Director Reputation, This study advances research on CEO...

Director Reputation,CEO-Board Power, andthe Dynamics ofBoard Interlocks

Edward J. ZajacNorthwestern UniversityJames D. WestphalThe University of Texasat Austin

© 1996 by Cornell University.0001-8392/96/4103-0507/$! .00.

Both authors contributed equally to thepaper. We are grateful to Jerry Davis,Ranjay Gulati, Paul Hirsch, Willie Ocasio.Toby Stuart, Brian Uzzi, and seminarparticipants at Carnegie MellonUniversity, the Harvard Business School,and UCLA for comments andsuggestions on earlier versions of thispaper. The paper has also benefited fromthe helpful comments of Mark Mizruchiand the anonymous reviewers for ASO,as well as the editorial assistance ofLinda Johanson.

This study advances research on CEO-boardrelationships, interlocking directorates, and directorreputation by examining how contests forintraorganizational power can affect interorganizationalties. We propose that powerful top managers seek tomaintain their control by selecting and retaining boardmembers with experience on other, passive boards andexcluding individuals with experience on more activeboards. We also propose that powerful boards similarlyseek to maintain their control by favoring directors witha reputation for more actively monitoring managementand avoiding directors with experience on passiveboards. Hypotheses are tested longitudinally usingCEO-board data taken from 491 of the largest U.S.corporations over a recent seven-year period. Thefindings suggest that variation in CEO-board powerrelationships across organizations has contributed to asegmentation of the corporate director network. Wediscuss how our perspective can reconcile contrary viewsand debates on whether increased board control hasdiffused across large U.S. corporations.*

Have boards of directors of large U.S. corporations recentlymoved toward greater control of top managers? Recentdiscussions of corporate governance practices in the U.S.seem to vary widely in their answer to this question. Forevery research study or business press editorial pointingtoward increasingly active boards, there appears to be acorresponding study or editorial arguing the opposite (Lorschand Maclver, 1989; Davis, 1991; Fortune, 1993; Wall StreetJournal, 1995a). The theoretical explanations used to supportone or the other perspective range from a belief that publicdiscourse and investor pressures have led to the diffusion ofgreater board independence across U.S. corporations(Useem, 1992; Wall Street Journal, 1995b) to themanagerialist belief that managerial entrenchment and thecooptation of boards by powerful chief executive officers(CEOs) remain as strong as ever (Pfeffer, 1992; Wall StreetJournal, 1995c).

Rather than debate the relative merits of such sweeping andrather one-sided perspectives on corporate governance, wesuggest instead that a theory is needed that cansimultaneously explain the coexistence and persistence ofboth board-controlled and CEO-controlled firms. We developand test a theory that can explain such variation, based onan analysis of director appointments, whereby powerfulactors in the CEO-board relationship affect the diffusion ofboard independence through the selection and retention ofdirectors whose prior directorship experiences suggestdifferential sympathy for their interests. For instance, wepropose that more powerful CEOs will avoid directorcandidates who have participated as directors in increasingthe level of board monitoring and control over CEOs onother boards, while favoring new director candidates withprior directorship experience in protecting or bolstering CEOcontrol.

With few exceptions (e.g., Mizruchi and Stearns, 1988;Davis, 1993), much of the corporate governance literature

507/Administrative Science Quarterly, 41 (1996): 507-529

has taken board composition as a given, seeking to examineits implications rather than its determinants, but threedistinct streams of research on boards could be invoked toexamine the selection or retention of individual directors.One stream of research in organization behavior has typicallyexamined the extent to which boards are a passive tool ofmanagement interests (Vance, 1968; Mace, 1971; Herman,1981). Typical for this line of inquiry is examining howspecific dimensions of board structure (e.g., the proportionof outside directors or the separation of CEO and boardchair roles) can affect relative power and decision-makingtendencies in the CEO-board relationship and thus affect theallegiance of board members to management's orshareholders' interests (Kosnik, 1987; Finkelstein andHambrick, 1989; Wade, O'Reilly, and Chandratat, 1990;Mallette and Fowler, 1992). This line of research wouldsuggest that the composition of boards will be determinedlargely by the attempts of powerful individual CEOs to cooptexisting directors and influence the selection and retentionof directors who are more likely to be sympathetic to thatparticular CEO's interests (Westphal and Zajac, 1995).

A second stream of research draws largely fromorganizational sociology and is concerned with determiningthe macro-structure of boards, i.e., the network ofinterlocking directorates (Ornstein, 1980), and itsimplications, such as the diffusion of particular organizationalpractices. From this perspective, internal power andinfluence dynamics between CEOs and boards are lessrelevant than how interlocking directorates may reflectinterorganizational dependencies (Pfeffer and Salancik, 1978;Pennings, 1980; Zajac, 1988) or unity among members ofthe elite class (Koenig and Gogel, 1981; Mizruchi, 1982;Useem, 1982). Thus the best predictor of the specificcomposition of boards of directors is likely to reflect eitherinterorganizational dependencies or social ties amongmembers of the elite class. Empirical studies examiningwhether "broken" interlocks between resource-interdependent firms are reconstituted reflect an attempt toaddress such issues (Ornstein, 1980; Palmer, 1983; Palmer,Friedland, and Singh, 1986; Stearns and Mizruchi, 1986).

A third perspective on corporate governance that canaddress the question of director appointments is grounded inthe agency conception of corporate boards as a generallyeffective, if imperfect control mechanism serving to protectshareholder interests (Fama, 1980), From this financialeconomics perspective, an efficient labor market forcorporate directors acts as a motivating and discipliningdevice (Fama and Jensen, 1983). Directors seek to developand maintain a favorable reputation as active representativesof shareholder welfare, thus enhancing their human capitalon the boards on which they sit and increasing theirattractiveness as candidates for board appointments at otherfirms.

While each of these largely nonoverlapping streams ofresearch contributes to some understanding of changingboard appointments, considering both the intraorganizationaland interorganizational dimensions of board membershiptogether would provide additional insights. We develop such

508/ASQ, September 1996

Board Interlocks

a theoretical framework in this study, linking both intra- andinterorganizational dynamics involving CEOs and boards ofdirectors. This study contributes to the literature onCEO-board relationships by showing how director experienceon other boards can affect decision making (i.e., directorselection and retention) on a given focal board. It contributesto the literature on interlocking directorates by proposing andtesting a new perspective on the formation and dissolutionof board interlock ties that differs from theinterorganizational resource dependence and intraclass unityapproaches described above. While we share the view ofintraclass theorists that interlocks can ser\/e to "guidemanagerial behavior, socialize new directors into [thecapitalist class] culture, and socially control deviant behavior"(Palmer, 1983: 42), we suggest that interlocks serve thesefunctions for distinct and competing subcultures within thelarger class of business elites. Finally, it contributes to theliterature on director reputation by suggesting that while adirector's prior experience can affect subsequent boardappointments, such appointments may reflect a politicalrather than an economic rationality, in which both active andpassive board members can thrive in a labor market fordirectors that is segmented by orientation towardmanagement.

We test our perspective using longitudinal data taken fromhundreds of large U.S. corporations over a recent seven-yearperiod. Some research has suggested that this periodcoincides with the spread of increased board activism andshareholder orientation (Useem, 1992; Davis and Thompson,1994; Zajac and Westphal, 1995) and an attendant increasein conflict between management and shareholder interests.In this context, a director's reputation for activism orpassivity, along with the relative power of CEOs and boards,should be particularly relevant in influencing directorselection and retention and thus the dynamics of boardinterlock formation and dissolution during this period.

INTERORGANIZATIONAL TIES ANDINTRAORGANIZATIONAL POWER

Although researchers have drawn on a variety of theoreticalperspectives to explain the diffusion of organizationalphenomena through interlock network ties, recentdiscussions have tended to emphasize the role of such tiesin spreading knowledge and awareness about specificorganizational changes. As articulated by Galaskiewicz andWasserman (1989: 456), network ties betweenboundary-spanning personnel, including board members,"act as a conduit to disseminate ideas and innovations."Empirical support for this argument has been found instudies investigating the likelihood of adopting poison pills(Davis, 1991), making specific kinds of corporate acquisitions(Haunschild, 1993), adopting the multidivisional form (Palmer,Jennings, and Zhou, 1993), or making campaigncontributions to particular political candidates (Mizruchi,1992).

On a deeper level, interlock ties may help spread morefundamental belief systems about corporate strategy,organizational structure, or the board's role in the


organization, providing the foundation for a variety oforganizational changes. Empirical research investigating thedevelopment of group norms has shown how the priorexperiences of individual group members in similar contextscan provide the foundation for norms of the focal group(Bettenhausen and Murnighan, 1985). Thus board memberswho have participated in various strategic or structuralchanges on other boards bring those beliefs and the various,more specific scripts associated with them to the focalboard, ultimately "negotiating" or advocating changes inboard norms consistent with their beliefs (Bettenhausen andMurnighan, 1985). While governance scholars have notedthe potential role of socialization experiences in affectingboard decision making (Alderfer, 1986), such processes arefrequently assumed to operate only within the focal board.Given that a large portion of corporate directors hold multipleboard seats, however, and have held other directorships inthe past (Useem, 1984; Mizruchi, 1992), the collectiveinfluence of directors' experiences on other boards mayfrequently outweigh the effect of socialization experienceson the focal board. Thus directors who have participated instructural changes that increase board control overmanagement may help spread increased board control to theother boards on which they sit, either by raising awarenessabout specific changes or by moving board norms toward abehavioral model that favors active monitoring and oversighton behalf of shareholders. In effect, by virtue of their priorexperiences, such directors "come to a normativeunderstanding" that the role of corporate director requiresmonitoring and controlling management decision making,prompting them to initiate and encourage changesconsistent with such behavior where they serve as outsidedirector (Burt, 1987: 1289; Ocasio, 1996). Conversely,directors who have participated in structural changes thatprotect or enhance the CEO's control over the board maycome to believe that outside directors should defer to theCEO's judgment on strategic issues, leading them to helpmaintain board norms that favor director passivity and loyaltyto the CEO.

Director Experience and Change in Board Interlocks

Given that directors' relative experience with increasedboard control at other companies can affect their willingnessto abide CEO control at the focal firm, powerful CEOs mayseek to sustain CEO control by (1) avoiding directorcandidates who have participated in greater board controlover management and (2) selecting, instead, directors whoseexperience on other boards reflects board passivity towardmanagement. Conversely, powerful boards can perpetuateor increase their control over management by (1) avoidingdirectors who have been socialized into a passive board role,as indicated by their participation in allowing greater CEOcontrol on other boards and (2) appointing individuals withprior experience in asserting board control. Moreover,individual board members should prefer activist newdirectors because, as Alderfer (1986) suggested, boardsrequire unanimity and a sense of common purpose tocontrol management decision making effectively, just as allgroups require unanimity and cohesion to manage actors in


Board Interlocks

their environment (Gladstein, 1984). Through selection andretention, then, powerful actors in the CEO-boardrelationship can "manage" board interlocks so as toreinforce or increase their control.

This argument is consistent more generally with thetheoretical and empirical literature on power in organizations,which cites control over employee selection as an efficientmeans of building political coalitions (Pfeffer, 1981: 163;1992). Managers can often build or protect their powerbases more easily and cheaply by hiring or promotingindividuals likely to support their personal, political intereststhan by trading favors with existing colleagues (Pfeffer,1981). Drawing on this literature, several governancescholars have suggested that CEO control over the directorselection process represents an important source ofmanagerial entrenchment (Mace, 1971; Kosnik, 1987;Fredrickson, Hambrick, and Baumrin, 1988; Wade, O'Reilly,and Chandratat, 1990). For instance, Lorsch and Maclver(1989) and Finkelstein and Hambrick (1989: 124) noted thatCEOs can "coopt" the board by favoring the appointment of"sympathetic" new directors. Although this process hastypically been associated with CEO behavior, Westphal andZajac (1995: 78) suggested that this process also applies toboards seeking to increase their monitoring and control byappointing new directors who are sympathetic to theorientation of existing directors. In effect, we propose thatdirectors' experiences on other boards may furnish arelatively direct indicator of their relative "sympathy" towardeither management or shareholder interests, thus providing abasis for director selection and retention.

An implication of this line of reasoning is that changes inrelative control by the CEO or board may spread through thenetwork of interlocking directorates in ways that have notbeen specified in prior research. We propose here that thediffusion of specific governance changes will depend onwhether powerful actors in CEO-board relationships seek todeflect or encourage such diffusion. For example, thediffusion of a governance change that diminishes CEOcontrol may be deflected by powerful CEOs by avoiding oreliminating interlock ties with prior "adopters." Similarly,powerful CEOs may also steer the diffusion of changes thatprotect or increase their control over the board by addingties to prior adopters of such changes. As one director citedin Lorsch and Maclver (1989: 77) suggested, "The CEOshapes the board very much the way he wants, not onlybringing people in, but also getting people he doesn't like offthe board." Similarly, directors who sit on powerful boardsare likely to prefer new director candidates who share theirbelief in active boards. As another director cited in Lorschand Maclver (1989: 5) noted, "Directors today don't wantcolleagues like the old ones who rubber-stampedmanagement's decisions."

In this way, contests for intraorganizational power can affectinterorganizational ties: The actions of powerful corporateleaders seeking to influence the selection and retention ofboard members generate a "segmented" interlock network(Kaufman, 1986), with relatively few ties between active andpassive boards and relatively dense ties between boards


For each of the hypotheses developedbelow, the two components of thehypothesis are related but distinct, suchthat a director can experience a changein appointments at low-control boards butnot experience any change inappointments at high-control boards.

with similar power structures. Here, we examine thedynamic process by which this segmentation is maintained.We propose that a director's prior experiences on boardsthat have made changes that increase or decrease boardcontrol over management provide a strong indication of thatdirector's likely subsequent behavior on other boards,affecting the relative attractiveness of that director in thefollowing ways:

Proposition la: Directors who sit on boards that have recentlymoved to increase board control over management will have (1)fewer subsequent appointments to boards having low control and(2) more subsequent appointments to boards having high control.

Proposition 1b: Directors who sit on boards that have recentlymoved to decrease board control over management will have (1)more subsequent appointments to boards having low control and(2) fewer subsequent appointments to boards having high control.

In the remainder of this section, we develop specifictestable hypotheses corresponding to the propositionsoffered above. While it is difficult to observe and measuredirectly the level of board control over management in largecorporations, a number of indicators have been used in priorgovernance research to capture such differences acrossorganizations. We consider five different indicators in thisstudy: changes in the ratio of outside to inside directors, inboard leadership structure, in diversification, in totalcompensation, and in compensation contingency. The firsttwo are measures of board structure that assess boardcontrol over management, and the latter three are measuresof strategy or policy that reflect attention to shareholderinterests or managerial interests. Given that all five variablesare visible to top managers and directors at othercompanies, they can provide a more salient basis for newdirector selection than less visible measures of actual boarddecision-making processes.

Change in the outsider ratio. Perhaps the most commonlyused indicator of board independence from management isthe ratio of outside to total directors. Governanceresearchers and champions of board reform have longcontended that nonemployee or "outside" directors arebetter positioned to control management decision makingthan insiders (Fama and Jensen, 1983). As subordinates tothe CEO, inside directors may be reluctant to challenge orquestion the CEO's position on an issue in board meetings,even when shareholder interests appear to be threatened(Fredrickson, Hambrick, and Baumrin, 1988; Hoskisson,Johnson, and Moesel, 1994). Outsiders should also be betterable than insiders to judge managerial performanceimpartially, raising the likelihood that a poorly performingCEO will be dismissed (Boeker, 1992). Thus, increases in theratio of outside to inside directors can be consideredindicative of increased board control over management,while decreases in this measure should indicate greaterboard passivity in monitoring management decision making:''

Hypothesis la: Directors who sit on boards that have recentlyincreased the ratio of outside to inside directors will have fewersubsequent appointments to boards having low control and moresubsequent appointments to boards having high control.


Board Interlocks

Hypothesis 1b: Directors who sit on boards that have recentlydecreased the ratio of outside to inside directors will have nnoresubsequent appointments to boards having low control and fewersubsequent appointments to boards having high control.

Change in board leadership structure. It is also commonlysuggested by organizational scholars and members of thebusiness press that separating the CEO and board chairpositions (i.e., allocating each position to separateindividuals) should greatly increase the board's capacity tocontrol management decision making (Vance, 1983;Harrison, Torres, and Kukalis, 1988; Crystal, 1991; Beattyand Zajac, 1994; Finkelstein and D'Aveni, 1994). Given thatthe board chair is nominally responsible for evaluating CEOdecisions, allocating both roles to the same individualpresents a conflict of interest (Kesner and Johnson, 1990;Zajac and Westphal, 1996). Moreover, board meetingsrepresent the primary forum for directors to challenge CEOs'proposals (Lorsch and Maclver, 1989). By dictating anagenda that offers a limited opportunity for open debate, aCEO serving as board chair can easily minimize the level ofboard monitoring behavior (Demb and Neubauer, 1992).Conversely, an independent chairperson is better able tocontrol management on behalf of shareholders. Thus, wehypothesize:

Hypothesis 2a: Directors who sit on boards that have recentlyseparated the CEO and board chair positions will have fewersubsequent appointments to boards having low control and moresubsequent appointments to boards having high control.

Hypothesis 2b: Directors who sit on boards that have recentlycombined the CEO and board chair positions will have moresubsequent appointments to boards having low control and fewersubsequent appointments to boards having high control.

Change in corporate diversification. According tomanagerialist and agency researchers, top managers havepersonal incentives to pursue diversification beyond the levelat which shareholder wealth is maximized (Hill and Snell,1988; Jensen, 1988; Baysinger and Hoskisson, 1990). Fromthe managerialist perspective, top managers may diversifyinto largely unrelated businesses to increase their personalpower, compensation, and status (Marris, 1964). Agencytheorists emphasize the link between diversification andreducing managerial risk (Amihud and Lev, 1981). Bydiversifying into unrelated businesses, managers canstabilize their investment portfolios while also reducing theiremployment risk. Shareholders should favor lower levels ofdiversification, because they can diversify their investmentportfolios more easily than CEOs can diversify theiremployment. Thus, as nominal representatives ofshareholder interests, relatively powerful boards of directorsshould resist managerial preferences for excessive corporatediversification (Hoskisson, Johnson, and Moesel, 1994).Conversely, as Jensen (1988) has suggested, board passivitymay be partly responsible for inefficient diversification levels.Therefore, reduced corporate diversification provides anindication of greater shareholder representation and lowermanagerial control over board monitoring activity.Participation on boards that have overseen reducedcorporate diversification would send a negative signal in the


market for directors at CEO-dominated boards and a positivesignal at companies with more active boards:

Hypothesis 3a: Directors yyho sit on boards that have recentlydecreased corporate diversification will have fewer subsequentappointments to boards having loyv control and more subsequentappointments to boards having high control.

Hypothesis 3b: Directors who sit on boards that have recentlyincreased corporate diversification will have more subsequentappointments to boards having low control and fewer subsequentappointments to boards having high control.

Change in total compensation. The managerial andacademic literatures on corporate governance have longattributed high levels of CEO compensation to boardpassivity and CEO entrenchment. Several studies havereported evidence that board independence from the topmanagement is associated with smaller increases in CEOcompensation (Finkelstein and Hambrick, 1989; Main,O'Reilly, and Wade, 1995; Westphal and Zajac, 1995).Changes in CEO compensation at large companies are wellpublicized and highly visible to managers and directors atother boards and thus may provide a salient indicator of therelative independence of the board from management.Participation on boards that have reduced the level of CEOcompensation would therefore send a negative signal in themarket for directors at CEO-dominated boards and a positivesignal among companies with more active boards:

Hypothesis 4a: Directors who sit on boards that have recentlydecreased total compensation will have fewer subsequentappointments to boards having low control and more subsequentappointments to boards having high control.

Hypothesis 4b: Directors who sit on boards that have recentlyincreased total compensation will have more subsequentappointments to boards having low control and fewer subsequentappointments to boards having high control.

Change in compensation contingency. Contingentcompensation contracts include a variety of differentincentive plans that link management pay to firmperformance, thus aligning the interests of CEOs with thepreferences of shareholders (Jensen and Meckling, 1976),According to agency theorists, long-term incentives such asstock options, performance shares, or restricted stock are aprimary mechanism by which corporate boards protectshareholder interests against management excesses (Kerrand Kren, 1992; Gibbs, 1993; Beatty and Zajac, 1994), andboth economic and behavioral literatures on executivecompensation suggest that, in general, CEOs should preferless long-term incentive compensation in their pay packages.From a normative agency theory perspective, CEOs (asrisk-averse agents) prefer less risk in their compensationcontracts (Harris and Raviv, 1979). By making compensationcontingent on future firm performance, long-term incentivesadd uncertainty to a CEO's compensation. Some empiricalevidence suggests that passive boards make only limiteduse of long-term incentive compensation in designing CEOcompensation contracts (Tosi and Gomez-Mejia, 1989), whileincreased board independence from management is typicallyfollowed by higher levels of contingent compensation(Westphal and Zajac, 1995). Membership on boards that


Board Interlocks

have increased CEO compensation contingency shouldtherefore decrease an individual's marketability amongCEO-dominated boards, while increasing his or herattractiveness to relatively active boards:

Hypothesis 5a: Directors who sit on boards that have recentlyincreased CEO compensation contingency will have fewersubsequent appointments to boards having low control and moresubsequent appointments to boards having high control.

Hypothesis 5b: Directors who sit on boards that have recentlydecreased CEO compensation contingency will have moresubsequent appointments to boards having low control and fewersubsequent appointments to boards having high control.

METHOD

Sample and Data Collection

The population for this study included top managers of thelargest U.S. industrial and service firms, as listed in the 1986Forbes and Fortune 500 indexes. We restricted our attentionin this study to directors who were also CEOs, becauserecent evidence shows that CEO-directors play a pivotal role(relative to other directors) in determining the extent towhich a board develops a passive or active orientation(Lorsch and Maclver, 1989: 18; Westphal and Zajac, 1997).The Forbes 500 uses multiple lists whose overlap dependson the specific size measure used; we included topmanagers of those firms that qualified according to two ormore size measures. Directors were excluded from the finalsample if complete diversification and compensation datawere unavailable. This procedure yielded a final sample of491 directors, f-tests revealed no significant differencesbetween this sample and the larger population across any ofthe director, board, or firm attributes for which data wereavailable on the larger population (i.e., director age,compensation level, board composition or leadershipstructure at the director's home company, or profitability orsize—measured as log of sales^of the director's homecompany).

Data were collected for the years 1985 to 1992, inclusive.We obtained data on interlock ties, board structure, and CEOand outside director characteristics from the Dun andBradstreet Reference Book of Corporate Management,Standard & Poor's Register of Corporations, Directors, andExecutives, and Who's Who in Finance and Industry;compensation data came from proxy statements. Finally, weobtained diversification data from Standard & Poor'sCOMPUSTAT Business Segment Tapes and CompactDisclosure; size and performance data were provided byCOMPUSTAT and the Center for Research in Security Prices(CRSP).

Independent Variables

We measured participation in increased board control acrossfive different indicators. For each indicator, we firstcalculated the measure for each board in the sample in eachyear and then created measures of participation in increasedboard control for each year by comparing—for all boards onwhich the focal individual sits—the value of the givenmeasure in year t with the value in year f - 1. Thus, for


example, we first measured the outsider ratio in each yearas the number of nonemployee directors divided by the totalnumber of board members. We then calculated participationin increased outsider ratio as the number of boards on whichthe focal individual sits that increased the ratio of outside toinside directors from year f - 1 to year t. Board leadershipstructure was measured dichotomously, coded as 1 if theCEO and board chair positions are separate in a given year,and 0 otherwise. To measure participation in CEO/boardchair separation, we calculated the number of boards onwhich the focal individual sits with combined CEO-boardchair position in year f - 1 and separated CEO-board chairpositions in year t.

In some cases, as in the outsider ratio, it was also possible tomeasure participation in increased board control as theaverage change across all boards on which the individualserves as director. We chose to treat each change as adichotomous event for theoretical and empirical reasons.Theoretically, for most of the five measures, any changeshould represent a distinctive and visible event to managersand directors on other boards (exceptions are discussedfurther below). Empirically, the size of change is veryrestricted for several of our independent variables, e.g.,boards rarely increase the outsider ratio by adding more thantwo individuals to the board, so that the magnitude of suchincreases is very similar across cases (Mizruchi and Stearns,1988). Thus the count variables based on dichotomouschange measures would be highly correlated with variablesindicating the average, continuous change. Nevertheless, toensure that our results did not hinge on one particularspecification, we also conducted a separate analysismeasuring participation as the average change in therelevant measure across all boards on which the individualserves as director (i.e., for those changes that can bemeasured continuously). The results were very similar tothose reported below, suggesting that our findings do notdepend on one particular specification of the independentvariables.

To measure participation in reduced diversification, we usedPalepu's (1985) entropy measure. This measure takes intoaccount the number of segments in which a firm operatesand weights each segment according to its contribution tototal sales. It is defined as follows:

( = 1

where P is the sales (dollar value) attributed to segment /and In(1/P,) is the weight for each segment i, or thelogarithm of the inverse of its sales. We calculated reduceddiversification as a decrease in the entropy measureexceeding one standard deviation, to capture relativelysignificant change and exclude change that reflects randomalterations in segment sales levels (Amit and Livnat, 1988).The results were substantively similar using alternativethresholds, such as a decrease of one-half of one standarddeviation. The standard deviation is based on change indiversification from year f - 1 to year t. Participation in


Board Interlocks

reduced diversification was then measured as the number offirms where the focal individual serves as director thatdecreased diversification from the prior year to the currentyear.

To assess participation in increased compensationcontingency, we first calculated contingency for each year asthe total value of long-term incentive grants divided by totaldirect compensation. Short-term bonuses were not includedin this measure because they are notoriously susceptible tomanipulation (Healy, 1985), but stock options were included.Although it is possible for companies to "swap"higher-priced options for lower-priced options, thus dilutingtheir incentive effect, this practice is extremely rare. Wevalued stock options using the Black-Scholes (1973) method,which estimates option value based on the historical pricevolatility of the underlying security. Other grants (e.g.,restricted stock, performance shares, etc.) were valuedaccording to the market price at date of grant (Crystal, 1984).All compensation values were adjusted for inflation torepresent 1990 constant dollars using the Consumer PriceIndex. This approach to measuring compensationcontingency is commonly used by compensation consultants(e.g.. Crystal, 1984; Ellig, 1984).

By contrast, the simple correlation between CEO pay andfirm performance is a very problematic measure ofcompensation contingency. The most serious problem ishow to treat long-term incentives in such a measure. Thevalue of long-term incentive grants (e.g., stock options,which confer the right to buy shares of common stock in thefuture) depends on future performance, while salary, bonus,and other forms of compensation depend on priorperformance. Thus total compensation could not becorrelated with performance over a single period.Researchers have sometimes attempted to avoid thisproblem by excluding long-term incentives from theirmeasures of compensation. This approach is problematic inassessing the correlation between pay and performance,however, because long-term incentive compensation is theprimary mechanism that companies use to make paycontingent on performance (cf., Ellig, 1984; Tosi andGomez-Mejia, 1989; Kerr and Kren, 1992; Westphal andZajac, 1995; Seward and Walsh, 1996).

We calculated participation in greater CEO compensationcontingency as the number of firms where the focalindividual serves as director that increased compensationcontingency from the prior year to the current year.Participation in reduced total CEO compensation wascalculated as the number of firms where the focal individualserves as director that reduced CEO total directcompensation from the prior year to the current year.

Finally, we created five parallel independent measuresgauging participation in reduced board independence:participation in decreased outsider ratio; participation inCEO/board chair combination; participation in increaseddiversification; participation in reduced compensationcontingency; and participation in increased total


compensation. Increased total compensation does notexactly parallel decreased total compensation. Rather, wemeasured it as a positive change in the CEO's total directcompensation exceeding one standard deviation. Thisenables us to capture relatively distinctive and noticeablechanges, since a large portion of companies in the sampleincreased total compensation in any given year. As with thediversification variables, the results were robust toalternative change thresholds, such as a positive change ofone-half of one standard deviation.

Several control variables were included in the analyses. First,an individual's reputation or attractiveness in the directorlabor market may depend on the performance of companieswhere he or she has served as director (Fama, 1980).Accordingly, we included a control variable indicating theaverage profitability of firms at which the focal individualserved as outside director in the prior year. We measuredprofitability as retum on assets, industry-adjusted at thetwo-digit Standard Industry Classification (SIC) code level.From a human capital perspective (Becker, 1975), topmanagers' compensation at their home company may alsoindicate their worth as a corporate leader and, by extension,their value as an outside director. Alternatively, the level of atop manager's compensation could be viewed as reflectingthe degree of board control at that individual's homecompany, thus providing another indicator of the individual'spersonal experience with greater or lesser board control.Thus we included the director's total direct compensation{total compensation) as CEO in his or her home company asa control variable in all models. This measure includes salary,short-term bonus, and the total value of long-term incentivegrants.

Furthermore, because an individual's age might be taken asan indication of his or her likely openness to new ideasabout board independence and control, we also controlledfor age in all models. In addition, the total number of boardappointments held by an individual could affect subsequentappointments in several ways. The total number ofappointments a person holds might independently affect hisor her prestige in the director labor market (Davis, 1993), butworkload limitations may prompt CEOs with manyappointments to decline further invitations. Accordingly, wecontrolled for the total number of directorships held in theprior year {number of prior appointments). Finally, we alsocontrolled for director participation in CEO succession events{participation in succession) on other boards. AlthoughBoeker (1992) suggested that succession is less indicative ofboard control than outright dismissal, which is more difficultfor individuals outside the organization to detect,participation in succession may be related to some of theother indicators of board control used in this study (Harrison,Torres, and Kukulis, 1988) and thus is included as a controlvariable in all models. This variable is defined as the numberof firms at which the focal individual serves as director thatexperienced CEO succession within the prior year.


Board Interlocks

Dependent Variables

Analyzing change in appointments to boards with high or lowcontrol over management required defining and identifyinghigh- and low-control boards. We addressed this bydeveloping an overall measure capturing the extent to whicha firm's board displays a relatively enduring orientationtoward board control (high control) or board passivity (lowcontrol). We began by standardizing the five indicators ofboard control discussed above and summing these fivemeasures together for each board in the sample. Thismeasure was calculated as the average value over a movingthree-year period (i.e., year f - 2 to year t) to avoidmeasurement error that would otherwise arise fromtemporary fluctuations in relative control. We thendistinguished between high- and low-control boards bytaking the median split for this measure, separately for eachyear. Using this information, we constructed two dependentmeasures: Additions/subtractions in appointments to boardswith low control over management and Additions/subtractions in appointments to boards with high controlover management.

We developed separate measures for additions andsubtractions in appointments to examine the robustness ofour results in predicting losses, as well as gains, insubsequent appointments and thus were able to assessdirector selection and retention, respectively. In creatingthese variables, we first calculated an overall changemeasure in each year as the number of appointments toboards with low or high control in year t + 2 minus thenumber of appointments to boards with low or high controlin year t. We measured change in board appointments overa multiyear interval because employment contracts fordirectors limit the number of openings that are available inany given year. From the two overall change measures wethen created four count variables indicating the net increaseor decrease in appointments to low- or high-control boardsover the two-year period. Table 1 provides the means,standard deviations, and bivariate correlations for all datapooled.

Analysis

We used poisson regression analysis to assess change inboard appointments (Maddala, 1983). This technique issuitable for estimating models predicting the number ofdiscrete occurrences (i.e., counts) of some event, in thiscase, additions or subtractions in board appointments.Separate models were created to analyze change inappointments to boards with high control or low control overmanagement. The individual models analyze the effect that aboard member's prior participation (from year f - 1 to yeart) in each of the five governance-related changes in boardstructure and strategy or policy decisions has on subsequentadditions or subtractions (from year t to year t -l- 2) inappointments to boards with high- or low-controlorientations. All control variables were lagged by one year.We observed relationships over five time periods, yielding2,455 CEO-years of data.


Table 1

Descriptive Statistics and Pearson Correlation Coefficients

Variables

1. Participation in:a. Increased outsider ratiob. CEO/board chair separationc. Reduced diversificationd. Greater compensation contingencye. Reduced total compensation

2. Participation in:a. Decreased outsider ratiob. CEO/board chair combinationc. Increased diversificationd. Reduced compensation contingencye. Greater total compensation

3. Average return on assets4. Total compensation (in millions)5. Age6. Number of prior appointments7. Change in appointments to boards with high control

a. Subtractionsb. Additions

8. Change in appointments to boards with low controla. Subtractionsb. Additions

Mean

.308

.206

.176

.439

.173

.173

.100

.137

.092

.091

.0011.875

58.5194.388

.7461.094

.873

.944

S.D.

.550

.449

.435

.611

.422

.454

.352

.413

.332

.321

.0592.2185.6803.212

.767

.940

111.947

1a

.15

.06

.19

.21

- .22- .23- .09- .20- .14- .09-.11- .18

.08

- .15.14

.27- .25

1b

.09

.16

.27

- .18- .16- .12- .17- .15

.15- .17- .05

.11

- .27.21

.32- .26

1c

.05

.19

- .19- .06- .18- .19- .05

.06- .08- .12

.12

- .23.25

.18- .15

Id

.28

- .13- .08- .06- .20- .26

.07- .13- .10

.07

- .16.26

.21- .23

We also conducted a series of separate analyses using twoother models: a generalized least squares (GLS) model and anegative binomial model. The results were substantivelyconsistent with the results of poisson regression analysisreported below (i.e., the tests of statistical significancesupported the same set of hypotheses).

RESULTS

Tables 2 and 3 report the results of poisson regressionanalysis of change in board appointments. The models inTable 2 predict change in appointments to boards with lowcontrol over management, while the models in Table 3predict change in appointments to boards with high boardcontrol. As noted earlier, we provide two models in eachtable to reflect the fact that while our hypotheses apply tochange in appointments, we can analyze such changesseparately in terms of additions or subtractions to furtherassess the robustness of the findings. Hypothesis la, forexample, predicted that participation in increasing theproportion of outside directors would lead to a decrease inappointments to boards with low control over managementand an increase in appointments to boards with high controlover management. The findings, as shown in Table 2,support this hypothesis: Participating in increasing theoutsider ratio is negatively related to subsequentappointments to boards with low control over management.This holds true both in terms of fewer additions to newboards and more subtractions from current boards. Theexpected converse pattern of results emerged in modelspredicting appointments to boards with high control overmanagement. The results in Table 3 show that participationin increasing the outsider ratio is positively associated withsubsequent appointments to boards with high control overmanagement (both in terms of more additions and fewersubtractions).


Board Interlocks

lable

1e

1 (continued)

2a 2b 2c 2d 2e 3 4 5 6 7a 7b 8a

301101253312191405

0708

.17

.11

.21

.09- .08

.02

.10

.14

.24- .26

.13

.36

.21- .08

.09

.1506

.29- .32

20.1906.11.09.14

.05- .04

.34

.09

.09

.10

.10

22- 2 3

- .12,21.08,09

.11- .15

.08

.07- .01

- .15.13

- .03.07

.16- .09

.19

20- .19

- .13.10

.22

.18- .16

.18- .31

.32- .07

.08- .17

.15- .08

.11.06

- .05- .20

.18- .17

.12- .10

.14

- .34

- .3833

.21- .25 - .37

Table 2

Poisson Regression Models of Change in(N = 2455)*

Independent Variables

1. Participation in:a. Increased outsider ratiob. CEO/board chair separationc. Reduced diversificationd. Greater compensation contingencye. Reduced total compensation

2. Participation in:a. Decreased outsider ratiob. CEO/board chair combinationc. Increased diversificationd. Reduced compensation contingencye. Greater total compensation

3. Average return on assets4. Total compensation5 Age6. Number of prior appointments7. Participation in succession

ConstantChi square

1 Appointments to Boards with Low Control overManagement

Change in Appointments

Subtractions

.408 (.100)—

.531 (.157)—

.312 (.138)*

.2701.102)-

.389 1.162)-

-.170 (.086)*-.594 (.135)—-.077 (.097)-.227 (.095)--.256 (.140)

.572 (.594)-.00002(00001)*

-.015 (.007)*-.021 (.017)

.195 (.389)

.273 (.175)465.69—

• p s .05; "p £ .01: •**p •& .001: t-tests are one-tailed for hypothesized effects, two-tailed for* Unstandardized coefficients are reported. Standard errors are in parentheses.

Additions

-.344 (. 118)--.417 (.174)-- .374 (.163)*- .408 (.151)--.450 (.226)*

206 (072)-393 (.142)—.040 (.089).208 (.081 ) -.280 (.154)

-.524 (.612).00003 (.00001 ) -

.010 (.005)*

.016 (.014)-.169 (.163)-

.755 (.312)-488.18—

control variables.

Hypothesis 1 b predicted opposite consequences fromparticipating in decreasing outsider ratios as an outsidedirector. Consistent with this hypothesis, the results indicatethat experience with decreasing the outsider ratio ispositively associated with subsequent appointments toiow-control boards (more additions and fewer subtractions)and negatively associated with subsequent appointments to


high-control boards (fewer additions and more subtractions).Taken together, the tests of hypotheses la and 1b are quiterobust in suggesting that a director's prior experiences witha governance change involving altering the outsider ratio canresult in predictable changes in subsequent appointments.

Table 3

Poisson Regression Models of Change in Appointments to Boards with High Control over Management(A/ = 2455)*

Change in Appointments

Independent Variables Subtractions Additions

1. Participation in:a. Increased outsider ratio - .237( .11 i r .214(.1O6)*b. CEO/board chair separation -.527 (.157)*^ .363 (.174)*

• c. Reduced diversification -.370 (.139)** .284 (.110)**d. Greater compensation contingency -.291 (.149)* .425 (.112)*^e. Reduced total compensation -.179(.171) .329(.211)

2. Participation in:a. Decreased outsider ratio .478 (.073)"* - .330 (.106)*~b. CEO/board chair combination .329 (.082)~* -.641 (.137)~*c. Increased diversification .043 (.094) -.035 (.113)d. Reduced compensation contingency .226 (.093)** - .362 (.135)**e. Greater total compensation .203 (.182) - .430 (.187)*

3. Average retum on assets -1.148 (.581) 1.169 (.621)4. Total compensation .00002 (.00001 )* - .00002 (.00001 )*5. Age .012 (.006)* - .016 (.007)*6. Number of prior appointments -.023 (.019) .019 (.017)7. Participation in succession -.213(.123) .144(.11O)

Constant .280 (.341) .643 (.421)Chi square 297.04*~ 334.57*~

• p £ .05; **p s .01; *^p £ .001; f-tests are one-tailed for hypothesized effects, two-tailed for control variables.• Unstandardized coefficients are reported. Standard errors are in parentheses.

A similar pattern of results emerged in support ofhypotheses 2a and 2b. As shown in Tables 2 and 3,participation in separating the CEO and board chair positionsis negatively associated with additional appointments toboards with low control over management and positivelyassociated with subsequent subtractions, while forhigh-control boards, such experience is positively associatedwith subsequent additions and negatively associated withsubsequent subtractions. Moreover, experience in combiningthe CEO and board chair positions is associated with fewersubsequent appointments to high-control boards (feweradditions and more subtractions) and more subsequentappointments to low-control boards (more additions andfewer subtractions).

Hypotheses 3-5 addressed the consequences ofparticipating in additional strategy and policy changes thatreflect greater (or lesser) board control over management.The findings generally support these hypotheses. Forexample, consistent with hypothesis 3a, participation indecreasing corporate diversification leads to fewersubsequent appointments to boards with low control overmanagement (i.e., fewer additions and more subtractions)and also leads to more subsequent appointments to boardswith high control (i.e., more additions and fewer


Board Interlocks

subtractions). The results of testing hypothesis 3b, however,show that participation in increasing corporate diversificationis not significantly related to subsequent change in boardappointments to either high-control boards or low-controlboards.

Hypotheses 4a and 4b, which addressed the effects onsubsequent board appointments of a director's priorparticipation in increasing or decreasing total CEOcompensation are generally supported. As hypothesis 4aargued, a director's prior participation in decreasing totalCEO compensation is negatively associated with subsequentappointments to boards with low control over management(both in terms of fewer additions and more subtractions) andalso is positively associated (p < .10) with subsequentappointments at high-control boards (but only in terms ofadditions). Hypothesis 4a is partially supported: Participationin increasing total CEO compensation is positively associatedwith subsequent appointments to low-control boards (moreadditions and fewer subtractions) and negatively associatedwith subsequent appointments to high-control boards (butonly for additions).

The results strongly support hypotheses 5a and 5b.Experience with increasing CEO compensation contingencyis negatively associated with subsequent appointments toboards with low control (fewer additions and moresubtractions) and positively associated with appointments tohigh-control boards (more additions and fewer subtractions).Conversely, experience with decreased compensationcontingency has the opposite effect. In general, therefore,the findings appear to provide robust evidence thatparticipation in specific changes in board structure, corporatestrategy, and CEO compensation that reflect change inrelative control of CEOs and boards can differentiallyinfluence subsequent appointments to boards with high orlow control over management.

The control variables yielded several interesting results. Theaverage profitability (return on assets) of firms at which thefocal individual served as outside director in the prior year isnot related to subsequent change in appointments to boardswith low control over management. Average profitabilitydoes increase appointments to high-control boards, however(significant at alpha = .10). The director's totalcompensation as CEO in his or her home company ispositively related to appointments at low-control boards andnegativeiy related to appointments at high-control boards.Moreover, older directors gain more subsequentappointments at low-control boards and fewer subsequentappointments to high-control boards.

We also examined the possibility that industry differencesmight affect our results. In separate analyses, we regressedthe changes in board structure and CEO compensation onindustry, designated as the company's two-digit SIC code.The industry factors did not significantly predict increasedboard control over management for any of the measures ofboard control; t-statistics ranged from .810 to 1.289. Thus itappears that our results are not driven by any tendency for


boards to choose directors from related industries (e.g., toalleviate resource dependencies).

DISCUSSION

Overall, the findings of this study strongly support our notionthat internal organizational politics in the CEO-boardrelationship can significantly affect the selection andretention of directors and, more generally, the formation anddissolution of board interlocks. The first set of resultsshowed how a director's prior participation on boardsengaging in specific changes in board structure and strategyor policy that reflect greater board control can affectsubsequent change in board appointments. In general, priorexperience with changes indicating increased board controlover top management enhanced an individual director'sattractiveness at companies with relatively high board controlwhile decreasing his or her attractiveness at companies withrelatively passive boards. This result was robust over fivedifferent indicators of increased board control: increases inthe ratio of outside to inside directors, separation of the CEOand board chair positions, reduced corporate diversification,increased CEO compensation contingency, and decreasedtotal CEO compensation. The results were also quite robustin terms of the type of change (i.e., additions andsubtractions) in board appointments.

A second set of results showed the effect on subsequentboard appointments of a director's prior participation inspecific governance-related changes in the oppositedirection, i.e., indicating greater CEO control overmanagement. Overall, prior experience with such changesreduced a director's attractiveness to firms with relativelyhigh board control while increasing his or her attractivenessto firms with more powerful CEOs and weak boards. Again,these results held for multiple indicators and for additionsand subtractions in subsequent appointments. Takentogether, these two sets of results provide strong evidencethat prior experience with increased or decreased boardcontrol over management provides an important basis forselecting or retaining directors.

The control variables used in this study provide additionalsupport for a sociopolitical perspective on director selectionand board composition. For example, director age is alsopositively associated with subsequent appointments toiow-control boards and negatively associated withappointments to high-control boards. Older directors may beperceived in the market for corporate directors as beingmore accepting of board passivity in controlling managementand less likely to embrace newer perspectives reflectingmore active board involvement and control in managementdecision making. Moreover, directors' own total executivecompensation at their home company is positively related tosubsequent appointments at companies with low boardcontrol and negatively related to appointments at companieswith high board control. This suggests that highly paid CEOsmay be perceived as individuals who are accustomed toweak board control, which would increase theirattractiveness at low board-control companies and decreasetheir attractiveness at high board-control companies.


Board Interlocks

Finally, the results show that the profitability at companieson whose boards the focal director sits is unrelated tosubsequent change in appointments at companies with lowboard control but is positively related to change inappointments to high-control boards. Thus it appears that anindividual's attractiveness in the market for directors may beincreased by sitting on boards of highly performingcompanies, as some financial economists have suggested(Fama, 1980), but this attractiveness applies only within thesubsegment of boards with relatively high control overmanagement. Our political perspective shares theeconomists' interest in the reputation of corporate directors,i.e., how prior director experience can provide a signal toother interested parties, but our theoretical framework ismore nuanced. It suggests that reputational effects may bemore complex than previously assumed. The market forcorporate directors is not driven by a simple "ex postsettling up" process in which "high-quality" (active andshareholder-oriented) directors are rewarded and"low-quality" (passive and management-oriented) directorsare punished (e.g., Fama, 1980). Our results suggest,instead, that the market for directors in U.S. corporationscan reward either type of director, and our theoreticalperspective identifies the mechanisms than can generateand sustain such a segmented market for corporatedirectors.

This study contributes to research on corporate boards ofdirectors by synthesizing three largely nonoverlappingtheoretical literatures pertaining to board behavior: theeconomic literature on director reputation, the sociologicalliterature on board interlocks, and the behavioral literature onCEO-board relationships. The findings show how powerfulactors in CEO-board relationships can manage the content ofboard interlocks to maintain or increase theirintraorganizational power. In this way, powerful CEOs caneffectively exclude individuals who might import norms ofactive board monitoring and shareholder representation (i.e.,at the expense of managerial preferences), while includingindividuals who have been socialized into a passive boardrole (Burt, 1987). In effect, our findings suggest thatdirectors' experience on other boards may serve as arelatively direct reputational indicator of their "sympathy"(Finkelstein and Hambrick, 1989: 124) toward managementor shareholder interests, thus providing a primary basis fordirector selection and retention. Moreover, this interpretationis consistent with the broader literature on power inorganizations, which emphasizes control over employeeselection and retention as an efficient means of building andmaintaining political coalitions (Pfeffer, 1981).

The findings also have implications for theon/ and researchon the formation of board interlocks. While theorists havetypically emphasized the role of interorganizationaldependencies or classwide interests in determining interlockties (Useem, 1982; Palmer, 1983; Mizruchi and Stearns,1988), the results of this study suggest that parochialpolitical interests within the firm may also be an importantdeterminant of interlock structure. In seeking to maintain orincrease their influence in CEO-board relationships, powerful


corporate leaders help maintain a "segmented" interlocknetwork with relatively few ties between active and passiveboards and relatively dense ties between boards with similarpower structures. For example, 82 percent of the totalnumber of ties observed are within-group ties, while only 18percent are between-group ties. Thus, while our resultscould be viewed as consistent with the notion that interlockties help elites to "close ranks" in the face of "deviant"behavior (Palmer, 1983: 42), interlocks appear to serve thisfunction for competing subcultures within the broader classof corporate elites. The "rise of shareholder power"documented by Useem (1992: 19), Davis and Thompson(1994), and others (e.g., Zajac and Westphal, 1995) appearsto have engendered a managerialist response that hasaffected the dynamic structure of board interlocks.

Our findings indicate that while organizational practices mayspread through the network of interlocking directorates,powerful actors in CEO-board relationships can block orredirect the diffusion of those changes that diminish theircontrol by cutting off interlock ties to other adopters ofthose practices and steering the diffusion of changes thatprotect or increase their control over the focal board byadding ties to prior adopters. The theoretical perspective andempirical findings of this study may generalize to thediffusion of other organizational phenomena by suggestinghow /nfraorganizational political interests can influence thecourse of /nferorganizational diffusion. Thus, while the role ofexisting social network ties in facilitating the diffusion ofvarious structural and strategic changes within anorganizational field is well understood (cf., Davis, 1991;Haunschild, 1993; Palmer, Jennings, and Zhou, 1993), futureresearch could consider further the endogeneity of such ties,i.e., how they are created or severed and how they mayreflect and reinforce the parochial political interests ofalready-powerful organizational decision makers.

Another question that merits additional research attention isthe origin of CEO-board power relationships and howreversals in that relationship might occur. Beatty and Zajac(1994) noted, for example, that the small and young firmsthey studied (i.e., initial public offering firms) havecompensation contracts and corporate governancemechanisms different from those found in large Fortune 500firms. Oualitative and quantitative historical approachesmight provide additional insight into how and why suchchanges occur over time and the implications of suchchanges on shifting the relative power balance in theCEO-board relationship.

We began our study by asking why increased boardindependence and control over top management may havespread to some organizations and not to others. We soughtto answer this question by developing a cross-level theory ofinterlock tie creation and dissolution that explains thesimultaneous coexistence and persistence of bothboard-controlled and CEO-controlled firms. Specifically, weshowed how powerful CEOs and powerful board membersseek directors whose prior director experiences (i.e.,reputation) suggest a shared belief about board passivity oractivity, respectively, and how this politically rational


Board Interlocks

"matching" and segmentation process of like-mindedindividuals can influence the dynamic formation anddissolution of corporate interlock ties. These dynamics helpto explain the fractured diffusion of increased boardindependence across large U.S. corporations. Moregenerally, the findings illustrate the promise of developing"meso-level" theoretical frameworks integratingmicropolitical factors and macro-social factors to explainorganizational behavior. While such "meso-level"frameworks may be more complex than frameworks thataddress only micro or macro factors, they may offer thegreatest potential to account for the simultaneous existenceof what appear to be contradictory or opposing organizationalphenomena.

REFERENCES

Alderfer, Clayton P.1986 "The invisible director on

corporate boards." HarvardBusiness Review, 64(6):38-52.

Amihud, Yakov, and Baruch Lev1981 "Risk reduction as a

managerial motive forconglomerate mergers." BellJournal of Economics, 12:605-617.

Amit, Raphael, and Joshua Livnat1988 "Diversification and the

risk-return trade-off."Academy of ManagementJournal. 31: 154-166.

Baysinger, Barry, and Robert E.Hoskisson1990 "The composition of boards of

directors and strategic control:Effects on corporatestrategy." Academy ofManagement Review, 15:72-87.

Beany, Randolph P., and EdwardJ. Zajac1994 "Top management incentives,

monitoring, and risk sharing:A study of executivecompensation, ownership andboard structure in initial publicofferings." AdministrativeScience Quarterly, 39:313-335.

Becker, Gary S.1975 Human Capital, 2nd ed. New

York: National Bureau ofEconomic Research.

Bettenhausen, Kenneth, and J.Keith Murnighan1985 "The emergence of norms in

competitive decision-makinggroups." AdministrativeScience Quarterly, 30:350-372.

Black, Fischer, and Myron Scholes1973 "The pricing of options and

corporate liabilities." Journalof Political Economy. 81:637-659.

Boeker, Warren1992 "Power and managerial

dismissal: Scapegoating at thetop." Administrative ScienceQuarterly. 37: 400-421.

Burt, Ronald S.1987 "Social contagion and

innovation: Cohesion versusstructural equivalence."American Journal ofSociology. 92: 1287-1335.

Crystal, Graef S.1984 Questions and Answers on

Executive Compensation.Englewood Cliffs. NJ:Prentice-Hall.

1991 In Search of Excess: TheQvercompensation ofAmerican Executives. NewYork: W. W. Norton.

Davis, Gerald F.1991 "Agents without principles?

The spread of the poison pillthrough the intercorporatenetwork." AdministrativeScience Quarterly, 36:583-613.

1993 "Who gets ahead in themarket for corporatedirectors: The politicaleconomy of multiple boardmemberships." Academy ofManagement Best PapersProceedings: 202-211.

Davis, Gerald F., and Tracy A.Thompson1994 "Asocial movement

perspective on corporatecontrol." AdministrativeScience Quarterly, 39:141-173.

Demb, Ada, and F.-FriedrichNeubauer1992 The Corporate Board:

Confronting the Paradoxes.Oxford: Qxford UniversityPress.

Ellig, Bruce R.1984 "Incentive plans: Qver the

long-term." CompensationReview. 16: 26-37.

Fama, Eugene F.1980 "Agency problems and the

theory of the firm." Journal ofPolitical Economy, 88:288-307.

Fama, Eugene F., and Michael C.Jensen1983 "The separation of ownership

and control." Journal of Lawand Economics, 26: 301-325.

Finkelstein, Sydney, and RichardA. D'Aveni1994 "CEQ duality as a

double-edged sword: Howboards of directors balanceentrenchment avoidance andunity of command." Academyof Management Journal, 37:1079-1108.

Finkelstein, Sydney, and DonaldC. Hambrick1989 "Chief executive

compensation: A study of theintersection of markets andpolitical processes." StrategicManagement Journal. 10:121-134.

Fortune1993 "The king is dead: Why more

boards are waking up andpushing out the CEQ/'January 11: 34-50.

Fredrickson, James W., Donald C.Hambrick, and S. Baumrin1988 "A model of CEQ dismissal."

Academy of ManagementReview, 13: 255-270.


Galaskiewicz. Joseph, and StanleyWasserman1989 "Mimetic processes within an

interorganizational field: Anempirical test." AdministrativeScience Quarterly, 34:454-479.

Gibbs, Philip A.1993 "Determinants of corporate

restructuring: The relativeimportance of corporategovernance, takeover threat,and free cash flow." StrategicManagement Journal, 14:51-68.

Gladstein. Deborah1984 "Groups in context: A model

of task group effectiveness."Administrative ScienceQuarterly, 33: 334-365.

Harris. Milton, and Artur Raviv1979 "Some results on incentive

contracts with imperfectinformation." Journal ofEconomic Theory, 20:231-259.

Harrison. J, Richard, David L.Torres, and Sal Kukalis1988 "The changing of the guard:

Turnover and structuralchange in thetop-management positions."Administrative ScienceQuarterly, 33: 211-232.

Haunschild. Pamela R.1993 "Interorganizational imitation:

The impact of interlocks oncorporate acquisition activity."Administrative ScienceQuarterly, 38: 564-592.

Healy. Paul M.1985 "The effect of bonus

schemes on accountingdecisions." Journal ofAccounting and Economics, 7:85-107.

Herman, Edward S.1981 Corporate Control, Corporate

Power. New York: CambridgeUniversity Press.

Hill. Charles W. L, and Scott A.Snell1988 "External control, corporate

strategy, and firmperformance inresearch-intensive industries."Strategic ManagementJournal, 9: 577-590.

Hoskisson. Robert E.. Richard A,Johnson, and Douglas D. Moesel1994 "Corporate divestiture

intensity in restructuringfirms: Effects of governance,strategy, and performance."Academy of ManagementJournal, 37: 1207-1251.

Jensen. Michael C.1988 "The takeover controversy:

Analysis and evidence." InJ. C. Coffee, L. Lowenstein,and S. Rose-Ackerman (eds.),Knights, Raiders, and Targets:314-354. Oxford: QxfordUniversity Press.

Jensen. Michael C, and WilliamH. Meckling1976 "Theory of the firm:

Managerial behavior, agencycosts, and ownershipstructure." Journal of FinancialEconomics, 3: 305-350.

Kaufman, Bruce E.1986 The Economics of Labor

Markets. Qrlando, FL: DrydenPress.

Kerr. Jeffrey L., and Leslie Kren1992 "Effect of relative decision

monitoring on chief executiveofficer compensation."Academy of ManagementJournal, 35: 370-397.

Kesner, Idalene F., and Roy B.Johnson1990 "An investigation of the

relationship between boardcomposition and stockholdersuits." Strategic ManagennentJoumal, 11: 327-336.

Koenig, Thomas, and RobertGogel1981 "Models of the significance of

interlocking corporatedirectorates." AmericanJoumal of Economics andSociology, 38: 173-186.

Kosnik. Rita D.1987 "Greenmail: A study of board

performance in corporategovernance." AdministrativeScience Quarterly, 32:163-185.

Lorsch. Jay W., and ElaineMaclver1989 Pawns or Potentates: The

Reality of America's CorporateBoards. Boston: HarvardBusiness School Press.

Mace, Myles L.1971 Directors: Myth and Reality.

Boston: Harvard BusinessSchool Press.

Maddala. G. S.1983 Limited-dependent and

Qualitative Variables inEconometrics. Cambridge:Cambridge University Press.

Main. Brian G.. Charles A, O'ReillyIII, and James Wade1995 "The CEO, the board of

directors and executivecompensation: Economic andpsychological perspectives."Industrial and CorporateChange, 4: 293-332.

Mallette. Paul, and Karen L.Fowler1992 "Effects of board composition

and stock ownership on theadoption of 'poison pills.' "Academy of ManagementJournal, 35: 1010-1035.

Marris, Robert L.1964 The Economic Theory of

Managerial Capitalism.London: MacMillan.

Mizruchi. Mark S.1982 The American Corporate

Network. Beverly Hills, CA:Sage.

1992 The Structure of CorporatePolitical Action: InterfirmRelationships and TheirConsequences. Cambridge,MA: Harvard University Press.

Mizruchi. Mark S.. and LindaBrewster Stearns1988 "A longitudinal study of the

formation of interlockingdirectorates." AdministrativeScience Quarterly, 33:194-210.

Ocasio, William1996 "Boards as normative

arenas." Working paper.Northwestern University, J. L.Kellogg Graduate School ofManagement.

Ornstein. Michael D,1980 "Assessing the meaning of

corporate interlocks: Canadianevidence." Social ScienceResearch, 9: 287-306.

Palepu, Krishna G,1985 "Diversification strategy, profit

performance, and the entropymeasure." StrategicManaqement Journal, 6:239-255.

Palmer, Donald1983 "Broken ties: Interlocking

directorates and intercorporatecoordination." AdministrativeScience Quarterly, 28: 40-55.

Palmer. Donald, Roger Friedland.and Jitendra V. Singh1986 "The ties that bind:

Organizational and class basesof stability in a corporateinterlock network." AmericanSociological Review, 51:781-796.

Palmer, Donald A.. P. DevereauxJennings, and Xueguang Zhou1993 "Late adoption of the

multidivisional form by largeU.S. corporations:Institutional, political, andeconomic accounts."Administrative ScienceQuarterly, 38: 100-131.


Board Interlocks

Pennings, Johannes H.1980 Interlocking Directorates. San

Francisco: Jossey-Bass.

Pfeffer, Jeffrey1981 Power in Organizations. New

York: Harper and Row.1992 Managing with Power.

Boston: Harvard BusinessSchool Press.

Pfeffer, Jeffrey, and Gerald R.Salancik1978 The External Control of

Organizations: A ResourceDependence Perspective.New York: Harper & Row.

Seward, James K., and Jannes P.Walsh1996 "The governance and control

of voluntary corporatespin-offs." StrategicManagement Journal, 17:25-39.

Stearns, Linda Brewster, and MarkS. Mizruchi1986 "Broken tie reconstitution and

the functions ofinterorganizational interlocks:A reexannination."Administrative ScienceOuarterly, 31: 522-538.

Tosi, Henry L., and Luis R.Gomez-Mejia1989 "The decoupling of CEO pay

and performance: An agencytheory perspective."Administrative ScienceOuarterly, 34: 169-189.

Useem, Michael1982 "Classwide rationality in the

politics of managers anddirectors of large corporationsin the United States and GreatBritain." AdministrativeScience Ouarterly, 27:199-226.

1984 The Inner Circle. Oxford:Oxford University Press.

1992 Executive Defense:Shareholder Power andCorporate Reorganization.Cambridge. MA: HarvardUniversity Press.

Vance, Stanley C.1968 The Corporate Director: A

Critical Evaluation.Homewood, IL: DowJones-lrwin.

1983 Corporate Leadership: Boardsof Directors and Strategy.New York: McGraw-Hill.

Wade, James B., Charles A.O'Reilly, and Ike Chandratat1990 "Golden parachutes: CEOs

and the exercise of socialinfluence." AdministrativeScience Ouarterly, 35:587-603.

IVa// Street Journat1995a "Despite poor returns.

Champion's chairman hangson for 21 years." October31: A l .

1995b "Irate shareholders targetineffective board members."November 6: BI .

1995c "How do you fire a CEO?Very, very slowly." January20: BI .

Westphal, James D., and EdwardJ. Zajac1995 "Who shall govern?

CEO/board power,demographic similarity, andnew director selection."Administrative ScienceOuarterly, 40: 60-83.

1997 "Defections from the innercircle: Social exchange,reciprocity, and the diffusionof board independence in U.S.corporations." AdministrativeScience Ouarterly, vol. 42 (inpress).

Zajac, Edward J.1988 "Interlocking directorates as

an interorganizational strategy:A test of criticalassumptions." Academy ofManagement Journal, 31:428-438.

Zajac, Edward J., and James D.Westphal1995 "Accounting for the

explanations of CEOcompensation: Substance andsymbolism." AdministrativeScience Ouarterly, 40:283-308.

1996 "Who shall succeed? HowCEO/board preferences andpower affect the choice ofnew CEOs." Academy ofManagement Journal, 39:64-90.


Director Reputation, This study advances research on CEO...

Documents

Transcript of Director Reputation, This study advances research on CEO...