a clash of governance logics: foreign ... employees as key firm stakeholders...

Strategic Management JournalStrat. Mgmt. J., 37: 349–369 (2016)

Published online EarlyView 4 December 2014 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/smj.2344Received 13 June 2013; Final revision received 29 September 2014

A CLASH OF GOVERNANCE LOGICS: FOREIGNOWNERSHIP AND BOARD MONITORING

KURT A. DESENDER,1* RUTH V. AGUILERA,2,3 MÓNICALÓPEZPUERTAS-LAMY,1 and RAFEL CRESPI4

1 Department of Business Economics, Universidad Carlos III, Getafe (Madrid), Spain2 International Business and Strategy Department, D’Amore-McKim School ofBusiness, Northeastern University, Boston, Massachusetts, U.S.A.3 Department of Strategy and General Management, ESADE Business School,Ramon Llull University, Barcelona, Spain4 Departament d’Economia de l’Empresa, Universitat de les Illes Balears, Mallorca,Spain

We ask whether and when shareholder-oriented foreign owners are likely to change corporate gov-ernance logics in a stakeholder-oriented setting by introducing shareholder-oriented governancepractices. We focus on board monitoring and claim that because the bundle of practices used in astakeholder context does not protect shareholder-oriented foreign owners’ interests, they seek tointroduce their own practices. Our results suggest that board monitoring is only activated whenshareholder-oriented foreign ownership is high and that the influence of foreign ownership is espe-cially strong in firms without large domestic owners, with high levels of risk and poor performance.Our findings uncover the possibility of the co-existence of different corporate governance logicswithin a given country, shaped by the nature and weight of foreign owners Copyright © 2014 JohnWiley & Sons, Ltd.

INTRODUCTION

Over the past two decades, institutional investors,especially from Anglo-American countries, beganto diversify in international equity and entered intostakeholder-oriented systems. As a result, localorganizations became increasingly exposed to for-eign norms that are often at odds with local norms(Campbell, 2004), raising the need to better under-stand the influence of foreign investors in shapinggovernance practices in different business systems.While prior studies have looked at the impact of for-eign ownership on firm outcomes (e.g., Ahmadjian

Keywords: corporate governance; ownership; board mon-itoring; board of directors; audit fees*Correspondence to: Kurt A. Desender, Department of BusinessEconomics, Universidad Carlos III, calle Madrid 126, 28903Getafe (Madrid), Spain. E-mail: [email protected]

Copyright © 2014 John Wiley & Sons, Ltd.

and Robbins, 2005; David et al., 2006), the effect offoreign ownership on organizational processes andchanges in corporate governance patterns is still farfrom known, and has received less attention fromscholars, practitioners, and policy makers.

Corporate governance (CG) is an institutionalelement of a nation’s business system and hencereflects economic and social structures and normsof key stakeholders in a society (Aguilera andJackson, 2003; Fiss and Zajac, 2004). One ofthe sharpest distinctions among CG systems isbetween the shareholder-oriented economies of theAnglo-American countries and the stakeholder-oriented economies typified by Germany andJapan (Aguilera et al., 2008; Hall and Soskice,2001; Streeck, 2001). Our objective is to ana-lyze to what extent and under what conditionsshareholder-oriented foreign owners shape corpo-rate governance in a stakeholder-oriented setting

350 K. A. Desender et al.

by introducing governance practices that are share-holder oriented. Specifically, we focus on testingthe influence of institutional Anglo-Americanforeign ownership on board monitoring behavior inJapanese firms. We examine monitoring behaviorby independent directors rather than solely lookingat board structural characteristics because thelatter may turn out to be cosmetic, symbolic, orcamouflaged, with the underlying director practicesand behavior left unchanged (Fiss and Zajac, 2004).

The general outline of our argument is as fol-lows. Shareholders’ interests vary across businesssystems. In stakeholder systems, shareholders tendto have other interests above and beyond theirequity investment, such as maintaining on-goingbusiness relationships. When foreign investorsfrom a shareholder system invest in a stakeholdersystem, their governance logics and interests clashwith those of (stakeholder-oriented) domesticshareholders. Because the bundle of governancepractices deployed by domestic stakeholders maybe unavailable or does not address the agencyconflict that foreign owners face, they seek toprotect their investment by introducing governancepractices that are common in the Anglo-Americancontext. In the shareholder-oriented context, twokey governance practices to alleviate the agencyproblems between shareholders and managers standout: board monitoring and contingency compensa-tion (Shleifer and Vishny, 1997). However, thesepractices have traditionally not been emphasizedin the stakeholder-oriented model. Instead, mainbanks, affiliated business partners, as well as seniorand retired executives play a significant monitoringrole (Deakin, 2010/2011). Then, we focus onboard monitoring to test our logic of whethershareholder-oriented foreign investors introduceforeign practices into the stakeholder-orientedcontext.

In order to examine whether board monitoringbehavior is contingent on foreign ownership, weanalyze the effect of independent directors on theamount of external audit fees1 in firms with dif-ferent levels of foreign ownership. The purpose ofthe external audit is to obtain reasonable assuranceon whether the financial statements as a whole arefree from material misstatement. To this end, the

1 Audit fees reflect the amount paid by the company for the pro-fessional examination and verification of the financial statementsfor the purpose of assessing their consistency, fairness, and con-formation to accepted accounting principles (Simunic, 1980).

auditor should obtain sufficient appropriate auditevidence, which depends on both qualitative andquantitative considerations. To gain qualitativeinformation, the external auditor generally meetswith the board to gather information about thebusiness and the quality of internal control systems.Through these meetings, directors who engage inmonitoring can provide information and expressconcerns about records, documentation, internalcontrol weaknesses, and other matters that arerelevant to the preparation and fair presentation ofthe financial statements (AU section 380; JICPAreport 260 (2009)). Moreover, they may alsoexpress their desire for a better and more thoroughauditing (Aguilera et al., in press). The externalauditor considers this qualitative information toplan and conduct the audit, which consecutivelydetermines the audit fee. Board members who seekto protect shareholders’ interests (potentially atodds with management), typically independentdirectors, have incentives to monitor. Hence, theytend to ask for more information and to share withthe auditor their concerns regarding internal controlweaknesses and other accounting issues as well asthe need to enhance overall shareholder protection,leading to a significant effect of independentdirectors on the level of audit fees (Carcello et al.,2002). In contrast, if independent directors do notplay a monitoring role, they will not have relevantinformation to assist the auditor, nor will they haveincentives to demand better and more thoroughauditing, leading to an insignificant effect ofindependent directors on audit fees. Therefore, webelieve that an effective way to capture board moni-toring behavior is to assess the relationship betweenboard independence and external audit fees, in linewith previous studies (e.g., Abbott et al., 2003;Carcello et al., 2002; Cohen, Krishnamoorthy, andWright, 2007; Desender et al., 2013).

We argue that the monitoring behavior of inde-pendent directors will be contingent on the degreeof foreign ownership. When foreign ownershipis high, we expect that independent directors willhave greater incentives to protect shareholders’interests by monitoring. However, we expect suchbehavior to be absent when foreign ownership islow, as independent directors do not have the sameincentives and ability to monitor because domesticshareholders employ a different set of governancemechanisms to protect their interests, of whichboard monitoring is not a key part. In addition,we contend that the effect of foreign investors is

Copyright © 2014 John Wiley & Sons, Ltd. Strat. Mgmt. J., 37: 349–369 (2016)DOI: 10.1002/smj

Foreign Ownership and Board Monitoring 351

unlikely to be homogenous across firms, and weexplore how large domestic owners, firm risk, andperformance impact foreign investors’ influence.

To test our arguments, we focus on Japan, whichis a particularly suitable setting for several reasons.First, the Japanese system contrasts sharply with theAnglo-American system, and Japanese firms havebeen resistant to change their corporate governancemodel towards the Anglo-American style—i.e.,Japan is the only major market in Asia that does notrequire a minimum number of independent direc-tors. Second, the presence of foreign investors hasincreased dramatically since the 1990s. While for-eign ownership accounted for less than 5 percent in1990, it rose to 28 percent in 2012—the vast major-ity being Anglo-American institutional investors(Bank of Japan, 2012). Ahmadjian and Robbins(2005) argue that to assess the influence of thesenew foreign actors it is necessary to look not only attheir direct effect, but also at how they are embed-ded in the existing governance system. In this sense,the growing presence of shareholder-oriented for-eign owners may give rise to hybrid forms of CG(Aguilera and Jackson, 2003). Finally, the falloutfrom recent Japanese corporate scandals such asOlympus, Daio Paper, Tokyo Electric Power, andKyushu Electric has revived the debate regardingthe monitoring role of the board and the relevanceof independent directors in Japan.

Considering the universe of listed Japanese firmsin the Tokyo Stock Exchange (TSE) over the period2006–2012, our results show that board indepen-dence and external audit fees are positively relatedonly when foreign ownership is high (above 20%).In addition, we reveal that the influence of foreignownership is especially strong in firms without largedomestic owners and in firms with high levels of riskand poor performance. Our findings are robust todifferent specifications and to the use of instrumentsto address endogeneity concerns.

Our study contributes to several lines of research.First, it sheds further light into the literature on cor-porate governance bundles. In particular, we extendRediker and Seth’s (1995) concept of governancebundles from a single decision-maker perspectiveto a setting where different types of shareholdersinfluence a (partial) set of practices embedded inan existing system. Furthermore, while research ongovernance bundles has been mostly conceptual, wehypothesize and empirically test the relevance offoreign ownership as a key contingency in explain-ing the boards’ monitoring function, uncovering a

hybrid governance system. Second, we contributeto the debate on the lack of convergence of CG sys-tems, by investigating the role of foreign investorsin shaping board monitoring practices in a sys-tem that is radically different. Whereas researchershave emphasized distinctions between governancesystems and the patterns through which these sys-tems evolve, there is less evidence on the mecha-nisms by which systems change. Third, our studyalso speaks to research on drivers of organiza-tional change. In the context of Japanese CG, stud-ies have looked at the influence of foreign owner-ship on downsizing (Ahmadjian and Robbins, 2005;Ahmadjian and Robinson, 2001), employee wages(Yoshikawa, Phan, and David, 2005), R&D andcapital investment (David et al., 2006), and cor-porate performance (Gedajlovic, Yoshikawa, andHashimoto, 2005; Miyajima and Kuroki, 2007).While these studies have mainly focused on out-comes, we examine when and how foreign owner-ship leads to changes in governance processes andcontribute to a better understanding of the mecha-nisms through which internationalization influenceslocal organizational practices.

JAPANESE CORPORATE GOVERNANCECONTEXT

Japan is usually categorized as a network-orientedcountry, where firms have maintained a strongstakeholder orientation. The Japanese systemis characterized by tight networks of verticaland horizontal groupings known for their cross-shareholdings and financial, human, and transac-tional ties (Lincoln and Gerlach, 2004). Instead ofowning stocks primarily as portfolio investmentsor for financial purposes, domestic investors areoften business partners or commercial banks, bothof which hold shares for the implicit purpose ofbusiness goodwill, information exchange, andmutual monitoring. Japanese investors typicallyfall under Aguilera and Jackson’s (2003) categoryof investors with strategic interests as opposed tofinancial interests. In addition, the Japanese systemconsiders employees as key firm stakeholders(Colpan et al., 2011; Yoshimori, 1995).

With stakeholders concerned about long-termrelationships, firms develop and implement strate-gies based on long-term goals, seeking to maxi-mize mostly market share and growth rather thanshort-term profits or share price. In this respect,



Deakin (2010/2011) argues that most large Japanesefirms are run on a variant of the “community firm”system in which executives see themselves as hav-ing a commitment to maintain the company as anentity in its own right, and view their obligationsto customers and employees, present and future,as taking priority over those owed to shareholders.While shareholders have the power to replace direc-tors with a majority vote and can initiate litigationagainst directors, in practice, Japanese shareholdersrarely exercise these rights.

The Japanese board of directors, like its U.S.counterpart, is vested with authority to make strate-gic decisions and monitor corporate activity. How-ever, Japanese boards have traditionally put littleemphasis on their monitoring role (Aoki, Jackson,and Miyajima, 2007; Gilson and Milhaupt, 2005).In part, this is because Japanese boards are oftenmainly composed of executives, former employ-ees, and a small number of affiliated or relatedoutsiders (Yoshikawa and McGuire, 2008). Sinceexecutive directors are viewed as a representativeof employees, they lack incentives and capabilitiesto monitor top executives to enhance shareholdervalue (Kubo, 2005). Deakin (2010/2011) comparesU.K. with Japanese boards and concludes that,in Japan, there is a belief that absence of deepknowledge of the company’s business makes itinappropriate for independent directors to makeinformed decisions concerning corporate strategy,justifying their limited monitoring role. Over thepast decade, however, independent directors havebecome more prevalent in Japanese boardrooms.Yoshikawa and McGuire (2008) state that thereis a possible evolution in the board’s role, givenchanging ownership pressures. Directors have aterm of two years in firms that use the corporateauditor system, while this is one year for firms thatadopted the committee system, but they may serveany number of consecutive terms if re-elected.

The limited presence of independent directors,however, does not equate to monitoring beingnonexistent. Instead, main banks, as well as affili-ated business partners and product markets, play asignificant monitoring role (Aoki and Patrick, 1994;Kaplan and Minton, 1994). In addition to monitor-ing by these relational stakeholders, there is strongpeer-based monitoring, with senior executives mon-itoring each other throughout their careers. Forexample, CEOs are typically company insiderswhose entire careers have been spent climbingthe corporate ladder. Retired executives also bear

a key role in monitoring, often through informalchannels (Deakin, 2010/2011). Therefore, the bun-dle of governance mechanisms used by domesticowners reflects their governance logic where moni-toring by independent directors is not fundamentalto the protection of their interests.

The audit of financial statements by an external(or accounting) auditor (kaikeikansinin) is manda-tory for all listed firms, and is equivalent to externalauditors elsewhere. The financial statement auditwas formally introduced in Japan in 1957 and isbased upon U.S. standards and practice. In Japan,as in most other countries, the board proposes theexternal auditors, and they are appointed at the gen-eral shareholders’ meeting. A series of reforms inearly 2000s made it mandatory for external audi-tors to declare whether a client faces a serious riskof going bankrupt within a year, and, more impor-tantly, auditors could be sued if found to have misledshareholders.

The rising influence of foreign investors in Japanoccurred concurrently to a decline in the bankingsystem that supported stakeholder systems, causedby the banking crisis and, to a smaller extent,the 1997 Asian financial crisis. These events ledfinancial institutions to sell off large proportions oftheir shareholdings (Ahmadjian and Robbins, 2005;Hoshi and Kashyap, 2004). As a result, the pres-ence of foreign shareholders sharply rose in thelate 1990s, and foreign investors, with very dif-ferent incentives, replaced domestic shareholderswho were more tightly bound to the stakeholdersystem. U.S. and U.K. investors, mainly institu-tional investors, jointly constitute around 70 percentof all foreign equity investments in Japan over thelast 15 years (Bank of Japan, 1996–2012). Foreigninvestors tend to hold small stakes at individual lev-els, and they rarely appear as strategic investors.

THEORETICAL FRAMEWORKAND HYPOTHESES

Corporate governance represents an interrelatedsystem in which some practices will be effectiveand relevant only in certain combinations, leadingto different patterns of CG (Aguilera et al., 2008).We establish that the heterogeneity in sharehold-ers’ logic and interests and their relative strengthwithin the organization is important to under-stand the influence of different types of ownersover existing governance arrangements, as well as



the emergence of hybrid bundles of governancepractices. Within an organization, there are differentgroups with diverse interests and value percep-tion, which represent and import into an organiza-tion the logics to which they have been primarilyexposed (Greenwood et al., 2011). While monitor-ing and contingency compensation are commonlyused in shareholder-oriented systems to alleviate theagency problem, a different set of mechanisms isused by domestic owners in a stakeholder-orientedsystem, which is specific to their agency problems.As a consequence, the bundle of governance mech-anisms employed by domestic shareholders in astakeholder-oriented setting is unlikely to solve theagency problems faced by shareholder-oriented for-eign owners. In this vein, we adopt a contingencyapproach to examine how shareholder-oriented for-eign investors protect their interests in a stakeholdercontext, focusing on whether (and when) foreignownership activates board monitoring behavior. Ourtheoretical arguments draw mainly on agency the-ory and on the resource dependence notion thatboards have distinct incentives and abilities to mon-itor management.

The monitoring role of boards has been the focusof extensive CG research (e.g., Adams, Hermalin,and Weisbach, 2010). Agency theorists argue that animpartial assessment of managers will occur morereadily if directors are independent from manage-ment (Hillman and Dalziel, 2003), and as a conse-quence board independence has often been used asa direct measure of board monitoring (Adams et al.,2010). Since executive directors report to the CEO,they will be less likely to perform comprehensivemonitoring tasks. Moreover, because independentdirectors are not part of the organization’s manage-ment team, they are less subject to the same poten-tial conflicts of interest that are likely to affect thejudgments of executive directors (Kosnik, 1987).Following this line of reasoning, prior researchhas highlighted the insufficient representation ofindependent directors as a possible explanation forboards’ failure to fulfill their monitoring role (e.g.,Finkelstein and Hambrick, 1996). However, crit-ics of the agency theory perspective have pointedout its “under-contextualized” nature and, hence,its inability to explain the diversity of governancearrangements across institutional contexts (Aguil-era and Jackson, 2003).

We draw on Hillman and Dalziel (2003) as theypush the agency theory logic further by bringingin a resource dependence component to propose

a more comprehensive view of board activities.Specifically, they suggest that, rather than assumingthat two boards with an identical proportion ofindependent directors will grant equal monitoringeffectiveness, it is critical to realize that these inde-pendent directors might have different incentivesas well as abilities to monitor. In addition, someindependent directors may focus less on the mon-itoring role and more on the advisory role, whichreduces both their incentives and ability to moni-tor. Similarly, Deutsch, Keil, and Laamanen (2011)note that independent directors should be under-stood as agents in their own right, with their ownindividual motivations as board members. Follow-ing this line of research, we argue that the primaryrole of independent directors is defined by the con-text in which the company is embedded. While themonitoring role of independent directors is stronglyemphasized in shareholder-oriented CG systems, ina stakeholder-oriented setting, like Japan, the moni-toring role has traditionally not been stressed, and asa consequence, higher ratios of board independencedo not necessarily correspond with higher monitor-ing effort. To capture board monitoring, and in linewith Desender et al. (2013), we focus on the effectsof independent directors on the amount of externalaudit fees (a board action) rather than on the levelof independent directors (board structure).

Our conceptualization of board monitoringbehavior builds on an extensive body of researchin the managerial accounting and audit field thatdiscusses the effect of board independence onexternal audit fees suggesting that, if independentdirectors engage in monitoring, this effect couldbe either negative or positive (e.g., Carcello et al.,2002; Cohen et al., 2007; Hay, Knechel, and Ling,2008). First, a negative relationship is expected ifindependent directors take an active role in improv-ing the design of internal controls and internalgovernance in general. To the extent that indepen-dent directors share information with the auditorabout the board’s effort to improve the firm’sinternal controls and overall reliability of financialreporting, the external auditor’s greater reliance oninternal controls results in less substantive audittesting and, hence, a lower audit fee (Cohen et al.,2007; Collier and Gregory, 1996). Second, a pos-itive relation is expected if independent directors,instead of actively improving internal controls andoverall reliability of financial reporting, ask formore information from management and sharetheir concerns about internal control weaknesses



and other accounting issues as well as their desirefor a better and more thorough auditing with theauditor. This information may lead the auditor toincrease the amount of audit evidence and hencethe audit fees. Furthermore, independent directors’commitment to oversight may also signal to theauditor that the expectations placed on the externalaudit are high. This second line of reasoning isgenerally supported by empirical evidence in theAnglo-American context (Abbott et al., 2003;Carcello et al., 2002; Hay et al., 2008).

In a stakeholder-oriented setting, like Japan, thepossibility of a negative effect of board indepen-dence on audit fees is less likely since the monitor-ing role of independent directors has, traditionally,not been emphasized, and independent directorshave, most likely, limited influence to enforce majorchanges in the internal organization of accountingcontrols. However, while one or two independentdirectors may not have enough power over a boarddecision that would require majority voting, even asmall number of independent directors may have aninfluence on the auditor’s perception regarding thequality of the internal control system and the over-all reliability of financial reporting. This makes theinformation exchange with auditors a feasible andeffective mechanism available to independent direc-tors to enhance the protection of shareholder rights.Therefore, in a stakeholder-oriented setting, if inde-pendent directors engage in monitoring, we expectto find a positive significant effect of board indepen-dence on audit fees, in line with previous literature(e.g., Abbott et al., 2003; Carcello et al., 2002). Incontrast, a nonsignificant relationship would sug-gest no monitoring behavior by independent direc-tors, i.e., they do not strengthen the internal controlenvironment, nor do they provide relevant infor-mation about internal control weakness or otheraccounting issues that would influence the auditscope, and in turn impact the audit fee. To testthe occurrence of board monitoring behavior in astakeholder-oriented setting, we establish our base-line hypothesis as follows:

Hypothesis 1: Board independence increasesexternal audit fees.

We contend that the relationship betweenindependent directors and audit fees (i.e., boardmonitoring) is contingent on the level of foreignownership. In particular, when foreign ownership is

low, we would not expect the relationship betweenboard independence and the amount of audit feesto be significant, because the monitoring role ofindependent directors has not been emphasized bydomestic owners, as they rely on other mechanismsto protect their interests. Yet, when foreign owner-ship is high, we argue that the relationship betweenindependent directors and audit fees becomescritical, as foreign investors seek to protect theirinvestment and exert pressure on independentdirectors to perform a monitoring role. Thus,independent directors in those firms may serve amonitoring role as there is a greater expectationfrom their monitoring.

There are three factors that explain this activationof the monitoring role of independent directors withthe relevant presence of foreign investors. First, theconflict of interest and asymmetry of informationbetween the (Japanese) management and the foreigninvestors are magnified by geographical distanceand cross-national differences (Buckley, 1997). Thedivergence in objectives and risks between foreignand domestic shareholders, combined with the tra-ditional informal governance practices employedby the domestic stakeholders, may force foreigninvestors to introduce CG practices that are closerto their own governance logic. Given the impor-tance of board monitoring and active collaborationby independent directors with the external auditorsin the shareholder-oriented context, foreign share-holders may put pressure on strengthening both themonitoring role of independent directors and theirinteraction with the external auditor above othergovernance practices. In this sense, the bundle ofgovernance practices that works for domestic share-holders is unlikely to address the agency conflictthat foreign owners face. In addition, the costs ofauditing are shared between all shareholders whilethe benefits in terms of reduced information asym-metries may be reaped by foreign investors, ratherthan domestic owners who have access to otherchannels of information and control.

Second, foreign investors tend to hold, individ-ually, a relatively small stake and rarely appear asstrategic investors, which reduces their incentivesof direct monitoring and their ability to introduceradical changes, such as the adoption of the boardcommittee system. Moreover, radical governancechanges may be met with strong resistance and,consequently, are difficult to implement. As aresult, foreign investors need to rely on internalCG practices. Fostering the monitoring role of



independent directors and their interaction withthe external auditor to express concerns aboutcertain accounting issues or internal control weak-nesses falls within the span of actions that foreigninvestors can adopt to protect their interests within astakeholder-oriented governance system like Japan.

Third, foreign shareholders rely on both voice,through direct meetings with management, and exitstrategies to make their interests clear (Ahmadjianand Robbins, 2005). For many years, foreigninvestors have pleaded for the regulatory adoptionof a transparent process of external supervisionof management through independent directors, inorder to protect shareholders’ interests (Gilson andMilhaupt, 2005). Hence, when foreign investorsare important, there may be a greater expecta-tion regarding the monitoring role of the board.Independent directors in those firms are likely tobe aware of foreign shareholders’ preferences fortheir more proactive monitoring role and may feelcompelled to take steps to enhance shareholderprotection (Colpan et al., 2011). When foreignownership is high, independent directors may alsohave higher incentives for monitoring to avoidreplacement.

For firms with a low degree of foreign ownership,we do not expect independent directors to exhibitthe same monitoring behavior for at least three rea-sons. First, domestic shareholders employ mainlyinformal channels to discipline managers and relyless on independent directors for monitoring. There-fore, in the presence of low foreign ownership, inde-pendent directors are less likely to feel pressure toexhibit monitoring behavior. In fact, asking morequestions and sharing this information and theirconcerns with the external auditor could representa shift from existing organizational routines andmay receive strong resistance from insiders (Han-nan and Freeman, 1984). Second, it is importantto note that, unlike in the United States, litigationagainst directors is very low in Japan (Numato andTakeda, 2010). This further diminishes the inde-pendent directors’ incentives to reduce their legalliability by monitoring and expressing accountingconcerns with the external auditor. Third, while rep-utation concerns may be linked to monitoring inAnglo-American firms, in a setting where the insti-tutional logic favors an active insider-oriented boardand monitoring through relational shareholders andpeers, the contribution of independent directors maybe more appreciated as providers of resources, suchas industry expertise or giving access to finance,

than as monitors overseeing general shareholders’interests. Taking these arguments together, ex ante,we would only expect to find a significant relation-ship between board independence and audit fees forfirms in a stakeholder-oriented setting when foreignownership is high. Therefore, we propose:

Hypothesis 2: Board independence increasesexternal audit fees more when foreign ownershipis high than when foreign ownership is low or notpresent.

The adoption of U.S.-inspired governance prac-tices by Japanese firms represents a departure fromorganizational routines (Han et al., 1998) and mayreceive strong resistance from inside actors (Han-nan and Freeman, 1984), reducing the ability offoreign shareholders to introduce board monitor-ing behavior. The effect of foreign ownership isunlikely to be homogenous across all firms, and itis likely to be driven by factors that influence theability and incentives of foreign owners to activatemonitoring by independent directors. We first con-sider the relevance of large domestic owners as aconstraining factor. In addition, we examine firmrisk and performance, which have been identified astwo key firm characteristics shaping board monitor-ing (Tuggle et al., 2010; Zajac and Westphal, 1994).

Our theoretical framework posits that within anorganization there are different groups with diverseinterests and value perception, who have access todifferent governance mechanisms. The ability offoreign owners to introduce changes in board mon-itoring, therefore, depends upon the extent to whichgroups have an interest in proposed reforms andtheir capacity to either support or resist them. Weargue that domestic owners in a stakeholder systemmay not share the interest of (shareholder-oriented)foreign owners as they rely on a governance bundlein which board monitoring does not play a promi-nent role, and the larger their ownership stake, themore likely that large domestic owners may limitforeign investors’ influence over the monitoring roleof the board. In this respect, Ahmadjian and Rob-bins (2005) find evidence that the impact of for-eigners with respect to corporate restructuring wasweaker in firms more deeply embedded in the localsystem, measured by the importance of domes-tic financial blockholders and corporate group ties.Therefore, we expect the influence of foreign own-ership on monitoring by independent directors to beless strong in the presence of large domestic owners.



In terms of firm risk, high levels of uncertaintymay widen information asymmetries between for-eign owners and managers and enable opportunis-tic behavior. Zajac and Westphal (1994) developand test a contingency cost/benefit perspective ongovernance decisions as resource allocation deci-sions, and identify how and why the observed levelsof managerial incentives and monitoring may varyacross organizations. Their findings suggest thatriskier firms can accrue greater benefits from higherlevels of board monitoring (relative to managerialincentives), and thus are likely to rely more on boardmonitoring. Following their logic, we argue that for-eign investors may have greater incentives to mon-itoring more closely their investment for firms withriskier operations, and we expect the influence offoreign ownership on board monitoring behavior tobe stronger for riskier firms.

Poor past performance indicates the ineffective-ness of existing organizational practices, and thusprovides strong and legitimate reasons for firmsto initiate reform (Chizema and Shinozawa, 2011).Previous research suggests that firms with poor per-formance face less resistance to the adoption of newpractices (Miller and Chen, 1994). In our setting,the ability of foreign owners to activate board mon-itoring behavior is likely to increase, as it repre-sents a justifiable governance alternative that carriesthe potential for better performance. In addition,Tuggle et al. (2010) find that negative deviationsfrom prior performance increase boards’ attentionto monitoring, while positive deviations from priorperformance reduce it (for a sample of US listedfirms). Underperformance could therefore affect theincentives and ability of foreign owners to enhanceboard monitoring. Considering these three contin-gency factors, we propose:

Hypothesis 3: The influence of foreign ownershipon board monitoring is stronger for firms (a)without large domestic owners, (b) with highrisk, and (c) with poor performance.

DATA AND METHODS

To test our hypotheses, we consider all listed firmson the TSE for the 2006–2012 period—Japanesefirms are only required to disclose external auditfees since March 2004. Financial firms are excludedbecause their accounts and audit process are sig-nificantly different. Our data comes from several

data sources. The governance and ownership struc-ture data was manually collected from the compa-nies’ annual corporate governance reports on theTSE website, eliminating any bias related to mea-surement errors. We gathered audit data as wellas all control variables from company financialstatements and Thomson Reuters Worldscope. Ourfinal sample contains 2,151 listed firms and 6,823firm-year observations.

To test our hypotheses, we estimate panel dataregressions extending the traditional audit feemodel (Abbott et al., 2003; Carcello et al., 2002;Simunic, 1980). The factors activating boardmonitoring behavior can be modeled as follows:

yit = 𝛼1zit−1 + 𝛽1xit−1 + vi, (1)

where yit is the total amount of audit fees, zit− 1is the vector of control variables, xit− 1 representsour variables of interest, i.e., board independenceand foreign ownership, and vi is the vector ofheteroskedastic-robust standard errors. All explana-tory variables are lagged by one period to mitigatepossible simultaneity. For our first hypothesis, weconsider the individual effects of board indepen-dence on audit fee, while for our second hypothe-sis, we introduce and focus on the interaction termbetween board independence and foreign owner-ship. Finally, to test Hypotheses 3a–c, we split oursample using the median values of domestic own-ers, risk, and performance and compare the interac-tion effect between board independence and foreignownership between the subsamples.

Our dependent variable, Total audit fees, is thenatural log of audit fees, following prior studieson the relationship between corporate governanceand audit services (Carcello et al., 2002; Hay et al.,2008). This variable represents the amount paidby the company for the professional examinationand verification of the financial statements for thepurpose of assessing their consistency, fairness, andconformation to accepted accounting principles.

We define Board independence as the propor-tion of outside board members (directors who havenever served as executive director, executive officer,employee of the company or any of its subsidiaries,as reported in the companies’ annual report) overthe total board size, similar to previous studies (e.g.,Carcello et al., 2002; Hay, Knechel, and Wong,2006). In addition, we also use a more conservativemeasure of board independence, taking advantageof the fact that since June 2006 the TSE requires



every listed company to prepare a report on cor-porate governance in which detailed disclosure ofthe relationship between an outside director and thecompany is mandatory. The TSE enumerates fivecategories (out of nine) of individuals who wouldnot be considered truly independent: (1) directorsrelated to the parent company, (2) directors relatedto other affiliated companies, (3) directors whoare major shareholders, (4) directors who share acompensation relationship, or (5) directors who arerelatives of executives. We went through every indi-vidual company’s annual CG report and calculatedthe variable Board independence (narrow) as theproportion of outsiders who qualify as truly inde-pendent directors over the total number of directors.

Foreign ownership reflects the percentage of totaloutstanding shares held by non-Japanese investors.The TSE reports the degree of foreign ownership infour categories: between 0 and 10 percent, between10 and 20 percent, between 20 and 30 percent, andmore than 30 percent. We use two main measuresof foreign ownership. First, we employ the fourcategories as defined by the TSE, which allowsus to test for nonlinearity and possible cut-offpoints. Second, we are able to establish the cut-offpoint of 20 percent foreign ownership as a criticalinflection point, and hence we use a dummy variabletaking the value 1 if foreign ownership is at least20 percent, and 0 otherwise. Moreover, we wereable to corroborate that our measure of foreignownership indeed captures institutional ownershipfrom shareholder-oriented countries (United Statesand the U.K.).

We start estimating our model (1) using the gen-eralized least squares (GLS) random effect (RE)technique with clustered standard errors by firm toaccount for within-firm error correlations, follow-ing the Baltagi and Wu (1999) procedure. This tech-nique is robust to first-order autoregressive, AR(1),disturbances (if any) within unbalanced panels andto cross-sectional correlation and/or heteroskedas-ticity across panels. It also has a number of advan-tages over fixed effects (FE) estimations. First, FEestimation requires significant within-panel varia-tion of the variable values to produce consistentand efficient estimates. Second, FE estimates mayaggravate the problem of multicollinearity if solvedwith least squares dummy variables (Baltagi, 2005).

We were concerned that the issue of endogenity,including omitted variables and simultaneity, couldbias the results. Endogeneity problems related tomeasurement errors are less of a concern given that

our data is derived directly from the TSE. Our mainissue is that foreign owners’ investments are notrandom and could be related to an unobserved oruncontrolled factor. Although it is difficult to com-pletely solve the endogeneity problem, we attemptto address this concern by (1) introducing ade-quate control variables, (2) estimating fixed effectsregressions to account for heterogeneity induced bytime-invariant factors and period effects, and (3)using instrumental variable techniques. We discusseach of them in turn below.

Multiple controls

A commonly used strategy for reducing concernsabout endogeneity is to saturate the regression witha large number of firm characteristics to captureas much of the error term as possible (Laevenand Levine, 2009). Following previous auditingliterature, we control for a wide range of factors thatcould bias the relationship between our variablesof interest:

Log of total assets

Larger companies are involved in a greater numberof transactions that necessarily require longer hoursfor an auditor to inspect (Carcello et al., 2002; Hayet al., 2006). We capture firm size as the naturallogarithm of total assets. Our results are robustto alternative measures of size based on sales oremployees.

Receivables and inventory/total assets

The evaluation of accounts receivable and invento-ries is complex and requires more in-depth inspec-tion, and therefore are considered risk categories inan audit. This variable is scaled by total assets andcaptures, partially, the complexity of the audit pro-cess (Hay et al., 2006).

Big 4 auditor

Higher audit fees are expected when an audit firmis recognized to be of superior quality to other firms(Hay et al., 2006). This variable takes the value 1if the client firm is working with one of the “Big4” auditors, i.e., AZSA & Co. (KPMG), Tohmatsu(Deloitte Touche), Aarata (PwC), and ShinNihon(Ernst & Young), and 0 otherwise.



Long-term debt/total assets

Total long-term debt divided by total assets is acommon measure of firm leverage. Highly lever-aged firms are more likely to fail, exposing the audi-tor to litigation risk, and hence are expected to beassociated with higher fees (Simunic, 1980).

Net income/total assets

Lack of client profitability is considered a concernfor the auditor because it reflects the extent to whichthe auditor may be exposed to loss in the event thata client is not financially viable (Simunic, 1980).To capture profitability, we focus on the return onassets, measured as net income over total assets.

Industry and year

We introduce industry (i.e., the 33 industrial sectorsused by the TSE) and year dummy variables tocontrol for industry and time effects.

Fixed effects

One limitation of adding control variables is that it isunlikely to effectively control for every relevant fac-tor, in part because some variables are unobserved.Using fixed effects allows us to control for unob-served time-constant firm heterogeneity. In addi-tion, we include dummies for the time periods inour FE estimator to control for endogeneity relatedto systemic shocks that lead to increases in auditfees in all firms. This approach permits us to assesswhether time-constant unobserved heterogeneity orperiod effects are creating bias.

Instrumental variables

To further mitigate reverse causality concernsor potential correlated omitted variables issuesand to increase confidence in the directionalityof our results, we implement a two-stage feasibleefficient generalized method of moments (GMM)estimation with validity-tested instruments. Tothis end, we define an instrumental variable (IV)that is correlated with foreign ownership, but isuncorrelated with the error term in regression 1. Inthe spirit of Laeven and Levine (2009), we generatean instrument by calculating the average levelof foreign ownership (excluding the contributionof the focal firm) for each industrial sector-size

pair.2 The intuition is that the level of foreignownership of other firms within the same sectorwith a similar size is likely to influence a focalfirm’s foreign ownership, but is unlikely to affectits audit fees. In fact, previous research has shownthat foreign ownership is linked to both industryand size (Kang and Stulz, 1997). Because thecontribution to the level of foreign ownership bythe focal firm is excluded, the instrument variesacross firms.3 For the interaction term betweenforeign ownership and board independence, oncewe verified that board independence can be treatedas exogenous, we follow Wooldridge (2002) andmultiply our instrument of foreign ownership withboard independence to create a second instrument.

To ensure the validity of our instruments, weperform diagnostic checks. First, we check that theinstruments are relevant (i.e., they are correlatedwith the included endogenous variables), using theF-statistic for joint significance of the instrumentsin the first-stage regression of each endogenousrepressors on the instruments and on the remainingexogenous regressors. As a second diagnostic,we use the Shea’s partial R2 in the first stage foreach endogenous variable. The Shea’s partial R2

records the additional explanatory power of theexcluded instruments taking the intercorrelationsof the instruments into account. We also test forunderidentification and report the Kleibergen-Paapunderidentification test (Kleibergen and Paap,2006). Finally, we report a test of endogeneity(GMM C-statistic) for the instrumented variables,in order to check whether the variables presumedto be endogenous in our model could instead betreated as exogenous.

To test our Hypotheses 3a–c, we split our sampleusing the median values of the ownership stake bythe largest domestic shareholder, firm risk, and poorperformance, respectively, and estimate our model(using instrumental variables) for both subsamples.Splitting our sample avoids the introduction ofnew endogeneity problems, while the comparisonbetween the subsamples enables us to contrastthe strength of foreign ownership on monitoring

2 We use 33 industrial sectors (following the TSE industryclassification) and four categories of size (using the quartilecut-off points Q1-Q2-Q3), based on total assets.3 An additional issue with our data is the presence of het-eroskedasticity. To address this issue, we specify a GMM optionin our implementation to make efficient estimation, valid infer-ence, and diagnostic testing, allowing for clustering the errors atthe firm level.



by independent directors contingent on these keyfactors, which we measure as:

Largest domestic shareholder

We define this variable as the total stake of thelargest domestic owner.

Firm risk

In line with Zajac and Westphal (1994) and Laevenand Levine (2009), we calculate firm risk as theannualized standard deviation of weekly equityreturns.

Poor performance

We focus on the deviation from prior performance,captured as the current performance compared withthe average of the prior two years’ performance,following Tuggle et al. (2010) where performanceis measured with return on assets.

RESULTS

In this section, we first provide descriptive statisticsof our data and we later test our proposed hypothe-ses. Table 1 gives an overview of the descriptivestatistics for the most important variables used inthis study as well as their correlations. The aver-age audit fee for the entire sample is ¥33.8 million($400,000). Regarding foreign ownership, morethan half of all firms have less than 10 percent for-eign ownership, while about one quarter have morethan 20 percent. Foreign portfolio investors arepredominantly institutional ones from the UnitedStates and the U.K. that on average constituted 47.0and 20.6 percent, respectively, of all foreign equityinvestments in Japan over the period 2006–2011(Bank of Japan, 1996–2011). In terms of boardcomposition, about 10 percent of the board mem-bers are outsiders (the average board size is 8.5),a finding that is substantially lower compared toAnglo-American or Continental European boards.Board independence grew steadily from 9.87 in2006 to 11.87 in 2011. About half of all firms inour sample have a board with only insiders, whileabout 22 percent have one outside director. Regard-ing the contingency factors that we believe are likelyto impact the influence of foreign ownership, theaverage stake of the largest domestic owner is about

20 percent, while the weekly stock return volatilityis around 3 percent and the average change in prof-itability is around zero.

Both board independence and foreign owner-ship are positive and significantly correlated withaudit fees. Furthermore, the correlation coefficientsbetween audit fees and our control variables showthe expected signs. In line with previous literature,the highest correlation coefficient is found for firmsize. We test for possible multicollinearity consid-ering our independent and control variables. Thevariance inflation factor gives a mean value of 1.40and a maximum value of 1.53, indicating no multi-collinearity problems.

Next, we discuss the multivariate analysis totest our three hypotheses. While Tables 2 and 3test for a general effect of board independenceon audit fees (Hypothesis 1) and whether boardmonitoring is activated in the presence of foreignownership (Hypothesis 2), Table 4 examines underwhat conditions the influence of foreign owner-ship on board monitoring is strongest (Hypothesis3). Table 2 presents the results obtained from theregression models with total audit fees as the depen-dent variable, when we consider board indepen-dence and the four categories of foreign ownership.Models 1–2 in Table 2 only include our controlvariables, with and without industry and year dum-mies, respectively, and explain a large proportionof the audit fee variance, confirming previous stud-ies with Anglo-American data (e.g., Abbott et al.,2003; Carcello et al., 2002). In these models, firmsize, receivables and inventory, long-term debt overtotal assets, and the presence of a Big 4 auditorare significantly associated with higher audit fees,while net income is not significantly related to auditfees, once we account for industry and year effects.

We next discuss Models 3–5 in Table 2, consider-ing the influence of board independence and foreignownership on audit fees. The specification of Model3 adds board independence and foreign ownershipto our model with controls. On the one hand, thecoefficients of foreign ownership reveal that auditfees in firms with more than 30 percent of foreignownership are significantly larger than audit fees infirms with less than 10 percent of foreign owner-ship, while intermediate levels are not significantlydifferent from firms with very low levels of for-eign ownership. On the other hand, the coefficientof board independence is positive and significant,in line with our Hypothesis 1. However, it is impor-tant to understand whether this direct positive effect



Tabl

e1.

Des

crip

tive

stat

istic

san

dco

rrel

atio

ns

Mea

nSD

12

34

56

78

910

1112

1314

15

1.A

udit

fees

(Log

)3.

860

0.89

41.

000

2.B

oard

inde

pend

ence

0.10

00.

140

0.18

8+1.

000

3.Fo

reig

now

ners

hip:

0–

10%

0.51

10.

500

−0.

309+

−0.

109+

1.00

04.

Fore

ign

owne

rshi

p:10

–20

%0.

252

0.43

40.

073+

−0.

015

−0.

498+

1.00

05.

Fore

ign

owne

rshi

p:20

–30

%0.

148

0.35

50.

183+

0.05

7+−

0.34

6+−

0.15

9+1.

000

6.Fo

reig

now

ners

hip>

30%

0.08

90.

285

0.26

9+0.

156+

−0.

281+

−0.

129+

−0.

090+

1.00

07.

Fore

ign

owne

rshi

p>

20%

0.23

70.

425

0.33

6+0.

152+

−0.

466+

−0.

214+

0.74

1+0.

602+

1.00

08.

Fore

ign

owne

rshi

p>

20%×

boar

din

depe

nden

ce0.

033

0.09

70.

278+

0.55

3+−

0.33

7+−

0.19

2+0.

389+

0.39

9+0.

593+

1.00

0

9.L

ogof

tota

lass

ets

18.3

941.

489

0.64

5+0.

008

−0.

368+

0.18

4+0.

238+

0.35

8+0.

248+

0.27

3+1.

000

10.R

ecei

vabl

esan

din

vent

ory/

tota

lass

ets

0.33

60.

170

−0.

074+

−0.

093+

0.05

6+0.

014

−0.

034+

−0.

068+

−0.

061+

−0.

073+

−0.

007

1.00

0

11.B

ig4

audi

tor

0.82

50.

380

0.22

7+0.

057+

−0.

115+

0.07

4+0.

062+

0.07

7+0.

041+

0.09

5+0.

205+

−0.

032+

1.00

012

.Lon

g-te

rmde

bt/to

tala

sset

s0.

102

0.11

70.

244+

0.03

2+−

0.01

1−

0.00

40.

022+

0.00

8−

0.01

40.

051+

0.18

2+−

0.27

5+−

0.02

4+1.

000

13.N

etin

com

e/to

tala

sset

s0.

055

0.06

4−

0.09

0+−

0.02

*−

0.15

3+0.

057+

0.05

2+0.

087+

0.06

6+0.

111+

−0.

032+

−0.

030+

0.15

6+−

0.12

5+1.

000

14.L

arge

stdo

mes

ticow

ner

0.20

10.

157

−0.

148+

0.07

5+0.

089+

−0.

026*

−0.

063+

−0.

035+

−0.

077+

−0.

038+

−0.

235+

−0.

061+

−0.

012

−0.

087+

0.10

5+1.

000

15.F

irm

risk

0.03

20.

022

0.22

1+0.

071+

−0.

242+

0.09

3+0.

110+

0.14

3+0.

188+

0.13

5+0.

201+

0.21

40.

023

0.09

3+0.

108+

−0.

035+

1.00

016

.Poo

rpe

rfor

man

ce−

0.00

30.

068

−0.

009

0.03

4+0.

012

−0.

019

−0.

001

0.00

90.

006

0.00

7−

0.01

40.

040+

−0.

018

−0.

013

0.36

5+0.

010

−0.

007

N=

6,82

3+

sign

ifica

ntat

0.01

perc

ent;

*sig

nific

anta

t0.0

5pe

rcen

t



Table 2. Results of regression analyses considering four levels of foreign ownership

GLS RE GLS RE GLS RE GLS RE GLS FEDependent variable: audit fees Model 1 Model 2 Model 3 Model 4 Model 5

Board independence 0.217*** 0.098 0.135(0.072) (0.080) (0.114)

Foreign ownership: 10–20% 0.020 0.016 0.018(0.020) (0.024) (0.0247)

Foreign ownership: 20–30% 0.031 −0.0102 −0.011(0.031) (0.039) (0.035)

Foreign ownership >30% 0.086* 0.041 −0.002(0.044) (0.048) (0.049)

Foreign ownership 10–20%× board independence 0.035 0.057(0.131) (0.136)

Foreign ownership 20–30%× board independence 0.338** 0.297**

(0.163) (0.153)Foreign ownership >30%× board independence 0.398** 0.357**

(0.174) (0.181)Log of total assets 0.363*** 0.407*** 0.398*** 0.398*** 0.201***

(0.009) (0.013) (0.014) (0.014) (0.034)Receivables and inventory/total assets 0.218*** 0.175** 0.204** 0.205** 0.186

(0.083) (0.087) (0.088) (0.078) (0.126)Big 4 auditor 0.342*** 0.272*** 0.271*** 0.270*** 0.278

(0.030) (0.028) (0.028) (0.029) (0.192)Long-term debt/total assets 0.540*** 0.421*** 0.448*** 0.436*** 0.369***

(0.142) (0.119) (0.108) (0.107) (0.129)Net income/total assets −1.343*** −0.180 −0.236 −0.229 −0.079

(0.169) (0.145) (0.148) (0.147) (0.130)Constant −3.009*** −5.201*** −5.033*** −5.014*** −0.669

(0.258) (0.287) (0.294) (0.295) (0.632)Year dummies No Yes Yes Yes YesIndustry dummies No Yes Yes Yes NoFirm dummies No No No No YesR2 - within 0.0260 0.4443 0.4392 0.4397 0.4444R2 - between 0.6063 0.7173 0.7206 0.7215 0.6163R2 - overall 0.4841 0.6686 0.6609 0.6618 0.5446

N = 6,823***p< 0.01; **p< 0.05; *p< 0.10, based on two-tailed tests, robust standard errors, clustered at the firm level in parentheses. Allindependent variables are lagged by one term.

holds across all firms, or whether it is driven bythe subsample of firms with higher levels of foreignownership (as suggested by our Hypothesis 2). Weexplore this issue in Model 4.

Model 4 in Table 2 uncovers whether board mon-itoring behavior is only present when the level offoreign ownership is high, by introducing the inter-action term between board independence and ourfour categories of foreign ownership. Interestingly,now the coefficient of board independence becomesnonsignificant, suggesting that the general effectof board independence disappears when introduc-ing the interaction terms. Thus, our results demon-strate that adding independent directors does notlead to more monitoring as a general rule. In line

with our Hypothesis 2, the coefficient of the inter-action terms is significant only when foreign own-ership is relatively high, lending support to the ideathat board monitoring behavior is contingent on thelevel of foreign ownership. Specifically, we find evi-dence that independent directors play a monitoringrole only when the level of foreign ownership isabove 20 percent. Our results show that, for firmswith low levels of foreign ownership, independentdirectors do not influence the audit fees. Finally,the coefficients of foreign ownership are not signif-icant, which is consistent with the idea that foreignshareholders do not interact directly with exter-nal auditors but, instead, focus on changing boardmonitoring.



Table 3. Results of regression analyses considering two levels of foreign ownership

GLS RE GLS FE IV -second stageDependent variable: audit fees Model 6 Model 7 Model 8

Board independence 0.115 0.109 0.120(0.083) (0.101) (0.084)

Foreign ownership >20% −0.015 −0.030 0.064(0.033) (0.027) (0.077)

Foreign ownership >20%× board independence 0.328** 0.293** 1.084***

(0.152) (0.128) (0.229)Log of total assets 0.404*** 0.206*** 0.414***

(0.013) (0.034) (0.062)Receivables and inventory/total assets 0.198** 0.183 0.216***

(0.088) (0.126) (0.062)Big 4 auditor 0.270*** 0.285 0.267***

(0.028) (0.192) (0.017)Long-term debt/total assets 0.424*** 0.366*** 0.489***

(0.101) (0.129) (0.082)Net income/total assets −0.206 −0.073 −0.311*

(0.146) (0.130) (0.166)Constant −5.111*** −0.669 −5.199***

(0.288) (0.632) (0.282)Year dummies Yes Yes YesIndustry dummies Yes No YesFirm dummies No Yes NoR2 - within 0.4396 0.4442R2 - between 0.7208 0.6130R2 - overall 0.6612 0.5430Centered R2 0.6619Kleibergen-Paap rk Wald F-statistic (weak identification test) 29.764***

F-statistic - foreign ownership >20% 45.763***

F-statistic - foreign ownership >20%× board independence 15.754***

Shea’s partial R2 - foreign ownership >20% 0.1617Shea’s partial R2 - foreign ownership >20%× board independence 0.3197Kleibergen-Paap rk LM statistic (underidentification test) 59.488***

Test of endogeneity (GMM C-statistic) 11.542***

N = 6,823***p< 0.01; **p< 0.05; *p< 0.10, based on two-tailed tests, robust standard errors, clustered at the firm level in parentheses. Allindependent variables are lagged by one term.

Model 5 in Table 2 employs a FE estimationwith time fixed effects, and yields very similarresults, demonstrating that board monitoring onlyoccurs when foreign ownership is above 20 percent,while no such behavior is observed for lower levelsof foreign ownership. Our results show a verysimilar effect of foreign ownership for the twotop categories using both random and fixed effectsestimations, which justifies a 20 percent cut-offpoint. For the following analyses, we use below andabove 20 percent to account for foreign ownership.

Models 6 and 7 in Table 3 consider only twocategories of foreign ownership, and the resultsfor both GLS RE and GLS FE are consistent. Inaddition, both estimations support our cut-off pointat 20 percent. Again, board independence is only

significantly related to audit fees when foreignownership is high, while the coefficients of boardindependence and foreign ownership are not signif-icant when foreign ownership is below 20 percent.

To mitigate additional endogeneity concerns orcorrelated omitted variables issues, and to increaseconfidence in the directionality of our results,Model 8 in Table 3 presents the results of the sec-ond stage regression using instrumental variables.For our data, the model is uniquely identified. TheF-statistic and Shea’s partial R-squared, obtainedfrom the first-stage regressions, are high, indicatingthat the instruments are relevant and strong.4

4 The tables of the first-stage regressions are available uponrequest.



Tabl

e4.

Res

ults

ofre

gres

sion

anal

yses

cons

ider

ing

cont

inge

ncie

s(I

Vse

cond

-sta

gere

sults

)

SH1

dom

estic

aSH

1do

mes

tica

Firm

risk

aFi

rmri

ska

Perf

orm

ance

aPe

rfor

man

cea

<m

edia

n>

med

ian

>m

edia

n<

med

ian

<m

edia

n>

med

ian

Dep

ende

ntva

riab

le:a

udit

fees

Mod

el9

Mod

el10

Mod

el11

Mod

el12

Mod

el13

Mod

el14

Boa

rdin

depe

nden

ce−

0.03

40.

147

0.07

20.

167

−0.

078

0.06

4(0

.179

)(0

.096

)(0

.152

)(0

.147

)(0

.137

)(0

.141

)Fo

reig

now

ners

hip>

20%

0.02

60.

054

0.07

2−

0.01

7−

0.05

8−

0.05

2(0

.149

)(0

.074

)(0

.134

)(0

.107

)(0

.127

)(0

.120

)Fo

reig

now

ners

hip>

20%

×bo

ard

inde

pend

ence

1.37

0***

0.64

5***

1.15

3***

0.59

2**1.

634**

*0.

751**

(0.3

91)

(0.2

17)

(0.3

17)

(0.2

77)

(0.4

24)

(0.3

61)

Log

ofto

tala

sset

s0.

462**

*0.

348**

*0.

414**

*0.

380**

*0.

418**

*0.

425**

*

(0.0

25)

(0.0

13)

(0.0

24)

(0.0

20)

(0.0

20)

(0.0

24)

Rec

eiva

bles

and

inve

ntor

y/to

tala

sset

s0.

250**

0.28

1***

0.35

8***

0.32

5***

0.40

9***

0.30

3***

(0.1

04)

(0.0

64)

(0.1

12)

(0.0

92)

(0.0

94)

(0.1

05)

Big

4au

dito

r0.

267**

*0.

273**

*0.

283**

*0.

251**

*0.

223**

*0.

314**

*

(0.0

23)

(0.0

21)

(0.0

28)

(0.0

21)

(0.0

24)

(0.0

25)

Lon

g-te

rmde

bt/to

tala

sset

s0.

438**

*0.

301**

*0.

301**

0.59

7***

0.26

8**0.

655**

*

(0.1

55)

(0.1

02)

(0.1

42)

(0.1

21)

(0.1

02)

(0.1

52)

Net

inco

me/

tota

lass

ets

−0.

429

−0.

074

−0.

221

−0.

528**

−0.

744

−0.

091

(0.2

80)

(0.1

90)

(0.2

78)

(0.2

68)

(0.4

57)

(0.1

41)

Con

stan

t−

6.07

8***

−4.

561**

*−

3.80

5***

−4.

579**

*−

5.22

7***

−5.

049**

*

(0.5

24)

(0.3

27)

(0.4

17)

(0.4

27)

(0.4

25)

(0.4

89)

Yea

rdu

mm

ies

Yes

Yes

Yes

Yes

Yes

Yes

Indu

stry

dum

mie

sY

esY

esY

esY

esY

esY

esN

obse

rvat

ions

3,25

33,

291

3,15

53,

291

3,28

73,

257

Cen

tere

dR

20.

6750

0.66

190.

6790

0.65

870.

4611

50.

5620

5K

leib

erge

n-Pa

aprk

Wal

dF

-sta

tistic

(wea

kid

entifi

catio

nte

st)

23.1

47**

*29

.243

***

23.1

47**

*26

.854

***

27.5

82**

*23

.658

***

F-s

tatis

tic-f

orei

gnow

ners

hip>

20%

44.1

26**

*45

.253

***

44.2

89**

*44

.926

***

46.0

24**

*45

.475

***

F-s

tatis

tic-f

orei

gnow

ners

hip>

20%

×bo

ard

Inde

pend

ence

14.8

54**

*15

.125

***

14.7

12**

*14

.787

***

15.9

85**

*15

.689

***

Shea

’spa

rtia

lR2

-for

eign

owne

rshi

p>

20%

0.15

750.

1421

0.16

250.

1383

0.17

200.

1469

Shea

’spa

rtia

lR2

–fo

reig

now

ners

hip>

20%

×bo

ard

Inde

pend

ence

0.32

980.

2894

0.31

420.

3159

0.32

730.

3151

Kle

iber

gen-

Raa

prk

LM

(und

erid

entifi

catio

nte

st)

54.1

13**

*54

.256

***

54.1

13**

*55

.128

***

61.3

45**

*53

.654

***

Test

ofen

doge

neity

(GM

MC

-sta

tistic

)14

.452

***

15.5

82**

*13

.245

***

17.1

25**

*14

.256

***

16.2

15**

*

aSH

1do

mes

ticis

mea

sure

das

the

tota

lsh

areh

oldi

ngs

byth

ela

rges

tdo

mes

ticsh

areh

olde

r,fir

mri

skis

the

annu

aliz

edst

anda

rdde

viat

ion

ofw

eekl

yeq

uity

retu

rns,

and

perf

orm

ance

isth

ede

viat

ion

from

prio

rpe

rfor

man

ce,c

aptu

red

asth

ecu

rren

tRO

Aco

mpa

red

with

the

aver

age

ofth

epr

ior

two

year

s’R

OA

.**

*p<

0.01

;**p

<0.

05;*

p<

0.10

,bas

edon

two-

taile

dte

sts,

robu

stst

anda

rder

rors

,clu

ster

edat

the

firm

leve

lin

pare

nthe

ses.

All

inde

pend

entv

aria

bles

are

lagg

edby

one

term

.



Finally, we also report a test of endogeneity forthe instrumented variables, which confirms thatthe variable presumed to be endogenous in ourmodel cannot be treated as exogenous. From theIV estimates in Model 8, in Table 3, we see that thecoefficients of the interaction term are positive andsignificant at the 1 percent level, with even largermagnitudes than the coefficient estimates fromthe GLS RE or FE regressions. Thus, the effect offoreign ownership on the board independence-auditfee relationship remains and is, in fact, strength-ened after addressing the potential endogeneityproblem. These results provide strong evidence toour Hypothesis 2 suggesting that the monitoringrole of independent directors is only activated whenforeign ownership is high (above 20%). Moreover,such monitoring behavior is absent when foreignownership is low, as the coefficient of boardindependence is not significant. In economic terms,an increase in board independence by one standarddeviation will lead to a 15 percent higher audit feewhen foreign ownership is high, while a similarincrease will have no effect on audit fees whenforeign ownership is low. Finally, our results revealno significant effect of foreign ownership on auditfees, which is consistent with the idea that foreigninvestors change the monitoring dynamics of theboard, rather than interacting directly with the exter-nal auditor. Overall, our findings give strong supportto our argument that foreign owners are likely tochange board monitoring in foreign-held firms toresemble those of their home-based CG system.

In Table 4, we test our Hypotheses 3a–c, in whichwe argue that the influence of foreign ownershipon board monitoring is unlikely to be homogenousacross firms. We focus on three main firm charac-teristics providing additional depth to our analysis.First, we examine the influence of domestic own-ership concentration as a key potential deterrentof changes in board monitoring. Our results, usinginstrumental variables, in Table 4 (Models 9 and 10)provide support for our hypothesis that the effect offoreign ownership on board monitoring is lower infirms with large domestic owners. While the overalleffect of foreign ownership on board monitoringremains in both subsamples, the effect is strongerin the absence of large domestic owners, comparedto when large domestic owners are present. Next,we focus on two key firm characteristics that couldimpact the incentives or ability of foreign own-ership to activate board monitoring: firm risk andperformance. Again, we split our sample in two,

i.e., above and below the median value of firm risk(Models 11 and 12) and firm performance (Models13 and 14), respectively. Our results show thatthe influence of foreign ownership on monitoringby independent directors is especially strong infirms with above-median risk and below-medianperformance. These results lend support to the ideathat a higher level of risk increases foreign owners’incentive to activate board monitoring, while a lowlevel of performance increases their ability andincentives to introduce changes in the monitoringrole of the board.

Robustness tests

We conduct a series of robustness tests. As a robust-ness test with respect to our instruments, we use the1997 Asian crisis as an exogenous event that led,in part, to the fast increase of foreign ownershipin Japan. After the banking crisis, and particularlyas of 1999, banks offset their losses by realiz-ing capital gains on long-held stocks (Hoshi andKashyap, 2004), reducing shareholding mainly byselling firms with high market valuations (Miyajimaand Kuroki, 2007). In line with these findings, weuse the accounting and stock market performanceover the period 1997–1998 as alternative instru-ments that explain the level of foreign ownership,and we find consistent results for both alternativeinstruments. Our results also hold for longer timeperiods, to account for both the banking and the1997 Asian crisis.

We next look for alternative explanations of ourresults. First, we examine the alternative hypothesisthat our results could be driven by higher levels ofindependent directors in firms with a high level offoreign ownership, i.e., that foreign owners enhanceboard monitoring by adding independent directorsto the board, rather than enhancing monitoring.Using the same control variables to explain auditfees, we have changed our dependent variable toboard independence, and we are particularly inter-ested in the effect of foreign ownership. Our resultsdo not reveal a significant relationship between for-eign ownership and board independence. In addi-tion, we test for possible endogeneity betweenboard independence and audit fees, and corroboratethat board independence is exogenous and that it isnot determined by audit fees.

In order to address the potential concern thatour results could be driven by a too-broad defini-tion of independent directors, we have constructed



a narrower definition of independent directors. Wehave gone through every firm’s corporate gover-nance report issued for our sample period to identifythe proportion of independent directors that maynot be truly independent—i.e., we excluded inde-pendent directors who fall into one of the five TSEcategories not considered truly independent. In linewith our previous results, we do not find that foreignownership explains the level of board independence(narrow), after controlling for other factors. Moreimportantly, our findings corroborate that the mon-itoring behavior of independent directors, narrowlydefined, is only present in firms with a large stake offoreign ownership. This finding holds when usingrandom and fixed effects models and when usinginstrumental variables.

We have replicated our analyses eliminating allfirms that adopted the board committee system(removing 126 observations, out of 6,767). Whileeven for these firms it is not mandatory to have themajority of the board be outside directors, the threecommittees must have a majority of outside direc-tors, which raises the level of board independenceindirectly. Our results are unaffected in all specifi-cation when removing these observations.

Finally, our results on the relative influence offoreign ownership on board monitoring are robust toalternative cut-off points and alternative definitionsof domestic ownership concentration, firm risk, andperformance.

DISCUSSION

This paper espouses the view that the effectivenessof corporate governance practices must be exam-ined in light of the institutional context, as wellas the ownership composition of the firm. Drawingon a contingency approach, which conceptualizescorporate governance as a system of interrelatedelements having strategic and institutional comple-mentarities, we explore how CG dynamics change,when shareholder-oriented foreign owners, withdistinct objectives and preferred governance prac-tices, coexist with (stakeholder-oriented) domesticshareholders, who rely on a different set of gover-nance practices. We contend that board monitoringdepends on the composition of shareholders as wellas the heterogeneity of the shareholders’ objectivesand influence over governance practices.

When investors from a shareholder systeminvest in a stakeholder system, we argue that their

interests clash with the stakeholder logic. Becausemonitoring channels used by domestic stakeholdersmay be unavailable or insufficient for foreigninvestors, they seek to protect their investmentby introducing corporate governance practicescommon in the Anglo-American context within theexisting stakeholder CG context. In particular, weanalyze differences in board monitoring behavior,which we capture as the effect of board indepen-dence on external audit fees. We claim that themonitoring behavior of independent directors willbe contingent on the degree of foreign ownership.When foreign ownership is high, we expect thatindependent directors will have greater incentivesto protect shareholders’ interests by monitoringand communicating concerns about internal controlweaknesses and other accounting issues to externalauditors. This exchange will broaden the auditscope and will result in higher audit fees. Incontrast, independent directors in firms with a lowproportion of foreign ownership will not show thesame monitoring behavior.

Using a large sample of Japanese listed firms,our findings support our claims by demonstratingthat monitoring behavior of independent directorsis contingent on the level of foreign ownership. Weare able to demonstrate that foreign shareholderscan change board monitoring dynamics when theyreach a critical mass. Our findings also uncoverthat the influence of foreign ownership on boardmonitoring is stronger in the absence of largedomestic owners, when firm risk is higher, andwhen firms become less profitable. To deal withendogeneity concerns, we use fixed effects regres-sion and instrumental variables and find consistentresults. We also conduct additional tests to rule outalternative explanations.

Our research offers critical insights for the com-parative corporate governance and strategic man-agement literature in several ways. First, we extendRediker and Seth’s (1995) concept of governancebundles from a single decision-maker perspectiveto a setting where multiple shareholders influencea (partial) set of practices embedded in an exist-ing system, demonstrating the relevance of for-eign ownership as a key contingency in under-standing boards’ monitoring role. In this sense, ouranalysis also expands on Hillman and Dalziel’s(2003) framework on monitoring effectiveness, asour findings demonstrate that exploring the natureof ownership allows for a better understanding ofboth directors’ incentives and ability to monitor



top management. We argue that it is important torecognize that the existing bundle of governancemechanisms designed to protect domestic owners’interest is unlikely to be equally effective for for-eign investors, and, as a result, they want to intro-duce additional governance practices. Our findingsuncover the possibility of hybrid systems, in whichJapanese corporations with a high degree of foreignownership combine elements common to both theJapanese CG context, e.g., a board with a major-ity of inside directors, and the Anglo-Americancontext, e.g., monitoring by independent directors.As such, we provide new insights on the substi-tute/complementary nature of CG practices.

Second, our results make an important contribu-tion to debates on globalization and convergence ofCG systems, by showing how foreign investors area channel through which convergence occurs. Ouranalysis reveals that convergence happens withinthe boundaries of the existing corporate governancesystem, as change in board monitoring behavioris still at work, even when the vast majority ofJapanese firms did not make the switch to thecommittee system common in the shareholder CGmodel. Our study also adds to the comparative CGliterature and the call by Aguilera, Desender, andKabbach de Castro (2011) to shift our conceptual-ization of governance systems beyond the dichoto-mous world of common law/shareholder-orientedsystem vs. civil law/stakeholder-oriented system. Ifforeign ownership is able to shape CG bundles, it isdifficult to continue to equate firm nationality withgovernance systems.

Third, prior studies on the influence of foreignownership have mainly focused on outcomes, ratherthan processes (e.g., Ahmadjianand Robbins, 2005;Ahmadjian and Robinson, 2001; David et al., 2006;Yoshikawa et al., 2005). Our study contributes toa better understanding of the mechanisms throughwhich foreign ownership influences local organiza-tional practices. Considering Japanese listed firmsalso allows for an exploration of the external valid-ity of the link between board independence andaudit fees in a stakeholder-oriented environment.

Our research also has implications for poli-cymakers and in particular for the CG reformsundertaken following the recent Japanese andworldwide accounting scandals. Our findings speakto the fact that universalistic policy prescriptionsmerely focusing on enhancing board independencemay not lead to better monitoring of managementin general. While it may be a necessary condition,

we uncover that it is not a sufficient conditionto enhance board monitoring. In practical terms,efforts could focus on greater information accessto strengthen monitoring processes and greateremphasis on the role of independent directorsand board monitoring, through the establishmentof a corporate code of best practice combinedwith a “comply or explain” requirement for listedcompanies. Recognizing the diversity of gover-nance needs among Japanese listed companiescould enable the development of a “mixed” or“hybrid” model that incorporates a greater elementof management supervision into Japan’s traditionalcorporate structure.

Our research also opens interesting venues forfuture research. First, it would be worthwhile toinvestigate the influence of foreign ownership onother aspects of corporate governance bundles.For example, since June 2010, Japanese firms arerequired to disclose compensation details for exec-utives and board members who receive more than¥100 million (about $1.1 million). This disclosurerequirement opens possibilities to evaluate howforeign ownership might influence the design ofcompensation packages. In addition, it would beintriguing to examine whether practices introducedby foreign investors create spillover effects, overtime, in firms without large levels of foreign owner-ship. Second, while we study how Anglo-Americaninstitutional foreign owners affect board practices infirms in a stakeholder economy, we establish, moregenerally, that heterogeneity in shareholders’ objec-tives (and logic) and their span of control over gov-ernance practices are important to understand theinfluence of different types of owners over existinggovernance arrangements.

Our framework is therefore not only usefulto analyze the influence of shareholder-orientedforeign ownership on board monitoring in astakeholder-oriented context, but it may also beused in other contexts. For example, acknowledgingdifferences in shareholder’s objectives and pref-erences for certain governance mechanisms mayhelp explain how the introduction of institutionalinvestors in family-controlled firms may lead tochanges in the firm’s governance practices. Whilefamily owners may have preferences over directmonitoring of managers, institutional investors maywant to reinforce board monitoring or link executivecompensation more closely to firm performance.Furthermore, we believe that the nature of foreignownership in terms of type and origin is likely



to affect their influence. Future research couldtherefore explore whether the effect of foreignownership on corporate governance is contingenton the type and degree of shareholder/stakeholderorientation of the foreign owner.

Our study has limitations as well. First,we focus on listed companies from a highlystakeholder-oriented context like Japan. Our resultsmay therefore not generalize to non-listed com-panies or to firms with a two-tier board system asin Germany. We expect our findings to be relevantin those settings where the bundle of governancepractices deployed by domestic stakeholders isunavailable or does not address the agency conflictthat shareholder-oriented foreign owners face.Given the potential influence of institutions in oursetting, future research could benefit from testingour framework in different institutional settings. Inaddition, while our study highlights the influenceof foreign ownership on the behavior of the boardof directors with respect to monitoring, we have notdiscussed other board functions such as advising.

CONCLUSION

We build on the contingency approach to CG,which proposes that effective CG depends uponthe alignment of interdependent organizationaland environmental characteristics, rather than onone universal set of relationships that hold acrossall organizations. Our study suggests that boardmonitoring must also be examined in light ofcontingencies related to firms’ foreign ownership.Our findings demonstrate that, for a large set oflisted Japanese firms, board monitoring behavior iscontingent upon the degree of foreign ownership.We uncover that the relationship between boardindependence and audit fees is positive only forhigh levels of foreign ownership, while this rela-tionship is not significant for lower levels of foreignownership. We also show that this relationship isstrengthened when domestic ownership concentra-tion is low, firm risk is high, and firm performanceis low. Our results highlight the possibility ofdifferent patterns of CG within a given country,shaped by the nature and weight of foreign owners.

ACKNOWLEDGEMENTS

The authors thank Christina Ahmadjian, theeditor Will Mitchell, two anonymous reviewers,

and the participants in seminars and workshopsat the Academy of Management, the Strate-gic Management Society, Wharton, Universityof Illinois, Kennesaw State University, North-eastern University, Singapore ManagementSchool, ESADE Business School, University ofLeuven, and Copenhagen Business School fortheir helpful comments on earlier drafts. Theauthors also acknowledge support from the Fun-dación Juan March, Fundación Ramon Areces,Projects ECO2010-22105-C03-03, ECO2010-21393-C04-01, ECO2010-21393-C04-02, andECO2013-48328-C03-3-P financed by the SpanishMinistry of Science and Innovation.

REFERENCES

Abbott L, Parker JS, Peters GF, Raghunandan K. 2003.An empirical investigation of the audit fees, nonauditfees, and audit committees. Contemporary AccountingResearch 20(2): 1–21.

Adams RB, Hermalin BE, Weisbach MS. 2010. Therole of boards of directors in corporate governance: aconceptual framework & survey. Journal of EconomicLiterature 48(1): 58–107.

Aguilera RV, Desender K, Bednar MK, Lee J. Connectingthe dots – bringing external corporate governance intothe corporate governance. Academy of ManagementAnnals in press.

Aguilera RV, Desender K, Kabbach de Castro LR. 2011. Aconfigurational approach to comparative corporate gov-ernance. In Sage Handbook of Corporate Goverance,Clarke T, Branson D (eds). Sage Publications: London,UK.

Aguilera RV, Filatotchev I, Gospel H, Jackson G. 2008.Contingencies, complementarities, and costs in corpo-rate governance models. Organization Science 19(3):475–492.

Aguilera RV, Jackson G. 2003. The cross-nationaldiversity of corporate governance: dimensions anddeterminants. Academy of Management Review 28(3):447–465.

Ahmadjian CL, Robbins GE. 2005. A clash of capi-talisms: foreign shareholders and corporate restructur-ing in 1990s Japan. American Sociological Review 70:451–471.

Ahmadjian CL, Robinson P. 2001. Safety in numbers:downsizing and the deinstitutionalization of permanentemployment in Japan. Administrative Science Quar-terly 46: 622–654.

American Institute of Certified Public Accountants. State-ments on auditing standards, AU Section 312-380.

Aoki M, Jackson G, Miyajima H. 2007. Corporate Gov-ernance in Japan: Institutional Change and Organiza-tional Diversity. Oxford University Press: Oxford, UK.

Aoki M, Patrick H. 1994. The Japanese Main BankSystem: Its Relevance for Developing and TransformingEconomies. Oxford University Press: Oxford, UK.



Baltagi BH. 2005. Econometric Analysis of Panel Data.John Wiley and Sons: Chichester, UK.

Baltagi BH, Wu PX. 1999. Unequally spaced panel dataregression with AR(1) disturbances. Econometric The-ory 15: 814–823.

Bank of Japan. 1996–2012. Honpo Taigai Shisan FusaiZandaka, Shoken Toshi Tou (Po-Toforio Toshi) Zan-daka Chiikibetsu Tokei (Japan’s Outstanding OverseasAssets and Liabilities: Statistics on Outstanding EquitySecurities (Portfolio Investment) by Region). Bank ofJapan: Tokyo, Japan.

Bank of Japan. 1996–2011. Honpo Taigai Shisan FusaiZandaka, Shoken Toshi Tou (Po-Toforio Toshi) Zan-daka Chiikibetsu Tokei (Japan’s Outstanding OverseasAssets and Liabilities: Statistics on Outstanding EquitySecurities (Portfolio Investment) by Region). Bank ofJapan: Tokyo, Japan.

Bank of Japan. 2012. Honpo Taigai Shisan Fusai Zandaka,Shoken Toshi Tou (Po-Toforio Toshi) Zandaka Chiiki-betsu Tokei (Japan’s Outstanding Overseas Assets andLiabilities: Statistics on Outstanding Equity Securi-ties (Portfolio Investment) by Region). Bank of Japan:Tokyo, Japan.

Buckley PJ. 1997. Cooperative form of transnational cor-poration activity. In Transnational Corporations andWorld Development, Dunning JH, Sauvant KP (eds).Thomson: London, UK; 473–493.

Campbell JL. 2004. Institutional Change and Globaliza-tion. Princeton University Press: Princeton, NJ.

Carcello JV, Hermanson DR, Neal TL, Riley RR Jr. 2002.Board characteristics and audit fees. ContemporaryAccounting Research 19(3): 365–384.

Chizema A, Shinozawa Y. 2011. The ‘company with com-mittees’: change or continuity in Japanese corporategovernance? Journal of Management Studies 48: 1–25.

Cohen JR, Krishnamoorthy G, Wright AM. 2007. Theimpact of roles of the board on Auditors’ risk assess-ments and program planning decisions. Auditing: AJournal of Practice & Theory 26(1): 91–112.

Collier P, Gregory A. 1996. Audit committee effectivenessand the audit fee. The European Accounting Review5(2): 177–198.

Colpan AM, Yoshikawa T, Hikino T, Del Brio EB. 2011.Shareholder heterogeneity and conflicting goals: strate-gic investments in the Japanese electronics industry.Journal of Management Studies 48: 591–618.

David P, Yoshikawa T, Chari MDR, Rasheed A. 2006.Strategic investments in Japanese corporations: do for-eign portfolio owners foster underinvestment or appro-priate investment? Strategic Management Journal 27:591–600.

Deakin S. 2010/2011. What directors do (and fail to do):some comparative notes on board structure and corpo-rate governance. New York Law School Law Review 55:525–541.

Desender KA, Aguilera RV, Crespi R, García-cestona M.2013. When does ownership matter? Board characteris-tics and behavior. Strategic Management Journal 34(7):823–842.

Deutsch Y, Keil T, Laamanen T. 2011. A dual agencyview of board compensation: the joint effects of outside

director and CEO stock options on firm risk. StrategicManagement Journal 32(2): 212–227.

Finkelstein S, Hambrick DC. 1996. Strategic Leadership:Top Executives and Their Effects on Organizations.West: Minneapolis /St. Paul, MN.

Fiss PC, Zajac FJ. 2004. The diffusion of ideas: the (non)adoption of a shareholder value orientation amongGerman firms. Administrative Science Quarterly 49:501–534.

Gedajlovic E, Yoshikawa T, Hashimoto M. 2005.Ownership structure, investment behaviour and firmperformance in Japanese manufacturing industries.Organization Studies 26: 7–35.

Gilson RJ, Milhaupt CJ. 2005. Choice as regulatoryreform: the case of Japanese corporate governance.American Journal of Comparative Law 53: 343–377.

Greenwood R, Raynard M, Kodeih F, Micellota E, Louns-bury M. 2011. Institutional complexity and organiza-tional responses. Annals of the Academy of Manage-ment 5(1): 1–55.

Hall PA, Soskice DW. 2001. Varieties of Capitalism: TheInstitutional Foundations of Comparative Advantage.Oxford University Press: Oxford, UK.

Hannan MT, Freeman JH. 1984. Structural inertia andorganisational change. American Sociological Review49: 149–164.

Han JK, Kim N, Srivastava RK. 1998. Market orientationand organizational performance: Is innovation a miss-ing link? Journal of Marketing 62: 30–45.

Hay D, Knechel WR, Ling H. 2008. Evidence on theimpact of internal control and corporate governance onaudit fees. International Journal of Auditing 12: 9–24.

Hay D, Knechel WR, Wong N. 2006. Audit fees: ameta-analysis of the effect of supply and demandattributes. Contemporary Accounting Research 23(1):141–191.

Hillman AJ, Dalziel T. 2003. Boards of directors and firmperformance: integrating agency and resource depen-dence perspectives. Academy of Management Review28(3): 383–396.

Hoshi T, Kashyap AK. 2004. Japan’s financial crisis andeconomic stagnation. Journal of Economic Perspec-tives 18(1): 3–26.

Japanese Institute of Certified Public Accountants. 2009.Accounting and auditing in Japan. Available at:http://www.hp.jicpa.or.jp/english/accounting/system/index.html (accessed 15 January 2014).

Kang J, Stulz R. 1997. Why is there a home bias? Ananalysis of foreign portfolio equity ownership in Japan.Journal of Financial Economics 46: 3–28.

Kaplan SN, Minton BA. 1994. Appointments of out-siders to Japanese board determinants and implica-tions for managers. Journal of Financial Economics 36:225–258.

Kleibergen F, Paap R. 2006. Generalized reduced ranktests using the singular value decomposition. Journalof Econometrics 133(1): 97–126.

Kosnik R. 1987. Greenmail: a study of board performancein corporate governance. Administrative Science Quar-terly 32(2): 163–185.



Kubo K. 2005. Executive compensation policy and com-pany performance in Japan. Corporate Governance: AnInternational Review 13: 429–436.

Laeven L, Levine R. 2009. Bank governance, regulationand risk taking. Journal of Financial Economics 93(2):259–275.

Lincoln J, Gerlach ML. 2004. Japan’s Network Economy:Structure, Persistence, and Change. Cambridge Univer-sity Press: Cambridge, UK.

Miller D, Chen MJ. 1994. Sources and consequences ofcompetitive inertia: a study of the US airline industry.Administrative Science Quarterly 39: 1–23.

Miyajima H, Kuroki F. 2007. The unwinding ofcross-shareholding in Japan: causes, effects, andimplications. In Corporate Governance in Japan, AokiM, Jackson G, Miyajima H (eds). Oxford UniversityPress: New York; 79–124.

Numato S, Takeda F. 2010. Stock market reactions to auditfailure in Japan: the case of Kanebo and ChuoAoyama.International Journal of Accounting 45(2): 175–199.

Rediker KJ, Seth A. 1995. Boards of directors and sub-stitution effects of alternative governance mechanisms.Strategic Management Journal 16: 85–99.

Shleifer A, Vishny R. 1997. A survey of corporate gover-nance. Journal of Finance 52: 737–783.

Simunic DA. 1980. The pricing of audit services: theoryand evidence. Journal of Accounting Research 18(1):161–190.

Streeck W. 2001. Introduction: explorations into the ori-gins of nonliberal capitalism in Germany and Japan.In The Origins of Nonliberal Capitalism, Streeck W,Yamamura K (eds). Cornell University Press: Ithaca,NY; 1–38.

Tuggle CS, Sirmon DG, Reutzel CR, Bierman L.2010. Commanding board of director attention:investigating how organizational performance andCEO duality affect board members’ attention tomonitoring. Strategic Management Journal 31(9):946–968.

Wooldridge JM. 2002. Econometric Analysis of CrossSection and Panel Data. MIT Press: Cambridge, MA,London, UK.

Yoshikawa T, McGuire J. 2008. Change and continuity inJapanese corporate governance. Asia Pacific Journal ofManagement 25: 5–24.

Yoshikawa T, Phan PH, David P. 2005. The impactof ownership structure on wage intensity inJapanese corporations. Journal of Management 31:278–300.

Yoshimori M. 1995. Whose company is it? The conceptof the corporation in Japan and the west. Long RangePlanning 28(4): 33–44.

Zajac EJ, Westphal JD. 1994. The costs and benefits ofmanagerial incentives and monitoring in the largestU.S. corporations: when is more not better? StrategicManagement Journal 15: 121–142.


a clash of governance logics: foreign ... employees as key firm stakeholders...

Documents