F. Hellwig Christian Laux Holger M. Müller

Abstract

Corporate …nance and investment theory is undergoing substantial changes,

away from the traditional paradigm where …rms consist of a single project and a

single owner/manager toward a richer, more complex setting where …rms comprise

multiple projects ¡ typically operated by di¤erent managers ¡ and headquarters,

an agent to whom owners have delegated authority and whose interests may or may

not coincide with those of the …rm’s owners. Along with this shift in focus a whole

host of questions has emerged. What guides the allocation of funds inside …rms?

Is the resulting allocation more e¢cient than the allocation via external capital

markets? What are the incentive e¤ects associated with the prospect of having

discretion over the retention and reinvestment of future returns? Last but not least,

in the background of all this lies the old, but important question of the boundaries

of the …rm. In an attempt to address these questions this paper discusses some

recent, selected research on internal capital markets and conglomeration.

¤Martin Hellwig and Christian Laux are from the University of Mannheim and Holger Müller is from

the New York University and the University of Mannheim. We are grateful for very helpful comments

from an anonymous referee and the editor of this issue, Günter Franke, as well as for …nancial support

from the Deutsche Forschungsgemeinschaft.

1

1 Introduction

Some twenty-…ve years ago Jensen and Meckling (1976) proposed to explain observed

…nancial contracts and …nancial institutions as optimal solutions to certain incentive

and information problems in relations between entrepreneurs and …nanciers. This was

the beginning of the modern institutional approach to corporate …nance. Jensen and

Meckling themselves used it to propose an explanation of external debt and equity

…nance, along the following lines:

² Decision making inside the …rm cannot be fully and costlessly controlled by outside

…nanciers. As their decisions are relevant to outside …nanciers, any form of outside

…nance gives rise to agency costs, i.e., ine¢ciencies that are due to decision makers

not taking su¢cient account of the external e¤ects of their choices on outside

…nanciers.

² Di¤erent …nancial instruments (debt, equity, etc.) give rise to di¤erent types of

agency costs.

² Observed …nancing patterns are the results of optimal …nancial contracting in that

they minimize total agency costs of outside …nance.

In this analysis, the …rm itself is considered as a homogeneous unit, a kind of black

box which transforms investment resources and managerial e¤ort into return prospects.

At the investment stage, an owner/manager running the …rm is looking for outside funds

to …nance a given investment project bearing a random return. With a clear distinction

between the investment stage and the return stage of the investment project, …nancial

contracting determines the division of returns between the owner/manager and his …-

nanciers at the return stage as well as the provision of outside funds at the investment

stage. Contracts are designed so as to minimize the negative e¤ects of outsiders’ claims

on the owner/manager’s incentives in running the …rm.

The literature that followed Jensen and Meckling relied on the same pattern of anal-

ysis.1 However, in many contexts, the identi…cation of a …rm with a single one-shot

project and a single managing person is misplaced. It neglects some phenomena which

1Prominent examples are Diamond’s (1984) analysis of …nancial intermediation as delegated moni-

toring and Grossman and Hart’s (1988) analysis of the alignment of dividend rights and voting rights

in publicly held corporations. For a critical assessment of this strand of literature see Hart (2001).

2

are important for the questions that the institutional approach to corporate …nance is

interested in:

² “Managerial e¤ort” is usually a matter of multi-person interaction rather than a

single person’s decision to work or shirk.

² Many …rms engage in multiple activities, some of them quite heterogeneous. The

proper degree of diversi…cation of activities is a major concern in the development

of business strategy.

² Firms’ activities involve multiple investment and return stages. These stages over-

lap, retentions of returns from earlier investments are a major source of funds for

new investments.

These concerns are particularly relevant for large publicly held corporations. Such

corporations often reach out from traditional activities into new …elds, using their earn-

ings from established activities to …nance these new ventures. A striking example is

provided by the transformation of Mannesmann from a steel producer into an engineer-

ing conglomerate and, more recently, into a telecommunications company. To come to

terms with such a phenomenon, the institutional approach to corporate …nance must

go beyond the simple paradigm of the …rm as a single one-shot project with a single

managing person.

The literature has recently begun to face up to this challenge. There has been a spate

of papers, both theoretical and empirical, which proposed to assess the phenomenon of

conglomeration of activities within given …rms. In this paper, we review some of this

work, using the occasion to highlight contributions made in a research project at the

University of Mannheim, which we try to place in a larger context.

Before we turn to the actual review, we explain why we consider the research pro-

gram of Jensen and Meckling and the institutional approach to corporate …nance to be

profoundly a¤ected by the three considerations listed above. We begin with the ob-

servation that the nature of agency problems changes dramatically when “managerial

e¤ort” is a matter of multi-person interaction rather than a single person’s decision to

work or shirk. In a …rm with many people, internal control mechanisms may substi-

tute for external control. The provision of “managerial e¤ort” may be enhanced by

mutual supervision or internal competition. It may also be impeded by internal pol-

iticking, rent seeking and the like. Further, when multiple projects are available, the

3

costs and bene…ts of diversi…cation must be a matter of concern even if there is just a

single owner/manager. With many people, the identi…cation of individual people with

individual projects serves to channel interactions, i.e., supervision, competition, poli-

ticking, etc. inside the organization. In a nutshell, the determination of “managerial

e¤ort” becomes a subject for organization theory, which must be fully integrated into

the institutional approach to corporate …nance.

On the …nancing side, it is important to note that the allocation of funds to individual

activities inside the …rm is largely handled as an internal concern of the organization,

without much say from outside …nanciers. This provides a new perspective on the

role of the …rm in the …nancial system. In the Jensen-Meckling approach, where each

…rm is identi…ed with a single one-shot project, the allocation of funds across projects is

altogether the result of …nancial contracting, with …nanciers having the power to approve

or disapprove each project. In reality, the allocation of funds across projects results from

(i) the allocation of funds across …rms and (ii) the allocations of funds across projects

inside the …rms. To the extent that …rms rely on external …nance, …nancial contracting

determines the allocation of funds across …rms, but, typically, it does not determine

the allocation of funds inside …rms. This is determined by the …rms’ “internal capital

markets”, which may therefore be seen as …nancial institutions of their own.

This point is reinforced by the role of retentions as a source of funds. The more a …rm

can rely on retained earnings, the less need it has to call on outside sources, and the more

discretion it will have over the use of funds. To be sure, in a publicly held corporation,

decisions on retentions and distributions of earnings must be approved by shareholder

meetings. However, the costs of organizing shareholders in opposition to management

are so high and the prospects of success so poor that such opposition rarely plays much

of a role. Shareholder meeting approval of management proposals may usually be taken

for granted.

The use of retentions as a source of funds raises additional questions about the …rm as

a multi-person/multi-project organization. To what extent does the allocation of funds

across divisions re‡ect the divisions’ contributions to earnings? To what extent should

it do so? To what extent should “cash cows” be milked in order to fund new ventures?

Making the internal allocation of funds re‡ect the di¤erent divisions’ contributions to

earnings may induce people to work harder for the success of their divisions. It may also

leave little room for funding new ventures and thereby make the overall structure of the

…rm’s activities in‡exible, making it hard to prepare for a time when current activities

4

will have their futures behind them. How should a …rm handle this tradeo¤? How is

this tradeo¤ handled in reality? To what extent are internal capital markets driven

by considerations of optimal incentive contracting, to what extent are they driven by

politicking and power play inside the organization? As we formulate these questions, we

see that concerns of internal organization are at the very heart of the problem of how

funds are allocated across activities.2

The paper proceeds as follows: Section 2 gives a brief introduction to the recent

literature, presenting in particular, two key empirical …ndings which serve as a focus

for current research on the pros and cons of conglomeration. Subsequently, Section 3

discusses the impact of conglomeration on optimal contracting between outside investors

and management (headquarters). We argue that incentive problems in relations between

outside investors and management may be a natural source of economies of scope in …rms

and that increasing headquarters’ discretion regarding the allocation of funds, i.e., cre-

ating an internal capital market may be bene…cial for some, but nor necessarily all types

of enterprises. Section 4 considers con‡icts of interest inside the multi-divisional …rm,

focusing on incentive problems between headquarters and division (project) managers.

Con‡icts of interest between headquarters and outside investors are neglected, and head-

quarters is autonomous in allocating capital. Even though, by assumption, headquarters

has no incentives to waste funds, conglomeration may be costly as both the redistribu-

tion of funds and the easier availability of funds in internal capital markets may weaken

division managers’ incentives to improve their investment opportunities. Section 5 leaves

the contract theoretic paradigm and considers the role of conglomeration in a setting in

which external control of the …rm is weak, and management has discretion over reten-

tions. In such a world, conglomeration is needed as a way to induce structural change,

not just for the …rm, but for the economy as a whole, from activities generating current

pro…ts, to activities promising to generate pro…ts in the future. Section 6 contains a few

concluding remarks.

2Other important issues, which are not addressed here, concern (i) the way in which information

is elicited inside the …rm through the capital budgeting process (see, in particular, Harris and Raviv

1997) and how the capital budgeting process is organized (see Laux 2001c) and (ii) the implications of

conglomeration for the informational e¢ciency of capital markets and the e¤ects of capital asset pricing

on the allocation of resources for investment (see Chang and Yu 2000).

5

2 Is Conglomeration Ine¢cient? A Review of Em-

pirical Findings

The literature on conglomeration has very much been driven by the question of whether

conglomeration is e¢cient or ine¢cient. Early theoretical work stresses e¢ciency gains

from conglomeration. Thus Alchian (1969), Williamson (1975), and Stein (1997) empha-

size the positive role of an improved capital allocation as a bene…t of conglomeration.3

In contrast, the empirical literature has been more skeptical about the e¤ects of con-

glomeration on e¢ciency. Current discussion focuses on the following two observations,

which seem to be at odds with the view that conglomeration creates value by allowing

for an improved allocation of capital:

(O1) Observation 1 (Conglomerate or Diversi…cation Discount). On av-

erage, stock market values of conglomerates are at a discount relative to the values of

comparable portfolios of stand-alone …rms.

(O2) Observation 2 (Ine¢cient Cross-Subsidization or “Socialism”). Less

productive divisions or divisions with bad investment opportunities tend to receive dis-

proportionately large shares of investment resources.

Evidence for (O1) is provided by, e.g., Lang and Stulz (1994), Berger and Ofek

(1995), Servaes (1996), and Lins and Servaes (1999).4 Evidence for (O2) is found in,

e.g., Lamont (1997), Shin and Stulz (1998), Scharfstein (1998), and Rajan, Servaes and

Zingales (2000).

When taken at face value, (O1) would suggest that, from the shareholders’ per-

spective, conglomeration of diverse activities destroys value. (O2) would suggest that

conglomeration of diverse activities goes along with a poor allocation of capital, which

would explain (O1). Conglomeration would thus seem to be the result of insu¢cient

outside control permitting corporate management to pursue its own interests at the ex-

pense of shareholders. Even before the systematic recent empirical work, this view was

prominent in discussions of the nineteen-eighties takeover wave. According to some in-

terpretations, the eighties takeover wave was a case of the “market for corporate control”

(Jensen and Ruback 1983) taking over in order to dismantle the ine¢cient conglomerates

3Other examples include Thakor (1990) and Matsusaka and Nanda (2000).4Note that conglomerates trade at a discount only on average. While the precise number varies from

study to study, roughly 40% of conglomerates are found to trade at a premium.

6

that had been formed in the sixties and to stop wasteful cross-subsidization, e.g., from

known wells to new exploration in the US oil industry (Jensen 1986, 1991).

The negative assessment of the sixties conglomeration wave is not uncontested. Mat-

susaka (1993) suggests that stock market responses to diversifying acquisition announce-

ments during that period were positive if bidder retained target management and neg-

ative when it replaced target management. According to Hubbard and Palia (1998)

the sixties witnessed relatively high bidder returns in diversifying acquisitions when a

…nancially “unconstrained” buyer acquired a …nancially constrained target or vice versa.

This would accord with the view of conglomeration creating value by relaxing …nanc-

ing constraints. This …nding is con…rmed by Klein (2001), who …nds that diversifying

mergers that created an internal capital market may have added value in the 60s.

Even for the more recent research, it not clear that either (O1) or (O2) should be

taken at face value. As there is likely to be a reason for why …rms diversify in the …rst

place, (O1) may be subject to sample selection bias, i.e., the index for “diversi…cation”

that is deemed to “explain” stock market valuations may itself be endogenous. To

…nd out whether diversi…cation destroys value one would ideally have to break up the

conglomerate, take each division public, sum the market values of the former divisions,

and compare this sum to the former market value of the conglomerate.5 As this is not

possible (and divisions in existing conglomerates are not publicly traded), researchers

have taken as a proxy for the divisional market value the value of the median stand-

alone …rm operating in the same industry. Clearly, if there is a systematic di¤erence

between the divisions forming a conglomerate and these benchmark …rms, then the

alleged discount may merely re‡ect this (overlooked) di¤erence and may have nothing

to do with an e¤ect of diversi…cation on value as such. Indeed, Campa and Kedia (1999)

…nd that, after controlling for the endogeneity of diversi…cation decisions, the discount

always is lower, and sometimes may actually be a premium. Similar criticisms are to

be found in Graham, Lemmon, and Wolf (2002), Hyland (1999), and Maksimovic and

Phillips (2002).

(O2) may also be due to selection bias. For instance, Whited (2001) and Chevalier

(2000) provide arguments which cast doubt on the appropriateness of industry Qs as

proxies for the investment opportunities of conglomerate divisions. Moreover, Chevalier

obtains the puzzling result that the cross-subsidization allegedly present in conglomer-

5There are studies examining spin-o¤s or bust-ups of conglomerates (e.g., Comment and Jarrell

1995), but these studies potentially su¤er from the same kind of selection-bias problem, for the spin-o¤

or bust-up decision is likely to be endogenous.

7

ates already exists between independent …rms prior to the diversifying merger, i.e., at a

time when there should be clearly no cross-subsidization. She therefore concludes that

(O2) is likely to be driven by selection and measurement problems.

From a theoretical perspective, it is not clear that (O1) and (O2) provide for a

suitable assessment of e¢ciency. If one thinks of structural change in the economy as

being …nanced alternatively either through disbursements of returns and reinvestments

through the market or through retentions and reinvestments by existing corporations,

then the relative e¢ciency of either regime will depend on the ratios of successes to

failures as one diversi…es into new activities. In this respect, the benchmark …rms in (O1)

are subject to survivorship bias; all those start-ups that disappear for lack of success

are ignored. In contrast, in conglomerate …rms, all those diversi…cation experiments

that fail, i.e., the analogues of the failing start-ups in a system of reinvestment of funds

through the market, are taken into account in market valuations. Similarly, (O2) may be

the result of experimentation inside the organization as well as the pursuit of managerial

sel…nterest, or it may be the result of political economy of rent-seeking and inter-division

in…ghting.

Going one step further, if one sees conglomeration as a result of management with-

holding earnings and e¤ectively expropriating shareholders, one would expect to see a

conglomerate discount even if the di¤erent divisions of the conglomerate …rm were com-

pletely up to standard for e¢ciency and pro…tability. After all, shareholder expropriation

or the prospect thereof will drive a wedge between the …rm’s market capitalization and

the discounted present value of the …rm’s returns.

Recent theoretical research on the e¢ciency implications of conglomeration has

broadly been divided between two lines: One line corresponds to the view that large

conglomerate …rms are ine¢cient because they o¤er more scope for rent-seeking biasing

the capital allocation in favor of low-productivity divisions (e.g., Rajan, Servaes, and Zin-

gales 2000; Scharfstein and Stein 2000) or simply in favor of “old”, established divisions

that happen to wield the most in‡uence in the organization (Hellwig 2000, 2001). An-

other line corresponds to the older view that conglomeration may improve the allocation

of capital as it provides more ‡exibility. According to this view, the observed phenom-

ena (O1) and (O2) may simply be explained from …rms’ optimal, endogenous choices

(e.g., Berkovitch, Israel, and Tolkowsky 2000; Inderst and Müller 2001a). In the follow-

ing two sections, we review some theoretical work along this second line. Subsequently,

in Section 5, we return to the implications of political economy for the assessment of

8

e¢ciency.

3 Incentive Problems between Headquarters and In-

vestors: The Impact of Conglomeration

In this section, we consider the potential impact of a conglomeration of projects on …-

nancial contracting between a …rm and its outside investors. As yet we abstract from

incentive problems inside the …rm. Starting from a standard e¤ort-incentive problem

with veri…able cash ‡ow, as in Jensen and Meckling (1976), Laux (2001a) considers how

the problem is a¤ected if there are multiple projects to be managed. The risk neu-

tral manager (headquarters) is protected by limited liability. Investors write incentive

contracts with headquarters to induce high e¤ort. In this situation a rationale for in-

vestors to combine multiple projects to carry them out within one …rm is provided: in

the absence of large monetary (and non-monetary) punishments an incentive scheme

that induces headquarters to exert e¤ort typically involves a rent as rewards have to

substitute for punishments. If headquarters manages multiple projects, punishments

for failed projects are possible as headquarters may now lose its reward from successful

projects. Combining projects therefore relaxes the limited liability problem and reduces

headquarters’ rent from each project. As a result, expected wage costs of inducing high

e¤ort decrease. Interestingly, the advantage of combining projects arises independently

of any synergies in managing the projects or improving the quality of the performance

measure. Instead, it is the possibility to allow for punishments that is advantageous.

Reducing expected wage costs by combining projects expands the range of project

types for which it is optimal for investors to provide incentives to induce high e¤ort. This

increases the set of parameters for which investors are willing to initiate and …nance the

projects. Hence, incentive problems are identi…ed as a natural source of economies of

scope in …rms where the limits to integration arise only from headquarters’ capacity

to oversee and manage di¤erent projects: with convex e¤ort costs, combining projects

involves a trade-o¤ between reducing the limited liability e¤ect and over-proportionally

increasing e¤ort costs.

Inderst and Müller (2001a) study the relation between internal and external capital

markets in a setting where cash ‡ow is nonveri…able. The question being asked is whether

centralized …rms where headquarters raises …nance on behalf of multiple projects are able

to write more e¢cient …nancial contracts than decentralized or stand-alone …rms. And

9

if yes, how does this feed back into …rms’ …nancing constraints?

Hints on the role which project bundling may play for …nancing constraints are

found in the literature on internal capital markets. There, headquarters either creates

or destroys value inside the …rm, e.g., by engaging in winner-picking, by redeploying

assets across projects, or by a¤ecting project managers’ incentives. Naturally, this value

creation or destruction will a¤ect the return to capital and hence also the …nancing

constraint. As none of these papers adopts an optimal contracting approach, the precise

nature and magnitude of the e¤ect remains unclear, however.6 On the other hand, (opti-

mal) …nancial contracting models, while deriving …nancing constraints and the associated

underinvestment problem from …rst principles, typically consider an entrepreneurial …rm

where an entrepreneur raises funds for a single project. In this setting questions of, e.g.,

organizational structure or the allocation of authority over multiple projects cannot be

addressed.

Inderst and Müller connect the internal capital markets literature with that on opti-

mal …nancial contracting, thus tying together internal and external capital markets. The

authors compare …nancial contracting between i) outside investors and individual project

managers (“decentralized borrowing”) and ii) outside investors and headquarters, who

borrows on behalf of several projects and subsequently allocates the funds to projects on

the …rm’s internal capital market (“centralized borrowing”). Financing constraints arise

endogenously as the paper assumes that part of the project cash ‡ow is nonveri…able.

The crux is to provide the …rm (i.e., the project manager or headquarters) with incen-

tives to pay out funds rather than to divert them. To distinguish the costs and bene…ts

of centralization resulting from …nancial contracting from other, previously studied costs

and bene…ts, Inderst and Müller assume that headquarters creates or destroys no value

per se, i.e., there is no winner-picking, etc.

The bene…t of centralization is that centralized …rms can write more e¢cient …nancial

contracts than decentralized …rms. To induce …rms to reveal their true cash ‡ow investors6Stein (1997) rules out optimal contracting by assuming that it is too costly to elicit managers’

private information. Scharfstein and Stein (2000) assume that outside investors can only decide on

the size of their investment and the …rm’s operating budget. In particular, contracts contingent on

(reported) cash ‡ows are not considered. Finally, Gertner, Scharfstein, and Stein (1994) consider

optimal contracting, but not between headquarters and outside investors. More precisely, they compare

contracting between project managers and investors under two scenarios: i) the manager owns the

project and ii) the investor owns the project. In the latter case they call the investor “headquarters”.

The possibility that headquarters itself may have to raise funds from investors is not explored.

10

must o¤er them a bribe (or information rent). Bribes can come either in the form

of a lower repayment or a higher continuation bene…t. Under centralized borrowing

a greater fraction of the bribe comes in the form of continuation bene…ts, which is

e¢cient as it involves undertaking positive net present value investments that would

have otherwise not been undertaken. E¤ectively, headquarters uses excess liquidity from

high cash-‡ow projects to buy continuation rights for low cash-‡ow projects. This allows

headquarters can make greater repayments to investors, which eases …nancing constraints

ex ante. The cost of centralization is that headquarters may pool cash ‡ows from several

projects, thereby accumulating internal funds, and make follow-up investments without

returning to the capital market. Absent any capital market discipline, however, it is

much harder for outside investors to force the …rm to pay out funds, which tightens

…nancing constraints ex ante. The point is reminiscent of Jensen’s (1986) free cash-‡ow

problem, where it is also di¢cult to force …rms to disgorge internally generated funds.

Anticipating that such a problem may arise in the future, investors are reluctant to

provide …nancing in the …rst place.

Connecting the costs and bene…ts of internal capital markets, Inderst and Müller

trace out the boundaries of the …rm. As the costs and bene…ts arise in di¤erent states of

nature, the boundaries depend on the ex-ante distribution of cash ‡ows. Holding every-

thing else …xed, centralization (or integration) is optimal for projects with a low expected

cash ‡ow while decentralization (or non-integration) is optimal for projects with a high

expected cash ‡ow. Cross-sectionally, this implies that conglomerates should on average

have a lower value than a comparable portfolio of stand-alone …rms, which is consistent

with (O1). While this is not the …rst, or only model generating this implication, it is

the …rst model showing that a discount may result from optimal …nancial contracting.7

Moreover, the discount arises from …rms’ optimal, endogenous choice to diversify, and

not from rent-seeking or agency problems inside the …rm.

By showing that the diversi…cation discount may emerge from low-productivity …rms’

endogenous choice to diversify, Inderst and Müller lend support to concerns that (O1)

may be prone to a selection bias. Evidence for this self-selection hypothesis is provided

by Hyland (1999) and Maksimovic and Phillips (2002), who both …nd that diversifying

…rms are less productive and have weaker operating performance than their stand-alone

counterparts.8 While Inderst and Müller (2001a) are silent about (O2), this point is

7Other models showing that conglomerates may trade at a discount are, e.g., Berkovitch, Israel, and

Tolkowsky (2000) and Fluck and Lynch (1999).8See Schoar (2000) for contrary evidence, however.

11

taken up in Inderst and Müller (2001b), who again connect internal with external capital

markets issues.

In Inderst and Müller (2001b) the fundamental advantage of conglomerates is that

they o¤er better diversi…cation with respect to the generation of investment opportuni-

ties. If a …rm operates in several, unrelated lines of business, it is more likely that at least

one of its divisions generates investment opportunities that can be pursued in the next

period. Given that many production factors are …xed in the short run (e.g., …rms have

administrative departments whose short-run size is given), bundling di¤erent projects

under the same roof alleviates the risk that production factors remain idle. Against

this diversi…cation bene…t must be counted the fact that a richer set of investment op-

portunities may entail greater scope for risk-taking. As is well known from Jensen and

Meckling (1976), shareholders of a levered …rm may have an incentive to choose riskier

projects even if these involve a lower net present value.

This raises several questions. Why should headquarters pursue shareholders’ (ex

post) interests and shift capital to riskier but possibly less productive divisions? And

why should …rms that are prone to risk-taking issue debt in the …rst place? Among other

things, Inderst and Müller consider the possibility that headquarters has a natural taste

for safe projects.9 In this case delegating the project choice to headquarters may act as

a commitment for shareholders vis-à-vis debtholders not to engage in risk-shifting. The

cost of letting headquarters have a free rein is that it may divert part of the …rm’s pro…t as

private bene…ts. Hence shareholders face a tradeo¤. If they remain passive headquarters

will pursue safe investments but also divert pro…ts. On the other hand, by keeping

headquarters on a tight leash shareholders may avoid shirking, but this undermines the

credibility of their commitment not to interfere with headquarter’s (conservative) project

decisions.

Still, the above tradeo¤ is only relevant if the …rm has debt. In an all-equity …rm

shareholders always select the e¢cient project, implying there is no reason to deliber-

ately refrain from monitoring headquarters (except that this saves monitoring costs, of

course). Consider therefore the problem of a founder (and sole shareholder) who is wealth

constrained, implying that he must raise outside …nance to pursue investments. If he is-

sues only equity there will be no risk-shifting, but his incentives to monitor headquarters

will be diluted, i.e., there will be undermonitoring. On the other hand, if he issues only

9Headquarter’s taste for safe projects may stem from career concerns (as in, e.g., Hirshleifer and

Thakor 1992), or from risk aversion which, if su¢ciently pronounced, also makes it prohibitively costly

for shareholders to align headquarter’s interests with their own through an incentive contract.

12

debt there will be overmonitoring as now the prospect of overruling headquarter’s con-

servative project decisions creates additional incentives to gain de facto control. Once

in control, the owner cannot help shifting capital to riskier but possibly less productive

divisions (or advise headquarters to do so), which constitutes an additional cost of debt

…nance.

Inderst and Müller endogenously derive the …rm’s optimal capital structure and

ownership concentration. In equilibrium the …rm will always have at least some debt,

implying that capital will be shifted to riskier, but possibly less pro…table divisions

with positive probability. Interestingly, dispersed ownership and leverage tend to go

hand in hand. While the latter improves the incentives for shareholders to monitor and

overrule headquarter’s decisions, the former acts as a counterbalance as it ampli…es the

free-rider problem in monitoring. As high leverage and concentrated ownership have

the same e¤ect as low leverage and dispersed ownership, the …rm’s capital structure is

indeterminate for a large range of parameter values.

4 Incentive Problems between Headquarters and Di-

vision Management

So far, we have focussed on relations between headquarters and investors without con-

sidering the internal organization of the …rm. Incentive problems inside the …rm are

sure to play a role when di¤erent projects are assigned to di¤erent people. In this sec-

tion we therefore consider the e¤ects of con‡icting interests between di¤erent managers

and headquarters. Momentarily neglecting the role of outside investors, we assume that

headquarters is autonomous in allocating funds between di¤erent projects, and project

managers derive a private bene…t from increased investments under their control. We

also assume that headquarters wants to maximize some measure of expected pro…ts.

This assumption is further discussed in Section 5 below.

One important potential advantage of an internal capital market is that it may allow

for an improved capital allocation. If headquarters allocates funds so as to maximize

(gross) pro…ts, any limit to this advantage of conglomeration must stem from a con‡ict

of interest between headquarters and division managers. Such a con‡ict arises if the

latter are only interested in the assets that they manage: In an internal capital market

residual rights of control over the …rm’s assets reside with headquarters rather than with

division managers. As …rst highlighted by Gertner, Scharfstein, and Stein (1994), this

13

may adversely a¤ect division managers’ incentives. Giving up discretion over the use of

funds in an internal capital market may, for example, reduce division managers’ initiative

to search for investment opportunities and their incentives to generate funds.10 To the

extent that headquarters cannot commit to a selective intervention and that the adverse

incentive e¤ects are large relative to the advantage of an optimal allocation, combining

projects may destroy value and therefore result in a conglomerate discount as de…ned in

(O1).

In addition, division managers may be tempted to engage in in‡uence activities or

rent seeking over the …rm’s resources.11 As a consequence, resources are wasted and

the capital allocations may be distorted, which may at least partly explain (O2). The

resulting ine¢ciencies may be so high that it is optimal not to combine projects. If they

are nevertheless combined (for whatever reason), the conglomerate trades at a discount,

i.e., we have (O1).

In both strands of literature (i.e., the ones dealing with e¤ort provision and in‡uence

activities) the positive side of an internal capital market is limited by the e¤ect of

integration on division managers’ incentives, which is always negative and traded o¤

against the potential positive e¤ect of integration.

Inderst and Laux (2001) consider a situation where division managers have to exert

costly e¤ort to increase the pro…tability of the divisions’ investment opportunities. Divi-

sions are endowed with …nancial resources and headquarters allocates total resources of

the divisions under its control so as to maximize the …rm’s NPV. There is no asymmetric

information with respect to the quality of investment opportunities and managers de-

rive a private bene…t that is increasing in the amount of funds invested in their division.

Hence there are two sources of potential incentives to managers: the capital allocation

as well as wages. If the sensitivity of capital allocation to the investment opportunities’

pro…tability increases, then incentives from the capital allocation increase and therefore

expected wage costs can be reduced. Clearly, in this setting an internal capital market

is bene…cial because scarce funds are put to their pro…t maximizing use. However, an

internal capital market has an additional e¤ect: division managers compete for scarce

corporate …nancial resources. While intuition suggests that competition increases divi-

sion managers’ incentives to exert e¤ort, Inderst and Laux show that the e¤ect may go

10On the …rst type of incentives see Gertner, Scharfstein, and Stein (1994) and Aghion and Tirole

(1997); on the second, see Brusco and Panunzi (2000) and Gautier and Heider (2000).11See Meyer, Milgrom, and Roberts (1992), Bagwell and Zechner (1993), Rajan, Servaes, and Zin-

gales (2000), Scharfstein and Stein (2000), Wulf (2000), and Inderst, Müller, and Wärneryd (2001).

14

either way depending on the types of projects that are combined.

Integration only unambiguously increases incentives if divisions are symmetric ex

ante. In this case integration increases the sensitivity of the capital allocation to the

investment opportunities’ pro…tability and managers’ incentives increase as well. Inter-

estingly, a relaxation of the resource constraint (e.g., due to the integration of a cash

cow division that only generates cash but has no pro…table investment opportunities on

its own) may decrease the sensitivity. The negative e¤ect on incentives may even over-

compensate the advantage of being able to carry out pro…table investment opportunities

on a larger scale.

If divisions are (ex ante) asymmetric with respect to their initial endowment or

the pro…tability of investment opportunities, the e¤ect of integration on the managers’

incentives di¤er. Incentives of the division manager who expects to receive more capital

under integration, i.e., whose initial endowment is relatively lower or whose investment

opportunities are relatively more pro…table, unambiguously increase. Incentives of the

manager of the division which has a higher endowment or which has less pro…table

investment opportunities may decrease under integration. The net e¤ect may be a

reduction of total expected pro…ts when combining projects.

The novelty of the paper is that it analyzes the implications of an internal capital

market on incentives to create investment opportunities and increase their pro…tability.

Thereby it shows that the e¤ect of integration on total …rm value may go either way

without having to resort to additional costs potentially arising from integration. Hence,

the same incentive con‡ict that may result in a conglomerate premium above and be-

yond the mere bene…t of an improved capital allocation may also be the source of a

conglomerate discount, depending on the characteristics of the units that are combined.

Incentives and total …rm value may decrease if units di¤er in their endowments and

investment opportunities. In this case the integrated …rm may trade at a lower value than

the sum of its parts, i.e., the …rm may trade at a discount, despite a pro…t maximizing

allocation of capital. Evidence for such an e¤ect of diversity on the size of the discount

is found by Lamont and Polk (2002) and Rajan, Servaes, and Zingales (2000). In Rajan,

Servaes, and Zingales, however, managers engage in ine¢cient rent-seeking activities,

which are largest if divisions are heterogeneous. Headquarters reduces the rent-seeking

incentives by distorting the capital allocation towards less pro…table projects. Note that

while ine¢cient cross-subsidization surely reduces incentives resulting from integration

in Inderst and Laux, the empirical evidence does not suggest that cross-subsidization

15

is so strong that incentives disappear, i.e., that the capital allocation sensitivity always

decreases under integration. For example, Shin and Stulz (1998), who are typically

quoted as providing evidence for cross-subsidization of smaller segments, do not …nd

such subsidization if the largest segment has itself good growth opportunities.

In a similar setting Stein (2000) analyzes how integration a¤ects incentives of division

managers to acquire privately costly information. By acquiring information, headquar-

ters and division managers can make more informed decisions. Importantly, in contrast

to Inderst and Laux, Stein only considers the symmetric case and does not allow for

monetary incentives. Instead, Stein shows that integration reduces division managers’

incentives to generate information when this information is soft. The reason is that

information cannot be used as a means to receive more funds. Rather headquarters may

take funds away and therefore division managers’ advantage of being informed is lower.

Whenever there is a con‡ict of interest between headquarters and division managers,

headquarters may use the …rm’s organizational and …nancial structures as a commit-

ment device. Bagwell and Zechner (1993) show that headquarters may use the capital

structure to reduce the necessity of divestitures in order to reduce divisional managers’

incentives to engage in in‡uence activities. In their model all projects are carried out

within one entity. Laux (2001b) shows that incorporation of a project in a subsidiary

and taking on external capital through the subsidiary increases headquarters’ incentive

to become informed about the project’s quality prior to making a continuation invest-

ment. Hence, headquarters may commit itself to monitoring the project’s quality prior

to making a continuation decision in situations where total pro…t maximization implies

carrying out the project without monitoring. The advantage of ex post excessive moni-

toring is that this provides the manager of the subsidiary, who derives a private bene…t

from continuing the project, with incentives to exert e¤ort.

5 Beyond Contract Theory: Conglomeration as a

Mechanism of Structural Change

Like Jensen and Meckling (1976) and the literature following them, the papers reviewed

in the preceding sections all follow a contract theoretic approach. Observed arrangements

are seen as the results of optimal initial contracting. For given allocation problems and

given constellations of interests, contractual and institutional arrangements are assumed

to be chosen so that overall agency costs are minimal.

16

However, the contractual and institutional arrangements that we observe in reality

are rarely the results of contracting. Contracts and institutional rules do play a role,

but often they are rewritten as the economic relation between the di¤erent participants

evolves. The changes that take place are shaped by currently prevailing interests and

currently prevailing power relations, which may be quite di¤erent from interests and

power relations at the time of initial contracting. Given this consideration, Hellwig

(2000) questions the appropriateness of the contract theoretic approach and suggests that

it may be more fruitful to explain observed contractual and institutional arrangements

as the results of evolution rather than initial contracting.

He exempli…es the argument with the history of Union Bank of Switzerland (UBS)

from the nineteen-seventies to the nineteen-nineties. In this period, there were three

changes of the corporate charter that a¤ected the standing of outside shareholders vis

à vis corporate management, namely (i) the introduction of registered shares with a

requirement of management approval for the registration of new shareholders, (ii) the

abolition of this latter requirement in combination with the introduction of a 5% voting

rights restriction, and (iii) against the opposition of a blockholder in registered shares,

the abolition of voting rights privileges for registered shares over bearer shares. Each of

these changes was proposed by corporate management and approved by the shareholder

meeting. Each change improved management’s scope for eliminating, disenfranchising,

or otherwise weakening the position of outside shareholders who wanted to mount an

opposition against management. Shareholder approval was ensured, among other things,

by votes cast under a Swiss law requiring banks representing their customers to vote with

management unless their customers give them explicit instructions to the contrary. By

the midnineties, institutional arrangements at UBS had largely been shaped by the

e¤ective power of incumbent management to change the rules.

Hellwig (2000, 2001) argues that such disenfranchisements of outsiders occurs in pub-

licly held corporations in many countries, albeit in di¤erent guises. Where Swiss cor-

porations used restrictions on the registration of name shareholders and, more recently,

voting rights restrictions, German corporations used cross-holdings enabling managers

to protect each other, and US American corporations used poison pill provisions to re-

duce or at least eliminate the prospects of outside interference. The common elements

in all these developments are: (i) Management is able to use its managing role to de-

termine the agenda, including the agenda for changes of rules; (ii) where shareholder

approval is needed, management can exploit the well known free-rider problem of share-

17

holder decision making; (iii) changes of rules that strengthen the power of incumbent

management tend to be supported by political and judicial systems as well as various

groups of “stakeholders” who prefer to deal with incumbent management without in-

terference from outside …nanciers.12 As a result, changes of rules tend to enforce the

entrenchment of incumbent management. An evolutionary approach to the analysis of

observed institutions would therefore view the modern corporation as the result of the

ability of incumbent management to emancipate itself from outside control and outside

interference.

This provides a new explanation of the well known “pecking order” of corporate

…nance whereby companies rely for their investment …nance …rst on retained earnings,

then on loans, and last on new share issues (Myers and Majluf 1984). From the per-

spective of incumbent management, internal …nance is attractive because it involves

a minimum of outside interference, approval of shareholder meetings being taken as a

matter of course. In contrast, external …nance su¤ers from a reluctance of outsiders

to surrender funds without e¤ective control as well as a reluctance of management to

surrender any e¤ective control. This is less of an issue for loans than for shares; out-

side lenders have well de…ned claims and, as long as the claims are honored, no control

rights. Outside shareholders in contrast have no well de…ned claims, but for whatever

it is worth, they have the right to vote at the shareholder meeting. It seems likely that

the pecking order of …nance re‡ects the desire and the e¤ective power of management

to withhold earnings from outside shareholders as well as the awareness of potential

outside shareholders of the moral hazard to which they would be exposed (Jensen 1986,

La Porta et al. 1997). In contrast, Myers and Majluf (1984), relying on a model in

which management acts in the shareholders’ interests, had argued that internal …nance

is preferred simply because it involves the lowest agency costs.

Empirically, the importance of inside …nance in many countries, for many companies,

has been well documented, see, e.g., Mayer (1988, 1990), Corbett and Jenkinson (1996,

1997), and Hackethal and Schmidt (1999). Cross-country comparisons show that the

weakness of outside share …nance is directly correlated with the weakness of shareholder

protection in the di¤erent legal systems. The empirical literature also …nds a signi…cant

positive correlation between cash ‡ow and investment, over time and across …rms, see,

e.g., Fazzari, Hubbard, and Petersen (1988) for the United States, Elston (1998) for

12On the workings of this mechanism in Germany, see Wenger (1996), in the United States, Jensen

(1991) and Roe (1994).

18

Germany, and Hall and Weinstein (1996) for Japan.13 Beyond the simple notion that

…rms invest only what they can …nance on their own, such a correlation may re‡ect the

fact that ceteris paribus, a …rm with a larger equity base will …nd loans to be more easily

forthcoming without too many restrictive covenants (Edwards and Fischer 1994).

Theoretical assessment of the prominence of internal …nance is divided. Myers and

Majluf (1984) stress the positive role of returns from previous investments in alleviating

…nancing constraints on new investments. In contrast, Jensen (1986) argues that man-

agement pursuing its own interests retains free cash ‡ow inside the company, e¤ectively

expropriating outside shareholders and choosing projects that involve private bene…ts

for management rather than returns for shareholders. From the perspective of Myers

and Majluf, a dependence of investment on the availability of internal funds must be

interpreted as evidence of underinvestment, a kind of rationing of pro…table investment

activities due to the unavailability or the agency costs of external funds. From the per-

spective of Jensen, the dependence of investment on the availability of internal funds

may also be interpreted as evidence of overinvestment by a management that is willing

to waste funds in unpro…table investment opportunities.

At this point it is instructive to think about the allocation of funds across …rms in

terms of the structure rather than the overall level of investment activity. If current cash

‡ows and current pro…table investment opportunities are not highly correlated, a system

based on internal …nance can exhibit overinvestment and underinvestment at the same

time, overinvestment at …rms with high cash ‡ows and poor investment opportunities

and underinvestment at …rms with low or negative cash ‡ows and good investment

opportunities. The Jensen and Myers-Majluf views both suggest that the latter will …nd

it hard to get the external …nance that they need. The Jensen view also suggests that

…rms with “free” cash ‡ow may be wasting their funds. The US oil companies in the

early eighties using revenues from known wells to fund unpro…table exploration would

seem to be a case in point.

However, in practice there exist mechanisms channeling funds from …rms with high

cash ‡ows and poor investment opportunities and underinvestment at …rms with low or

negative cash ‡ows and good investment opportunities. Jensen’s (1986) suggestion that

management is willing to waste free cash ‡ow rests on the presumption that management

is tied to a given line of business and is either unable or unwilling to exploit more

pro…table investment opportunities elsewhere. Why should this be so? Why shouldn’t

13For a critical perspective on this literature see, however, Kaplan and Zingales (1997).

19

the management of a …rm with free cash ‡ow be buying up new ventures and then

…nancing their new developments internally? After all, such practices are common place

in many industries.

For a mature economy with large corporations earning and retaining signi…cant pro…t

incomes, it may be useful think of current pro…ts as a source of funds in its own right.

The task of the …nancial system then is not so much to channel funds from households

to …rms, but to channel funds from the places where pro…ts are currently earned to the

places where investment …nance is needed. One way to achieve this is to have earnings

distributed to investors and then reinvested through capital markets. Another way is

to have earnings retained and then reallocated through internal capital markets inside

the corporation. If management power over retention decisions forecloses a mechanism

based on distributions of funds to investors and reinvestment through capital markets,

a mechanism based on conglomeration and investment …nance through internal markets

may be the only way to get funds from where they are generated to where they are

needed. In a corporate system based on management autonomy, conglomeration may

be unavoidable, at least if the economy is to have room for structural change.

How does a system in which structural change is …nanced through internal markets

inside a conglomerate corporation compare to a system in which structural change is …-

nanced through external markets? This brings us back to the question of what incentives

are driving the allocation of funds inside a …rm. Jensen’s notion of investment …nanced

by retained earnings seems to rest on the presumption that management is biased be-

cause the costs of unpro…table investments are borne by outside investors. However, this

presumption is unconvincing if one considers that management discretion over the use

of funds will prevail in the future as well as today. Management has no reason to neglect

the e¤ects of poor investment choices today on its own room for manoeuvre tomorrow

(Hellwig 2000, 2001).14 The assumption made in the preceding section that headquarters

is interested in some measure of expected gross pro…ts may therefore be not too far o¤

the mark, not because headquarters has the same interests as shareholders, but because

headquarters sees itself as a kind of residual claimant to the …rm’s returns. For e¢-

ciency considerations, the e¤ective expropriation of outside shareholders by incumbent

management may then be irrelevant if funding available inside the corporate system is

su¢cient and the allocation of these funds is e¢cient.14Or, as Scharfstein and Stein (2000, p. 2538) succinctly put it: “...although agency-prone CEOs may

want big empires, it also seems reasonable that, holding size …xed, they will want valuable empires”.

20

There are of course some caveats: First, returns to the company are not the same

as returns to oneself, a distinction that may be particularly relevant for a CEO shortly

before retirement who realizes that the returns to current investment strategies will ben-

e…t his successor rather than himself. Second, investment strategies may be shaped by

a desire to buttress management control even if this foregoes some pro…t opportunities.

Indeed, if we abandon the notion that headquarters itself can be treated like a single

person, we have to consider the possibility that investment strategies may be shaped by

desires to enhance individual positions in internal politicking. Here we link up with the

view mentioned in Section 2 that rent-seeking in large conglomerate …rms may bias the

capital allocation in favor of low-productivity divisions (e.g., Rajan, Servaes, and Zin-

gales 2000; Scharfstein and Stein 2000) or simply in favor of “old”, established divisions

that happen to wield the most in‡uence in the organization (Hellwig 2000, 2001).

6 Conclusion

Corporate …nance and investment theory is currently undergoing substantial changes,

away from the traditional paradigm where …rms consist of a single project and a single

owner/manager towards richer, more complex settings where …rms comprise multiple

projects ¡ typically operated by di¤erent managers ¡ and headquarters, an agent to

whom authority is delegated and whose interests may or may not be fully aligned with

those of the …rm’s owner(s). Given the possibly complex interaction between the dif-

ferent parties, theoretical research usually focuses on a subset of the “grand problem”,

i.e., either on the agency problem between investors and headquarters, or on the prob-

lem between headquarters and multiple project/division managers. Clearly, to gain a

more comprehensive understanding of the interplay between external and internal capi-

tal markets theoretical research eventually must …nd a framework where both layers of

agency problems can be addressed at the same time.15

The investor(s)-headquarters-multiple project managers setting is not the end of the

story. Corporate boards usually consist of multiple executives with diverse, possibly

con‡icting opinions and interests. Therefore treating headquarters as a homogeneous

entity is likely to be as misleading as assuming that …rms consists of a single project

with a single manager. Indeed a nonnegligible fraction of the investment ine¢ciencies

15E¤orts along these lines have been made by Scharfstein and Stein (2000), who consider both layers

of agency problems at the same time in the context of intra-…rm rent-seeking.

21

allegedly found in empirical studies may result from clashes of interest between di¤erent

coalitions of top executives. In its simplest form, headquarters may consist of a bunch of

former division managers having climbed the corporate ladder who still have their power

bases in their former divisions. In such a world, the rent-seeking of division managers is

carried right into corporate headquarters; ine¢cient patterns of capital allocation may

then result from power struggles and coalition formation within headquarters.16

Empirical work, while having inspired much of the recent theoretical research on

conglomeration and internal capital markets, faces challenges of its own. Most notably

there is the lingering doubt that studies …nding a conglomerate discount or ine¢cient

cross-subsidization of divisions with poor investment opportunities may su¤er from an

endogeneity problem and the associated sample selection bias. Better proxies for divi-

sional investment opportunities need to be found before these doubts can be dispelled.17

This problem is relevant not only for empirical, but also for theoretical research which

currently expends substantial time and energy trying to explain phenomena such as

the conglomerate discount or ine¢cient cross-subsidization while implicitly taking the

empirical correctness and validity of these phenomena for granted.

References

[1] Aghion, P., and J. Tirole: Formal and Real Authority in Organizations, Journal of

Political Economy 105, 1997, 1-29.

[2] Alchian, A.A.: Corporate Management and Property Rights, in H. Manne, ed.,

Economic Policy and the Regulation of Corporate Securities, Washington, D.C.:

American Enterprise Institute, 1969, 337-360.

[3] Bagwell, L.S., and J. Zechner: In‡uence Costs and Capital Structure, Journal of

Finance 48, 1993, 975-1008.

[4] Berger, P., and E. Ofek: Diversi…cation’s E¤ect on Firm Value, Journal of Financial

Economics 37, 1995, 39-65.

16This points to a bene…t of narrow business strategies, or disintegration, which is complementary to

that in Rotemberg and Saloner (1994, 1995).17An alternative route is taken by Lamont and Polk (2002), who try to avoid the endogeneity problem

by looking at exogenous industry shocks a¤ecting the degree of …rm diversi…cation.

22

[5] Berkovitch, E., R. Israel, and E. Tolkowsky: The Boundaries of the Firm: The

Choice between Stand-Alone and Integrated Firms, Working Paper, Tel Aviv Uni-

versity, 2000.

[6] Brusco, S., and F. Panunzi: Reallocation of Corporate Resources and Managerial

Incentives in Internal Capital Markets, CEPR Discussion Paper No. 2532, 2000.

[7] Campa, J.M., and S. Kedia: Explaining the Diversi…cation Discount, Working Pa-

per, Harvard University, 1999.

[8] Chang, C., and X. Yu: Investment Opportunities and the Informational Bene…ts

and Costs in Conglomerate Mergers, mimeo University of Minnesota, 2000.

[9] Chevalier, J.A.: What do We Know About Cross-Subsidization? Evidence From

the Investment Policies of Merging Firms, Working Paper, University of Chicago,

2000.

[10] Comment, R. and G. Jarrell: Corporate Focus and Stock Returns, Journal of Fi-

nancial Economics 37, 1995, 67-87.

[11] Corbett, J., and T. Jenkinson: The Financing of Industry, 1970-1989: An Interna-

tional Comparison, Journal of the Japanese and International Economies 10, 1996,

71-96.

[12] Corbett, J., and T. Jenkinson: How is Investment Financed? A Study of Ger-

many, Japan, The United Kingdom and the United States, The Manchester School

Supplement, 1997, 69-93.

[13] Diamond, D.W.: Financial Intermediation and Delegated Monitoring, Review of

Economic Studies 51, 1984, 393-414.

[14] Edwards, J. and K. Fischer: Banks, Finance, and Investment in Germany, Cam-

bridge University Press, Cambridge, 1994.

[15] Elston, J.: Investment, Liquidity Constraints and Bank Relationships: Evidence

from German Manufacturing Firms, in: Competition and Convergence in Financial

Markets: The German and Anglo-Saxon Models, ed. by S.W. Black (Advances in

Finance, Investment, and Banking, Vol. 5), Elsevier Science Publishers, 1998.

23

[16] Fazzari, S., G. Hubbard, and R. Petersen: Financing Constraints and Corporate

Borrowing, Brookings Papers on Economic Activity, 1988, 141-195.

[17] Fluck, S. and A.W. Lynch: Why Do Firms Merge and Then Divest? A Theory of

Financial Synergy, Journal of Business 72, 1999, 319-346.

[18] Gautier, A., and F. Heider: What do Internal Capital Markets Do? Incentives vs.

Redistribution, Working Paper, LSE, Financial Markets Group, 2000.

[19] Gertner, R.H., D.S. Scharfstein, and J.C. Stein: Internal versus External Capital

Markets, Quarterly Journal of Economics 109, 1994, 1211-1230.

[20] Graham, J.R., M.L. Lemmon, and J. Wolf: Does Corporate Diversi…cation Destroy

Value?, Journal of Finance 57, 2002.

[21] Grossman, S.J., and O.D. Hart: One Share/One Vote and the Market for Corporate

Control, Journal of Financial Economics 20, 1988, 175-202.

[22] Hackethal, A. and R.H. Schmidt: Financing Patterns: Measurement Concepts and

Empirical Results, University of Frankfurt, Working Paper Series: Finance and

Accounting, Discussion Paper No. 33, 1999.

[23] Hall, B., and D. Weinstein: The Myth of the Patient Japanese: Corporate Myopia

and Financial Distress in Japan and the United States, NBER Discussion Paper,

1996.

[24] Harris, M., and A. Raviv: The Capital Budgeting Process: Incentives and Informa-

tion, Journal of Finance 51, 1996, 1139-1173.

[25] Hart, O.: Financial Contracting, NBER Working Paper 8285, 2001.

[26] Hellwig, M.: On the Economics and Politics of Corporate Finance and Corporate

Control, in: X. Vives (ed.), Corporate Governance, Cambridge University Press,

Cambridge 2000, 95-134.

[27] Hellwig, M.: Corporate Governance and the Financing of Investment for Struc-

tural Change, in: Deutsche Bundesbank (ed.), Investing Today for the World of

Tomorrow: Studies on the Investment Process in Europe, Springer Verlag, Berlin,

Heidelberg, New York 2001, 203-223.

24

[28] Hirshleifer, D., and A.V. Thakor: Managerial Conservatism, Project Choice, and

Debt, Review of Financial Studies 5, 1992, 437-470.

[29] Hubbard, R.G., and D. Palia: A Re-Examination of the Conglomerate Merger Wave

in the 1960s: An Internal Capital Markets View, NBER Working Paper 6539, 1998.

[30] Hyland, D.C.: Why Firms Diversify: An Empirical Examination, Working Paper,

University of Texas at Arlington, 1999.

[31] Inderst, R., and C. Laux: Incentives in Internal Capital Markets: Capital Con-

straints, Competition, and Investment Opportunities, Working Paper, University

of Mannheim, 2001.

[32] Inderst, R., and H.M. Müller: Corporate Borrowing and Financing Constraints,

CEPR Discussion Paper No. 2784, 2001a.

[33] Inderst, R., and H.M. Müller: Ownership Concentration, Capital Structure, and

Risk-Taking in Internal Capital Markets, Working Paper, University of Mannheim,

2001b.

[34] Inderst, R., H.M. Müller, and K. Wärneryd: Distributional Con‡ict in Organiza-

tions, Working Paper, University of Mannheim, 2001.

[35] Jensen, M.C.: Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers,

American Economic Review 76, 1986, 323-329.

[36] Jensen, M.C.: Corporate Control and the Politics of Finance, Journal of Applied

Corporate Finance 4, 1991, 13-33.

[37] Jensen, M.C., and W.H. Meckling: Theory of the Firm: Managerial Behaviour,

Agency Costs and Capital Structure, Journal of Financial Economics 3, 1976, 305-

360.

[38] Jensen, M.C., and R.S. Ruback: The Market for Corporate Control: The Scienti…c

Evidence, Journal of Financial Economics 11, 1983, 5-50.

[39] Kaplan, S., and L. Zingales: Do Investment-Cash Flow Sensitivities Provide Useful

Measures of Financing Constraints, Quarterly Journal of Economics 112, 1997,

169-215.

25

[40] Klein, P.G.: Were the Acquisitive Conglomerates Ine¢cient?, mimeo University of

Georgia, 2001, forthcoming Rand Journal of Economics.

[41] La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny: Legal Determinants

of External Finance, Journal of Finance 52, 1997, 1131-1150.

[42] Lamont, O.: Cash Flow and Investment: Evidence From Internal Capital Markets,

Journal of Finance 52, 1997, 83-109.

[43] Lamont, O., and C. Polk: Does Diversi…cation Destroy Value? Evidence from

Industry Shocks, Journal of Financial Economics 63, 2002.

[44] Lang, L.H.P., and R.M. Stulz: Tobin’s Q, Corporate Diversi…cation, and Firm

Performance, Journal of Political Economy 102, 1994, 1248-1280.

[45] Laux, C.: Limited Liability and Incentive Contracting with Multiple Projects, Rand

Journal of Economics 32, 2001a, 514-526.

[46] Laux, C.: Project Speci…c External Financing and Headquarters Monitoring Incen-

tives, Journal of Law, Economics, and Organization 17, 2001b, 397-412.

[47] Laux, C.: Delegated Information Acquisition and Capital Budgeting: On the Sepa-

ration of Project Evaluation and Project Management, Journal of Institutional and

Theoretical Economics 157, 2001c.

[48] Lins, K, and H. Servaes: International Evidence on the Value of Corporate Diver-

si…cation, Journal of Finance 54, 1999, 2215-2239.

[49] Maksimovic, V., and G. Phillips: Do Conglomerate Firms Allocate Resources Inef-

…ciently?, Journal of Finance 57, 2002.

[50] Matsusaka, J.: Takeover Motives During the Conglomerate Merger Wave, Rand

Journal of Economics 24, 1993, 357-379.

[51] Matsusaka, J., and V. Nanda, Internal Capital Markets and Corporate Refocusing,

Working Paper, University of Southern California, 2000, forthcoming in Journal of

Financial Intermediation.

[52] Mayer, C.: New Issues in Corporate Finance, European Economic Review 32, 1988,

1167-1188.

26

[53] Mayer, C.: Financial Systems, Corporate Finance, and Economic Development, in

R.G. Hubbard, ed., Asymmetric Information, Corporate Finance, and Investment,

Chicago: Chicago University Press, 1990, 307-332.

[54] Meyer, M., P.R. Milgrom, and D.J. Roberts: Organizational Prospects, In‡uence

Costs, and Ownership Changes, Journal of Economics and Management Strategy

1, 1992, 9-35.

[55] Myers, S.C., and N.S. Majluf: Corporate Financing and Investment Decisions When

Firms Have Information that Investors Do Not Have, Journal of Financial Eco-

nomics 13, 1984, 187-221.

[56] Rajan, R., H. Servaes, and L. Zingales: The Cost of Diversity: The Diversi…cation

Discount and Ine¢cient Investment, Journal of Finance 55, 2000, 35-80.

[57] Roe, M.J.: Strong Managers, Weak Owners: The Political Roots of American

Corporate Finance, Princeton: Princeton University Press, 1994.

[58] Rotemberg, J.J., and G. Saloner: Bene…ts of Narrow Business Strategies, American

Economic Review 84, 1994, 1330-1349.

[59] Rotemberg, J.J., and G. Saloner: Overt Interfunctional Con‡ict (and Its Reduction

through Business Strategy), Rand Journal of Economics 26, 1995, 630-653.

[60] Scharfstein, D.S.: The Dark Side of Internal Capital Markets II: Evidence from

Diversi…ed Conglomerates, NBER Working Paper 6352, 1998.

[61] Scharfstein, D.S., and J.C. Stein: The Dark Side of Internal Capital Markets: Divi-

sional Rent-Seeking and Ine¢cient Investment, Journal of Finance 55, 2000, 2537-

2564.

[62] Schoar, A.S.: E¤ects of Corporate Diversi…cation on Productivity, Working Paper,

University of Chicago, 2000.

[63] Servaes, H.: The Value of Diversi…cation during the Conglomerate Merger Wave,

Journal of Finance 51, 1996, 1201-1225.

[64] Shin, H.-H., and R.M. Stulz: Are Internal Capital Markets E¢cient?, Quarterly

Journal of Economics 113, 1998, 531-552.

27

[65] Stein, J.C.: Internal Capital Markets and the Competition for Corporate Resources,

Journal of Finance 52, 1997, 111-133.

[66] Stein, J.C.: Information Production and Capital Allocation: Decentralized vs. Hi-

erarchical Firms, NBER Working Paper 7705, 2000, forthcoming in Journal of Fi-

nance.

[67] Thakor, A.V.: Investment “Myopia” and the Internal Organization of Capital Al-

location Decisions, Journal of Law, Economics, and Organization 6, 1990, 129-154.

[68] Whited, T.M.: Is It Ine¢cient Investment that Causes the Diversi…cation Dis-

count?, Journal of Finance 56, 2001, 1667-1691.

[69] Williamson, O.E.: Markets and Hierarchies: Analysis and Antitrust Implications,

Collier Macmillan Publishers, Inc., New York, 1975.

[70] Wenger, E.: Institutionelle De…zite am deutschen Kapitalmarkt, mimeo, Universität

Würzburg 1996.

[71] Wulf, J.: In‡uence and Ine¢ciency in the Internal Capital Market: Theory and

Evidence, Working Paper, University of Pennsylvania, 2000.

28