® Academy of Management Journal1999, Vol. 42, No. 1, 7-24.

COLLABORATION IN THE BOARDROOM:BEHAVIORAL AND PERFORMANCE CONSEQUENCES OF CEO-

BOARD SOCIAL TIES

JAMES D. WESTPHALUniversity of Texas at Austin

Empirical research has typically rested on the assumption that board independencefrom management enhances board effectiveness in administering firms. The presentstudy shows how and when a lack of social independence can increase board involve-ment and firm performance by raising the frequency of advice and counsel interactionsbetween CEOs and outside directors. Hypotheses were tested with original survey datafrom 243 GEOs and 564 outside directors on behavioral processes and dynamics inmanagement-board relationships.

In recent years, corporate stakeholders have ex-pressed increased concern about the inflnence ofboards of directors over corporate affairs. Large in-stitutional investors, for instance, have stronglycriticized the relationships that are thonght to existbetween top managers and outside directors inmany corporations. In particular, advocates of boardreform have commonly suggested that boards lackindependence from top management and tbat tbeirdependence fosters board passivity in the decision-making process. This perspective is reflected in thepopular media, where it is frequently assumed thatsocial ties between top managers and outside direc-tors (e.g., friendships) diminish board effectiveness;such relationships are often described in pejorativeterms as "chummy" or even "collusive" [Wall Streetfournal, 1993: Bl; 1995, 1996).

Academic research on boards has also devotedincreased attention to how CEO-board relation-ships influence board effectiveness. Empirical re-searchers have often assumed that a lack of socialindependence from management can compromise

This study was generously funded by the State FarmGompanies Foundation. I am also indebted to EdwardZajac for providing financial support and assistance onthis project. Thanks also to Gautam Ahuja, Mason Gar-penter, Jennifer Ghatman, Alison Davis-Blake, JamesFredrickson, Ranjay Gulati, Paul Hirsch, Robert Hoskis-son, Herminia Ibarra, Fiona Lee, Priti Shah, Mark Shan-ley, Ann Tenbrunsel, Brian Uzzi, and seminar partici-pants at the Massachusetts Institute of Technology, theUniversity of Michigan, the University of North Garolinaat Ghapel Hill, the University of Notre Dame, the Univer-sity of Texas at Austin, the University of Washington,and Washington University for their helpful commentson an earlier version of this article. Mike Frandsen andDennis Zhang provided valuable assistance in data col-lection.

board effectiveness in the strategy-making process.It has been proposed, for instance, that CEOs keeptheir boards largely passive and uninvolved in stra-tegic decision making through cooptation, or pack-ing boards with their supporters (e.g., Herman,1981; Mace, 1986; Wade, O'Reilly, & Chandratat,1990). Outside directors are thought to engage inless vigilant monitoring and to exert less controlover top managers with whom they have close per-sonal ties (e.g., Fredrickson, Hambrick, & Baumrin,1988; Spencer, 1983; Walsh & Seward, 1990).

The present study departs from this dominantview of how a lack of board social independencefrom management affects a board's contribution tostrategic decision making. The perspective devel-oped in this study first builds on prior literature inwhich a board's role in an organization has beenbroadly conceived as including two different formsof administration: the provision of advice andcounsel and the exercise of oversight and control(cf. Pfeffer & Salancik, 1978). I then develop a the-oretical framework that draws from the literatureson advice seeking and social ties in organizations toconsider how social factors such as trust and per-ceived social, obligations in CEO-board relationshipsmay promote rather than hinder boaid involvementand effectiveness in administering a firm. In particu-lar, although existing perspectives tend to suggestthat a lack of independence should reduce a board'sinvolvement and effectiveness in firm administra-tion, this study examines how social ties between topmanagers and outside directors may facilitate boardinvolvement by encouraging the provision of adviceand counsel in the strategy-making process.

The theoretical framework developed in thisstudy also addresses the issue of when this alterna-tive model of board social independence and in-volvement is more, or less applicable than the more

Academy of Management fournal February

traditional empirical model that focuses primarilyon how board independence affects control. In par-ticular, I explore how managerial incentive align-ment may affect whether social ties between a CEOand board enhance involvement by encouragingCEO-board collaboration rather than reduce boardinvolvement by hindering vigilant control. Theproposed theoretical framework was tested with acomprehensive data set that combines longitudinalarchival data on board structure, CEO compensa-tion, and performance with primary survey datafrom a large sample of both top managers and out-side directors regarding processes and dynamics inCEO-board relationships. The following sectionfirst describes the traditional perspective on howsocial ties in CEO-board relationships affect boardinvolvement in strategic decision making and thendevelops an alternative "collaboration model" ofthe relationship between CEO-board social ties andboard involvement in firm administration.

CEO-BOARD SOCIAL TIES AND BOARDINVOLVEMENT

The Independent Board Model

Much of the empirical literature examining howCEO-board relationships influence board involve-ment in firm governance is predicated on the as-sumption that effective boards influence corporatestrategy and performance primarily by monitoringmanagement on behalf of shareholders. Forinstance, Walsh and Seward (1990) describedthe board as an internal control mechanism, andKosnik (1987, 1990) emphasized the role of outsidedirectors in disciplining managerial decision mak-ing. Moreover, governance researchers havestressed that effective corporate directors serveshareholders by actively evaluating managerial per-formance (Boyd, 1994; Rechner & Dalton, 1991;Westphal & Zajac, 1995).

Theoretical support for the importance of boardmonitoring as a form of involvement is rooted inagency theory (Jensen & Meckling, 1976). Accord-ing to this perspective, the function of boards is toreduce agency costs resulting from the delegationof strategic decision making, or "decision manage-ment," to top executives by exercising "decisioncontrol," which involves monitoring managerial de-cision making and performance (Fama & Jensen,1983: 303). In doing so, boards rely crucially uponoutside directors, who are considered less likely thaninsiders to "collude with managers to expropriateresidual claimants" (Fama & Jensen, 1983; 315). Theformal independence possessed by outsiders is as-sumed to permit more objective evaluation. Several

studies have examined the influence of board inde-pendence on director involvement in strategic deci-sion making. Johnson, Hoskisson, and Hitt (1993) andJudge and Zeithaml (1992) provided evidence thatincluding more outside directors on a board in-creased board involvement (cf. Pearce & Zahra, 1991).These studies have suggested that the presence ofmore outside directors might proinote involvementby raising the level of monitoring and control.

Researchers have also recognized variation in theextent to which outside directors are truly indepen-dent. Although outside board members are for-mally independent of top management, powerfulsocial and psychological factors are thought tocompromise their willingness and ability to objec-tively monitor managerial performance. From thisperspective, CEOs use their influence over directorselection to render boards passive by favoring theappointment of personal friends and other individ-uals with whom they share close social ties (Finkel-stein & Hambrick, 1988; Johnson et al., 1993; Kim-berly & Zajac, 1988). Researchers have alsotheorized that social ties are created through theappointment process itself. Board appointmentsconfer prestige and status, as well as financial re-wards and perquisites. Thus, given norms of reci-procity, outside directors should feel socially obli-gated to support CEOs who favored theirappointment (Johnson et al., 1993; Wade et al.,1990). Overall, the dominant perspective in priorresearch on CEO-board relations suggests that per-sonal social ties and obligations between managersand directors critically impair a board's capacity tomonitor and control management decision makingand performance, thus diminishing effective boardinvolvement in the strategy-making process.

The Collaborative Board Model

Although much of the empirical literature onboard involvement is based on the assumption thatboards contribute to strategy primarily by monitor-ing management, the larger literature on boardssuggests an additional way in which directors caninfluence strategy. Pfeffer and Salancik's (1978;170) influential discussion of the different possiblefunctions performed by outside directors distin-guished between a board's role as an administrativebody and its role in linking an organization with itsenvironment. They further identified two distinctfunctions within the broader administrative role;the provision of expert advice and counsel and theexercise of oversight and control.

Several other scholars have also suggested thatboards can extend their involvement beyond mon-itoring to the provision of ongoing advice and

1999 Westphal

counsel on strategic issues (e.g., Baysinger & Butler,1985; Gomez-Mejia & Wiseman, 1997; Johnson,Daily, & EUstrand, 1996; Whisler, 1984; Zahra &Pearce, 1989), Bacon and Brown (1975: 18) andLorsch (1989) described how directors may serve asa sounding hoard for management in addition toexercising control, and Alderfer (1986) suggestedthat the hest CEO-board relationships included on-going collaboration in the decision-making process(cf. Spencer, 1983), Although several prior studieshave thus considered the board role in providingadvice, there is some disagreement about the role ofoutside directors in exercising this function,Baysinger and Butler (1985) suggested that outsidedirectors serve primarily to exercise control and thatinside directors are the main source of advice onstrategic issues. Daily and Dalton (1994) argued, how-ever, that outside directors can provide access tovaluable information—about, for instance, how to se-cure needed resources from the envirormient—in ad-dition to exercising control (cf. Judge & Zeithaml,1992; Pearce & Zahra, 1991; Pfeffer & Salancik, 1978),

Thus, the provision of advice has been recog-nized as a potentially important form of board in-volvement, but empirical researcbers have neitherexplicitly modeled advisory relations nor exam-ined how social factors may enhance a board's abil-ity to exercise tbis function, Wbereas tbe indepen-dent board model focuses on bow social ties in tbeCEO-board relationsbip may diminish board in-volvement in firm decision making by reducingmonitoring activity, it is suggested bere that boardsmay provide advice and counsel as well as engagein control and tbat social ties may increase theprominence of advisory interactions as a form ofinvolvement. Social ties with outside directorsshould enhance the propensity of top managers tosolicit their advice on strategic issues while alsoincreasing the outside directors' tendency to offersuch advice. According to the larger literature onadvice seeking, a primary inhibitor to seeking ad-vice is the perceived effect it could have on theadvice seeker's status. Employees tend to believethat others will view their need for assistance as anadmission of uncertainty or dependency and as anindication that they are less than fully competent orself-reliant (Blau, 1955; Rosen, 1983), Moreover,individuals seeking advice from a superior mayneed to disclose the existence of problems and toadmit tbeir own limitations in solving them—im-plicitly or explicitly—thus relinquishing power de-rived from information asymmetry in the agencyrelationship (Jensen & Meckling, 1976),

This concern about losing status has been shownto inhibit a variety of different kinds of adviceseeking, Allen (1977) showed that engineers sought

relatively few technical ideas from others becausetbey feared that such requests would engenderskepticism about tbeir competence, Asbford andNorthcraft (1992) found evidence that feedhackseeking by managers was impeded by impressionmanagement concerns, or a fear of appearing un-certain or dependent, and tbese concerns were par-ticularly salient in an evaluative context, wben tbeinformation providers (e,g,, directors) were in aposition to evaluate tbe person seeking information(e.g,, a CEO). Studies have shown, however, thatpersonal relationships increase employees' ten-dency to seek advice from either friends or othersin their work units by creating a sense of socialsecurity tbat reduces the perceived risk (Anderson& Williams, 1996; Fischer, 1982; Rosen, 1983), Per-sonal ties encourage advice seeking by enbancingmutual trust. Trust bas been described as tbe will-ingness to take risks, or to make oneself vulnerable(cf, Mayer, Davis, & Scboorman, 1995). Tbus, socialties sbould encourage CEO advice seeking by in-creasing a CEO's willingness to take tbe perceivedsocial and professional risks associated with suchbehavior. Much research in organizational behaviorhas shown that interpersonal trust promotes coop-erative problem-solving activity in groups (Zand,1972) and enhances upward communication in supe-rior-subordinate dyads (Roberts & O'Reilly, 1974),

Tbe literature on belping bebavior also suggeststbat people feel more comfortable offering advice toindividuals with whom they have a social relation-ship, and research on friendship has shown thatpeople feel socially obligated to give advice to friendswho appear in need of assistance (Shah & Jehn, 1993),Moreover, dyadic social ties may have additional,hoard-level effects. The small-groups literature sug-gests that as the proportion of individuals offeringadvice increases, social conformity pressures may in-crease the likelihood that remaining group memberswill engage in similar behavior (Hackman, 1992).

•Thus, the perspective developed in this studysuggests that although boards may engage in bothmonitoring and advice giving, CEO-board socialrelations can determine the relative emphases theyplace on these roles; according to the independentboard model, social ties should decrease involve-ment by reducing monitoring activity; in contrast,according to tbe collaboration model, such rela-tions should increase involvement by encouragingthe provision of advice and counsel on strategicissues. I next develop specific hypotheses,

CEO-board friendship ties. Silver characterizedfriendship as "protypical [of] personal relationsthat are normatively free of instrumental and cal-culative orientations" (1990: 1474), Krackhardt(1992) noted that friendship implies trust, or the

10 Academy of Management Journal February

expectation of personal loyalty. Similarly, Segal(1979) noted that certain social obligations are nor-matively part of the friendship relationship: forinstance, a friend is expected to come to one's aidor defense when needed. The friendship relation isgoverned by communal norms, whereby individu-als are obligated to care for eacb otber's welfare,ratber tban exchange-hased norms, whereby indi-viduals are concerned with reciprocation of bene-fits (Clark & Mills, 1982; Jebn & Sbaw, 1997), Tbus,friendsbip ties between a CEO and outside direc-tors sbould increase a board's loyalty to tbe CEO,Altbougb the independent hoard model suggeststhat such loyalty should diminish board monitor-ing activity, tbe collaboration model discussedabove suggests tbat perceived friendship ties mayincrease a CEO's advice-seeking bebavior by en-hancing his or her trust in the board's support,wbile also enbancing tbe board's perceived socialobligation to provide assistance. Thus,

Hypothesis la. Friendship ties between a CEOand a board will be negatively associated withthe board's monitoring of the CEO.

Hypothesis 2a. Friendship ties between a CEOand a board will be positively associated withinteractions between them involving adviceand counsel on strategic issues.

Director appointment by a CEO. As discussedabove, norms of reciprocity may cause outsideboard members to feel socially obligated to supportthe CEO who was responsible f'or nominating them toa board (Boeker, 1992; Daily & Dalton, 1995), In fact,several empirical studies have provided evidencethat the portion of a board appointed by a CEO canenhance the board's support for tbe CEO's leadersbip.Studies bave linked tbe percentage of CEO appoint-ments to the size of CEO compensation packages andexplanations for CEO incentive compensation as wellas to the adoption of antitakeover provisions (Lam-bert, Larcker, & WeigeU, 1993; Main, O'Reilly, &Wade, 1995; Zajac & Westphal, 1995), Thus, the con-ventional independent board model would suggestthat boards comprised largely of a CEO's appointeeswould be less likely to engage in vigilant monitoringof CEO performance, because most of tbe directorsfeel socially obligated to return the favor of appoint-ment hy supporting the CEO's leadership. Accordingto the collaboration model, bowever, a bigb propor-tion of outside directors appointed by a CEO mayengender a closer, more collaborative working rela-tionsbip by enhancing the CEO's trust in the supportof his or her directors. Thus,

Hypothesis lb. The portion of outside directorsappointed after a given CEO's appointment

will be negatively associated with board mon-itoring of the CEO.

Hypothesis 2b. The portion of outside directorsappointed after a given CEO's appointmentwill be positively associated with interactionsbetween the CEO and board involving adviceand counsel on strategic issues.

Incentive Alignment and Board Involvement

The discussion thus far has developed two mod-els of how CEO-board relations affect board in-volvement in managerial decision making. Thissection considers how normative agency theorymight help determine when each model is mostapplicable. On the one band, the costs of lowerlevels of board control may depend on whetherpartial substitutes for monitoring are available, Al-tbougb some level of board monitoring is necessaryand valuable, agency tbeorists recognize financialincentives as a partial substitute for monitoringactivities, suggesting that the need for monitoringdeclines as agent incentives are aligned moreclosely with the interests of principals (Gomez-Mejia & Balkin, 1992; Gomez-Mejia & Wiseman,1997; Kosnik, 1990), Empirical evidence suggeststhat relatively large reductions in agency costs arederived from the mere introduction of financialincentives, or from relatively small levels of incen-tive alignment (Rediker & Seth, 1995; Zajac & West-phal, 1994), Tbus, when CEO incentives have beenintroduced, the agency costs associated with CEO-board social ties sbould be relatively small,• Further, CEO financial incentives should not

only reduce the costs of CEO-board social ties, butshould also enhance tbe benefits of sucb ties. Froman agency perspective, incentive alignment moti-vates a CEO to use corporate resources to tbe ad-vantage of sbarebolders (Jensen & Murphy, 1990).Several studies in the small groups literature haveshown that when mechanisms are in place to mo-tivate groups' memhers to pursue stakeholder ob-jectives, social cobesion facilitates productivegroup interaction; the absence of such mechanisms,however, can lead to a negative relationship be-tween cobesion and productive interaction (Guzzo& Shea, 1992; Schachter, EUerston, McBride, &Gregory, 1951; Seashore, 1954), Thus, incentivealignment should motivate CEOs to use their socialcapital as a resource in developing more effectivepolicies for shareholders. Accordingly, financialincentive alignment should moderate the effects ofCEO-board social ties on board involvement: whenincentive alignment is present, the collaborationmodel is more applicable, and wben incentive

1999 Westphal 11

alignment is ahsent, the independent hoard modelis more germane. Thus,

Hypothesis 3. The higher the level of CEO in-centive alignment, the weaker the negative re-lationship between (a) CEO-board friendshipties and board monitoring of a CEO and (b) theportion of outside directors appointed after theCEO and board monitoring of the CEO.

Hypothesis 4. The higher the level of CEO in-centive alignment, the stronger the positive re-lationship between (a) CEO-board friendshipties and interactions between the CEO andboard involving advice and counsel on strate-gic issues and (b) the portion of outside direc-tors appointed after the CEO and these inter-actions.

Board Involvement and Firm Performance

Prior evidence regarding the ultimate perfor-mance consequences of hoard independence frommanagement has been inconsistent (cf. Finkelstein& Hambrick, 1996). This inconsistency can be re-solved, in part, by specifying the different behav-ioral processes (that is, forms of board involvementin strategic decision making) that could mediatethis relationship (Cook & Camphell, 1979). Accord-ing to the collaboration model, social ties may ul-timately enhance firm performance by enablingboards to extend their involvement heyond deci-sion control to decision management (Fama &Jensen, 1983: 303). From this perspective, boardadvice and counsel complement board monitoringas a form of involvement: when involvement ex-tends beyond control to include advisory interac-tions, directors participate in formulating strategicdecisions as well as in evaluating them (throughperiodic CEO performance reviews, for instance),so that boards influence each major phase of thedecision-making process.

Several researchers have recognized the potentialvalue of outside director advice and counsel ingenerating strategic proposals. Johnson, Daily, andEllstrand (1996) suggested that advice and counselinteractions may enable top managers to tap thebreadth of knowledge possessed by outside direc-tors, thus complementing the depth of firm-specificknowledge held by insiders. Pfeffer and Salancik(1978) also discussed how advice and counsel fromoutside directors can hroaden the range of strategicoptions considered by management, and Judge andZeithaml (1992) suggested that fresh perspectivesand new information provided by outside directors

can help managers identify promising strategic op-portunities. This view is consistent with the largerliterature on group decision making, which hasshown that a closed leadership style characterizedby a failure to incorporate outside perspectives inmaking decisions can lead to truncated consider-ation of alternatives and adherence to faulty orobsolete assumptions (Janis, 1982; Tetlock, Peter-son, McCuire, Chang, & Feld, 1992). Conversely,the solicitation of outside opinions can help exposeunrealistic assumptions or alternatives and preventpremature conclusions.

Moreover, as agency theorists have acknowl-edged, although board monitoring can help par-tially resolve the agency problem between manag-ers and shareholders, its effectiveness is limited byinformation asymmetries in the CEO-board rela-tionship (Jensen & Meckling, 1976). Outside direc-tors may not become aware of organizational prob-lems in a timely fashion (cf. Baker, Jensen, &Murphy, 1988). To the extent that CEOs must oftendisclose information ahout organizational proh-lems in seeking advice from outside directors, ad-vice seeking reduces this information asymmetryand results in raore informed board involvement.This formulation is consistent with prior researchon leader-member exchange showing that superior-subordinate relationships characterized hy fre-quent advisory interactions (in addition to supervi-sory control interactions) yield the highestperformance (Bauer & Green, 1996). Thus, the qual-ity of board involvement may be higher wheremonitoring activity is complemented hy relativelyhigh levels of director advice and counsel.

This discussion suggests two additional hypoth-eses. In the first, a positive relationship betweenboard monitoring and subsequent firm perfor-mance is predicted; this hypothesis, together withHypotheses la and lb, addresses whether socialties ultimately reduce firm performance by lower-ing the level of board monitoring, as suggested bythe independent board model. The second hypoth-esis, together with Hypotheses 2a and 2b, addresseswhether social ties ultimately enhance firm perfor-mance by raising the level of hoard advice andcounsel, as the collaboration model suggests.

Hypothesis 5. Board monitoring of CEOs willbe positively associated with subsequent firmperformance.

Hypothesis 6. CEO-board interactions involv-ing advice and counsel on strategic issues willbe positively associated with subsequent firm

. performance.

12 Academy of Management fournal February

Incentive Alignment, Board Involvement, andFirm Performance

Finally, agency theory also suggests that CEOincentive alignment moderates the effects of boardinvolvement on firm performance. To the extentthat CEO incentives can partially suhstitute forhoard monitoring as a solution to the agency prob-lem (Jensen & Meckling, 1976), higher incentivealignment should lessen the need for monitoringactivity, reducing the relationship between moni-toring and performance. In effect, where CEO inter-ests are already well aligned with shareholder in-terests, there is less incremental benefit from amarginal increase in monitoring activity. In addi-tion, incentives should also increase the relation-ship hetween CEO advice seeking and perfor-mance. From an agency perspective, incentivealignment should focus managerial attention andeffort toward using corporate resources in a waythat henefits shareholders. Thus, incentives shouldmotivate managers not only to seek advice fromoutside directors (as argued above), but also to fol-low through and incorporate useful informationand perspectives gained from those interactionsinto strategic decisions. Thus,

Hypothesis 7. The higher the level of CEO in-centive alignment, the weaker the positive re-lationship between board monitoring of a CEOand subsequent firm performance.

Hypothesis 8. The higher the level of CEO in-centive alignment, the stronger the positive re-lationship between CEO-board interactions in-volving advice and counsel on strategic issuesand subsequent firm performance.

Overall, then, incentive alignment was expectedto moderate relationships between CEO-hoard so-cial ties, board involvement, and firm performance.As incentive alignment increases, the hehavioraland ultimate performance consequences of CEO-board social ties should hecome more consistentwith the collaborative board model and less consis-tent with the independent hoard model.

METHODS

Sample and Data Collection

The sample frame for this study consisted of 600companies randomly selected from the Eorbes1,000 index of U.S. industrial and service firms. Isent a questionnaire to all 600 CEOs from thesecompanies and, to permit interrater reliability as-sessments, sent a second survey to the outside di-rectors of each company whose CEO responded to

the first survey [N = 1,312). Directors who sat onmore than one hoard in the survey sample wereasked to respond for only one company (the com-pany was randomly selected and specified in thecover letter); similarly, CEOs who sat on anotherboard in the survey sample were not surveyedtwice. The surveys were distributed in April 1995.

Surveys of corporate top managers have oftensuffered from response rates of less than 25 percent,and response rates above 40 percent are excep-tional (Judge & Dobbins, 1995). To ensure the high-est possihle number of responses in this case, I tookthe following steps (Forsythe, 1977; Fowler, 1993;Croves, Cialdini, & Couper, 1992): (1) an in-depthpretest (described below) was used to streamlinethe survey, making it easier and more appealing tocomplete, (2) requests for participation linked thecurrent study with an ongoing series of surveys ontop management issues conducted hy a major busi-ness school to which hundreds of the surveyedCEOs' and directors' peers had responded, and (3)nonrespondents were sent a new questionnaireabout 21 days afrer the initial mailing. The re-sponse rate for CEOs was 44 percent (JV = 263), andfor directors, it was 43 percent {N = 564). Data onincentive alignment and ownership were obtainedfrom proxy statements, and data on hoard compo-sition and board structure were obtained fromproxies. Standard and Poor's Register, the Dun andBradstreet Reference, and Who's Who in Einanceand Industry. I used the COMPUSTAT and CRSP(Center for Research on Securities Prices) databasesto obtain performance, size, and industry data. Ar-chival data were unavailable for 20 of the compa-nies with responding CEOs, leaving a final sampleof 243 CEOs and an effective response rate of 41percent.^ Firms in the final sample represented 27different two-digit Standard Industrial Classifica-tion (SIC) categories in both the manufacturing andservice sectors, including financial services firmsand electric utilities, and ranged from single-prod-uct firms to highly diversified conglomerates. Inaddition, these firms ranged from 260 million to 58billion in sales, from 500 to 200,000 employees,and from 500 million to 195 billion in assets.

For the 543 companies for which complete archi-val data were available, I examined whether re-spondents and nonrespondents differed signifi-cantly on variables derived from archival sourcesusing the Kolmogorov-Smirnov test (Siegel & Castel-

^ On average, 2.14 directors responded per firm (min-imum of 0, maximum of 5).

1999 Westphal 13

Ian, 1988].^ This test assesses whether significant dif-ferences exist in the distribution of respondents andnonrespondents for a given variable. This is a strin-gent test, because if the sample distributions differsignificantly at any point, it is concluded that thesamples come from different populations. The resultsof this test provided consistent evidence across mul-tiple variables that respondents and nonrespondentscame from the same population; the variables wereincentive alignment, firm performance, size, and thenumber of a firm's directors appointed after its CEO[p = .32-.77).

Dependent Measures

To enhance the construct validity of the surveymeasures, I conducted a pretest involving in-depthpilot interviews with 22 top managers and boardmembers (cf. Fowler, 1993: 102). After each indi-vidual had completed the pilot questionnaire, Iasked him or her to identify questions that wereunclear, difficult to answer, or potentially subjectto bias. I also used these interviews to ensure thatquestions were interpreted as expected, to identifyimprovements to the format of the survey, and tomodify its length. To reduce response bias, multi-ple response formats were used, and items measur-ing each construct were scattered throughout thesurvey (DeVellis, 1991). Moreover, drawing on in-put from the pilot interviews, I carefully wordedthe final questions to minimize the likelihood ofsocial desirability bias.

Advice and counsel interactions and board mon-itoring were assessed with two multi-item scales inthe CEO survey (see Appendix A for the texts ofthese scales). In developing the wording of thequestions, I drew from available qualitative re-search suggesting how top managers and directorsdescribe CEO-board interaction and a board's rolevis-a-vis management; feedback from the pilot in-terviews was used to further improve the clarityand face validity of each question. The items were

^ Although this test could not be applied to boardleadership structure, a difference of proportions testshowed that respondents and nonrespondents did notdiffer significantly on the proportion of firms with sepa-rate CEO and board chair positions (.29 versus .27, p =.37). Respondents and nonrespondents were also not sig-nificantly different with respect to the number of hoardmemberships they held (p = .69), and separate analysesconfirmed that the subset of firms with a responding CEOand at least one responding outside director (n = 188)was also not significantly different from other firms inthe sample frame on the independent and dependentvariables measured with archival data.

factor-analyzed with the iterated principal factorsmethod. A "scree" test indicated two common fac-tors, and "promax" rotation verified that the adviceand monitoring items loaded on different factors asexpected, with loadings for each item greater than.5 on one factor and less than .2 on the other.Cronbach's alpha was .88 for the advice and coun-sel scale and .92 for the monitoring scale, suggest-ing acceptable interitem reliabilities (Nunnally,1978). I estimated advice and monitoring factorsusing the Bartlett method.

I examined interrater reliability by comparingCEO and outside director responses for the moni-toring and advice items, calculating kappa coeffi-cients for each item. Kappa is a correlation coeffi-cient that corrects for the expected level ofcorrelation between raters. The sample includedcompanies with a responding CEO and at least oneresponding outside director [n = 188). As shown inTable 1, kappa coefficients exceeded .75 for all butone survey item, and tbe overall kappa was .82,which represents an exceptional level of agreementaccording to criteria set forth by Fleiss (1981).^

Two measures of firm performance were used:return on equity (an accounting-based measure)and the market-to-book value of equity (a market-based measure). The latter measure gauges a firm'seffectiveness in creating value for shareholders bycomparing its market value with the cost of capitalcontributed by shareholders. Following Johnsonand colleagues (1993) and Judge and Dobbins (1995),I adjusted each measure for industry differences bysubfracting the average value for the firm's primaryindusfry. Performance was measured two years afterthe survey date^ {.yearf + 2)- hi other analyses, I mea-sured performance in year^ + 1, and the results re-ported below were substantively unchanged.

Independent Measures

The portion of a board appointed after a CEOwas calculated as the number of outside directorsappointed during a CEO's tenure divided by thetotal number of outside board members. To assessthe level of CEO-board friendship ties, CEOs wereasked to consider their personal relationships withoutside board members and to indicate (1) how

^ Correlations between the monitoring variable andseveral control variables provided further evidence forconstruct validity. The monitoring measure was posi-tively associated with archival variables that have heenused in prior studies as causal indicators of monitoring,including CEO-board chair separation (p = .34), directorstock ownership (p = .27), hlockholder ownership (p =.17), and institutional ownership (p = .20).

14 Academy of Management Journal February

TABLE 1Results of Interrater Reliability Assessment"

Agreement

Item'' Observed Expected

Board monitoring1. To what extent does the board monitor top management strategic decision 86.50% 24.86% .82. 24.80

making?2. To what extent does the board formally evaluate your performance? 85.65 21.16 .82 24.973. To what extent does the board defer to your judgment on final strategic 87.76 25.82 .84 25.47

decisions?Advice and counsel interactions

1. To what extent do you solicit board assistance in the formulation of corporate 79.32 24.00 .73 21.75strategy?''

2. To what extent are outside directors a "sounding board" on strategic issues? 85.73 21.11 .81 24.653. How often have directors provided advice and counsel in discussions outside 87.09 24.65 .83 25.14

of board/committee meetings (by telephone or in person)?Overall K .82 24.98

" N= 188. Where multiple outside directors responded for the same company, director responses were averaged. This procedure helpedto ensure that reliability estimates were not inflated by common perspectives derived from respondents' holding the same position.

^ The phrasing of each item is taken from the CEO survey; most items were altered appropriately for the director survey. For purposesof comparison across items, I calculated kappas for continuous-scale items by converting them into categorical variables (i.e., dividing intoquartiles).

" Z-statistics for all kappas are highly significant.'^ Although this item restricts advice and counsel to CEO-initiated interactions, the results reported for the regression analyses were

substantively unchanged when this item was excluded.

many they considered to be acquaintances but notfriends and (2) how many they considered to befriends. Excluding mere acquaintances allowed amore precise measure of perceived friendship (cf.Chatman & Brown, 1995). The number of perceivedfriends was then divided by the total number ofoutside board members. This approach to measur-ing friendship is commonly used in organizationalresearch (e.g.. Brass, 1984; Krackhardt, 1992). Itwas not necessary to validate the survey measure offriendship ties against some alternative, behavioralmeasure, because the theoretical arguments regard-ing advice seeking required only that CEOs per-ceived the presence of friendship. Moreover, a sep-arate analysis of interrater reliability showed a highlevel of agreement between CEOs and respondingboard members about the number of outsider direc-tors who were friends of the CEO (K = .76).*

I used two measures of CEO incentive alignment

"* Although high interrater reliability should allay con-cerns about response bias, I also conducted an analysis inwhich friendship was measured with the CEO surveyand advice was measured with the director survey; theresults were unchanged. Since the advice items precededthe friendship items on the director survey, these rela-tionships cannot be attributed to response bias. I alsoconducted an analysis in which friendship ties excludedrelationships formed after the director(s) joined theboards and, again, the results were unchanged; this find-

commonly used in the governance literature (e.g.,Bergh, 1995; Kosnik, 1990; Johnson et al., 1993).First, CEO ownership was measured as the numberof common shares owned by the CEO divided bythe total amount of common stock outstanding.This definition captures the extent to which anagency problem exists by measuring the extent towhich ownership and control are separated. It alsoreflects evidence in social psychology and organi-zational behavior for the "mere ownership effect"(Beggan, 1992: 229), which suggests that individu-als feel more motivated to protect property as theirownership stake increases, independent of theproperty's financial value. Second, long-term in-centive plan compensation indicated the extent towhich CEO compensation was contingent upon theachievement of specific performance goals. Coalspecificity has been identified as a critical determi-nant of motivation (Locke, Shaw, Saari, & Latham,1981), and Larcker (1983) and )arrell (1993) bothdiscussed how incentives can lengthen executives'time horizons and focus their attention on creatingshareholder value. Long-term incentive plan com-pensation was calculated as the total value of long-term incentive grants made in the year prior to the

ing rules out the possibility that advice interactions en-couraged the formation of friendship, rather than thereverse.

1999

survey date divided by total compensation in thatyear (cf. Westphal & Zajac, 1995).

Control Variables

Several control variables were included in theanalyses. First, "problemistic" search could leadfirms' leaders to react to poor performance withhigher levels of monitoring or advice interactions(Cyert & March, 1963). Prior studies have offeredmixed evidence on the relationship between firm per-formance and changes in board composition thoughtto increase monitoring capacity (Daily & Dalton,1995; Hermalin & Weisbach, 1988), but two recentstudies have provided evidence for a negative rela-tionship between firm performance and overall boardinvolvement (Johnson et al., 1993; Judge & Zeithaml,1992). Thus, I controlled for firm performance in theyear prior to the time the survey was administered inmodels of board monitoring and advice. I also con-trolled for the prior value of firm performance inmodels predicting subsequent performance, consis-tent witb prior research (e.g., Haveman, 1992).

Although prior evidence regarding the effect offirm size on board decision making is also mixed(Finkelstein & Hambrick, 1996), one might expectthe greater complexity involved in governing largeorganizations to make directors at such firms moreinclined to delegate decision-making responsibil-ity, reducing their involvement (Zajac & Westphal,1994). Thus, I controlled for sales (measured as alogarithm) in all models. Prior research has sug-gested that directors lack the structural power tocontrol management decisions when the CEO alsoserves as board chair, and some evidence linksboard leadership structure to responsible decisionmaking (Mallette & Fowler, 1992; Westphal &Zajac, 1995); thus, a variable indicating separationof the CEO and board chair positions was includedin the monitoring models (this variable was notnecessarily expected to influence advice interac-tions, and separate analyses confirmed that it wasunrelated to advice and did not affect the hypoth-esized results). I also controlled for CEO tenure.CEOs are most likely to consider major strategicchanges early in their tenures (Gabarro, 1985), solength of tenure may be related to board involve-ment in strategy making. Daily (1996) suggestedthat directors who are affiliated with a focal firmthrough a family or business relationsbip may baveaccess to valuable sources of information, raisingthe possibility that director affiliations could influ-ence the level of board involvement. Thus, I con-trolled for the percentage of outside directors whowere affiliated with a focal firm using two mea-sures: affiliated directors: family ties and affiliated

' Westphal 15

directors: business ties. The measures were basedon the Securities and Exchange Commission's def-initions of the relevant relationships.

Given some prior evidence that ownership byoutside directors or other external blockholdersmay increase a board's motivation and/or power toexercise control, I included both ownership mea-sures in the analyses (Bergh, 1995). Director own-ership was measured as the percentage of total com-mon equity held by outside directors, andblockholder ownership was measured as the per-centage held by owners of 5 percent or more of afirm's common voting shares, excluding managers,directors, and institutions (e.g., Bergh, 1995). Al-though prior evidence is mixed (e.g.. Bethel &Liebeskind, 1993; Black, 1998; Wabal, 1996), someauthors have suggested that institutional owners, inparticular, are active in compelling higher levels ofboard control (Davis & Thompson, 1994). Thus, Icontrolled for institutional ownership, measured asthe percentage of total common stock held by pen-sion funds, banks and trust companies, savings andloans, mutual fund managers, and labor unionfunds (Kochhar & David, 1996; Hill & Hansen,1991; Kochhar & David, 1996).^ Given that power-ful owners could also affect management decisionmaking independent of their influence over boardinvolvement, these variables were included in themodels of performance calculated for this study, aswell as in the models of board monitoring andadvice. Finally, I controlled for change in corporatestrategy in the models of performance, given thatstrategic change can disrupt organizations in a waythat impairs performance or can facilitate adapta-tion that enhances performance (cf. Hannan & Free-man, 1989; Zajac & Shortell, 1989). Finally, I mea-sured change in corporate strategy as the change inthe entropy measure of diversification over thetwo-year period prior to survey administration.

Analysis

Since the proposed relationships between CEO-board social ties, monitoring, and advice interac-tions were recursive, I used ordinary least squares(OLS) multiple regression analysis to estimatemodels of monitoring and advice (Johnston, 1984).The performance models were estimated usingtwo-stage least squares regression. To the extent

^ In separate analyses, I restricted this measure to own-ership by institutions classified as "pressure-resistant"hy Kochhar and David (1996), including public pensionfunds, mutual funds, and foundations. The results weresuhstantively unchanged from results presented below.

16 Academy of Management Journal February

that the models of board monitoring, advice, andperformance encompassed a larger system of equa-tions in wbicb some variables tbat predict boardmonitoring and/or advice might be independentlyrelated to performance, OLS regression could yieldbiased estimates in modeling performance(Jobnston, 1984), Two-stage least squares regres-sion corrects for this bias by generating reduced-form estimates of monitoring, advice, and prior per-formance and tben including predicted values f̂ omthese equations as instruments in a second-stageequation estimating subsequent firm performance,

RESULTS

Descriptive statistics and bivariate correlationsare displayed in Table 2, The results of multipleregression analyses of board monitoring and adviceand counsel interactions are provided in Table 3.Tbe first set of results was inconsistent with theindependent board model, Neitber tbe level ofCEO-board friendship ties nor the portion of afirm's board appointed after its CEO was signifi-cantly related to the level of board monitoring ac-tivity. The main effects did support tbe collabora-tive board model, however. Consistent withHypothesis 2, friendship ties and subsequent boardappointments were botb positively related to tbelevel of advice and counsel interactions on strategicissues,

Tbe results also generally supported tbe bypoth-

esized effect of tbe interaction between social tiesand incentive alignment on board monitoring andadvice interactions (Hypotbeses 3-4), As CEOownersbip or long-term incentive plan compensa-tion increased, the negative effect of friendship tieson board monitoring grew weaker (i,e,, more posi-tive), and tbe positive effect of such ties on adviceinteractions grew stronger. Similarly, for two of theinteraction terms, as incentives increased the neg-ative effect of subsequent board appointments onmonitoring grew weaker, and tbe positive effect ofsubsequent board appointments on advice interac-tions grew stronger,

A further analysis of simple effects (Jaccard, Tur-risi, & Wan, 1990) indicated that botb social tievariables were insignificant in predicting adviceinteractions at low levels of CEO ownership orlong-term incentive plan compensation (for exam-ple, at zero), and the significant effects in model 6indicate that social ties were positively related toadvice and counsel interactions at average levels ofeach incentive variable. Conversely, altbough theeffects of social ties on board monitoring were in-significant at average levels of CEO ownersbip andlong-term incentive plan compensation, tbey be-came significant at low levels of incentive align-ment. In general, these results support the proposi-tion that incentive alignment moderates the effectsof CEO-board social ties on board involvement.When incentive alignment is absent, the indepen-dent board model is more appropriate in predicting

TABLE 2Descriptive Statistics and Pearson Correlation Coefficients"

Variable Mean

0.390,31

7.590.050.33

s.d.

0.350.26

1,520,060.27

1

.23

-.05.04

-.16

2

.15

.02

.03

3

-.04.09

4

.09

10 11 12 13 14 15 16

1. CEO-board friendsbip ties2. Portion of board appointed

after CEO3. Sales''4. CEO ownership5. Long-term incentive plan

compensation6. Director ownersbip7. Blockbolder ownership8. Institutional ownersbip9. CEO tenure

10. Affiliated directors: Family ties11. Affiliated directors: Business ties

0.02 0.06 -.03 .00 -.14 .13 .100.17 0.22 .01 -.04 .18 -.11 -.09 -.080.29 0.16 -.06 .04 .16 -.07 -.12 -.11 -.226.99 6.02 .08 .27 .06 -.22 -.15 -.17 -.03 -.020.08 0.11 .12 .01 -.03 -.05 -.09 .03 .02 -.04 .110.35 0.24 .09 .01 .14 -.07 -.06 -.18 -.11 -.05 .04 -.03

12. Separation of CEO and board cbair 0.29 0.46 -.27 -.26 ,13 .15 .13 .02 -.01 .05 -.22 .02positions

0.01 0.65 .02 -.03 -.05 .03 .01 -.04 .060.00 0.87 -.10 -.08 -.04 -.16 -.19 .27 .170.00 0.81 .24 .26 -.06 .14 .06 .12 .060.00 0.07 .14 .11 -.06 .05 .02 .08 .070.03 0.76 .20 .13 -.03 .08 .04 .10 .12

.01

13. Cbange in corporate strategy14. Board monitoring15. Advice and counsel interactions16. Return on equity17. Market-to-book value of equity

.09 -.08 -.05 .02 .04

.20 -.08 -.12 -.09 .34 .03

.07 -.13 .11 .19 -.05 .05 -.14

.14 .01 .03 .04 .02 .05 .22 .25

.08 -.02 .06 .01 .03 .07 .18 .33 .19

N = 243. Correlations greater than .13 are significant at p < .05.Logarithm.

1999 Westphal 17

TABLE 3Results of OLS Regression Models"

1.2.3.4.5.6.7.

8.9.

10.11.12.13.14.

15.16.17.

18.

Variable

CEO-board friendship tiesPortion of board appointed after CEOPrior return on equityPrior market-to-book value of equitySales'"CEO ownershipLong-term incentive plan°compensationDirector ownershipBlockholder ownershipInstitutional ownershipCEO tenureAffiliated directors: Family tiesAffiliated directors: Business tiesSeparation of CEO and board chairpositionsFriendship ties X CEO ownershipFriendship ties X LTIP compensationPortion of board appointed afterCEO X CEO ownershipPortion of board appointed afterCEO X LTIP compensation

Constant

FR^AR2

Model 1

-0.77 (0.50)-0.06 (0.05)-0.01 (0.04)-1.47 (0.65)*-0.47 (0.22)*

2.46 (1.00)*0.46 (0.26)0.55 (0.33)

-0.01 (0.01)-1.05 (0.56)-0.40 (0.28)

0.50 (0.13)***

0.20 (0.37)

5.19***.18

Monitoring

Model 2

-0.15 (0.16)-0.27(0.21)-0.76(0.50)-0.06 (0.05)-0.01 (0.04)-1.46 (0.65)-0.46 (0.22)*

2.43 (1.00)*0.36 (0.26)0.56(0.33)

-0.01 (0.01)-1.00(0.54)-0.36(0.27)

0.50(0.13)***

0.24 (0.38)

5.04***.19.01

Model 3

-0.15 (0.16)-0.26(0.21)-0.75 (0.50)-0.60 (0.05)-0.01 (0.04)-1.41 (0.64)*-0.45 (0.21)*

2.33 (1.00)*0.34 (0.25)0.56(0.33)

-0.00 (0.01)-1.00(0.54)-0.36 (0.26)

0.49 (0.13)***

7.32 (2.34)***1.29 (0.61)*7.02 (3.63)*

1.26(0.82)

0.30 (0.36)

9.27***.38.19***

Advice

Model 4

-0.94 (0.45)-0.03 (0.05)-0.06 (0.04)

0.57 (0.58)0.26 (0.20)

1.29 (0.89)0.09 (0.23)0.16 (0.28)

-0.01 (0.01)0.64 (0.51)0.55 (0.24)

-0.04 (0.32)

2.56*.08

and Counsel Interactions

Model 5

0.36 (0.15)*'0.41 (0.20)*

-0.86 (0.45)-0.03 (0.05)-0.05 (0.03)

0.55 (0.58)0.24 (0.19)

1.21 (0.88)0.08 (0.23)0.16 (0.28)

-0.01 (0.01)0.59 (0.50)

* 0.52 (0.24)*

-0.05 (0.31)

4.32***.16.08**

Model 6

* 0.37(0.15)**0.41 (0.20)*

-0.72 (0.44)-0.02 (0.05)-0.05 (0.03)

0.54 (0.57)0.23 (0.19)

1.41 (0.88)0.08 (0.22)0.16(0.27)

-0.01 (0.01)0.58 (0.49)0.50(0.24)*

5.36 (2.12)**1.58 (0.55)**4.36 (3.37)

2.00 (0.72)**

-0.04 (0.30)

10.04***.40.14***

" N = 243. Standard errors are in parentheses. T-tests were one-tailed for hypothesized effects and two-tailed for control variables.^ Logarithm." Abbreviated below as "LTIP."

* p £ .05** p < .01

***p<.001

board involvement, and when incentive alignmentis average to high, the collaborative board modelfits better.

The results of a two-stage least squares regressionanalysis of firm performance are presented in Table4. Hypothesis 5 predicted a positive relationshipbetween board monitoring of CEOs and subsequentfirm performance. The results support this hypoth-esis for both return on equity and market-to-bookvalue. Hypothesis 6 is also supported: the level ofadvice and counsel interactions was positively re-lated to each measure of subsequent firm perfor-mance. Hypotheses 7-8 predicted that incentivealignment would determine when each kind of in-volvement is most effective. Neither kind of incen-tive alignment interacted with board monitoring toinfluence firm performance, but the results do sup-port Hypothesis 8: The relationship between adviceinteractions and subsequent firm performance be-came stronger as the level of CEO ownership or the

level of CEO long-term incentive plan compensa-tion increased for both measures of performance.^

DISCUSSION

The findings of this study appear to have impor-tant implications for the corporate governance lit-erature. The first set of results showed not only thatsocial ties typically fail to reduce the level of board

" I also conducted separate analyses of board advice,monitoring, and performance for the smaller sample offirms with a responding CEO and at least one respondingoutside director (AT = 188). Results for the hypothesizedeffects were substantively the same as the results re-ported in Tables 3 and 4. Results were also unchangedwhen friendship ties, advice, and/or monitoring weremeasured with responses to the director survey, whichreflects the high levels of interrater reliability reportedabove.

18 Academy of Management Journal February

TABLE 4Results of Two-Stage Least Squares Regression Models of Firm Performance"

Variable

1. Advice and counselinteractions

2. Board monitoring3. Prior return on equity4. Prior market-to-book

value of equity5. Sales'̂6. CEO ownersbip7. Long-term incentive

plan*̂ compensation8. Director ownersbip9. Blockbolder ownersbip

10. Institutional ownersbip11. Cbange in corporate

strategy12. Advice interactions x

CEO ownersbip13. Advice interactions X

LTIP compensation14. Board monitoring X

CEO ownersbip15. Board monitoring x

LTIP compensationConstant

FR^

Model 1

0.43 (0.03)***

-0.00 (0.00)0.03 (0.04)0.01 (0.02)

0.11 (0.07)0.02 (0.02)0.04 (0.02)0.00 (0.01)

0.05 (0.03)*

17.07***.36

Return on Equity

Model 2

0.01 (0.01)**

0.01 (0.01)**0.43 (0.03)***

-0.00 (0.00)0.03 (0.05)0.01 (0.02)

0.11 (0.07)0.02 (0.02)0.04 (0.02)0.00 (0.01)

0.05 (0.03)*

20.26***.47.11**

Model 3

0.01 (0.01)**

0.01 (0.01)**0.44(0.03)***

-0.00 (0.00)0.04 (0.05)0.02 (0.02)

0.11 (0.07)0.02 (0.02)0.04 (0.02)0.00 (0.01)

0.17(0.06)**

0.05 (0.02)**

-0.10 (0.06)

-0.03 (0.02)

0.04 (0.03)

25.74***.63,16***

Market-to-Book Value

Model 4

0,62 (0.04)**'

-0.01 (0.03)0.80 (0.52)0.19(0.17)

1.13 (0.80)0.29 (0.20)0.34 (0.26)0.07 (0.07)

0.03 (0.29)

19.33***.39

Model 5

.19 (0.06)**'

0.12 (0.05)*

0.62(0.04)**'

-0.03 (0.03)0.79 (0.78)0.19 (0,17)

1.11 (0.80)0.28 (0,20)0.34 (0.26)0.07 (0.07)

0.02 (0.35)

22.69***.50.11**

of Equity

Model 6

0.20(0.06)***

0.12 (0,06)*

0.46(0.04)***

-0.02 (0.03)0.78 (0.53)0.18(0.18)

1.16(0.81)0.28 (0.21)0.34 (0.26)0.07 (0.08)

1.56 (0.77)*

0.70(0.22)***

-0.71 (0.72)

-0.37(0.22)

0.03 (0.38)

29.37***.65.15***

° A? = 243. Standard errors are in parentbeses. T-tests are one-tailed for bypotbesized effects, two-tailed for control variables.^ Logaritbm." Abbreviated below as "LTIP."

* p < .05**p < .01

*** p < .001

monitoring activity, but also that such ties enhancethe provision of advice and counsel from outsidedirectors on strategic issues. These results offer per-haps the first large-sample empirical evidence thataddresses how boards can extend their involve-ment in corporate governance beyond monitoringto provide ongoing advice and counsel to manage-ment on strategic issues (Baysinger & Butler, 1985;Johnson et al,, 1996; Zahra & Pearce, 1989), Muchof the empirical literature on boards of directorshas proceeded from the assumption that social tiesin CEO-board relationships diminish the involve-ment and effectiveness of outside directors by re-ducing their tendency to control management de-cision making (e.g,, Johnson et al,, 1993; Tosi &Gomez-Mejia, 1989; Wade et al., 1990), The find-ings of this study suggest instead that such ties,rather than promoting a passive board role, mayincrease board involvement by encouraging collab-

oration between top managers and outside direc-tors in strategic decision making.

The findings are also consistent with results ofbehavioral research on advice seeking at lower lev-els of organizations. This literature has shown thatpersonal relationships between colleagues in awork unit increase the level of advice-seeking be-havior by alleviating impression management con-cerns, such as the fear of appearing uncertain ordependent, and concerns about revealing sensitiveinformation, especially when the colleagues are ina position to evaluate the advice seekers' perfor-mance. The results support the view that CEOs aremore willing to take the social and professionalrisks associated with advice-seeking behavior whenthey can rely on the loyalty of their boards. Similarly,the results are also consistent with research on friend-ship in organizations, which has shown that friend-ship ties among group members lead to higher levels

1999 Westphal 19

of task-related communication and mutual assistanceby enhancing interpersonal trust.

Further results showed that CEO incentive align-ment moderated relationships between social tiesand board involvement: Higher levels of CEO own-ership or long-term incentive compensation furtherenhanced the positive relationship between socialties and board advice while reducing the negativerelationship between social ties and board monitor-ing. This moderation is consistent with the viewthat management incentives decrease the need forboard monitoring as a control mechanism and alsomotivate. CEOs to use their social capital (their so-cial ties to directors) as a resource in strategic de-cision making. Although some prior research haslinked management incentives to corporate strategyand overall firm performance (e,g,, Abowd, 1990;Gibbs, 1993), the present study extends this litera-ture by addressing how incentives affect behavioralprocesses (here, advice seeking) that may mediatethese relationships.

The last set of results addresses the ultimate per-formance consequences of each kind of board in-volvement, showing that CEO-board collaborationand control are independently and positively re-lated to subsequent firm performance. These find-ings provide systematic empirical support for theview that boards can contribute to strategy not onlyby independently evaluating managements, butalso by serving as sounding boards for top manag-ers who are in the process of formulating strategy(Bacon & Brown, 1975: 18; Johnson et al., 1996;Lorsch, 1989). Together with the first set of results,these findings challenge the dominant assumptionsthat underlie most empirical research on boardsand that characterize social ties narrowly as fric-tions that reduce the efficiency of corporate gover-nance (Fama & Jensen, 1983; Hirsch, Michaels, &Friedman, 1987), Rather than impairing corporategovernance by reducing the vigilance of board mon-itoring, social ties can ultimately contribute to boardeffectiveness and firm performance by fostering col-laboration between CEOs and directors in the strate-gy-making process without reducing board control.

It might be suggested that CEOs are more com-fortable seeking advice from their friends and ap-pointees because those individuals may be less ex-perienced or knowledgeable than themselves, andthus less capable of challenging them. To explorethe validity of this interpretation, I analyzed theeffect of director expertise and experience on theobserved relationships. I used four indicators ofexpertise, averaged across directors (Finkelstein,1992; Wiersema & Bantel, 1992): the number ofdifferent management positions a director hadheld, the number of different functional areas the

director had worked in, the number of years thedirector had served as a top manager, and the di-rector's level of education. Factor analysis con-firmed that all four indicators loaded on the samefactor, A separate regression analysis showed thatthe effects of social ties on advice and counselinteractions were unchanged when director exper-tise and experience were controlled for. Also, asshown in Appendix B, the interaction between di-rector expertise and social ties had a positive effecton advice interactions, indicating that social tieswere especially likely to increase advice when thelevel of director expertise was relatively high.These results are inconsistent with the view thatsocial ties increase advice seeking because boardsthat have more social ties with CEOs are populatedby less experienced or knowledgeable directorswhose input might be less threatening to the CEOs,

Future Research Directions

Although this study focused on the administra-tive role of boards, future research should examinehow CEO-board cooperation and control affect aboard's ability to exercise its other major functions,including its role in managing resource depen-dence and enhancing organizational legitimacy.One could investigate, for instance, whether adviceand counsel interactions between CEOs and man-ager-directors (outside directors who serve as topmanagers at another firm) might help managers toidentify opportunities for interorganizational coop-eration (for instance, strategic alliances) that secureneeded resources and resolve interdependenciesbetween firms. Conversely, CEO-board relation-ships characterized only by independent boardcontrol might reduce the likelihood that CEOs andmanager-directors will identify and pursue suchopportunities. Similarly, although in prior researchon interlocking directorates the assumption hasbeen that managers use outside directors to secureneeded resources from the environment (for a re-view, see Mizruchi [1996]), the effectiveness of thiscooptation mechanism may hinge on the particularcontent of CEO-board relationships.

Studies are also needed that examine how differ-ent kinds of CEO-board relationships are perceivedby internal and external stakeholders and the con-sequences of these relationships for organizationallegitimacy. There is evidence that institutional in-vestors respond positively to changes in boardstructure that are thought to increase a board'smonitoring capacity (for a review, see Westphaland Zajac [1998]), It is uncertain, however, whetherexternal constituents also assess the actual behav-ioral processes occurring in CEO-board relation-

20 Academy of Management Journal February

ships (for example, monitoring and advice giving)in evaluating corporate governance. Given the po-tential decoupling of formal board structure and ac-tual behavior (Meyer & Rowan, 1977; Westphal &Zajac, 1994), it is important to examine the depth atwhich external stakeholders evaluate corporateboards. According to an institutional perspective,firms with strong CEO-board social ties could satisfyexternal constituents by adopting formal structuresthat implied independent control (the socially legiti-mate CEO-board relationship) but could decouplethese structures from actual board processes.

The findings of this study may also suggest newdirections for research on management incentives.The confirmed interactions between incentivealignment and social ties suggest that incentivescan be more effective in raising performance whenmanagers have more social connections to drawupon for information and advice. Compensationresearchers should further explore the role of socialcapital as an important contingency factor in theeffectiveness of incentive compensation plans. Re-searchers investigating management incentives inparticular might explicitly consider how managersare either constrained or enabled in responding toeconomic incentives by the social relationships inwhich they are embedded. Research is also neededthat explores in more detail how incentives affectCEO behavior such as advice seeking. Stewardshiptheory suggests that managers are motivated by theintrinsic satisfaction their work provides ratherthan by extrinsic rewards (Davis, Schoorman, &Donaldson, 1997), This view might appear incon-sistent with evidence that incentive compensationincreases the tendency for CEOs to draw on theirsocial ties for advice. However, it is possible thatincentives influence CEOs' behavior primarily bycommunicating goals and directing the CEOs' at-tention toward them, and not strictly by motivatingthe CEOs to increase their financial compensation.Future studies should examine the psychologicalmechanisms that mediate the effects of incentiveson CEO behavior toward boards.

Limitations and Managerial Implications

Several limitations of the study should be ac-knowledged. First, this research focused on adviceinteractions initiated by CEOs, A somewhat differ-ent set of social and economic factors may leaddirectors to initiate such interactions. Separatequestions in the survey indicated that a large por-tion of such advice interactions (approximately 92percent, on the average) were initiated by the CEOs,a finding that is consistent with prior descriptivesurveys suggesting that such interactions are typi-

cally CEO initiated (Alderfer, 1986; Bacon &Brown, 1975; Demb & Neubauer, 1992), Neverthe-less, board-initiated advice giving does occur, andit could become a more significant form of involve-ment in the future. Second, this study did not ad-dress the specific content or quality of board ad-vice—for instance, I did not examine whether andwhen directors furnished information about devel-opments in the external environment, as opposedto giving opinions about strategic options for cop-ing with them. In this study, I did not seek toexplain variation in the value or quality of adviceor the accuracy of information furnished by outsidedirectors, qualities that could moderate the effectsof collaboration on performance; unique informationor experience provided by outsiders could representan especially valuable resource (Pfeffer & Salancik,1978). Finally, as noted above, in this research I didnot investigate how board collaboration and controlaffected a board's ability to perform certain externalroles, such as managing resource dependence orbuilding organizational legitimacy.

The present findings may have important norma-tive implications for corporate boards. Institutionalinvestors, management consultants, and the popu-lar press have all strongly advocated greater boardindependence from management. Directors withclose ties to management are routinely character-ized in cynical terms as "pals of the CEO" or "hand-picked appointees," and their relationships toCEOs are described as overly "chummy" (e,g,. WallStreet Journal, 1993: Bl; 1995: Bl; 1996), Thisstudy suggests that in fact board effectiveness andultimately, firm performance may be enhanced byclose, trusting CEO-board relationships combinedwith moderate to high levels of CEO incentivealignment. Thus, rather than dividing top managersand outside directors into independent groups, afirm might profit by using team development tech-niques that unify managers and directors into asingle, cohesive team. Interestingly, although suchinterventions are commonly used to develop team-work among lower-level employees, they have notbeen applied at the highest levels of organizations.

Overall, this study contributes to the corporategovernance literature by providing empirical evi-dence on how CEOs and outside directors collabo-rate in the strategic decision making process andthen demonstrating that such collaboration inde-pendently and positively contributes to firm per-formance. The results challenge dominant assump-tions in the prior empirical literature on boards byshowing how social ties in CEO-board relation-ships can enhance rather than diminish board in-volvement and firm performance by encouragingadvice seeking in the strategic decision making pro-

1999 Westphal 21

cess. More generally, the findings suggest the valueof conducting rigorous empirical research that ex-amines the behavioral processes and dynamics thatunderlie corporate governance.

REFERENCES

Abowd, J, M, 1990, Does performance-based managerialcompensation affect subsequent corporate perfor-mance? Industrial and Labor Relations Review,43: 552-573,

Alderfer, C, P, 1986, The invisible director on corporateboards. Harvard Rusiness Review, 64: 38-52,

Allen, T, J, 1977, Managing the flow of technology:Technology transfer and the dissemination of tech-nology information within the R&D organization.Cambridge, MA: MIT Press,

Anderson, S, E,, & Williams, L, J, 1996, Interpersonal, job,and individual factors related to helping processes atwork. Journal of Applied Psychology, 81: 282-296,

Ashford, S, J,, & Northcraft, G, B, 1992, Conveyingmore (or less) than we realize: The role of impres-sion-management in feedback-seeking. Organiza-tional Rehavior and Human Decision Processes,53: 310-334,

Bacon, J,, & Brown, J, K, 1975, Corporate directorshippractices: Role, selection and legal status of theboard. New York: Conference Board,

Baker, G, P,, Jensen, M, G,, & Murphy, K, J, 1988, Gom-pensation and incentives: Practice vs, theory. Jour-nal of Finance, 43: 593-616,

Bauer, T, N,, & Green, S, G, 1996, Development of leader-member exchange: A longitudinal test. Academy ofManagement Journal, 39: 1538-1567,

Baysinger, B,, & Butler, H, N, 1985, Gorporate governanceand the board of directors: Performance effects ofchanges in board composition. Journal of Law, Eco-nomics, and Organization, 1: 101-124,

Beggan, J, K, 1992, On the social nature of nonsocialperception: The mere ownership effect. Journal ofPersonality and Social Psychology, 62: 229-237.

Bergh, D, D, 1995, Size and relatedness of units sold: Anagency theory and resource-based perspective. Stra-tegic Management Journal, 16: 221-239,

Bethel, J, E., & Liebeskind, J, 1993, The effects of owner-ship structure on corporate restructuring. StrategicManagement Journal, 14: 15-32,

Black, B, S, 1998, Shareholder activism and corporategovernance in the United States, In P, Newman (Ed,),The new Palgrave dictionary of economics and thelaw: Forthcoming,

Blau, P, M, 1955, The dynamics of bureaucracy. Ghi-cago: University of Ghicago Press,

Boeker, W, 1992, Power and managerial dismissal:

Scapegoating at the top. Administrative ScienceQuarterly, 37: 400-421,

Boyd, B, K, 1994, Board control and GEO compensation.Strategic Management Journal, 15: 335-344,

Brass, D, J, 1984, Being in the right place: A structuralanalysis of individual influence in an organization.Administrative Science Quarterly, 29: 518-539,

Ghatman, J, A,, & Brown, R, A, 1995, It takes two totango: Understanding the personal and situationalantecedents of friendship at the individual anddyadic levels. Paper presented at the annual meet-ing of the Academy of Management, Vancouver,

, Glark, M, S,, & Mills, J, R, 1982, Gommunal and exchangerelationships. In L, Wheeler (Ed,), i?evjeiv of per-sonality and social psychology. Beverly Hills, GA:Sage,

Gook, T, D,, & Gampbell, D, T, 1979, Quasi-experimen-tation: Design and analysis issues for field set-tings. Ghicago: Rand McNally,

Gyert, R, M,, & March, J, G, 1963, A behavioral theory ofthe firm. Englewood Gliffs, NJ: Prentice-Hall,

Daily, G, M, 1996, Governance patterns in bankruptcyreorganizations. Strategic Management Journal,17: 355-375,

Daily, G, M,, & Dalton, D, R, 1994, Bankruptcy and cor-porate governance: The impact of board compositionand structure. Academy of Management Journal,37: 1603-1617,

Daily, G, M,, & Dalton, D, R, 1995, GEO and directorturnover in failing firms: An illusion of change?Strategic Management Journal, 16: 393-400,

Davis, J, H,, Schoorman, F, D,, & Donaldson, L, 1997,Toward a stewardship theory of management. Acad-emy of Management Review, 22: 20-47,

Davis, G, F,, & Thompson, T, A, 1994, A social movementperspective on corporate control. AdministrativeScience Quarterly, 39: 141-173.

Demb, A,, & Neubauer, F, 1992, The corporate board:Confronting the paradoxes. Oxford: Oxford Univer-sity Press,

DeVellis, R, F, 1991, Scale development: Theory andapplications. Newbury Park, GA: Sage,

Fama, E, F,, & Jensen, M, G, 1983, The separation ofownership and control. Journal of Law and Eco-nomics, 26: 301-325,

Finkelstein, S, 1992, Power in top management teams:Dimensions, measurement, and validation. Acad-emy of Management Journal, 35: 505-538,

Finkelstein, S,, & Hambrick, D, G, 1988, Ghief executivecompensation: A synthesis and reconciliation. Stra-tegic Management Journal, 9: 543-558,

Finkelstein, S,, & Hambrick, D, G, 1996, Strategic lead-ership: Top executives and their effects on organi-zations. St, Paul: West,

22 Academy of Management fournal February

Fischer, C. S. 1982. What do we mean by "friend'? Aninductive study. Social Networks, 3: 287-306.

Fleiss, J. L. 1981. Statistical methods for rates and pro-portions. New York: Wiley.

Forsythe, J. B. 1977. Obtaining cooperation in a surveyof business executives. Joumal of MarketingResearch, 14: 370-373.

Fowler, F. J. 1993. Survey research methods. NewburyPark, CA: Sage.

Fredrickson, J. W., Hambrick, D. C, & Baumrin, S. 1988.A model of CEO dismissal. Academy of Manage-ment Review, 13: 255-270.

Gabarro, J. G. 1985. When a new manager takes charge.Harvard Business Review, 63(3): 110-123.

Gibbs, P. A. 1993. Determinants of corporate restructur-ing: The relative importance of corporate gover-nance, takeover threat, and free'cash flow. StrategicManagement Journal, 14: 51-68.

Gomez-Mejia, L. R., & Balkin, D. B. 1992. Determinants offaculty pay: An agency theory perspective. Academyof Management Journal, 35: 921-955.

Gomez-Mejia, L. R., & Wiseman, R. M. 1997. Reframingexecutive compensation: An assessment and out-look. Journal of Management, 23: 291-374.

Groves, R. M., Gialdini, R. B., & Couper, M. P. 1992.Understanding the decision to participate in a sur-vey. Public Opinion Quarterly, 56: 475-495.

Guzzo, R. A., & Shea, G. P. 1992. Group performance andintergroup relations in organizations. In M. D. Dun-nette & L. M. Hough (Eds.), Handbook of industrialand organizational psychology, vol. 3: 271-313.Palo Alto, GA: Consulting Psychologists Press.

Hackman, J. R. 1992. Group influences on individuals inorganizations. In M. D. Dunnette & L. M. Hough(Eds.), Handbook of industrial and organizationalpsychology, vol. 3:199-267. Palo AKo, GA: Consult-ing Psychologists Press.

Hannan, M. T., & Freeman, J. 1989. Organizational ecol-ogy. Gambridge, MA: Harvard University Press.

Harman, H. H. 1976. Modern factor analysis (3rd ed.).Ghicago: University of Ghicago Press.

Haveman, H. A. 1992. Between a rock and a hard place:Organizational change and performance under con-ditions of fundamental environmental transforma-tion. Administrative Science Quarterly, 37: 48-75.

Hermalin, B. E., & Weisbach, M. S. 1988. The determi-nants of board composition. Rand Journal of Eco-nomics, 19: 589-606.

Hill, G. W. L., & Hansen, G. 1991. A longitudinal study ofthe causes and consequences of changes in diversi-fication in the U.S. pharmaceutical industry 1977-1986. Strategic Management Journal, 12: 187-199.

Herman, E. S. 1981. Corporate control, corporatepower. New York: Gambridge University Press.

Hirsch, P. M., Michaels, S., & Friedman, R. 1987. "Dirtyhands" versus "clean models": Is sociology in dangerof being seduced by economics? Theory and Soci-ety, 16: 317-336.

Jaccard, J., Turrisi, R., & Wan, G. K. 1990. Interactioneffects in multiple regression. Newbury Park, GA:Sage.

Janis, I. L. 1982. Victims of groupthink. Boston: Hough-ton Mifflin.

Jarrell, G. A. 1993. An overview of the executive com-pensation debate. Journal of Applied CorporateFinance, 5(1): 76-82.

Jehn, K. A., & Shah, P. P. 1996. Interpersonal relation-ships and task performance: An examination ofmediating processes in friendship and acquain-tance groups. Paper presented at the annual meetingof the Academy of Management, Gincinnati.

Jehn, K. A., & Shah, P. P. 1997. Interpersonal relation-ships and task performance: An examination ofmediating processes in friendship and acquain-tance groups. Working paper. University of Penn-sylvania.

Jensen, M. G., & Meckling, W. H. 1976. Theory of thefirm: Managerial behavior, agency costs, and owner-ship structure. Joumal of Financial Economics, 3:305-350.

Jensen, M. G., & Murphy, K. J. 1990. Performance pay andtop-management incentives. Journal of PoliticalEconomy, 98: 225-263.

Johnson, J. L., Daily, G. M., & Ellstrand, A. E. 1996.Boards of directors: A review and research agenda.Joumal of Management, 22: 409-438.

Johnson, R. A., Hoskisson, R. E., & Hitt, M. A. 1993.Board of director involvement in restructuring: Theeffects of board versus managerial controls and char-acteristics. Strategic Management Journal, 14 (spe-cial issue): 33-50.

Johnston, J. 1984. Econometric methods. New York:McGraw Hill.

Judge, W. Q., Jr., & Dobbins, G. H. 1995. Antecedents andeffects of outside director's awareness of GEO deci-sion style. Journal of Management, 21: 43-64.

Judge, W. Q., Jr., & Zeithaml, G. P. 1992. Institutional andstrategic choice perspectives on board involvementin the strategic decision process. Academy of Man-agement Journal, 35: 766-794.

Kimberly, J., & Zajac, E. J. 1988. The dynamics of GEO-board relationships. In D. G. Hambrick (Ed.), Theexecutive effect: Concepts and methods for study-ing top managers: 179-204. Greenwich, GT: JAIPress.

Kochhar, R., & David, P. 1996. Institutional investors andfirm innovation: A test of competing hypotheses.Strategic Management Journal, 17: 73-84.

Kosnik, R. D. 1987. Greenmail: A study of board perfor-

1999 Westphal 23

mance in corporate governance. Administrative Sci-ence Quarterly, 32: 163-185.

Kosnik, R. D. 1990. Effects of board demography anddirectors' incentives on corporate greenmail deci-sions. Academy of Management Journal, 33:129-150.

Krackhardt, D. 1992. The strength of strong ties: Theimportance of philos in organizations. In N. Nohria &R. G. Eccles (Eds.), Networks and organizations:Structure, form, and action: 216-239. Boston: Har-vard Business School Press.

Lambert, R. A., Larcker, D. F., & Weigelt, K. 1993. Thestructure of organizational incentives. Administra-tive Science Quarterly, 38: 438-461.

Larcker, D. F. 1983. The association between perfor-mance plan adoption and corporate capital invest-ment. Journal of Accounting and Economics, 5:3-30.

Locke, E. A., Shaw, K. N., Saari, L. M., & Latham, G. P.1981. Goal setting and task performance: 1969-1980.Psychological Bulletin, 90: 125-152.

Lorsch, J. W. 1989. Pawns or potentates: The reality ofAmerica's corporate hoards. Boston: Harvard Busi-ness School Press.

Mace, M. L. 1986. Directors: Myth and reality (2nd ed.).Boston: Harvard Business School Press.

Main, B. G. M., O'Reilly, C. A. Ill, & Wade, J. 1995. TheGEO, the board of directors and executive compen-sation: Economic and psychological perspectives.Industrial and Corporate Change, 4: 293-332.

Mallette, P., & Fowler, K. L. 1992. Effects of board com-position and stock ownership on the adoption of"poison pills." Academy of Management Journal,35: 1010-1035.

Mayer, R. G., Davis, J. H., & Schoorman, F. D. 1995. Anintegrative model of organizational trust. Academyof Management Review, 20: 709-734.

Meyer, J. W., & Rowan, B. 1977. Institutional organiza-tions: Formal structure as myth and ceremony.American Journal of Sociology, 83: 340-363.

Mizruchi, M. S. 1996. What do interlocks do? An analy-sis, critique, and assessment of research on inter-locking directorates. In Annual review of sociology,vol. 22: 271-298. Palo Aho, GA: Annual Reviews.

Nunnally, J. G. 1978. Psychometric theory. New York:McGraw Hill.

Pearce, J. A., & Zahra, S. A. 1991. The relative power ofGEOs and boards of directors: Associations with cor-porate performance. Strategic Management Jour-nal, 12: 135-153.

Pfeffer, J., & Salancik, G. R. 1978. The external control oforganizations. New York: Harper & Row.

Rechner, P. L., & Dahon, D. R. 1991. GEO duality andorganizational performance: A longitudinal analysis.Strategic Management Journal, 12: 155-160.

Rediker, K., & Seth, A. 1995. Boards of directors andsubstitution effects of alternative governance mech-anisms. Strategic Management Journal, 16:85-100.

Roberts, K. H., & O'Reilly, G. A. 1974. Failures inupward communication in organizations: Threepossible culprits. Academy of Management Jour-nal, 17: 205-215.

Rosen, S. 1983. Perceived inadequacy and help-seeking.In B. DePaulo, A. Nadler, & J. Fisher (Eds.), Newdirections in helping, vol. 2: 3-12. New York: Aca-demic Press.

Schachter, S., Ellerston, N., McBride, D., & Gregory, D.1951. An experimental study of cohesiveness andproductivity. Human Relations, 4: 229-238.

Seashore, S. E. 1954. Group cohesiveness in the indus-trial work group. Ann Arbor: University of Michi-gan Press.

Segal, M. W. 1979. Varieties of interpersonal attractionand their interrelationships in natural groups. SocialPsychology Quarterly, 42: 253-261.

Shah, P. P., & Jehn, K. A. 1993. Do friends perform betterthan acquaintances? The interaction of friendship,conflict, and task. Group Decision and Negotiation,2: 149-165.

Siegel, S., & Gastellan, N. J., Jr. 1988. Nonparametricstatistics for the hehavioral sciences (2nd ed.). NewYork: McGraw-Hill.

Silver, A. 1990. Friendship in commercial society: Eigh-teenth-century social theory and modern sociology.American Journal of Sociology, 95: 1474-1504.

Spencer, A. 1983. On the edge of the organization. NewYork: Wiley.

Sundaramurthy, G. 1996. Gorporate governance withinthe context of antitakeover provisions. StrategicManagement Journal, 17: 377-394.

Tetlock, P. E., Peterson, R. S., McGuire, G., Ghang, S., &Feld, P. 1992. Assessing political group dynamics: Atest of the groupthink model. Journal of Personalityand Social Psychology, 63: 403-425.

Tosi, H. L., & Gomez-Mejia, L. R. 1989. The decoupling ofGEO pay and performance: An agency theory per-spective. Administrative Science Quarterly, 34:169-189.

Wade, J. B., O'Reilly, G. A., & Ghandratat, I. 1990.Golden parachutes: GEOs and the exercise of so-cial influence. Administrative Science Quarterly,35: 587-603.

Wahal, S. 1996. Pension fund activism and firm perfor-mance. Journal of Financial and QuantitativeAnalysis, 31: 1-23.

Wall Street Journal. 1993. Less-than-watchful eyesdidn't oversee expenses of a utility's chairman. June6: Bl.

24 Academy of Management fournal February

Wall Street Journal. 1995. How do you fire a GEO? Very,very slowly. January 20: Bl.

Wall Street Journal. 1996. Beware of the Trojan horse inthe boardroom. March 7: Bl.

Walsh, J. P., & Seward, J. K. 1990. On the efficiency ofinternal and external corporate control mechanisms.Academy of Management Review, 15: 421-458.

Westphal, J. D., & Zajac, E. J. 1994. Substance and sym-bolism in GEOs' long-term incentive plans. Admin-istrative Science Quarterly, 39: 367-390.

Westphal, J. D., & Zajac, E. J. 1995. Who shall govern?GEO/board power, demographic similarity, and newdirector selection. Administrative Science Quar-terly, 40: 60-83.

Westphal, J. D., & Zajac, E. J. 1998. The symbolic man-agement of stockholders: Gorporate governance re-forms and shareholder reactions. AdministrativeScience Quarterly, 43: 127-153.

Whisler, T. L. 1984. The rules of the game: Inside thecorporate board room. Homewood, IL: Dow Jones-Irwin.

Wiersema, M. F., & Bantel, K. A. 1992. Top managementteam demography and corporate strategic change.Academy of Management Joumai, 35: 91—121.

Zahra, S. A., & Pearce, J. A. 1989. Boards of directorsand corporate financial performance: A reviewand integrative model. Journal of Management,15: 291-334.

Zajac, E. J., & Shortell, S. M. 1989. Ghanging genericstrategies: Likelihood, direction, and peri^ormanceimplications. Strategic Management Journal, 10:413-430.

Zajac, E. J., & Westphal, J. D. 1994. The costs and benefitsof incentives and monitoring in the largest U.S. cor-porations: When is more not better? Strategic Man-agement Journal, 15 (special issue): 121-142.

Zajac, E. J., & Westphal, J. D. 1995. Accounting for theexplanations of GEO compensation: Substance andsymbolism. Administrative Science Quarterly, 40:283-308.

Zand, D. E. 1972. Trust and managerial problem solving.Administrative Science Quarteriy, 17: 229-240.

APPENDIX A

Survey Scales

Possible responses for the board monitoring and ad-vice and counsel items were 1, "minimally"; 2; 3, "mod-erately"; 4; and 5, "very much so."

Board Monitoring

1. To what extent does the board monitor top manage-ment strategic decision making?

2. To what extent does the board formally evaluateyour performance?

3. To what extent does the board defer to your judg-ment on final strategic decisions?

Advice and Gounsel Interactions

1. To what extent do you solicit board assistance in theformulation of corporate strategy?

2. To what extent are outside directors a "soundingboard" on strategic issues?

3. How often have directors provided advice andcounsel in discussions outside of board/committeemeetings (by telephone or in person)? Times

Friendship Ties

How many of the outside directors would you considerto be acquaintances or friends?Acquaintances but not friends: (number of di-rectors).Friends: (number of directors).

APPENDIX B

Results of Supplementary OLS Regression Model"

Independent Variable

1. CEO-board friendship ties2. Portion of board appointed after CEO3. Prior return on equity4. Prior market-to-book vaiue of equity5. Sales'"6. CEO ownership7. Long-term incentive plan compensation8. Director ownership9. Blockhoider ownership

10. Institutional ownership11. CEO tenure12. Affiliated directors: Famiiy ties13. Affiliated directors: Business ties14. Director expertise and experience15. Director expertise and experience

X friendship ties16. Director expertise and experience

X appointments after CEO17. Constant

FR^

Advice andCounsel

Interactions

0.36(0.15)**0.42 (0.20)*

-0.83 (0.43)-0.03 (0.05)-0.05 (0.03)

0.55 (0.57)0.24 (0.19)1.20 (0.88)0.08 (0.22)0.16 (0.28)

-0.01 (0.01)0.58 (0.50)0.51 (0.24)*0.10 (0.05)0.44 (0.15)**

0.51 (0.21)*

-0.06 (0.30)

7.65***.27

" N = 243. Standard errors are in parentheses.*" Logarithm.

* p s .05** p s .01

*** p S .001

James D. Westphal is an assistant professor of manage-ment in the Graduate School of Business at the Univer-sity of Texas at Austin. He received his Ph.D. from North-western University. His current research interestsinclude boards of directors and executive compensation,with a focus on social processes in corporate governance.