supervising across borders: the case of multinational ... · yue maggie zhou stephen m. ross ......

This article was downloaded by: [141.213.162.246] On: 16 February 2015, At: 07:37Publisher: Institute for Operations Research and the Management Sciences (INFORMS)INFORMS is located in Maryland, USA

Organization Science

Publication details, including instructions for authors and subscription information:http://pubsonline.informs.org

Supervising Across Borders: The Case of MultinationalHierarchiesYue Maggie Zhou

To cite this article:Yue Maggie Zhou (2015) Supervising Across Borders: The Case of Multinational Hierarchies. Organization Science26(1):277-292. http://dx.doi.org/10.1287/orsc.2014.0934

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial useor systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisherapproval, unless otherwise noted. For more information, contact [email protected].

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitnessfor a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, orinclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, orsupport of claims made of that product, publication, or service.

Copyright © 2015, INFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, managementscience, and analytics.For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org

http://pubsonline.informs.org

http://dx.doi.org/10.1287/orsc.2014.0934

http://pubsonline.informs.org/page/terms-and-conditions

http://www.informs.org

OrganizationScienceVol. 26, No. 1, January–February 2015, pp. 277–292ISSN 1047-7039 (print) � ISSN 1526-5455 (online) http://dx.doi.org/10.1287/orsc.2014.0934

© 2015 INFORMS

Supervising Across Borders:The Case of Multinational Hierarchies

Yue Maggie ZhouStephen M. Ross School of Business, University of Michigan, Ann Arbor, Michigan 48109, [email protected]

This paper examines how multinational corporations (MNCs) selectively assign supervisory responsibilities to units incountries with varying levels of institutional quality. Arbitraging across institutional contexts is an important function of

MNCs, but it also creates coordination challenges. The choice of organization structure, such as the differential assignmentof supervisory responsibilities, is an important tool for managing these coordination challenges. Using data on the businessactivities and supervision relationships within U.S. multinational manufacturers in 1996–2008, I find that frontline sub-sidiaries in countries with weaker institutions are more likely to be supervised by foreign rather than domestic supervisoryunits. Foreign supervision is even more likely when subsidiaries in weak-institution countries conduct activities that aremore central to or interdependent with their parents’ global operations. These findings confirm that MNCs use differentialsupervision to enhance global coordination. The paper highlights one of the most unique features of MNCs: a multinationalhierarchy that resides within a firm’s boundary but across national borders. It also connects MNCs’ hierarchical structurewith institutional imperfections that give rise to the emergence of the firm in the first place.

Keywords : organization structure; coordination; delegation; institutions; multinational corporations; global strategyHistory : Published online in Articles in Advance September 11, 2014.

IntroductionScholars in law, finance, development economics, andstrategy have long argued that firm growth is under-mined in countries with weak institutions, such as coun-tries with insufficient transparency, inefficient judicialsystems, and ambiguous property rights. This is largelybecause weak institutions increase domestic firms’ costsfor sourcing inputs, distributing products, and mobiliz-ing resources (North 1990). Interestingly, however, weaknational institutions do not impose an equally bind-ing constraint on multinational corporations (MNCs). Infact, MNCs often employ their unique organizationalform to arbitrage between varying institutional con-straints and reallocate resources across national borders,effectively putting “sovereignty at bay” (Kobrin 2001;Vernon 1971, p. 3). For example, MNCs can circum-vent trade barriers through foreign direct investment andinternal sourcing (Caves 1996) or leverage differencesin tax regimes by redistributing operations and profitsamong host countries (Desai et al. 2004). They can alsocompensate for underdeveloped local financial marketswith internal capital markets (Antràs et al. 2009) or mit-igate appropriability risks in countries with weak intel-lectual property rights protection (IPR) by sourcing fromtheir units in weak IPR countries innovations that areof greater value internally than to potential competitors(Zhao 2006). Although these arbitrage strategies providea basis for competitive advantage of MNCs vis-à-vistheir domestic counterparts, it also imposes coordinationchallenges. In this paper, I investigate the organization

structures that MNCs employ to manage these coordina-tion challenges.

I direct the lens at an important but understudiedstructural choice available to MNCs: differential super-vision, or employing local or foreign supervision fortheir overseas subsidiaries. I argue that differential super-vision facilitates coordination of subsidiaries in hostcountries with varying levels of institutional quality.Institutions affect firm coordination in two importantways (North 1990). First, institutions shape the availabil-ity of complementary information, which affects jointdecision making. Second, through the prevailing ruleof law and enforcement mechanisms, institutions affectthe clarity of property rights and, consequently, the riskthat assets valuable for joint tasks may be expropri-ated. In response, MNCs operating in weak-institutioncountries (WICs) can manage coordination challengesby specifying internal channels of information flowand reallocating decision rights through an organiza-tional hierarchy. Reducing the supervisory responsibil-ities of units in WICs and increasing the supervisoryresponsibilities of units in strong-institution countriesallows an MNC to (1) integrate scarce local informationwith complementary regional and international data and(2) reduce expropriation risks by limiting the exposureof corporate resources to a host country’s weak propertyrights.

This is illustrated in Figure 1. This U.S. MNC oper-ates in two foreign countries, A and B. If both coun-tries have strong institutions, then each subsidiary in A

277

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Zhou: Supervising Across Borders: The Case of Multinational Hierarchies278 Organization Science 26(1), pp. 277–292, © 2015 INFORMS

Figure 1 MNC Structures Under Different Institutional Environments

Country BSUBCountry A

SUA

United Statesheadquarters

U1 U2 U3 U4 U5 U6

and B is expected to report to supervisory units in itsown country (SUA or SUB) to ensure fit with the localenvironment (represented by the solid lines of command)or to report to corporate headquarters or regional super-visory units, with A and B having equal probability tohost regional supervisory units. However, if country Ais a weak-institution country and country B is a strong-institution country, the MNC may “reallocate” supervi-sory responsibilities for some subsidiaries in country A(e.g., U3) to supervisory units in country B (representedby the dotted line of command). Moving supervision ofU3 away from country A allows the MNC to engageU3 in global operations while mitigating its exposure tocountry A’s institutional risks, giving the MNC a uniquecompetitive advantage relative to domestic firms operat-ing in any single country.

In addition to institutional differences at the coun-try level, this paper explores differences in the impactof institutions at the firm level. In particular, becausecoordination is more important when activities are inter-dependent, this paper examines the extent to which anMNC unit is engaged in tasks that are interdependentwith its parent firm’s global activities. Greater task inter-dependence generates greater demand for coordination,which in turn magnifies the impact of institutions on theallocation of supervisory responsibilities.

I tested my hypotheses using data on the businessactivities and organization structures of U.S. multina-tional manufacturers from 1996 to 2008. My resultsshow that MNC frontline subsidiaries in host countrieswith weaker institutions are more likely to be supervisedby a foreign (as opposed to domestic) supervisory unit,especially when the subsidiaries’ tasks are more cen-tral to their MNC parents’ global operations or whentheir tasks are interdependent with a greater number oftasks performed by other subsidiaries.1 This suggeststhat selectively allocating supervisory responsibilities isan important managerial lever for MNCs in coordinatingglobal operations.

This paper relates to the literature on the trade-offbetween adaptation and coordination within multidivi-sional firms. In particular, it relates to studies aboutheadquarters–subsidiary relationships within MNCs(Birkinshaw et al. 2006). Whereas most previous studiesfocused on the allocation of decision rights between theheadquarters and a subsidiary as an independent dyad,this paper examines whether supervisory responsibilitiesfor a subsidiary are assigned to a local or foreign super-visory unit. It identifies the mechanism of differentialsupervision—that is, local supervision being granted tosome subsidiaries but not to others—for selective inter-vention/coordination within MNCs. Local supervision ofsubsidiaries implies local autonomy for the local super-visory unit and subsidiaries as a group. In contrast, for-eign supervision of frontline subsidiaries implies lesslocal autonomy. By shifting the focus of analysis to theallocation of supervisory responsibilities, I hope to high-light one of the most unique features of MNCs: a multi-national hierarchy that resides within a firm’s boundarybut across national borders.

This paper also relates to the literature on MNCs’institution strategy, a topic that prior MNC head-quarters–subsidiary relationship studies have not suf-ficiently explored. Those studies more often focusedon product market conditions and held the institutionalenvironment constant, whereas in reality, MNC sub-sidiaries’ institutional environments are no less criticalthan their product markets. The MNC institution strat-egy literature explicitly examines MNCs’ global strategyin dealing with weak institutions, albeit mainly throughthe choice of subsidiary location or ownership type.By extending the analysis to MNCs’ use of differen-tial supervision, this paper connects MNCs’ hierarchi-cal structure with institutional imperfections or voids(Khanna and Palepu 2000) that give rise to the emer-gence of the firm in the first place. It also comple-ments an emerging body of work that examines howMNCs circumvent institutional obstacles when their

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Zhou: Supervising Across Borders: The Case of Multinational HierarchiesOrganization Science 26(1), pp. 277–292, © 2015 INFORMS 279

location and ownership choices are limited (Alcácer2006, Zhao 2006).

In addition to the academic literature, the studyalso has implications for public policy and interna-tional business. One of the most significant featuresof MNCs is the extent to which they move resourcesacross national borders (Dunning 2001). According toKobrin (2001), because of these movements, MNCsare viewed as a compensating instrument for intrin-sic cross-border market failures and are uniquely posi-tioned to take advantage of the asymmetry between anincreasingly integrated global economic system and astill segmented political system. The MNCs’ ability tooperate worldwide systems against the limited reachof any national authority “creates asymmetries of bothinformation and jurisdiction” (Kobrin 2001, p. 187).The increasing interdependencies among MNC activitiesaround the globe weaken national governments’ controlover their national economic actors and economic pol-icy. MNCs have emerged as a source of private authorityand gained increasing decision-making power vis-à-visnational states. This study sheds light on one impor-tant mechanism for MNCs to adapt their organization inresponse to national institutions: the allocation of super-vision responsibilities across national borders.

Related LiteratureHeadquarters–Subsidiary RelationshipWithin MNCsMNC headquarters and their overseas subsidiaries bothface trade-offs in managing their relationships. Like inall multidivisional firms, MNC headquarters need to bal-ance the opposing demands for adaptation and coordina-tion (Lawrence and Lorsch 1967). In fact, this trade-offis particularly salient for MNCs. On the one hand,product demand and production conditions are usuallymore heterogeneous across countries than within coun-tries, requiring a greater level of adaptation (Bartlettand Ghoshal 1989). On the other hand, to realizetheir competitive advantage of global arbitrage, MNCsneed to coordinate their subsidiaries across a num-ber of countries (Kogut 1983). This trade-off createsa dilemma for MNC headquarters when it comes tothe design of headquarters–subsidiary relationships: themore autonomous a firm’s subsidiaries are, the betterthey can adapt to idiosyncratic local conditions, but theless they can be coordinated around broad corporateobjectives.

MNC subsidiaries also face trade-offs. On the onehand, they need to adapt to the local business envi-ronment to remain locally competitive (Rugman andVerbeke 2001). On the other hand, they need accessto unique resources possessed by MNC headquarters,such as knowledge, management skills, global productreputation, production technology, and financial capital

(Bartlett and Ghoshal 1989). Autonomy enables localadaptation and optimization of local decision making,but an overly local focus might hinder subsidiaries’ abil-ity to access corporate knowledge and resources.

The literature has studied in great detail the benefitsand costs of subsidiary autonomy. Autonomy is usefulfor firms operating in diverse product markets (Bartlettand Ghoshal 1989). It allows subsidiaries to betteradapt to their immediate task environment (Lawrenceand Lorsch 1967). It provides flexibility and creativ-ity in local problem solving (Eisenmann and Bower2000). It also saves time by localizing communica-tion and information processing (Radner 1993). At thesame time, an overly local focus hinders coordination.Locally autonomous teams may become isolated (Hass2010, Monteiro et al. 2008). The benefits of local auton-omy also vary by contingencies. For example, a moreautonomous structure enhances a firm’s ability to adaptto its environment, but only if the environment is sim-ple and the interdependencies between organization unitsare low. If the level of interdependencies is high, greaterautonomy for each unit may jeopardize efforts in otherunits, hamper coordination, and hurt performance at thefirm level (Rivkin and Siggelkow 2003).

Whereas the literature has proposed many coor-dination mechanisms to manage MNC headquarters–subsidiary relationships (Martinez and Jarillo 1989), oneparticular mechanism has been understudied: the inter-mediary units in MNCs’ formal hierarchical structure.This alone is unsatisfactory because (1) individuals (suchas chief executive officers, or CEOs) and organiza-tion units (such as the general office at headquarters)face limits in their cognitive capacity (Cyert and March1963), (2) the diversity and complexity of global busi-nesses taxes the attention of the top management teams(Bouquet and Birkinshaw 2008), and (3) multilayer hier-archical structures with intermediary units are a preva-lent phenomenon within real MNCs. These intermediaryunits connect corporate headquarters with subsidiaries atthe front line of businesses; they reduce span of con-trol for the corporate headquarters and provide greaterattention to resource allocation within a subgroup of sub-sidiaries (Birkinshaw et al. 2006).

To study the role of intermediary supervisory units,I draw insights from the organization theories. The lit-erature suggests that intermediary units assist in infor-mation processing and communication (Tushman andNadler 1978). They solve problems emerging from mul-tiple subordinate units (Eisenmann and Bower 2000),make joint decisions for multiple subordinate units(Marschak and Radner 1972), and exercise authorityover assets useful for joint tasks—tasks to be jointly car-ried out by multiple subordinate units (Hart and Moore2005). They set priorities when subordinates have differ-ent opinions (Hart and Moore 2005) and resolve conflict-ing expectations (Simon 1991). Together, these functionsalleviate the coordination burden on top management.

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Unfortunately, prior studies on MNCs have largelyfocused on the MNC headquarters–subsidiary dyads andoverlooked the intermediary supervisory units. They donot examine which unit (foreign or local) immediatelyabove the frontline subsidiary in the corporate hierar-chy has been granted the supervisory responsibility forthe subsidiary. Therefore, the literature missed the mostunique dimension of MNC structure: the allocation offormal supervisory responsibilities across national bor-ders, where institutional environments differ.

Dealing with Weak InstitutionsAlthough the MNC literature has a long tradition ofstudying integration and subsidiary autonomy based ona host country’s environment, much of this prior workhas focused on cross-country differences in product mar-ket, knowledge endowment, and technology capabili-ties (Ambos et al. 2010, Rugman and Verbeke 2001),with the subsidiaries’ institutional environments keptconstant. This treatment is at odds with the reality ofinternational business, in which MNCs with global net-works of economic activities are constantly expandinginto countries with institutions that are vastly differentfrom those in their home countries. How could MNCsleverage their organization structure to manage the het-erogeneous institutional environments facing their over-seas subsidiaries is therefore an urgent question for theorganization and strategy scholars.

Institutions are humanly devised constraints thatstructure human interactions (North 1990). They includeformal rules, laws, and constitutions, as well as infor-mal norms of behavior and conventions. This studyfocuses on formal institutions. Strong institutions helpfirms grow. For example, law and order promotes thedevelopment of local financial markets, which in turnsupply capital for firms’ investments (Demirgüç-Kuntand Maksimovic 1998, Rajan and Zingales 1998). Effi-cient judicial systems increase the incentive for innova-tion: firms in research and development (R&D)-intensiveindustries are larger in countries with better patent pro-tection (Kumar et al. 2001).

The literature on MNC strategies for combatting weakinstitutions may be usefully partitioned into two relatedthemes: location and ownership. Accordingly, it has beensuggested that MNCs can either (1) select locations withthe greatest market or production opportunities and theleast institutional constraints on creating value (Changand Park 2005) or (2) pick ownership types (e.g., whollyowned subsidiaries versus joint ventures or alliances)that help safeguard value (Anderson and Gatignon 1986,Oxley 1999). Both strategies have limitations, however.Location choices are limited primarily because arbitrageopportunities and institutional constraints often accom-pany each other: if MNCs only go to locations whereinstitutions are strong, then their comparative advantage

over domestic firms trading across borders will be sig-nificantly dampened. Ownership choices are similarlyinsufficient. MNCs’ ownership choices may be affectednot only by governance considerations but also by reg-ulatory or normative pressures in their host countries(Svejnar and Smith 1984, Yiu and Makino 2002). There-fore, we need to broaden the search to include otherdesign choices such as hierarchical structure.

In sum, the MNC headquarters–subsidiary literatureand the MNC institution strategy literature offer insightsas well as opportunities. The opportunities arise fromthe as-yet missing link between the varied institutionalenvironments MNCs operate in and their deployment ofhierarchical structure across those environments.

Hypothesis DevelopmentEven though the link between institutional quality andMNCs’ hierarchical structure has yet to be formallyestablished in the literature, anecdotal evidence of sucha link can be found in a number of classic books onMNCs (Galbraith 2000, Ghemawat 2007, Gupta et al.2008, Hill 2012).

According to these books, MNCs organize their sub-sidiaries along a number of dimensions. Some MNCsorganize by global functional areas. For example, IBMlocated its global procurement center in China, globalservice delivery center in India, and global internal Webdesign centers in Brazil and Ireland. Hyundai establishedR&D centers in Germany, Japan, and the United Statesto supervise R&D subsidiaries in Europe, Asia, andNorth America, respectively. Some MNCs reorganize byglobal product divisions. For example, Eaton Corpora-tion located its global center for light and medium trucktransmissions in Amsterdam, the Netherlands (oversee-ing subsidiaries in Argentina, Brazil, India, and Mexico)and its global center for automotive control business inStrasbourg, France. MNCs can also organize by cus-tomer profiles or technologies. Based on each sub-sidiary’s function, business segment, customer portfolio,technology, or geographic location, it can report to dif-ferent supervisory units.

MNCs allocate supervisory and coordination respon-sibilities across host countries mostly based on the coun-tries’ location-specific advantages. For example, whenProcter & Gamble (P&G) was selecting a place to locateits headquarters for Global Business Services, it pickedCosta Rica over other low-cost places such as Mexico(which ranks lower than Costa Rica in terms of insti-tutional quality). P&G explained that the selection wasmade based on Costa Rica’s political stability, businessclimate, and telecommunications infrastructure (Luxner2001). Microsoft is yet another example. Both Chinaand India are among the fastest-growing markets forMicrosoft, and its research centers in the two countrieshave become major powerhouses for research programs

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


such as language and speech technologies, which areinvaluable for Microsoft to localize its products. How-ever, neither the Chinese nor the Indian research centerreports to Microsoft’s country headquarters in China orIndia; they both report to Microsoft Research, based inRedmond, Washington, mostly as a result of concernsover high piracy and appropriation risks in these coun-tries (Khanna and Choudhury 2007). These anecdotalexamples motivate my theoretical development in thissection.

Cross-Border Supervision inWeak-Institution CountriesOver the last two decades, MNCs have increasinglydispersed their production networks to take advantageof locational advantages and global production scale,causing “the increasing interconnectedness of productionprocesses in a vertical trading chain that stretches acrossmany countries, with each country specializing in partic-ular stages of a good’s production sequence” (Hummelset al. 2001, p. 76). Such a global value chain allowsMNCs to exploit differences between their local units’business environments for synergies at the corporatelevel (Ghemawat 2007). It also requires effective coordi-nation across national borders. One way to coordinate isto selectively grant autonomy to local units. Followingprior studies (e.g., Rivkin and Siggelkow 2003), I con-ceptualize autonomy as the degree of freedom enjoyedby managers in making decisions and as the amount ofreporting they need to present to, and approval they needto seek from, parties above them in the corporate hier-archy. Also following prior studies (Ambos et al. 2010,Bouquet and Birkinshaw 2008), I assume that super-vision decisions are partially reflected in the structuralposition of the units and that a lower level of direct mon-itoring is associated with a higher level of subsidiaryautonomy.

Coordination means managing interdependent tasksacross business units (Malone and Crowston 2001,Puranam et al. 2012). It entails making joint decisionsand synchronizing joint actions for units undertakingthese interdependent tasks. Both joint decision makingand joint action synchronization are subject to institu-tional influences.

Joint decision making requires the gathering, inter-preting, and synthesis of information (Tushman andNadler 1978), and institutions affect the quality of jointdecision making by influencing the availability of infor-mation (Arrow 1959, North 1990). Both governmentand corporate information affect business decision mak-ing (Gelos and Wei 2005). First, national and regionaleconomic indicators published by governments on con-sumption, production, capacity utilization, and inflationhelp firms smooth their production cycles across mul-tiple plants and synchronize procurement, production,and delivery. Without these complementary data, local

information—about a locally contained demand or sup-ply shock, for example—becomes less valuable. Second,without effective regulations for financial and accountingdisclosure, fair competition, and IPR protection, firmswill share less information with their investors, cus-tomers, and industry peers; they will instead practicemore trade secrecy in their host country. This furtherreduces the amount of business information availablein the local environment. Finally, in addition to domes-tic government and corporate information, institutionsmay regulate the availability of international informationthrough censorship.

The lack of reliable information in WICs hurts localbusinesses. For example, when a number of transitionaleconomies first opened their borders for foreign invest-ment, there were few regulations that promoted infor-mation disclosure. Market intelligence was difficult tocollect. It has been shown that a severe lack of informa-tion (or the perception of it) encourages economic agentsto herd in their behavior based on observed patterns ofothers rather than on fundamentals (Bikhchandani et al.1992, Garcia-Pont and Nohria 2002). As a result, domes-tic firms gambled with perceived business opportunitiesand followed each other’s past successful moves, ofteninto overly crowded markets with thin profit margins.

In response to the lack of local information, bothdomestic firms and MNCs can increase their localinformation-seeking efforts. For MNCs, they can hirelocal managers or station their foreign managers locallyso that these managers are on the ground and embeddedin the local context to develop the ability to under-stand the nuances of local politics and market informa-tion. However, if MNCs merely increase their effort inseeking local information, as their domestic competi-tors do, they will not gain much competitive advantage.This is because MNCs’ competitive advantage comesfrom global or regional coordination to move resourcesacross national borders (Dunning 2001), which needsmore than isolated local information. Cross-border coor-dination also needs complementary information, whichis often not available in WICs but available in strong-institution countries where policies are more transparent,rules and norms for disclosure are stronger, and censor-ship is less prevalent. MNCs have to strike a balancebetween seeking both local and complementary infor-mation. One solution is to have local subsidiaries inweak-institution countries specialize in collecting localinformation but add a layer of supervisory units instrong-institution countries to synthesize local infor-mation in weak-institution countries with complemen-tary information collected in strong-institution countries.Firms with units in both weak-institution and strong-institution countries will have more tools for informationsynthesizing and coordination. MNCs enjoy this uniqueorganizational advantage.

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


In short, MNCs can partially solve the problem ofinformation scarcity in a subset of their external envi-ronment by changing internal channels of information.Local supervision should decrease as the reliabilityof local information—relative to the information avail-able elsewhere in the organization—decreases (Harrisand Raviv 2005). For example, MNCs sometimes setup strong local presence or hire consultants in WICsto collect local information. However, they almostalways combine local information with additional local,regional, or global data in regional hubs such asHong Kong and Singapore, where better institutionssupport knowledge and information sharing. Armedwith a superior quality and quantity of information,MNC supervisory units in these hubs coordinate MNCs’regional activities, including reporting, competitor intel-ligence, and strategy formation (Enright 2000). Thecolocation of resource allocation and strategic decisionmaking with information gathering and sharing makescoordination more efficient (Benito et al. 2011).

Whereas making joint decisions depends on sufficientinformation, synchronizing joint actions between organi-zation units requires that each unit have control over itsresources to take the necessary actions. As today’s globalcompanies build up increasingly tightly yet broadly con-nected global production networks, they rely more andmore on the control of their core resources to fulfillquality and speedy delivery on a global scale. MNCsubsidiaries carrying out adjacent productive processesalong a global value chain increasingly rely on eachother’s input to deliver their own output. Institutionsaffect the coordination of joint actions by influencingthe clarity of property rights, which provide the ulti-mate right of control over assets (Alchian and Demsetz1972). By protecting property rights against expropri-ation and corruption, institutions help to align owner-ship and control rights, thereby facilitating coordination(North 1990).

Guarding against expropriation risk is not just aboutprotecting the underlying resources. More importantly,it is about protecting the integrity of the decision rightswith respect to the use of these assets (e.g., decisionsabout applying resources to new geographic markets orabout launching new products). Even though local super-vision gives local units flexibility to allocate resourcesand better adapt to local environments, it also exposeslocal managers to influence from the local environment.Therefore, units that operate in countries with weakproperty right protection and high levels of corruptionare at greater risk of losing their operations than unitsin strong institution countries. Losing operations in onecountry may negatively affect subsequent operations ofMNC units in other countries. In addition, the lack ofpredictability and reliability inherent to assets in WICsmakes it difficult to coordinate joint productive activities.

Each MNC subsidiary has a double personality: it isboth a local corporation and a unit in a multinational

network under the control of its MNC parent (Kobrin2001). When local units are delegated supervisoryresponsibilities, they are expected to conform more withhost country institutions than with an MNC’s inter-nal anticorruption practices. They will be subject togreater pressure from corrupt local entities. Corrupt enti-ties in the host country will demand more bribery fromautonomous local units than from units under foreignsupervision (Spencer and Gomez 2011). Therefore, reas-signing supervisory responsibilities to a unit in a foreigncountry will limit an MNC’s exposure to expropriationrisks in the subsidiary’s host country. In addition, as cen-tralization demonstrates to internal and external stake-holders that headquarters’ policies are being enforced(Gates and Egelhoff 1986), nonlocal supervision sig-nals the MNC’s overall commitment to global objectivesand reinforces its anticorruption reputation, discouragingbribery requests from corrupt entities in the host country.

In sum, operating simultaneously across countrieswith a variety of institutional quality allows MNCsto strategically allocate supervisory responsibilities. Tocoordinate with activities in WICs, MNCs are morelikely to employ foreign supervision.

Hypothesis 1 (H1). MNC subsidiaries in host coun-tries with weaker institutions are more likely to be underforeign supervision.

Task Interdependence andCross-Border SupervisionTask interdependence in this study refers to interme-diate inputs being supplied from one unit to anotheror passed back and forth in successive stages of pro-duction (Thompson 1967). It exists in the networkof value-chain activities interrelated through physicalinput–output feedback loops that transfer and transforminformation and materials (Baldwin 2008, Porter 1985,Sturgeon 2002). MNCs’ global productive network canbe viewed as a system of interdependent tasks. Forexample, Toyota’s global pickup truck production pro-cess will collect common engines and manual transmis-sions from Asian plants to assembly bases in Asia, LatinAmerica, and Africa and distribute to almost all majormarkets around the world (Ghemawat 2007). As MNCskeep segmenting their value chains and dispersing eachof their value chain activities to optimal location, theirsubsidiaries are becoming more and more diverse andinterdependent at the same time, demanding “collabora-tive information sharing and problem solving, coopera-tive support and resource sharing, and collective actionand implementation” (Bartlett and Ghoshal 1987, p. 47).Among all the interdependent subsidiaries, those withtasks more central to MNCs’ global operations, or thosewith tasks interdependent with a greater number of tasksperformed by other subsidiaries, require more multilat-eral coordination.

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Task interdependence makes autonomy in decisionmaking less effective. When decision variables arehighly interdependent, autonomy implies that individ-ual decisions will be made based on partial informa-tion and will not be globally optimal (Marschak andRadner 1972). Similar arguments can be made of auton-omy causing joint action to be less effective. Accord-ing to Galbraith and Lawler (1993), the more imperativethe need for lateral coordination between subunits, thegreater the need for hierarchical intervention that dealswith increasingly aggregated levels of the organizationsystem. For example, if the subsidiaries are highly inter-dependent because they share resources, technology, orcustomers, it is important for their common supervisorsto forge business directions (regarding common tech-nology and product strategy) so that the subsidiaries’operations do not conflict with each other. In addition,for issues that cannot be resolved laterally between sub-sidiaries, a common boss helps to speed up or finalizejoint decisions and allows uncertainties to be resolvedmore quickly.

The coordination challenges in WICs as a result ofinformation scarcity will be more detrimental for anMNC if its WIC subsidiaries also perform more cen-tral tasks. A lack of complementary information makesit harder for supervisory units in WICs to aggregate,benchmark, evaluate, and synthesize discrete local infor-mation so as to make decisions. Decision errors for acentral subsidiary will have a greater adverse effect onother subsidiaries than decision errors for a peripheralsubsidiary. Even if these other subsidiaries are outsideWICs, they will find it harder to predict the decisionsmade in WICs and schedule their own actions accord-ingly. Therefore, assigning supervisory responsibilitiesfor central subsidiaries to a country outside the WICwhere information is more abundant will help managecoordination challenges for the MNC.

Similarly, the coordination challenges in WICs as aresult of property rights ambiguity will be more harm-ful to an MNC if its WIC subsidiaries also performmore central tasks. A lack of clear property rights pro-tection makes it harder for supervisory units in WICs tofend off expropriation and maintain full control of MNCassets and resources in the host countries. Protecting theassets and resources of a central subsidiary is more crit-ical for other subsidiaries than protecting the assets andresources of a peripheral subsidiary. Therefore, assigningsupervisory responsibilities for central subsidiaries to acountry outside the WIC where expropriation risk is lowwill help manage coordination challenges for the MNC.

In sum, a central subsidiary’s failure to perform atask can have a cascading effect on all dependent sub-sidiaries. Therefore, when a subsidiary’s task is moretightly integrated into its MNC parent’s global net-work, less local autonomy should be granted. When sub-sidiaries located in WICs are also responsible for tasks

that are more central to MNCs’ global operations, theMNC parent should be even more likely to employ for-eign supervision for them.

Hypothesis 2 (H2). MNC subsidiaries in countrieswith weak institutions are more likely to be under for-eign supervision when their tasks are more central totheir parent’s global operations.

Empirical Research DesignThe hypotheses were tested based on the geographiclocation and business activities of U.S. manufacturingMNCs, and the supervision relationships within them,in 1996–2008. The level of analysis is MNC subsidiary-year: I estimate the probability that an MNC subsidiarywill be supervised by a supervisory unit in the samehost country rather than a supervisory unit in a differentcountry. This empirical setting is suitable because, first,manufacturing often entails multiple stages of produc-tion and requires large quantities of intermediate inputs.This provides large variation in business activities acrossfirms and their subsidiaries in the same primary indus-try. Second, firms in the manufacturing sector face fierceglobal competition and intense pressure to outsource andrestructure, making their decisions about firm scope andstructure critical to firm growth. For example, in the auto-motive industry, Toyota beat General Motors (GM) infirst-quarter global car sales in 2007, ending more than 75years of GM dominance (Chozick and Shirouzu 2007).Sliding market shares and profits put U.S. automakersunder tremendous pressure to restructure their overlycumbersome production systems and relocate more com-ponent production and processes to overseas affiliates.One of Ford’s key restructuring initiatives was to sell 17U.S.-based component plants and 6 component facilities(Ford Motor Company 2007). At Renault–Nissan, worse-than-expected earnings reports raised so much skepti-cism about CEO Carlos Ghosn’s ability to manage thecompany’s complex global businesses that Ghosn wasforced to turn over responsibilities for North Ameri-can markets to another executive and establish multi-ple regional offices (Morse and Shirouzu 2007). Finally,by focusing on U.S. MNCs only, I controlled for het-erogeneity in home-country institutions that could eitheraffect MNCs’ political capability or strategy (Holburnand Zelner 2010).

Data and SampleTesting the hypotheses requires data on (1) MNC sub-sidiaries’ business activities and their interdependencewith other units, (2) each subsidiary’s geographic loca-tion and the quality of institutions at those locations,and (3) the assignment of supervisory responsibilitiesfor each subsidiary. Information about MNC subsidiarieswas drawn from the Directory of Corporate Affilia-tions (DCA), provided by LexisNexis. For firms with

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


more than 300 employees and $10 million in rev-enue, DCA describes reporting/supervision relationshipsbetween their units (groups, departments, divisions, sub-sidiaries, etc.) to the seventh level of corporate link-age (LexisNexis 2005). Based on my conversation withDCA analysts, “reporting” includes mainly hierarchicalauthority in supervision. This is because a major usageof the database is for potential suppliers, customers,acquirers, investors, and other business partners to easilyidentify which supervisory unit makes the relative deci-sions for a subordinate unit in a business segment and/orgeographic area. LexisNexis collects information fromthe companies, annual reports, and business publicationsin the LexisNexis database; it also contacts each com-pany to verify the information. Its analysts extensivelyedit and validate the content to prevent errors beforedatabase entry. DCA also reports the segments (four-digit Standard Industrial Classification codes, or SICs)of each subsidiary and supplies detailed street addressesfor most of them. (A small number of missing streetaddresses were added by searching company websites.)The DCA data set for publicly traded U.S. firms in1996–2008 contains 1,902 MNC parent companies withprimary industries in the manufacturing sector.

Here is an example from the DCA data set. FordMotor Company’s main industry is SIC 3711 (“motorvehicles and passenger car bodies”). In addition, it hassubsidiaries operating in SIC 6141 (“personal creditinstitutions”). Among those operating in SIC 6141, theBelgian and Spanish subsidiaries report locally, whereasthe Italian and Brazilian subsidiaries report to divisionsin the United States. This is consistent with the fact thatBelgium and Spain rank higher than Italy and Brazilin terms of institutional quality. In addition, Ford hassubsidiaries operating in SIC 3724 (“aircraft enginesand engine parts”). In contrast to the Spanish subsidiaryoperating in SIC 6141, the Spanish subsidiary operat-ing in SIC 3724 reports not locally but to supervisoryunits in Sweden. This is consistent with my task cen-trality arguments. Engine manufacturing is more centralto Ford’s global operations than the personal credit cardbusiness; therefore, we observe that the Spanish sub-sidiary in the engine business reports to Sweden, a coun-try with higher institutional quality than Spain, but theSpanish subsidiary in the personal credit card businessreports locally.

Financial information about MNC parents wasextracted from Compustat. The data sets were matchedby company names. Ambiguous matches were furtherverified using company websites. A total of 1,602 (84%)MNC parent companies were matched.

Macroeconomic and institutions data about each hostcountry were collected from the World Bank and othermultinational organizations. Among the 125 countrieshosting MNC subsidiaries in the sample period, datawere available for 111. One hundred small MNCs had no

operations in these 111 countries and were dropped fromthe sample. To control for the possibility that some unitswere established to facilitate tax evasion, I excludedunits located in tax havens and units that had super-vising units in tax havens.2 Because my measure oftask centrality relies on intersegment relationships, I alsodropped 94 MNCs that operated in one segment only.Finally, I dropped observations with missing values.

I focused on organization units that represent the low-est level of profit-center responsibility and, therefore,can be compared across firms. Consequently, I includedonly frontline subsidiaries (subsidiaries that have nosubordinate subsidiaries). For each subsidiary I identi-fied a supervisory unit based on the corporate hierarchyreported in DCA. In all, my final sample includes 1,332MNCs with 14,886 subsidiaries in 96 foreign countries,for a total of 70,901 subsidiary-year observations.

Unfortunately, existing measures of institutional qual-ity are mostly comprehensive and do not break downinto information availability or property rights. There-fore, as a robustness check, I collected regional mea-sures of transparency and the rule of law from the WorldValue Surveys (WVSs) in order to relate foreign super-vision more directly to information and property rights.WVS is an ongoing, cross-country project coordinatedby the Institute of Social Research at the University ofMichigan. The WVS samples from populations repre-senting more than 88% of the world total to assess thesocial, moral, religious, and political values of differ-ent cultures across regions within each surveyed country.Therefore, the regional survey also allows me to con-trol for unobserved heterogeneity across countries. Thesurvey is carried out in three- to five-year cycles. Themost recent cycles covered 1994–1998, 1999–2004, and2005–2008. I manually matched each subsidiary loca-tion with regions covered by the WVS. I was able tomatch locations for 7,595 (49%) frontline subsidiariesfrom 1,087 (81%) MNC parents to WVS regions, result-ing in a total of 20,561 subsidiary-year observations.These matched WVS regions are in 45 non-tax-havenand non-island countries.

VariablesMy dependent variable is local supervision. I estimatedthe probability that a subsidiary shares a host countrywith its supervisory unit, indicating more local super-vision. The dummy variable, LocalSupervision, is 1 ifa subsidiary is supervised by a supervisory unit in thesame host country as opposed to a supervisory unit in adifferent country.

The main independent variable, QualityofInstitutions,is the average value of the Worldwide Governance Indi-cators (WGIs) developed by the World Bank Group(Kaufmann et al. 2010). WGIs are reported annuallyalong six dimensions of governance: voice and account-ability, political stability and absence of violence, gov-ernment effectiveness, regulatory quality, rule of law,

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


and control of corruption. Because these indicators arehighly correlated (� = 0068–0.95), I used their averagevalue to measure a country’s overall quality of institu-tions in a given year. As would be expected, there ishigh correlation between current and lagged estimatesof a country’s governance. Nevertheless, many countriesshow significant governance changes over time. WGIauthors report that, between 2000 and 2009, 18 countriesexperienced changes significant at the 75% confidenceinterval in each of the six indicators, and 54 countriesexperienced a significant change in at least one of thesix indicators (Kaufmann et al. 2010). Table 1 presentsthe list of 81 foreign countries hosting my sample MNCunits during 2008, ranked by average WGIs.

For estimations based on WVSs at the regional level,I measured the quality of institutions based on localresidents’ confidence in the rule of law (the policeand the justice system) and information transparency(the press). Respondents chose from one of the fouranswers: a great deal of confidence, quite a lot of con-fidence, not very much confidence, or no confidence atall. Following prior studies that use the WVS to measureinstitutions (Bloom et al. 2012), I used the percentageof respondents who reported “a great deal” or “quite alot” of confidence.

TaskCentrality reflects the centrality of the focal sub-sidiary’s task in the MNC’s task system. For eachsubsidiary, I calculated the number of same-MNC sub-sidiaries (in or outside the United States) whose primary

Table 1 Countries Hosting Sample MNCs Ranked by World Governance Indicators (2008)

Rank Country WGI Rank Country WGI Rank Country WGI

1 Denmark 10796 28 Latvia 00651 55 Zambia −002952 Finland 10764 29 Poland 00629 56 Thailand −002983 Sweden 10745 30 Israel 00592 57 Peru −002984 New Zealand 10715 31 Greece 00570 58 Bosnia and Herzegovina −003295 The Netherlands 10674 32 Costa Rica 00556 59 Guyana −003816 Norway 10666 33 Italy 00547 60 Colombia −003837 Australia 10652 34 South Africa 00407 61 Ukraine −003958 Canada 10648 35 Croatia 00379 62 China −004659 Austria 10624 36 Malaysia 00263 63 The Philippines −00484

10 Iceland 10600 37 Bulgaria 00256 64 Indonesia −0050111 Germany 10503 38 Trinidad and Tobago 00185 65 Fiji −0052512 United Kingdom 10478 39 Romania 00178 66 Egypt −0052813 Belgium 10260 40 Ghana 00065 67 Honduras −0053514 France 10241 41 Brazil 00039 68 Guatemala −0054515 Japan 10203 42 Jamaica −00038 69 Vietnam −0055516 Chile 10153 43 Turkey −00053 70 Gabon −0060817 Portugal 10086 44 Tunisia −00054 71 Kenya −0068118 Estonia 10036 45 Suriname −00061 72 Paraguay −0069519 Slovenia 00976 46 Serbia and Montenegro −00081 73 Russia −0072720 Spain 00949 47 El Salvador −00090 74 Bolivia −0074121 Czech Republic 00888 48 Mexico −00137 75 Cambodia −0078522 Hungary 00813 49 India −00171 76 Ecuador −0086523 Mauritius 00782 50 Senegal −00250 77 Angola −0098824 Slovakia 00778 51 Saudi Arabia −00251 78 Nigeria −1004225 South Korea 00703 52 Argentina −00266 79 Pakistan −1008626 Lithuania 00687 53 Dominican Republic −00269 80 Venezuela −1014527 Uruguay 00673 54 Morocco −00277 81 Ivory Coast −10359

segments have significant input–output flows to and fromthe focal subsidiary’s primary segment as a percentageof the number of all subsidiaries of the same MNC par-ent. In robustness checks, I used a dummy variable tomeasure whether the subsidiary and its MNC parent havesignificant input–output flows to and from one another;the results are similar.

To construct the measure, I used the BenchmarkInput–Output (IO) “Use” tables provided by the U.S.Bureau of Economic Analysis (BEA). The tables containthe value of pairwise commodity flows among IO indus-tries and can be converted to commodity flows amongSIC industries through an IO–SIC concordance (Fan andLang 2000). They are updated every five years. Becausethe BEA changed the IO industry coding system in1997, I used the 1992 tables to ensure comparability.Except for the code change, coefficients in the tableshave been fairly stable over time (Fan and Lang 2000).The use of IO table coefficients as proxies for inter-segment relationships within diversified firms has beenadopted by studies in finance, economics, and man-agement (Schoar 2002, Villalonga 2004, Zhou 2011).In their study, Alfaro and Charlton (2009) used theinput–output tables to identify subsidiaries that provideinputs to their parent firms.

For each MNC-year, I constructed a task matrix. Ifan MNC has N subsidiaries (U.S. and foreign) in agiven year, the task matrix is an N × N matrix whoseentries (i1 j) and (j1 i) are set to x’s if subsidiary i’s and

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


subsidiary j’s primary segments on average contributemore than 1% of the input to one another according tothe IO tables. Based on the task matrix, I then countedthe total number of x’s in row i divided by N as a mea-sure of task centrality for subsidiary i.

Besides the factor variables of institutional qualityand task centrality, I added several control variables.I included year dummies to capture macroeconomic,political, and cultural factors that could change thepropensity to local supervision over time. At the host-country level, I controlled for the gross domestic product(GDP) (in constant year 2000 dollars), which reflects acountry’s general level of development (such as humancapital) and is expected to have a positive correlationwith local autonomy. I also controlled for telecommuni-cation using the number of Internet users per 100 peoplein the population. Better telecommunication technologywould make it easier to transfer complementary informa-tion collected throughout the organization to local units,increasing the efficiency of local autonomy. Addition-ally, to further dilute the impact of tax regimes on orga-nization structure, I controlled for the effective corporateincome tax rate.

At the level of MNC parent firms, I first controlledfor firm age and size. Older firms have more experiencedealing with heterogeneous environments and are morelikely to develop coordination knowledge and routines(Nelson and Winter 1982), facilitating autonomy. Topmanagement teams at larger firms are also more likelyto be overloaded and to prefer more delegation (Aghionand Tirole 1997). In addition, to capture each MNC’sexperience and scale of operations by country, I con-trolled for the number of frontline subsidiaries an MNChas in each country. An MNC’s country-specific expe-rience enhances its local political capability to manageweak institutions (Holburn and Zelner 2010), facilitatinglocal supervision. A larger number of local operationsalso raises the importance of local coordination and deci-sion making, encouraging local supervision.

At the subsidiary level, I first controlled for the busi-ness need for local supervision. A subsidiary that oper-ates in a different business segment than its MNC parentbenefits more from local adaptation and therefore localsupervision. This is because when an MNC parent doesnot possess expertise in a subsidiary’s business, the sub-sidiary needs to source more skills from the local envi-ronment (Chang and Rosenzweig 2001). It is thereforemore likely to be supervised locally. I then controlledfor subsidiary age. Older subsidiaries should be moreexperienced with local institutions and should thereforerequire less supervision from outside the host country.Older subsidiaries are also more likely to develop highercapabilities and require less resources from their parentheadquarters; therefore they will demand more auton-omy (Birkinshaw and Hood 1998). I do not have data onsubsidiary size. Instead, I controlled for a subsidiary’s

product scope (number of four-digit SICs in which itoperates). Subsidiaries undertaking a wider range ofactivities are expected to be larger and be supervisedlocally.

Table 2 provides descriptive statistics of the sample.There is large variation across countries in the qualityof institutions. The WGI has a mean value of 0.26 anda standard deviation of 0.86. The host countries in thesample have an average GDP of $0.27 billion in constant2000 dollars, an average effective corporate tax rate of30%, and an average of 19 Internet users per 100 people.At the MNC parent level, an average MNC in the sam-ple has about $1.8 billion (exp(7.48)) in sales, is about51 (exp(3.94)) years old, and operates 1.4 (exp(0.32))frontline subsidiaries in each host country. At the sub-sidiary level, the average age is 32.5 (exp(3.48)) yearsold. Local supervision is observed in about 6% of thesubsidiary-year observations. Table 2 also shows that anaverage subsidiary operates in 1.24 segments and hassignificant input–output flows with 30% of its MNC par-ent’s other subsidiaries. Among subsidiaries, 63% oper-ate in a primary segment different from their corporateparent.

Table 3 presents two preliminary analyses of the spanof control for supervisory units in countries of differ-ent institutional quality. First, if coordination is indeedmore difficult in WICs as I argued in the theory section,we would expect supervisory units in WICs to have anarrower span of control (controlling for the total num-ber of subsidiaries in a host country). Based on thisintuition, column (1) in Table 3 estimates the maximumspan of control (the maximum number of subsidiariessupervised by any supervisory unit of an MNC in ahost country) against institutional quality. The resultsshow that, indeed, supervisory units in weaker institu-tion countries have a narrower span of control. Sec-ond, if coordinating WIC subsidiaries is more difficult,we would expect supervisory units in strong-institutioncountries to have a narrower span of control if a largerproportion of their subordinate subsidiaries are in WICs.Accordingly, column (2) estimates the span of controlfor each supervisory unit in strong-institution countries.I split the countries into two groups: strong-institutioncountries whose institutional quality is above the medianlevel of institutional quality for all countries in a givenyear and weak-institution countries whose institutionalquality is at or below the median level. I then includedonly supervisory units in strong-institution countries forthis analysis. For each supervisory unit, I calculated itsspan of control as the number of subsidiaries across allcountries that directly report to it. As an independentvariable, I used the percentage of the supervisory unit’ssubsidiaries that are in WICs. The results show that,indeed, supervisory units in strong-institution countrieshave a narrower span of control if a larger proportionof their subordinate subsidiaries are in weak-institution

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Table 2 Variable Definitions and Summary Statistics

Definition Mean SD Min Max

Country-level variablesa

QualityofInstitutions Average value of Worldwide Governance Indicators 0026 0086 −1070 1096HostCountryGDP GDP per World Development Indicators, in billions of constant

year 2000 dollars0027 0063 00001 5021

HostCountryTaxRate Effective corporate income tax rate 30037 6094 5 50HostCountryTelecommunication Internet users per 100 people per World Development

Indicators18064 22051 00001 92014

MNC parent-level variablesMNCSizeb1e log(sales in million dollars) 7048 2004 −2099 12048MNCAgeb1e log(years since MNC establishment) 3094 0084 −2099 5045LocalScalec1e log(number of MNC subsidiaries in the country) 0032 0052 0005 4023

Subsidiary-level variablesd

LocalSupervision 41105 Equals 1 if the MNC subsidiary is supervised by a unit withinthe same host country

0006 0023 0 1

TaskCentrality Percentage of MNC subsidiaries whose primary segments areinterdependent with the focal subsidiary’s primary segment

0030 0027 0 0099

SubsidiaryProductScope Number of segments (four-digit SICs) in which the subsidiaryoperates

1024 0067 1 10

Subsidiary-MNC_DifferentBusinesses 41105 Equals 1 if the subsidiary operates in a different four-digit SICthan its MNC parent 41105

0063 0048 0 1

SubsidiaryAgee log(years since subsidiary establishment) 3048 0088 −2099 6014

aN = 11067 country-year observations.bN = 131490 MNC-year observations.cN = 451968 MNC-host country-year observations.dN = 701901 MNC subsidiary-host country-year observations.eLog value.

countries. Therefore, Table 3 confirms my intuition andprovides indirect support for H1.

Model SpecificationI adopted the following logit model:

E6LocalSupervisionjict = 17

= �0 +�1Ict +�2Kjict +�3Ict ×Kjict +Ujictê

+ Fitâ +Cctå1 (1)

where LocalSupervisionjict is a dummy variable thatequals 1 if subsidiary j of MNC i in host country c andyear t is supervised by a unit in the same host country,3

Ict is the quality of institutions in country c and year t,and Kjict is task centrality of subsidiary j of MNC i inyear t. The subsidiary-, MNC parent-, and host-country-specific characteristics are denoted by Ujict , Fit , and Cct ,respectively. Hypothesis 1 predicts that �1 > 0; H2 pre-dicts that �3 > 0.

Because the residuals of a given MNC may be corre-lated across countries as a result of unobserved firm het-erogeneity, and the residuals of a given country may becorrelated across firms as a result of unobserved coun-try heterogeneity, I adjusted standard errors to accountfor these two dimensions of within-cluster correlation(Petersen 2009).

ResultsTable 4 presents estimations based on Equation (1). Col-umn (1) contains only control variables at the country

level. As expected, subsidiaries in higher GDP coun-tries were more likely to be supervised locally. Telecom-munication positively affected local supervision and taxrate negatively affected local supervision, although thesecoefficients are not statistically significant. Column (2)adds quality of institutions; it had a significant and pos-itive impact on local supervision.

Column (3) adds MNC parent characteristics. Asexpected, subsidiaries of larger MNCs were more likelyto be supervised locally. The impact of MNC age waspositive but not significant. Also as expected, the moresubsidiaries an MNC operated in a host country, themore likely they were to be supervised locally. Col-umn (4) adds unit characteristics. The coefficients werenot statistically significant. With all control variables inplace, the quality of institutions continued to have a pos-itive impact on local supervision. Thus, H1 is supported.

Column (5) adds task centrality, and column (6) addsits interaction with institutional quality. Results showthat task centrality was negatively correlated with localsupervision: the more a subsidiary was interdependentwith an MNC’s global operations, the less likely it wassupervised locally. This negative effect was amplified forunits in countries with weaker institutions. Thus, H2 issupported.

The impact of institutional quality on local supervi-sion became weaker when more control and independentvariables were added, but it remained economically andstatistically significant. A marginal-effect analysis based

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Table 3 Span of Control and Institutions

(1) (2)Maximum span of control for MNC Span of control for MNC supervisorysupervisory units in a host country units in strong-institution countries

QualityofInstitutions 00015∗∗∗

4000055Number of subordinate subsidiaries in WICs −130002∗∗∗

as a percentage of total number of subordinate 4004465subsidiaries in all countries

MNCSize 00014 000334000095 4000925

MNCAge −00043∗∗ −00594∗∗

4000175 4002715LocalScale 00372∗∗∗ 10131∗∗∗

4000055 4000815HostCountryTaxRate −00001 00037

4000005 4000225HostCountryTelecommunication 00001∗∗∗ −00018

4000005 4000145HostCountryGDP 00004∗ 2E-14

4000025 (1E-13)Constant 00023 60110∗∗∗

4000935 4105355

Year dummies Yes YesMNC fixed effects Yes NoUnit fixed effects No YesObservations 45,968 16,719Adjusted R2 00232 00715

Notes. Column (1) shows linear estimates of the maximum span of control for MNC supervisory units in a host country, based on datafrom U.S. multinational manufacturers in 1996–2008. The unit of analysis is MNC-country-year. Column (2) shows linear estimates of spanof control for MNC supervisory units in strong-institution countries. The unit of analysis is MNC supervisory unit-year. Standard errors thataccount for clustering at both the firm and country level appear in parentheses.

∗Significant at the 10% level; ∗∗significant at the 5% level; ∗∗∗significant at the 1% level (two-tailed tests).

on column (6) suggests that increasing the quality ofinstitutions by one point, while keeping all other vari-ables at their mean values, increased the probability oflocal supervision by 1.4%.

Finally, Table 5 moves the logit analysis in Equa-tion (1) from the country level to the regional level.Variables are comparable to those in Table 4, but insti-tutional quality is measured by residents’ confidence inregional transparency and the rule of law. Column (1)includes the same country-level control variables as inTable 4. Column (2) uses country dummies to controlfor unobserved country-level heterogeneity. The resultsare similar: though statistically weaker than the country-level results, they are still supportive of H1 and H2.

Overall, the results in Tables 3–5 show that, con-sistent with my hypotheses, subsidiaries located inweak-institution countries or regions were less likely tobe supervised locally and that this effect was strongerwhen the subsidiaries’ activities were more central totheir MNC parents’ global operations. In addition tothese main results, I ran a host of robustness checksto control for additional factors that might influencelocal supervision, including factors at the MNC level

(total levels of hierarchy, R&D intensity, etc.), the indus-try level (growth, capital and R&D intensity, competi-tion, etc.), and the country level (language; distance tothe United States in terms of knowledge, globalization,geography, financial development, demography, admin-istration; country dummies; etc.); the results were simi-lar. I also ran a conditional logit model with MNC fixedeffects; the sample size is smaller, but the results aresimilar.

Discussion and ConclusionThis paper examines how MNCs may use organizationstructure to manage the effects of institutional qual-ity on business activities across countries. The supervi-sion relationships with respect to overseas subsidiariesof U.S. multinational manufacturers in 1996–2008 showthat MNCs do, indeed, strategically assign supervisoryresponsibilities to enhance coordination across diverseglobal operations.

This paper’s core theoretical contribution is establish-ing differential supervision as a mechanism for selec-tive intervention and managing coordination challenges

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Table 4 Supervision, Institutions, and Task Centrality

(1) (2) (3) (4) (5) (6)

LocalSupervision 41105

QualityofInstitutions (H1) 00747∗∗∗ 00626∗∗∗ 00625∗∗∗ 00636∗∗∗ 00288∗∗

4002405 4001935 4001935 4001935 4001545TaskCentrality 41105 −00717 −20676∗∗∗

4004695 4009605QualityofInstitutions×TaskCentrality (H2) 10396∗∗

4005585Subsidiary-MNC_DifferentBusinesses 41105 −00057 −00188 −00194

4002615 4002075 4002085SubsidiaryProductScope −00089 −00098 −00099

4001255 4001205 4001215SubsidiaryAge −00032 −00026 −00026

4000925 4000955 4000935MNCSize 00210∗∗ 00210∗∗ 00230∗∗ 00233∗∗

4001005 4000995 4000965 4000975MNCAge 00092 00109 00105 00104

4001595 4001555 4001535 4001535LocalScale 00661∗∗∗ 00655∗∗∗ 00647∗∗∗ 00641∗∗∗

4001355 4001375 4001365 4001355HostCountryTaxRate −00003 −00004 −00005 −00005 −00005 −00005

4000125 4000095 4000065 4000075 4000065 4000065HostCountryTelecommunication 00027∗∗∗ 00004 00003 00003 00003 00003

4000065 4000065 4000045 4000045 4000045 4000045HostCountryGDP 00254∗∗∗ 00278∗∗∗ 00211 00214 00217 00228

4000795 4000645 4001895 4001855 4001755 4001795Constant −30373∗∗∗ −40162∗∗∗ −60855∗∗∗ −60667∗∗∗ −60605∗∗∗ −60161∗∗∗

4004805 4004265 4100695 4101285 4101225 4101255

Year dummies Yes Yes Yes Yes Yes YesSector dummies Yes Yes Yes Yes Yes YesObservations 70,901 70,901 70,901 70,901 70,901 70,901Pseudo R2 00090 00096 00180 00180 00183 00186Log-likelihood −141065 −131966 −121672 −121666 −121619 −121581

Notes. This table shows the logit estimates of the likelihood that a subsidiary is supervised by a supervisory unit in the same country versusa foreign country, based on data from U.S. multinational manufacturers in 1996–2008. Standard errors that account for clustering at boththe firm and country level appear in parentheses.

∗Significant at the 10% level; ∗∗significant at the 5% level; ∗∗∗significant at the 1% level (two-tailed tests).

across heterogeneous institutional environments. It alsosupports a view of the firm as a complex system of inter-dependent activities that must be actively coordinatedto realize benefits from integration (Zhou 2011, 2013).Although a primary function of MNCs is to exploit arbi-trage opportunities arising from transaction costs acrossinstitutional environments, internalizing transaction costscreates coordination challenges as well. A hierarchi-cal structure across national boundaries allows differen-tial supervision, thereby balancing the trade-off betweenadaptation and coordination. The finding that MNCsmay design their organization structures to mitigate insti-tutional obstacles also complements existing studies onMNCs’ location and ownership choices.

This study offers implications for managers as wellas policy makers. It highlights a channel through whichMNCs “redistribute” managerial responsibilities awayfrom WICs. During the past two decades, governments

in developing countries have been working to improve“hard” conditions—building infrastructure, giving spe-cial tax breaks or subsidies to MNCs, and raising theeducational level of their labor force—to attract for-eign direct investments. These incentives may attract for-eign investments, but MNCs in these countries may onlyengage in fragmented business activities aimed mainlyat leveraging a cheaper labor force, engineering tal-ent, or market potential, without delegating substantialcorporate or regional responsibilities to local manage-ment teams. To the extent that managers make decisionsabout resource allocation on a daily basis, institutionalquality will have a profound impact on the sustainabledevelopment of the host country’s economy.

This study has a few limitations that invite futureresearch. First, it treats subsidiaries’ locations and tasksas predetermined and studies their impact on organiza-tion structure. It does not further investigate why some

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Table 5 Foreign vs. Local Supervision: World Value Survey

(1) (2)

LocalSupervision 41105

QualityofInstitutions (H1) 20988∗ 10983∗

4105625 4100575TaskCentrality 41105 −10150∗ −20062∗

4007005 4100885QualityofInstitutions× 40045∗ 40394∗∗

TaskCentrality (H2) 4203895 4200145Constant Yes YesSubsidiary controls Yes YesMNC parent controls Yes YesCountry controls Yes NoCountry dummies No YesObservations 20,561 19,313Pseudo R2 00226 00160Log-likelihood −11578 −31101

Notes. This table shows the logit estimates of the likelihood that asubsidiary is supervised by a supervisory unit in the same coun-try versus a foreign country, based on data from U.S. multinationalmanufacturers in 1996–2008. Control variables are the same asthose included in Table 4. Standard errors that account for cluster-ing at both the firm and country levels appear in parentheses.

∗Significant at the 10% level; ∗∗significant at the 5% level;∗∗∗significant at the 1% level (two-tailed tests).

firms choose to integrate certain productive activitiesat certain locations while others standardize and out-source them. Although firms can certainly alter theirtask systems—rather than their organization structures—to make coordination easier, the literature suggests thatfirms often make decisions about tasks based on fac-tors other than coordination. For example, firms mayintegrate certain activities to leverage their core com-petencies into adjacent value chain activities (Leibleinand Miller 2003), to accommodate differential position-ing strategies for their products (Argyres and Bigelow2010), or to preserve an integral knowledge of productarchitecture that deters imitation (Ethiraj et al. 2008).These corporate, product, and R&D strategies may con-strain firms from adopting independent task systems andpresent opportunities for organization design. How firmsendogenously choose their organization structures andtheir tasks at each location is left for future study.

In addition, the measure of “differential supervision”is based on the physical location of the supervisory units.I conceptualized that if the subsidiary reports to a for-eign rather than local unit, then less autonomy is given tosupervisory units and frontline subsidiaries in the coun-try as a group. My data set does not allow me to capturethe exact decisions that are made by the subsidiaries andthose that are made by the supervisory units. Althoughthe paper is not about the delegation of specific deci-sion rights with respect to each subsidiary, it will still beimportant to learn exactly what activity types are super-vised by foreign versus local supervisory units. This isa topic left for future study, when detailed data aboutallocation of decision-making rights between supervi-sory units and subsidiaries become available.

Despite its caveats, this paper connects organizationstructure with institutional imperfections that give riseto the emergence of the firm in the first place. It the-orizes and quantifies the relationships between interde-pendence, organization structure, and institutions using alarge sample of firms. This effort will hopefully deepenour understanding of the firm and its integration mech-anisms, and motivate future research exploiting the richand complex reality of the firm.

AcknowledgmentsThis paper is based on the author’s Ph.D. dissertation. Fortheir invaluable guidance, the author is indebted to GautamAhuja, Sendil Ethiraj, Scott Page, Jagadeesh Sivadasan, andJan Svejnar. The author gratefully acknowledges the sugges-tions provided by Wilbur Chung, Bennet Zelner, MinyuanZhao, senior editor Phanish Puranam, and three reviewers. Allerrors remain the author’s own.

Endnotes1In this study, unless otherwise specified, subsidiaries refer toMNC subsidiaries at the front line of business. In contrast,supervisory units refer to divisions, departments, groups, ornon-frontline subsidiaries that the frontline subsidiaries reportto. They include, but are not limited to, regional headquarters.2Tax havens are those listed by Hines and Rice (1994,Appendix 2) and Organisation for Economic Co-operation andDevelopment (2002).3I ran a robustness check to estimate the probability that a sub-sidiary reports to a supervisory unit in a country with strongerinstitutions than the subsidiary’s host country on a smallersample where institution data for the supervisory unit’s hostcountry are available. Results were similar.

References

Aghion P, Tirole J (1997) Formal and real authority in organizations.J. Political Econom. 105(1):1–29.

Alcácer J (2006) Location choices across the value chain: Howactivity and capability influence collocation. Management Sci.52(10):1457–1471.

Alchian AA, Demsetz H (1972) Production, information costs, andeconomic organization. Amer. Econom. Rev. 62(5):777–795.

Alfaro L, Charlton A (2009) Intra-industry foreign direct investment.Amer. Econom. Rev. 99(5):2096–2119.

Ambos TC, Andersson U, Birkinshaw J (2010) What are the conse-quences of initiative taking in multinational subsidiaries? J. Inter-nat. Bus. Stud. 41(6):1099–1118.

Anderson E, Gatignon H (1986) Modes of foreign entry: A transactioncost analysis and propositions. J. Internat. Bus. Stud. 17(3):1–26.

Antràs P, Desai MA, Foley CF (2009) Multinational firms, FDIflows, and imperfect capital markets. Quart. J. Econom. 124(3):1171–1219.

Argyres N, Bigelow L (2010) Innovation, modularity, and verticaldeintegration: Evidence from the early U.S. auto industry. Organ.Sci. 21(4):842–853.

Arrow KJ (1959) Rational choice functions and orderings. Economica26(102):121–127.

Baldwin CY (2008) Modularity, transactions, and the boundaries offirms: A synthesis. Indust. Corporate Change 17(1):155–196.

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Bartlett CA, Ghoshal S (1987) Managing across borders: New orga-nizational responses. MIT Sloan Management Rev. 29(1):43–54.

Bartlett CA, Ghoshal S (1989) Managing Across Borders: TheTransnational Solution (Harvard Business School Press, Boston).

Benito GRG, Lunnan R, Tomassen S (2011) Distant encounters ofthe third kind: Multinational companies locating divisional head-quarters abroad. J. Management Stud. 48(2):373–394.

Bikhchandani S, Hirshleifer D, Welch I (1992) A theory of fads,fashion, custom, and cultural-change as informational cascades.J. Political Econom. 100(5):992–1026.

Birkinshaw J, Hood N (1998) Multinational subsidiary evolution:Capability and charter change in foreign-owned subsidiary com-panies. Acad. Management Rev. 23(4):773–795.

Birkinshaw J, Braunerhjelm P, Holm U, Terjesen S (2006) Why dosome multinational corporations relocate their headquarters over-seas? Strategic Management J. 27(7):681–700.

Bloom N, Sadun R, Reenen JV (2012) The organization of firmsacross countries. Quart. J. Econom. 127(4):1663–1705.

Bouquet C, Birkinshaw J (2008) Weight versus voice: How foreignsubsidiaries gain attention from corporate headquarters. Acad.Management J. 51(3):577–601.

Caves RE (1996) Multinational Enterprise and Economic Analysis,Cambridge Surveys of Economic Literature, 2nd ed. (CambridgeUniversity Press, Cambridge, UK).

Chang S-J, Park S (2005) Types of firms generating network exter-nalities and MNCs’ co-location decisions. Strategic ManagementJ. 26(7):595–615.

Chang S-J, Rosenzweig PM (2001) The choice of entry mode insequential foreign direct investment. Strategic Management J.22(8):747–776.

Chozick A, Shirouzu N (2007) GM slips into Toyota’s rearview mir-ror. Wall Street Journal (April 25) http://online.wsj.com/news/articles/SB117739853275580259.

Cyert RM, March JG (1963) A Behavioral Theory of the Firm(Prentice-Hall, Englewood Cliffs, NJ).

Demirgüç-Kunt A, Maksimovic V (1998) Law, finance, and firmgrowth. J. Finance 53(6):2107–2137.

Desai MA, Foley CF, Hines J Jr (2004) Foreign direct investment ina world of multiple taxes. J. Public Econom. 88(12):2727–2744.

Dunning JH (2001) The key literature on IB activities: 1960–2000.Rugman AM, Brewer TL, eds. The Oxford Handbook of Interna-tional Business (Oxford University Press, Oxford, UK), 36–68.

Eisenmann TR, Bower JL (2000) The entrepreneurial M-form: Strate-gic integration in global media firms. Organ. Sci. 11(3):348–355.

Enright MJ (2000) Regional clusters and multinational enterprises:Independence, dependence, or interdependence? Internat. Stud.Management Organ. 30(2):114–138.

Ethiraj SK, Levinthal D, Roy RR (2008) The dual role of modularity:Innovation and imitation. Management Sci. 54(5):939–955.

Fan JPH, Lang LHP (2000) The measurement of relatedness: Anapplication to corporate diversification. J. Bus. 73(4):629–660.

Ford Motor Company (2007) Ford Motor Company annual report2006. Report, Ford Motor Company, Dearborn, MI. http://corporate.ford.com/doc/2006_AR.pdf.

Galbraith JR (2000) Designing the Global Corporation (Jossey-Bass,San Francisco).

Galbraith JR, Lawler EE (1993) Organizing for the Future (Jossey-Bass, San Francisco).

Garcia-Pont C, Nohria N (2002) Local versus global mimetism: Thedynamics of alliance formation in the automobile industry. Strate-gic Management J. 23(4):307–321.

Gates SR, Egelhoff WG (1986) Centralization in headquarters-subsidiary relationships. J. Internat. Bus. Stud. 17(2):71–92.

Gelos RG, Wei S-J (2005) Transparency and international portfolioholdings. J. Finance 60(6):2987–3020.

Ghemawat P (2007) Redefining Global Strategy: Crossing Borders ina World Where Differences Still Matter (Harvard Business SchoolPublishing, Boston).

Gupta AK, Govindarajan V, Wang H (2008) The Quest for GlobalDominance: Transforming Global Presence into Global Compet-itive Advantage (Jossey-Bass, San Francisco).

Harris M, Raviv A (2005) Allocation of decision-making authority.Rev. Finance 9(3):353–383.

Hart O, Moore J (2005) On the design of hierarchies: Coordinationversus specialization. J. Political Econom. 113(4):675–702.

Hass MR (2010) The double-edged swords of autonomy and exter-nal knowledge: Analyzing team effectiveness in a multinationalorganization. Acad. Management J. 53(5):989–1008.

Hill CWL (2012) International Business: Competing in the GlobalMarketplace (McGraw-Hill/Irwin, New York).

Hines JR Jr, Rice EM (1994) Fiscal paradise: Foreign tax havens andamerican business. Quart. J. Econom. 109(1):149–182.

Holburn GLF, Zelner BA (2010) Political capabilities, policy risk,and international investment strategy: Evidence from the globalelectric power generation industry. Strategic Management J.31(12):1290–1315.

Hummels D, Ishii J, Yi K-M (2001) The nature and growth of verticalspecialization in world trade. J. Internat. Econom. 54(1):75–96.

Kaufmann D, Kraay A, Mastruzzi M (2010) The Worldwide Gov-ernance Indicators: Methodology and analytical issues. WorldBank Policy Research Working Paper 5430, World Bank,Washington, DC.

Khanna T, Choudhury P (2007) Microsoft in China and India, 1993–2007. HBS Case 9-708-444, Harvard Business School, Boston.

Khanna T, Palepu K (2000) Is group affiliation profitable in emerg-ing markets? An analysis of diversified Indian business groups.J. Finance 55(2):867–891.

Kobrin SJ (2001) Sovereignty@bay: Globalization, multinationalenterprise, and the international political system. Rugman AM,Brewer TL, eds. The Oxford Handbook of International Business(Oxford University Press, Oxford, UK), 181–205.

Kogut B (1983) Foreign direct investment as a sequential process.Kindleberger CP, Audretsch D, eds. The Multinational Corpora-tion in the 1980s (MIT Press, Cambridge, MA), 38–56.

Kumar KB, Rajan RG, Zingales L (2001) What determines firm size?CEPR Working paper, Center for Economic Policy Research,London.

Lawrence PR, Lorsch JW (1967) Differentiation and integration incomplex organizations. Admin. Sci. Quart. 12(1):1–47.

Leiblein MJ, Miller DJ (2003) An empirical examination oftransaction- and firm-level influences on the vertical boundariesof the firm. Strategic Management J. 24(9):839–859.

LexisNexis (2005) Content FAQs. Accessed September 15, 2005, http://www.corporateaffiliations.com/nonsub/aboutContentFAQ.asp.

Luxner L (2001) Procter & Gamble decided to consolidate itsback-office operations for all the Americas in one place. LatinCEO: Executive Strategies Americas (April) http://www.ticonet.co.cr/costa-rica/readings/0104.html.

Malone TW, Crowston K (2001) The interdisciplinary study of coor-dination. Olson GM, Malone TW, Smith JB, eds. CoordinationTheory and Collaboration Technology, Computers, Cognition,and Work (Lawrence Erlbaum Associates, Mahwah, NJ), 7–50.

Marschak J, Radner R (1972) Economic Theory of Teams (Yale Uni-versity Press, New Haven, CT).

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Martinez JI, Jarillo JC (1989) The evolution of research on coordina-tion mechanisms in multinational corporations. J. Internat. Bus.Stud. 20(3):489–514.

Monteiro LF, Arvidsson N, Birkinshaw J (2008) Knowledge flowswithin multinational corporations: Explaining subsidiary isolationand its performance implications. Organ. Sci. 19(1):90–107.

Morse A, Shirouzu N (2007) Nissan’s Ghosn gives up U.S. duties inreshuffle. Wall Street Journal (March 17) http://online.wsj.com/news/articles/SB117403773086539264.

Nelson RR, Winter SG (1982) An Evolutionary Theory of EconomicChange (Harvard University Press, Cambridge, MA).

North DC (1990) Institutions, Institutional Change, and EconomicPerformance (Cambridge University Press, New York).

Organisation for Economic Co-operation and Development (2002)Towards global tax co-operation. Report, OECD, Paris.

Oxley JE (1999) Institutional environment and the mechanisms ofgovernance: The impact of intellectual property protection onthe structure of inter-firm alliances. J. Econom. Behav. Organ.38(3):283–309.

Petersen MA (2009) Estimating standard errors in finance panel datasets: Comparing approaches. Rev. Financial Stud. 22(1):435–480.

Porter ME (1985) Competitive Advantage: Creating and SustainingSuperior Performance (Free Press, New York).

Puranam P, Raveendran M, Knudsen T (2012) Organization design:The epistemic interdependence perspective. Acad. ManagementRev. 37(3):419–440.

Radner R (1993) The organization of decentralized information pro-cessing. Econometrica 61(5):1109–1146.

Rajan RG, Zingales L (1998) Financial dependence and growth. Amer.Econom. Rev. 88(3):559–586.

Rivkin JW, Siggelkow N (2003) Balancing search and stability: Inter-dependencies among elements of organizational design. Manage-ment Sci. 49(3):290–311.

Rugman AM, Verbeke A (2001) Subsidiary-specific advantages in multi-national enterprises. Strategic Management J. 22(3):237–250.

Schoar A (2002) Effects of corporate diversification on productivity.J. Finance 57(6):2379–2403.

Simon HA (1991) Organizations and markets. J. Econom. Perspect.5(2):25–44.

Spencer J, Gomez C (2011) MNEs and corruption: The impact ofnational institutions and subsidiary strategy. Strategic Manage-ment J. 32(3):280–300.

Sturgeon TJ (2002) Modular production networks: A new Ameri-can model of industrial organization. Indust. Corporate Change11(3):451–496.

Svejnar J, Smith SC (1984) The economics of joint ventures in lessdeveloped countries. Quart. J. Econom. 99(1):149–168.

Thompson JD (1967) Organizations in Action (McGraw-Hill,New York).

Tushman ML, Nadler DA (1978) Information processing as an inte-grating concept in organizational design. Acad. Management Rev.3(3):613–624.

Vernon R (1971) Sovereignty at Bay (Basic Books, New York).

Villalonga B (2004) Diversification discount or premium? New evi-dence from the business information tracking series. J. Finance59(2):475–502.

Yiu D, Makino S (2002) The choice between joint venture andwholly owned subsidiary: An institutional perspective. Organ.Sci. 13(6):667–683.

Zhao M (2006) Conducting R&D in countries with weak intellectualproperty rights protection. Management Sci. 52:1185–1199.

Zhou YM (2011) Synergy, coordination costs, and diversificationchoices. Strategic Management J. 32(6):624–639.

Zhou YM (2013) Designing for complexity: Using divisions and hier-archy to manage complex tasks. Organ. Sci. 24(2):339–355.

Yue Maggie Zhou is an assistant professor of strategy atStephen M. Ross School of Business, University of Michigan.She received her Ph.D. from the University of Michigan. Herresearch interests include the theory of the firm, organizationstructure, and multinational corporations. Her recent studiesinvestigate the role of coordination challenges in setting limitsto firm growth and in organization design.

Dow

nloa

ded

from

info

rms.

org

by [

141.

213.

162.

246]

on

16 F

ebru

ary

2015

, at 0

7:37

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

supervising across borders: the case of multinational ... · yue maggie zhou stephen m. ross ......

Documents