The Econometrics Of Corporate Governance Studies Sanjai .

The Econometrics of Corporate Governance StudiesSanjai BhagatandRichard H. Jefferis, Jr.

iiTable of ContentsPrefaceiiiChapter 1:Introduction1Chapter 2:Econometrics of corporategovernance studies17Sample construction and data87Chapter 3:Chapter 4:Joint distribution of takeovers,managerial turnover and takeover defense104Chapter 5:Bootstrap regression results116Chapter 6:Probit models126Chapter 7:Summary and ity AnalysisEndnotesNon-Published AppendicesAvailable from MIT Press WebsiteAppendix A: Takeover Defenses at sample firmsAppendix B: Change in Corporate Control insample firmsAppendix C: Management Turnover in sample firmsAppendix D: Construction of the PerformanceVariables and performance in the sample firmsAppendix E: Longitudinal Analysis of PerformanceVariables and Takeover/Turnover

iiiPrefaceA vast theoretical and empirical literature incorporate finance considers the inter-relationshipsbetween corporate governance, takeovers, tof the extant literature considers the weenarguecorporatethattakeoverdefenses, takeovers, management turnover, wnership structure are interrelated. Hence, from aneconometric viewpoint, the proper way to study ousequations is non-trivial. To illustrate this problemin a meaningful manner we consider the following twoquestions that have received considerable attention

ons:Do antitakeover measures prevent takeovers?Do antitakeover measures help managers enhancetheir ,athatperhaps,sometimes, impossible). Also, more often than not,subsequenttoamanagement-opposed takeover, thesemanagers of the target company usually “leave” ment-opposedmanagersoftheparticular company in increasing (the security orporate

vtakeoverdefensesontakeoveractivityandmanagerial turnover. Our focus is the efficacy cipline poor performers in themanagerial ranks, and also suggests that corporatetakeover defenses are designed to shield incumbentmanagers from these forces. If this is in fact thecase, and the belief that motivates the adoption oftakeover defenses is rational, the presence of g management’s job-tenure. We do observe over defense that is statistically rformance of the company, we do not observe vermodelthatactivityandallowsthe

y to vary with takeover defense, we find thatdefensive activity is ineffective.In the case of management turnover, our resultsare even stronger.The frequency of CEO departuresis uncorrelated with the status of takeover nt with both simple correlations, and withthe estimates from probit models, where we find thatturnover is related to performance.poisonpilldefenses,thereisAt firms withastatisticallysignificant relationship between management turnoverand performance.We stress that these results do not imply thatdefensive activity is costless to shareholders.Itmay well be the case that managers who are keoverlesswelldefensesthannottheybeeninplace.This hypothesis is consistent with both eevidencethatdoes,takeoverdefenses are not completely effective in te

viifinancial performance.AcknowledgmentsWe thank workshop participants at the heir helpful comments on earlier drafts of chapters5 and 6 of this monograph. We also thank ElizabethMurry andVictoria Richardson Warneck of MIT Pressfor their constant encouragement in getting us tocomplete this book.

Chapter 1Introduction1.1. Historical perspectiveCorporate managers are the dominant power-brokers inlarge, U.S. corporations. Roe (1991) notes that ourparticular political and economic history might beresponsible for the dominance of corporate managers.A substantialMeans(1932)literature going back to Berle uedwouldbeimproved if corporations had monitors to oversee themanagers;seeJensenandMeckling(1976).AfterWorld War II through the early 1970s, U.S. was hisperiodiscorporategovernance and power structure that had evolved herewas appropriate for U.S., that is, corporate Americawas delivering thegoods. Hence, there was no needto reconsider the corporate power structure. Othersmight argue that our global economic dominance inthis period was a direct result of the War, which

2haddestroyedthephysicalandeconomicinfrastructure of most other major economic playersin the levenU.S.competitiveedge. Observers in the popular media argued that eThe argument went that corporate managersmoreinterestedinincreasingandmanagingtheir empires; serving the shareholder interest wasof secondary importance.These observers noted thatthe reason managers were successful in engaging insuch behavior was lack of meaningful oversight oftheir decisions, and lack of an alternate power withdisciplining cern about the role of such raiders onthe long-term impact on corporations, and the nearterm impact on other stakeholders was raised; seeBhagat, Shleifer, and Vishny (1990). Sometime in thelate1980s,hostiletakeoversbecamemuchrarer;

3Comment and Schwert (1995) provide a discussion tartingpopularinandtheacademiccommentators started emphasizing the monitoring Blair (2001).1.2. Corporate antitakeover devicesSome have suggested that corporate antitakeoverdevices (such as antitakeover amendments, and poisonpills) played a role in diminishing the e1980s.Antitakeoverbycorporateboardsandapproved by shareholders; these amendments amend oardoftheexistingamendmentprovides for the election of typically a third ofthe board in any annual election; this extends thetime required to elect a majority in the board. ysharesduringacertain period. Some corporations have amended their

4charter to reincorporate in to Delaware – a statethat is generally considered to be yapproval.Whileadoptedpoisonwithoutpillscomeinmany flavors, they typically impose a very high coston a potential acquirer that the board disapprovesof; for example, the pill may require the eacquirer’s equity, or/and lessen the voting power 8), and Bruner (1991) contain a description ofthese antitakeover provisions.1.3. The econometric problem of measuring the impactof antitakeover provisionsA vast theoreticaland empirical literature incorporate finance considers the inter-relationshipsbetween corporate governance, takeovers, tof the extant literature considers the time–forownershipandbetweencorporate

5governance and takeovers.The following is just a sampling from theabovementioned literature: Pound (1987) and Commentand Schwert (1995) consider the effect of takeoverdefenses on takeover activity; Morck, Shleifer, andVishny (1989) examine the effect of corporateownership and firm performance on takeover activityand management turnover; DeAngelo and DeAngelo(1989), Martin and McConnell (1991), Denis andSerrano (1996), and Mikkelson and Partch (1997)consider the effect of firm performance onmanagement turnover; Denis, Denis and Sarin (1997)consider the effect of ownership structure onmanagement turnover; Bhagat and Jefferis (1991)consider the impact of corporate ownership structureon takeover defenses; Ikenberry and Lakonishok(1993) investigate the effect of firm performance ontakeover activity; Berkovitch, Israel and Spiegel(1998) examine the impact of capital structure onmanagement compensation; Mahrt-Smith (2001) studiesthe relationship between ownership and capitalstructure; Garvey and Hanka (1999) investigate theimpact of corporate governance on capital structure;McConnell and Servaes (1990), Hermalin and Weisbach

6(1991), Loderer and Martin (1997), Cho (1998),Himmelberg, Hubbard and Palia (1999), and Demsetzand Villalonga (2001); and DeAngelo and DeAngelo(2000) and Fenn and Liang (2001) focus on ownershipstructure and the corporate payout ement turnover, corporate performance, nterrelated. Hence, from an econometric viewpoint,the proper way to study the relationship between anytwo of these variables would be to set up a ionships between these six variables. ultaneous equations is of

cequationindex models, where the functional form is unknownand the explanatory variables in that equation arecontinuous, known functions of a basic parametervectorisdiscussedbyIchimura and Lee (1991).Estimation of a system of equations in the absenceof strong restrictions on both the functional formoftheequationsandthejointdistributionoferror terms is, to the best of our knowledge, anunsolved problem.We are unaware ofa model of takeover defensethat implies specific functional forms.If thesefunctions are linear, identification may be attainedthrough either strong distributional assumptions orexclusionrestrictions.Amemiya xclusion theanderrorabsenceofBut these restrictions areincentive-basedexplanationsoftakeover defense, since unobservable characteristics

8of managerial behavior or type will be reflected inall of the error terms.Exclusion restrictions aretherefore the most likely path to akeover defense affects the likelihood of lusionrestrictionsjustify.Intuitively,variables that affect the likelihood of a ingfulmannerweeconometricconsiderthefollowing two questions:Do antitakeover measures prevent takeovers?Do antitakeover measures help managers enhancetheir ership structure and corporate governance (whichincludes corporate antitakeover devices) on takeoveractivity and managerial turnover.Our focus is onthetakeovertakeover

9defense on the relationship between performance andmanagerial turnover.A vast literature suggests that takeovers andthe managerial labor market serve to discipline anks,defensesareandalsoproposedbyincumbent managers to shield themselves from ze this literature.1 DeAngelo and Rice agerial entrenchment tionisthatoftheycorporaterepresentanagreement that alters the distribution of ofdirectors and outsiders, but not necessarily in sinsuchwithfirm-specifichuman capital, and/or negotiate a higher bid premiumin a takeover; DeAngelo and Rice (1983) ypotheses:Heparadox"notesthat

10proponents of the managerial entrenchment keovers which are a voluntary transaction betweentarget and bidder shareholders.Knoeber argues thattakeover defenses are also a voluntary transactionamong target shareholders, board of directors, s must still answer to a board of directors;both management and the board may be vulnerable topressure from quarters other than the direct ican Express, IBM and General Motors illustratesthis eovers(and/orbetweenmanagerialturnover) at firms that have takeover defenses and/or managerial turnover) at firms that do encewhethermanagersfromthisinvestigation complements the indirect evidence fromannouncement returns.Our effort builds on the work of Palepu (1986),

keovers.Weincorporate their insights into a model that influencecontroloftakeoveractivity.Wecontribute to the growing literature on the effectof corporate governance on firm performance: Bhagat,Carey and Elson (1999), Bhagat and Black (2001), zes the endogeneity in the relationship ion. We also contribute to the literatureon the effect of corporate performance on managementturnover: Warner, Watts and Wruck (1988), Weisbach(1988), and Denis and Denis (1995).2We control forthe influence of ownership and takeover defense inevaluating the effect of performance on yareof

12takeover activity.Thedistinctionbetweenourworkearlier authors is significant.inferencethattakeoverandthatofWe show that thedefensesdecreasethefrequency of takeover activity, which is consistentwith the correlations reported by Pound, is spuriousand attributable to the omission of performance fromthe econometric model.We also demonstrate that theomission of takeover defenses from a model of rnover) and performance results in a specificationerror that biases inference about the influence ofperformance on takeover activity (and/or managementturnover).selectionFinally, our results suggest that selfplaysanimportantroleinmodelsthatrelate takeover defenses to ice-based sample of firms during the years , the array of takeover defenses in place atsample firms during this time period varies ied board provisions, poison pills, and fairpriceamendments.Thisvariation,whichenhances