Diffusion and Variation of a Controversial Practice

23 Pages·2012·0.42 MB·English

OrganizationScience
Vol.23,No.4,July–August2012,pp.1077–1099
ISSN1047-7039(print)(cid:151)ISSN1526-5455(online) http://dx.doi.org/10.1287/orsc.1110.0685
©2012INFORMS
How Golden Parachutes Unfolded:
Diffusion and Variation of a Controversial Practice
(s).rg/. Peer C. Fiss, Mark T. Kennedy
ro
hos. MarshallSchoolofBusiness,UniversityofSouthernCalifornia,LosAngeles,California90089
utm {fi[email protected],[email protected]}
aor
enf Gerald F. Davis
ths.i StephenM.RossSchoolofBusiness,UniversityofMichigan,AnnArbor,Michigan48109,[email protected]
oal
tn
r
syou We contribute to a growing focus on variation in diffusion processes by examining the ways in which contested
courtehttp://j eexxtaemnsinipvereatncheteiscsseesainsasdrueseismm,wioldaerifiistteyuddoyafsvaatphrireaaytcitosincperienwad“ilglaovmldaoreynngipnaarrdaeocsphptuoetnress”.ecEtooxnpbtaroantchdtisnp,goapocuonlantptiroroinov-reradsniidaffluocsrogiorapnnoirazacatcteioogunon-vltesev,renwlaenmcaeercgphuraaencitstihmcaest.tthThaoet
aat emerged and spread widely during the hostile takeover wave of the 1980s. Using a concept network approach to analyze
asble the composition of parachute plans, we find evidence of mechanisms that both increase and decrease extensiveness and
a variation of golden parachutes. Our findings hold implications for accounts of practice diffusion over contested terrain by
pyail revealingsubstantialvariationinthecourseofdiffusion.
ov
ca
s Keywords: diffusion;practices;variation;institutionaltheory;corporategovernance;goldenparachute;conceptnetwork
thises,i History: PublishedonlineinArticlesinAdvanceAugust24,2011.
dci
butepoli Introduction (e.g., Fiss and Zajac 2004, Schneiberg and Soule 2005,
rin
distsio Recent work on the diffusion of organizational practices Sanders and Tuschke 2007). For instance, Ingram and
dmis hasrevealedagrowinginterestinthewaysinwhichcor- Rao (2004, p. 448) have argued that extant research has
aner porate practices are modified during the diffusion pro- either emphasized political contestation or the diffusion
p cess(e.g.,Ansarietal.2010,Djelic1998,Westphaletal. ofpractices“withoutjointlyconsideringtheirinterdepen-
e
clnd 1997). Such a focus on practice variation carries consid- dencies.”Furthermore,priorworkhastendedtooverlook
tia
ars erablepromiseforadvancingourunderstandingofdiffu- themultilevelcharacterofsuchcontests(IngramandRao
sht sion processes because it redirects the study of diffusion 2004,SchneibergandSoule2005).Yetpracticevariation
thirig towardfiner-grainedmechanismsthatcreateheterogene- islikelytobeaffectedbyprocessesoccurringatboththe
tong ity among organizational populations (Lounsbury 2007). population and organization levels (Ansari et al. 2010).
htudi Understanding how practices vary allows insights into As such, the diffusion of contested practices presents an
rigncl both the relations between actors and the nature of the appealingyetlargelyuncharteredcontextforunderstand-
pyn,i practiceitself(Ansarietal.2010).Asanessentialaspect ing practice variation.
oo of the implementation process, the emergence of prac- Toward an understanding of the role of practice vari-
scmati tice variation is likely to be the rule, not the excep- ation in contested diffusion, we study patterns of prac-
d
olor tion (Rice and Rogers 1980, Johnson and Rice 1987). tice variation in the spread of the “golden parachute,”
hnf Yet as Campbell (2005) points out, most diffusion sto- a controversial contract that provides compensation to
i
MSal ries assume that practices are adopted uncritically and top executives in the event that the firm they run is
n
Ro in toto, and thus they pay little or no attention to how taken over. Golden parachutes generated intense debate
FOditi practices are modified as they diffuse. If diffusion is to among shareholders, policy makers, and scholars. More
Nd
A be more than a “mindless mechanical transfer of infor- recently, they have become controversial again as the
I
mation from one place to another” (Campbell 2005, U.S. government has assumed the role of executive pay-
p. 55), then we need to explore how diffusion affects master for companies it rescued in 2008. This is par-
practices. ticularly interesting because golden parachutes had dif-
To accomplish this goal, it is helpful to consider the fused to a substantial majority of the U.S. Fortune 500
nature of the practice in relation to its adoption en- companies by the late 1980s. As golden parachutes dif-
vironment. Although many prior diffusion studies have fused, contract features varied widely both among firms
examined situations in which practices are adopted andovertime,indicatingthatadoptionwasaccompanied
uncritically, several recent studies have pointed to the byconsiderablevariationinimplementation.Thegolden
importanceofexaminingdiffusionasacontestedprocess parachute is therefore a particularly attractive site for
1077
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
1078 OrganizationScience23(4),pp.1077–1099,©2012INFORMS
research because it provides an opportunity to examine in even greater numbers. By the late 1980s, golden
howpracticesvaryacrossthecourseofthediffusionpro- parachutes had become part of the compensation pro-
cess in response to both population- and organization- gram of the majority of large American corporations.
level mechanisms. The context of this successful diffusion of parachutes
We take a twofold approach to examining patterns of is the acquisitions wave of the 1980s, which was pre-
practice variation. First, following a recent framework cipitated by a confluence of several factors, including
for analyzing practice variation (Ansari et al. 2010), high inflation and depressed stock prices, thus making
we theorize the causes and consequences of practice it attractive for firms to buy assets instead of build-
s).g/. extensiveness and similarity during the course of diffu- ing them by means of organic growth. This wave of
(r
hors.o sion. Second, we develop a concept network approach takeovers and mergers, particularly in the period from
utm to more clearly understand which elements of golden 1984 to 1989, introduced hitherto unprecedented lev-
aor parachutesarecentralversusmoreperipheraltowhatthe els of uncertainty into the life of America’s top man-
enf
ths.i practice came to mean. By constructing co-occurrence agers. Whereas smaller corporations had always been at
oal matrices of practice elements, we are able to employ a higher risk of acquisition, the largest U.S. firms had
yturn network visualization techniques to explore the rela- by and large felt protected from being taken over. How-
urtesp://jo tiinoenhaolwstrtuhcistusrteruocftupraeracchhauntgeeedleomveernttismaen.dTofugretthheerr,exthaemse- eavwera,ywittohfithneanecmeertgheenacceqoufishiitgiohn-yoiefldFo(“rtjuunnek”5)0b0onfidrsmass,
cohtt approaches allow us to conduct a fine-grained analysis topmanagersofeventheseblue-chipfirmshadgoodrea-
aat of how golden parachute plans vary across the diffusion son to be concerned about their positions. This concern
asble process as well as to develop hypotheses related to the was enhanced by a series of high-profile cases in which
a population- and organization-level influences that affect
pyail top executives—and even those who had previously per-
coav both heterogeneity and homogeneity in the implementa- formed well—were fired immediately following their
sis tion of this practice. firm’s acquisition. The result was “a fair amount of ter-
thies, rorinexecutivesuites”(Nossiter1982,p.15).Thisterror
butedpolici TAmheeriCcaanseexoecfuGtivoeldpeaynpPraacrtaiccehsuwteerse the subject of wHaoswneollttwhhatolcliyrcuunlfaoteudndaemd.oAngstbuodayrdbymceomnsbuelrtsanincdyiWcaaterdd
rin that most CEOs of acquired companies occupied posi-
sto increasing public scrutiny during the 1980s. Until that
dissi time, large pay packages—and particularly those that tions of significantly reduced responsibility within a few
dmi yearsofbeingacquired,withmostformerCEOsplacing
aner paidmorethan$10millioninasingleyear—werearare notevenamongthe20highest-paidofficersoftheacquir-
p exceptionforchiefexecutivesofnonfinancialinstitutions
clend (Blair 1994). However, over the next two decades, the ingfirm(Dwyer1983).BecauseCEOstookanaverageof
rtia average compensation of U.S. top executives increased 23 years to ascend to the top of the corporate ladder, the
sahts about sixfold, making American CEOs the most highly notionoflosingcontrolandstatusafteryearsoftoilpre-
thirig paid managers in the world. sented a considerable threat to earning prospects, career
og Among the most contested aspects of CEO compen- options, and even self-esteem. As a response to these
tn
tdi sation was the diffusion of the “golden parachute,” the new uncertainties, top executives and corporate boards
hu
gcl high-profile severance packages that emerged in the late quickly embraced golden parachutes, spurring the dif-
yriin 1970s among a handful of firms. With the 1980s hostile fusion of this practice. Consequently, golden parachutes
opon, takeover wave, this practice spread rapidly. By the late becameanexpectedelementofseniorexecutivecompen-
cati 1980s, the majority of large public corporations in the sation.Asamergerbrokerstated,“Executivesaresaying:
sm
oldor United States had golden parachute contracts for their If you can’t protect my power, then at least protect my
hnf most senior executives. In the event that their compa- wallet” (Kleinfield 1982, p. D1).
i
MSal nies were acquired, these contracts contained clauses Agreements could vary tremendously. Although the
Ron thatprovidedexecutiveswithcashpayments,accelerated term“goldenparachute”impliesasinglecoherententity,
FOditi stock option vesting, continued health insurance, retire- 1980schange-in-controlcontractsvariedgreatlyinprac-
NAd mentplaneligibility,andthelike.Themosttypicalform ticealonganumberofdimensions.Forinstance,whereas
I
provided a lump-sum payment of three years’ salary in some parachutes covered only the CEO and perhaps
the event that the executive’s employment ended follow- a handful of the most senior executives, others came
ing a “change in control” (usually a takeover). to include dozens of managers, such as the parachute
The spread of golden parachutes among the top strata ofAMF,whichprotected“28officersholdingtherankof
of American firms is quite remarkable. By 1981, some corporate vice president or higher” (AMF Inc. 1981), up
15% of the 250 largest U.S. corporations had such plans to the example of Beneficial Corporation, which imple-
in place, covering about one-third of the management mentedaplancoveringsome500employeesandprovid-
contractsofthosefirms.By1986,thisnumberhadgrown ingthemwiththreeyears’salaryintheeventofahostile
toabout33%,andsmallercompaniesadoptedparachutes takeover. Such variation is even more common today.
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
OrganizationScience23(4),pp.1077–1099,©2012INFORMS 1079
Likewise, parachutes differed in terms of what ben- (cid:48)(cid:48)(cid:48)threetimestheCorporation’sannualmatchingcontri-
efits they provided. Some included only a lump-sum butions on behalf of the officer to the Salary Employee
cash payment, whereas others extended to stock grants, SavingsPlan;thecashvalueoftheofficers’targetestab-
lished under the Long Term Performance Plan perfor-
options, health insurance, pension plans, consultancy
mance cycles as in effect on the date of termination;
arrangements, the payment of legal fees, and use of the
andtheequivalentcashvalueofprovidingcertainhealth
corporate jet. There was also considerable variation in
and dental insurance plans and other fringe benefits as
what events could trigger a parachute. Some parachutes
in effect prior to the change in control for a three-year
protected not only against being laid off by the new periodfollowingtermination(cid:48)(cid:48)(cid:48)(cid:48)Additionalbenefitspro-
s).g/. management but also against substantive changes of videdbythe agreementsincludethevestingof allretire-
(r
uthorms.o tohfetheexiercauutitvheosr’ityd,uftioersc,edsutcrhanasfseras,soigrnoifipcparenstsirveedutcrativoenl mseernvticbeeunnedfietsrathnedstahlearaydrdeittiiroenmoefntthprleaensyeaanrds tohfecvreedstiitnegd
aor schedules. Going even further, Superior Oil’s parachute of all benefits under the Salary Employee Savings Plan.
enf (Lockheed Corporation 1987; pp. 9–10; italics added)
ths.i allowed covered executives to quit for any reason except
oal death, disability, willful misconduct, or normal retire- The Lockheed agreement offers an example of a “gold-
trn ment. Finally, parachutes also differed in terms of the plated” parachute. It provides an extensive package of
yu
so reasons given for their adoption—whether as a device benefitsthatexpandsuponmeresalarybyaddingunpaid
urtep://j to retain key employees, a means to align the incentives incentives, matching contributions, and health and den-
cohtt of executives with those of shareholders in a takeover tal insurance. In addition to the provisions shown here,
aat situation, or a strategy for preventing takeovers entirely the plan furthermore included the immediate vesting
asble by making them particularly costly to the acquirer. of stock option plans for management, thus making
a The following examples illustrate the variation in them immediately exercisable as part of the payout to
pyail management.
ov golden parachute provisions. Hammermill Paper, one of
ca In short order, outside audiences and certain share-
s the earliest adopters in 1976, covered the CEO and five
si holders came to regard more extensive parachutes such
thies, other members of the top executive team in its early asLockheed’sasexcessivepayandbenefits.Oneofear-
butedpolici parEaacchhuotef:Messrs. Duval, Herbutzheimer, Kilgore, Leslie, loifesBtecnodnitxro,vwehrsoiarleccaesiveesdwmasortehatthoanf W$4illmiaimllioAnguepe,oCnEhOis
strion Stolley, and Volanakis has an agreement with the Com- departure in 1983, with the total parachute payments
disi pany providing that, if a change of control of the com- to Bendix executives amounting to some $15.7 million.
s
dmi panyoccurswhileheisanemployeeoftheCompany,his The details of these compensation arrangements caused
aner employment by the Company shall continue for at least widespread outrage when they became public knowl-
clendp tthortaelecyoemarpseantsaantioannnfuoarl1r2atemoofncthosmppreencseadtiionngetqhueaclhtoanhgies edge in the wake of Bendix’s high-profile takeover bat-
tia tle with Martin Marietta. Many shareholders saw this
ars of control. as payment for failure: well-managed firms do not get
sht (Hammermill Paper 1980, pp. 7–8; italics added)
othigrig This early parachute is an example of the simple “bare- tsaukcehnaovreicr,hsorewwahryd?shWouiltdhintheshCoErtOorodferB,esnudcixh rgeocledievne
ttdin bones” type, providing payment of the salary only parachute agreements were called a “web of enrichment
ghclu with no further benefits or elaborating provisions. It is devices” that were “grossly excessive and wholly unjus-
rin also fairly limited in covering only six top executives, tified” (Berry 1988, p. 28), a “waste of corporate assets”
yi
pn, although some other parachutes were even more restric- and “breach of fiduciary duty” (Nossiter 1982, p. 15),
oo
cati tive and covered only the CEO. and “a fraud upon the corporation” (Kleinfield 1982,
dsm In contrast to this fairly simple parachute, consider a p. D1).
holnfor parachute adopted by the Lockheed Corporation: In response to this critique, proponents of golden
i parachutes asserted that they were needed to protect
MSal Inordertoencouragecertainexecutiveofficerstoremain the interests of shareholders because CEOs would be
n
FORditio wtioitnhttohethCeoCrpoorrpaotiroantioann’dstboucsoinnetsinsuientothedeevvoetnetfualnlaetftfeonr-t mifothreeyprdoidnenotot haallvoewtoecwoonrormyiacbaolluyt btheenierfiocwianl etackoenoovmerics
INAd is made to obtain control of the Corporation through a well-being afterward; indeed, a company’s share price
tender offer or otherwise, the Corporation has entered
tended to rise in the wake of the announcement of a
into severance agreements with certain executive offi-
parachute, suggesting that they benefited shareholders
cers,includingMessrs.Kitchen,Fuhrman,Marafino,and
on balance (Lambert and Larcker 1985). Nonetheless,
Tellep. These severance agreements provide for certain
despite the vigorous defense of golden parachutes given
payments and benefits in the event of the termination of
by many managers and compensation consultants, some
the officer’s employment within three years of a change
in control (as defined in the severance agreements)(cid:48)(cid:48)(cid:48)(cid:48) directors were not convinced of their value or legiti-
The payments and benefits include cash payments of macy. For instance, Felix Rohatyn, a seasoned corpo-
threetimestheofficer’sbaseannualsalaryatthetimeof rate director serving on six boards, stated, “If an execu-
thechangeincontrolortermination,whicheverishigher; tive needs a multimillion-dollar contract to get his mind
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
1080 OrganizationScience23(4),pp.1077–1099,©2012INFORMS
clear in a takeover situation, then maybe he should see otherwise be thwarted by executive resistance. In effect,
a psychiatrist” (Davis and Greve 1997, p. 31). Simi- golden parachutes fit the logic of de-diversification that
larly, the chairman of another company stated, “I and became institutionalized over this period (Davis et al.
my board hold the opinion that golden parachutes are an 1994).Inessence,apracticethathademergedtoaddress
unconscionable rape of a shareholder’s assets” (Morri- the employment uncertainty of top executives had come
son 1982, pp. 83, 87). to be frequently justified as safeguarding shareholder
Eventually, the golden parachute debate became so interests, thus shifting the logic of adoption from bene-
controversial that the U.S. Congress intervened. As part fitting management to benefitting shareholders.
s).g/. of the Deficit Reduction Act of 1984, Congress revised Inthebroadercourtofpublicopinion,however,golden
(r
hors.o the tax code to deny tax benefits to “excess golden parachutes continued to be widely viewed as inappro-
utm parachute payments.” The law defined these as pay- priate payoffs for an abdication of stewardship. Rather
aor mentswhosepresentvalue“equalsorexceedsanamount than viewing golden parachutes as incentives that lubri-
thes.inf equal to three times the base amount” of executives’ catevalue-maximizingchangesinownership,analterna-
oal prior total compensation (see §280G of the law). This tivesphereandsegmentofsocietyconceptualizesgolden
trn threshold of three times the base amount had two unin- parachutes as seven- and eight-figure payoffs given to
yu
so tended effects. First, it legitimated payments below and executives for standing aside to let their companies be
urtep://j up to the threshold (e.g., Berry 1988), with amounts of acquired and, in some cases, broken up in ways that
cohtt 2.99 times the base salary becoming widely prevalent— meant thousands of workers lost their jobs. It is clear
aat “the government maximum quickly became an industry from both the debates of the 1980s and the most recent
asble norm” (Reed-Lajoux 2005, p. 269). Second, it shifted furor over executive pay that golden parachutes remain
a innovation in parachutes to additional provisions that controversial in the political arena to this day.
opyvail werenotsalaryrelated,suchasstockoptions,retirement Overall, therefore, golden parachute agreements offer
ca
s plans, or health insurance. However, shifting to other a particularly attractive context for examining how prac-
si
dthicies, bbeenreofiftscworapsornaottioansunsiivmerpslayl dtreecnidd;eda ctoon“sgidroersasbluep”nutmhe- tdiecsecsrivbaerdy aobvoevret,hethceoyurpsreesoefntthae pdriaffcutiscioenwpirthocecsosn.siAds-
butepoli pimarpaocsheudtebyplCanosn—grtehsastoins,ptaoymcoevnetrsaenxyceaedddinitgiotnhaelthtarxeees- eorvaebr,ledveatariilastioonnagloonldgena pnaurmacbheurteosf adriemerneqsiuoinresd. Mtoorbee-
rin
sto base-salaries threshold. disclosed in a firm’s proxy statement each year, thus
disi
s Courtsalsobegantoquestiongoldenparachuteagree- allowing easy access to their content for other poten-
dmi
aner ments. For example, after the Beatrice Companies in tial adopters (Rogers 1995). Thus, parachutes allow for
p 1985 awarded golden parachute contracts totaling $23.5 imitation as well as customization in the diffusion pro-
e
clnd million to six of its top executives, its decision was cess. Finally, golden parachutes were also highly con-
tia
ars challenged by a slew of shareholder suits. Likewise, in troversial from the very beginning, thus allowing us to
sht the wake of AMF’s takeover bid by Minstar in 1985, examine practice variation in the context of active con-
thirig Florida shareholder Edith Citron filed a class action suit testationandconflictingrationales(FissandZajac2004,
tong on behalf of all former AMF shareholders, alleging that Lounsbury 2007).
tdi thegoldenparachuteawardedtoCEOThomasYorkand
hu
gcl other AMF executives presented a breach of fiduciary
yriin duty on the part of the firm’s board of directors. Theory and Hypothesis Development
pn,
oo Lawsuitssuchasthesebecamemorenumerousduring Asourhistoryofadoptionandimplementationindicates,
cati the mid-1980s, including cases against companies such the widespread adoption and use of golden parachutes
sm
oldor asLockheed,RCA,Revlon,Signal,andWesternPacific, among top executives of U.S. corporations did not
hnf indicatingthatanewphaseofcontestationwithregardto coincide with taken-for-grantedness or even acceptance
i
MSal parachutes had arrived. In fact, by the end of the 1980s, beyond the executive suite. Yet widespread adoption is
n
Ro the adoption of new parachutes had dwindled, and by frequently taken as evidence of institutionalization (e.g.,
FOditi 1994,newadoptionshaddecreasedtoahandfulperyear Tolbert and Zucker 1983, Westphal et al. 1997). This
NAd among the largest U.S. corporations, with the overall puzzle reveals a unique and theoretically important fea-
I
adoption rate at about 60%. ture of our study site: the wide use of a practice in one
Thus, the overall pattern of diffusion is accelerating sphere of social life need not coincide with or lead to
adoption throughout the 1980s but declining adoption acceptance in other spheres. This situation is consistent
after1991becausethemajorityoftheFortune500popu- withthenotionthatdistinctspheresofsocietyhavetheir
lationhadalreadyadoptedbythen.Asgoldenparachutes owninstitutionallogicsforelaboratingwhatisappropri-
diffused and changed, so too did the reasons for adop- ate (Boltanski and Thévenot 1991, Friedland and Alford
tion. In particular, financial economists and investment 1991). Thus, what becomes legitimate among top exec-
bankers succeeded in linking the ability to maximize utives may not meet with approval in the larger spheres
shareholder value with takeover deals—deals that might of politics, family, religion, or education. Indeed, as the
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
OrganizationScience23(4),pp.1077–1099,©2012INFORMS 1081
history of golden parachutes illustrates, the diffusion of Population-Level Mechanisms Affecting
a practice may come with considerable contestation and Practice Variation
countermobilization.
Diffusion Trajectory and Time of Adoption. Although
Intraditionalaccountsofinstitutionalization,diffusing
prior research has mostly focused on the adoption event
practices came to be universally accepted and taken for
itself, several recent studies have begun to consider
granted (e.g., Tolbert and Zucker 1983, Westphal et al.
issues relating to the extensiveness of practice imple-
1997). Here, organizations were expected to adopt or
mentation (e.g., Fiss and Zajac 2006, Lounsbury 2001,
facealossoflegitimacy;forinstance,afirmthatdidnot
s).g/. adopt total quality management could lose contracts or Westphal et al. 1997, Westphal and Zajac 2001). In this
(r regard, factors influencing practice diffusion are likely
hors.o facenegativepublicity.Incontrast,wecomplementthese to reach beyond the mere adoption event to also affect
utm priorstudiesbyfocusinghereonasituationofcontested
aor diffusion where this was not at all the case. Throughout patterns of practice implementation. Developing a cog-
thes.inf the diffusion process, golden parachutes were fiercely nafiftievcetlythegreoxutnednetdofmiomdpellemofenhtoawtionad,oKpetinonnedmyoatinvdatiFoinsss
oal criticized by shareholders and in the media and were
trn also the subject of frequent court challenges. Accord- (2009) argue that motivations to achieve gains will be
syou ingly,ourstudyshedslightnotonasituationofgrowing associated with more extensive practice implementation,
urtep://j institutionalization but rather on the diffusion trajectory whereas motivations to avoid losses will be associated
cohtt ofahighlycontroversialpracticethat—althoughbecom- with less extensive practice implementation. Their argu-
aat ing expected among executives—never came to be fully ments suggest that the extensiveness of practice imple-
asble institutionalized in the public realm. mentationistoaconsiderableextentaffectedbywhether
a What, then, is the effect of such contestation on adopters perceive a diffusing practice as legitimate and
opyvail the diffusion trajectory of a practice? What differences aligned with their interest or as imposed by the institu-
ca should we expect in diffusion patterns? We suggest tional environment.
s
thises,i tihtsaetlfvaarreiatainonismpinorttahnet iwmipnldeomwenintattoiounndoefrstthaendpinrgactdiicfe- tesFtaotriothneacdrioffsussitohnedoifffguoslidoennppraorcaecshsuitnedsi,ccaotenstitnhuaetdadcoopn--
dci
butepoli ftiuosnioanllyprfoocceussseeds.oAnlththoeugdhemdiafnfuds-isoidnestcuhdaireaschtearvisetitcrsadtio- tlieognitimwaatsenboutt rsaothmerucbhy tdhreiviennterbeystsaodfetsoirpeetxoecauptipveeasr.
rin
sto predict the diffusion of a relatively invariant practice, Goldenparachutesevidentlybenefittedtopmanagement,
dissi recent work has suggested shifting the focus to under- and such severance agreements came to be increas-
dmi
aner standing practice variation during the diffusion process ingly expected among executives (Kleinfield 1982). As
p (Ansari et al. 2010, Frenkel 2005, Lounsbury 2001). parachutes became “part of the fabric of compensa-
e
clnd Regarding the current case of patterns of practice vari- tionprogramsatmostlargecompanies”(Proeksch1986,
tia
ars ation, we follow Ansari et al. (2010) in focusing on p. F1), there was a self-serving motivation on the part
sht two fundamental dimensions to explain variation in dif- of executives to engage in elaboration and extension of
thirig fusing practices: extensiveness and similarity.1 Here, the practice to include more elements. To complement
tong extensivenessmeasuresthedegreeofpracticeimplemen- this development, major law firms increasingly estab-
htudi tation, and similarity measures how alike these imple- lished in-house experts on severance packages to guide
gcl mentations are to previous versions of the practice. Both companies through the implementation process (Berry
rin
pyn,i present conceptually different dimensions in that exten- 1988). In addition, the continued contestation of golden
oo siveness assesses change in degree whereas similarity parachutes also led to a shift in which these packages
scmati assesses change in kind: “Practices are high [in simi- focused less on salary and instead on more provisions
d
olor larity] but not extensive when they are truer to the pre- for additional benefits, such as the covering of legal fees
hnf vious version—but not comprehensively implemented. or health insurance. Such benefits still benefited exec-
i
MSal Practices are extensive but low [in similarity] if com- utives but had two advantages: first, they were harder
n
Ro prehensively implemented—but not true to the previous to quantify and thus did not invite contestation to the
FOditi version” (Ansari et al. 2010, p. 73). We believe using sameextent;andsecond,theywereeconomicallyadvan-
Nd
IA these two dimensions of diffusing practices allows for a tageousbecausetheydidnotfacethesametaxpenalties.
more fine-grained analysis of how practices vary as they Given a situation of external contestation but with
spreadbecausetheycaptureboththenatureofapractice
the presence of internal champions, this suggests a pat-
and its actual implementation.
tern in which golden parachute plans become increas-
This argument suggests hypotheses that distinguish
ingly expansive over time, which brings us to our first
population- and organization-level mechanisms of prac-
hypothesis.
tice variation (Ansari et al. 2010, Love and Cebon 2008,
Schneiberg and Soule 2005). We now turn to translat- Hypothesis 1A (H1A). A later time of adoption will
ing the argument into hypotheses regarding variation in have a positive effect on the extensiveness of golden
golden parachutes. parachute plans.
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
1082 OrganizationScience23(4),pp.1077–1099,©2012INFORMS
Regarding the similarity to prior adoptions, extant practice becomes available, potential adopters are bet-
institutional arguments primarily suggest decreasing ter able to understand the practice and which features
variation in response to isomorphic pressures toward are particularly problematic, and this knowledge should
conformity. In these arguments, customization is largely enable them to implement more extensive versions of a
restricted to early adopters, whereas later adopters are practice (Glick and Hays 1991, Mooney and Lee 1999).
more likely to mimic the normative model or defini- Regarding the availability of information, the media
tion established by earlier adopters (e.g., Westphal et al. play an important role in disseminating information by
1997, p. 387). facilitating the sharing of information and knowledge
s).g/. When practices are contested, however, isomorphic (Gamson et al. 1992, Pollock and Rindova 2003, Sine
(r
hors.o pressures are weaker, and there can be nothing taken for et al. 2005). Specifically, media discourse is a central
utm granted. This suggests a somewhat different logic for venue for conversations about new products and prac-
aor practice variation. Specifically, we can no longer expect tices and acts as a key source of sensemaking (Fiss
thes.inf variation to be anchored to a widely embraced nor- and Hirsch 2005, Kennedy 2008); thus the media not
oal mative model. Instead, contested diffusion both allows only cover but also construct social realities and agen-
yturn and requires later adopters to discover which forms of das(Zilber2006).WefollowStrangandMacy(2001)in
urtesp://jo tlheaestpraacccteicpetabarleetmooostthearpsp.eFailnindgingtothsiosmbealaaundceiegnacienssatnhde atossluemarinngfroammolidmeilteodfibnofuonrmdeadtiloynraatbioonuatlthaecteoxrspesreieeknicnegs
cohtt approval of the audiences that are for the practice while of others, particularly by emulating the most successful
aat avoiding protest from the audiences that are against it. peers.Regardinggoldenparachutes,increasingdiffusion
asble In the case of golden parachutes, there is considerable was associated with a considerable growth in the cover-
a evidence that there were two distinct reactions to them.
opyvail On the one hand, managers desired golden parachutes atigael oafdosupctehrssevweirtahncceonasgirdeeermabelnetsi,ntfhoursmpartoiovnidianbgopuottethne-
ca
(e.g.,Kleinfield1982)and,atleasttosomeextent,sodid
s nature of these practices and “success stories” regard-
si
thies, bnoitaierdssthoaft ddierpeecntodresd. Oonnctohmepoatnhieershfaonrdj,oblasb,oarn,dcoemlecmteud- ing the experiences of peer firms. Building on these
dci arguments, we expect growing media discourse to be
butepoli odfimficmiaelrsvwiehwo orefpgroelsdeenntepdatrhaochseutecso.mAmcuonnittiiensueadllintoteorkesat associated with greater knowledge about the diffusing
rin practice,thusreducinguncertaintyandallowingadopters
sto in adoption, together with a desire to minimize the rep-
disi to implement more extensive and potentially less offen-
s utational cost to the firm, should lead firms to converge
dmi sive versions of a practice, particularly when controlling
aner on packages that are less offensive to outside audiences. for whether the media coverage presents a practice as
p Again, this process is likely to be aided by the emer-
e contested (Green 2004, Sine et al. 2005).2 This leads us
ticland gence of compensation specialists at outside law firms to our next hypothesis.
ars guiding firms through the implementation process, help-
sht ing these firms to create potentially less offensive pack- Hypothesis 2A (H2A). Media coverage of golden
thirig ages. Drawing on these arguments, this suggests that parachutes will have a positive effect on the extensive-
tong over time, parachute plans should come to increasingly ness of golden parachute plans.
tdi resemble those of prior adopters. Hence, we propose the
hu
gcl following hypothesis. At the same time, greater information availability
yriin about a practice also provides potential adopters with
opon, Hypothesis 1B (H1B). A later time of adoption will more leeway to match the practice to their specific
cati have a positive effect on the similarity of golden para- needs. Regarding the effect of media coverage on the
dsm chute plans. adoption of a contested practice, greater information
olor
hnf Information Availability. An important issue regard- availabilitymightprovidepotentialadopterswithknowl-
i
MSal ing practice variation relates to the availability of infor- edge as to what features resulted in positive experiences
n
Ro mation about a diffusing practice. Specifically, greater and what features prompted resistance from outside
FOditi availability of information is likely to be associated constituents. Once workable arrangements have been
NAd with greater extensiveness (Ansari et al. 2010). In the identified,therenotonlymaybecostsofcontinuedvari-
I
absence of detailed information, risk aversion is likely ation, but there also may remain few possible efficiency
to negatively affect deviation. Particularly during the increases. Accordingly, availability of information about
early period of practice diffusion, when there is limited prior adopters’ experiences will likely promote con-
information about the effectiveness of a new practice vergence on implementations that worked well else-
or when it is likely to generate political controversy, as where. Controlling for the contested nature of the media
golden parachutes did, we expect organizational deci- discourse, this argument would suggest that increased
sion makers to proceed more cautiously and incremen- media coverage will lead newly adopted plans to be
tally with adoption (Mooney and Lee 1999, Rice and more similar to those previously adopted. Accordingly,
Rogers 1980). However, as more information about the we suggest the following hypothesis.
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
OrganizationScience23(4),pp.1077–1099,©2012INFORMS 1083
Hypothesis 2B (H2B). Media coverage of golden will have a negative effect on the similarity of golden
parachutes will have a positive effect on the similarity parachutes.
of golden parachute plans. On the other hand, however, a strategy of plan differ-
entiation might require significant changes to plans and
Regulative Environment and Legal Challenges. Vari-
considerable effort in setting up severance agreements
ation in diffusing practices is furthermore likely to be
that may be dissimilar enough to avoid court challenges.
affectedbytheregulativeenvironmentinwhichadoption
When faced with considerable uncertainty regarding the
and implementation take place (e.g., Sine et al. 2005,
best course of action, prior work has pointed to the
s).g/. Suchman 1995). In this regard, contested practices and importance of imitation processes (e.g., DiMaggio and
r(or innovations are particularly likely to be challenged in Powell 1983, Haunschild and Miner 1997). In the reg-
authoorms. ctoougratinbyfrosmtakeeihthoeldreirnshiobritiontghedrifofurgsaionnizoatrioanffsectthiantgswtahnadt utilcaetotrhyatdroemseaminb,leadthoopstiengofvoetrhseiornasdoopftearscohnatsesttheedapdrdaecd-
enf version of the practice is adopted. Regarding the current benefit of also providing a “safety in numbers” effect
othals.i context, such lawsuits—both those already decided and because an increasing number of similar plans makes
tn those still pending—introduce greater uncertainty into legal challenges less feasible. In contrast, adoption of a
r
syou the process of creating parachutes. fairly unique plan that differs considerably from stan-
urtep://j As noted above, golden parachutes became the target dardplansinusemakesitmorelikelyacorporationwill
cohtt of several highly publicized legal suits in which both attract unwanted attention from the opponents of golden
aat shareholders and acquirers challenged the legitimacy of parachutes.Adoptingplanssimilartothoseadoptedpre-
asable tahensuemsbeevreroafncpeotepnatcikaalgteesstincasceosurwt.erHeotwrieevde,r,thaelthcoouugrths vsiinogulseldyothuutsforerdluecgeasl tchoentreissktatthioant.the corporation will be
pyail generally did not take a definite position on the enforce- Given considerable arguments suggesting both nega-
ov
ca ability of golden parachute contracts, and considerable tiveandpositiveeffectsofcourtcasesonplansimilarity,
s
si legal uncertainty persisted even during the later 1980s it is more difficult to make a clear prediction. Where
thies, (Bress 1987). In particular, application of the business there is legal uncertainty about a practice, however, the
butedpolici jpuadragcmheuntetsruwleoumldadbeeitupmhoerled.liNkeevlyertthhaetlecsasr,etfhuellyprdersaefntecde choigshteor,fbdoetvhiaitnintgermfrosmofprreepvuiotautsiomnoadnedlsinisreltiakienliyngtooubte-
strion of these court cases and the associated indeterminacy side counsel capable of developing new types of con-
disi regardingthelegalstandingofgoldenparachutespresent tracts that are likely to be upheld. Thus, we cautiously
s
dmi a dilemma for boards of directors because they lead to adopt the following hypothesis.
anper considerable uncertainty regarding what versions of the Hypothesis 3B (H3B). Court cases against golden
e
clnd parachute are acceptable. In response, the risk of the parachutes will have a positive effect on the similarity
rtia unwanted publicity associated with a legal challenge as of golden parachute plans.
as
sht well as the threat of legal liability should make directors
thirig less willing to award large parachutes to management. Organization-Level Mechanisms Affecting
og As the head of a senior consultancy noted, “The suits Practice Variation
tn
htudi have people skittish” (Proeksch 1986, p. F1). We there- FirmVisibility. Patternsofpracticevariationarelikely
gcl fore expect contestation in the regulative environment to
yriin inhibittheadoptionofextensivegoldenparachuteplans. to be influenced by organization-level factors. In par-
pn, ticular, prior research suggests that the specific form
coatio Accordingly, we hypothesize the following. an adopted practice takes is likely to be influenced by
sm Hypothesis 3A (H3A). Court cases against golden the visibility of the adopting organization. This is even
d
olor parachutes will have a negative effect on the extensive- more likely if the practice in question is contested, as
hinf ness of golden parachute plans. is the case in the current context. As suggested by
MSal Oliver (1991), exposure to or isolation from a variety
n
Ro In a similar vein, the presence of legal contestation of diverse stakeholder groups is likely to affect a firm’s
FOditi of golden parachutes should also affect how far corpo- response to institutional pressures. Similarly, Meznar
INAd rations look to prior parachutes as models for their own and Nigh (1995) have argued that visibility of firms in
severance agreements. On the one hand, the presence of the media makes them more responsive to social and
courtchallengessuggeststhatfirmsmayconsideradjust- political pressures. Visibility of firms has been shown to
ing their own plans to avoid the risk of having them affect issues such as corporate philanthropy (Brammer
invalidatedbyanegativejudgment.Intherelatedcaseof and Millington 2006), the framing of strategic change
poison pills, private law firms additionally acted as legal (Fiss and Zajac 2006), and profit margins (Erfle and
entrepreneurs that developed and marketed new prod- McMillan 1990).
ucts (Powell 1993), leading to the emergence of new This suggests that the visibility of an adopting firm
legaldevicesforrespondingtocourtchallenges.Accord- will likely also affect the extensiveness of practice im-
ingly, one might argue that the presence of court cases plementation. Regarding golden parachutes, because of
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
1084 OrganizationScience23(4),pp.1077–1099,©2012INFORMS
the increasing attention and opposition to them on the arises.Theylooklesssuspecttooutsidersiftheyarefix-
partofshareholdersandthepublic,boardswouldusually tures before bids crop up” (Kleinfield 1982, p. D1). For
prefer that these agreements did not come under close instance, when Garfinckel, Brooks Brothers, Miller &
scrutiny (Berry 1988). Because visibility in the media Rhoades Inc. implemented a quite generous plan only a
creates exposure to constituent groups such as share- day before a hostile bid by Allied Stores became public,
holders, visibility should lead firms to adopt versions of a number of questions arose about whether this was a
a practice that are less extensive rather than more exten- legitimateplan.Becausesuchparachutesaremuchmore
sive in order to avoid antagonizing these constituents. difficult to justify, we would expect boards to approve,
s).g/. Conversely, firms with a lower profile may face fewer on average, less generous severance agreements when in
(r
hors.o constraints in offering their executives more elaborate the middle of a takeover battle. Thus, we propose the
utm parachute packages. This suggests the following effect following hypothesis.
eanfor of visibility of practice extensiveness. Hypothesis 5A (H5A). The presence of a takeover
ths.i Hypothesis 4A (H4A). Greater visibility of a firm bid for the focal firm will have a negative effect on the
tonal will have a negative effect on the extensiveness of its extensiveness of its golden parachute plan.
r
syou golden parachute plan. Golden parachute plans adopted in the midst of a
urtep://j Furthermore, the visibility of an adopting organiza- takeover battle are, furthermore, likely to be more sim-
cohtt tion is likely to affect not only the extent of the adopted ilar to other plans previously adopted. Specifically, the
aat practice but also whether the version of a practice it evident difficulty of justifying the adoption of a golden
asble implements resembles prior adoptions. Actors that are parachute plan at that time should lead boards to imple-
a ment plans similar to those that have already been
copyavail mfacoeregcreeantterarllryesltoriccattieodnsinonnetthweiorrkasctoiofnsso(cDiaalvrisela1t9io9n1s, adopted by other large corporations. Adopting a fairly
standard plan allows the boards to point toward prece-
sis Leblebici et al. 1991, Strang and Soule 1998). Similarly,
thies, organizationswithgreatervisibilityarelikelytobemore dReonbtinassonwe2l0l0a1s),“wsahfiecthy ainppneaurmsbpearsrt”ic(uAlahrmlyadimjiapnortaanndt
dci limited in their ability to adopt custom-tailored versions,
butepoli making them more likely to implement versions of a gthiveetnimthinegcohfaltlheengaedsopintiornat.ioSnuaclhizliansgt-tmheinuptleanpabraascehdutoens
rin practice that are also adopted by other organizations.
sto have also been criticized as contracts that are legally not
disi Given the contested nature of golden parachutes, this
s enforceable because the executives had not done any-
dmi suggeststhathighlyvisibleorganizationswouldbemore
aner likely to adopt “plain vanilla” versions of the practice in thing to warrant the payout (e.g., Greenhouse 1985), but
p prior adoption and payout of similar plans reduces the
clend order to avoid negative press coverage rather than stand- likelihood of court challenges for the board. Accord-
tia ing out with unique versions of the plan. Accordingly,
ars more visible organizations are likely to implement the ingly, we expect the following relationship.
thisright kind of practice that is commonly implemented by prior Hypothesis 5B (H5B). The presence of a takeover
og adopters during that time of the diffusion process. This bid for the focal firm will have a positive effect on the
ttdin suggeststhefollowinghypothesisregardingsimilarityin similarity of its golden parachute plan to those of prior
ghclu diffusing golden parachute plans. adopters.
rin
pyn,i Hypothesis 4B (H4B). Greater visibility of a firm Managerial Influence. Variation among severance
scomatio wpailrlahchauveteapplaonsittoivetheofsfeecotfopnritohreasidmoipltaerristy. of its golden aaggreeres’maebnitlsityistofubrteheurnmitoedreinliktheleyirtiontbeereasftfse.cFteodribnystamnacne-,
d
olor executives who have made firm-specific investments
hnf Takeover Exposure. Firms also differ in terms of with their careers will expect some protection from
i
MSal their exposure to hostile bids for a change in control, the hazard of loss of employment (Singh and Harianto
n
Ro an issue particularly germane with regard to golden 1989). Beyond such human capital arguments, however,
FOditi parachute arrangements. A key argument for the adop- the balance of power between executives and boards of
NAd tion of parachutes is that they allow the firm to recruit directors also governs the nature of golden parachute
I
andretaintalentedexecutives.However,ifafirmalready payments. Negotiations about golden parachute pack-
faces a takeover bid, there is no need for the parachutes’ ages are generally dealt with at the board level and
recruiting attributes, and executives are also unlikely do not require shareholder approval, thus increasing the
to leave their employment in the midst of a takeover likelihood that powerful managers will aim to use social
battle. Accordingly, boards of directors that deploy a influencemechanismstoaffecthowagreementsareactu-
golden parachute after a takeover bid face consider- ally implemented (Belliveau et al. 1996, Fiss 2006).
able challenges in justifying the expense (Bress 1987). Regarding the ability of managers to negotiate their
For these reasons, “it is also considered bad form to severance agreements, CEOs hired from the outside
hastily draw these agreements right after a hostile bid may have a significant advantage vis-à-vis the board in
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
OrganizationScience23(4),pp.1077–1099,©2012INFORMS 1085
such negotiations. In particular, the availability of other Figure1 GoldenParachuteContracts:YearlyAdoptionswith
optionsshouldincreasethebargainingpowerofoutsider ExtensivenessandSimilarityofImplementations
CEOs,whereasinsidecandidatesareusuallyseenbythe 60 320e
boardtohavefeweroptions(LawlerandBacharach1979, Adoptions or
Hermalin and Weisbach 2003). Furthermore, boards are ns50 SEixmteilnasriitvyeness 290y sc
lciaknedlyidatotesbebemcaourseepthowereerfaurel liinkenleygsoetviaetriaolnpsowteinthtiailncsaidne- doptio40 260milarit
hor(s).s.org/. dmbiedenaattbeslsesu(tEgoglsneaesigtdothetiatattaeel.xfe2oc0ru0mt9iv)oe.rseInheixcreotedmnfsbriiovnmeatpitoahnrea,cothuhutestsieedepaarwgciukl--l Number of a123000 122703000nsiveness/si
autorm ahgyepsotthheasnizweothueldfoCllEoOwsinhgi.red from the inside. Thus, we 0 140Exte
enf 19801981198219831984198519861987198819891990
ths.i Hypothesis 6A (H6A). HavinganoutsiderCEOwill Year
tonal have a positive effect on the extensiveness of a firm’s
r
yu golden parachute plan. insurance), or coverage of legal fees, as long as these
so
courtehttp://j askSifmoirlaurnaiqrguuemeelenmtseanltssoinaptphley ptoartahcehuCteE.OB’soaarbdislittyentdo pcoacnktraogle.s were specifically triggered by a change in
aat to make fairly extensive use of benchmarking in set-
Adoptions and Periodization
asble ting executive compensation (Porac et al. 1999). With U.S. firms adopted 314 parachute agreements between
a uncertainty about parachute agreements, boards should
pyail generally opt for severance packages that are similar to 1980 and 1990. After excluding cases with incom-
coav thosepackagesadoptedbyalargenumberofotherfirms plete data, 294 golden parachute agreements were avail-
thises,is rhaotohderotfhastnakfoerhoulndieqrureepsiasctaknacgeesa,ntdhulsawresduuitcsi.nTghteheneliekdeltio- apbalreacfhourteanaaglryeseims.enFtisgupruet 1insphloawcesetahcehnyuemarb.eTrhoefpneeawk
dci years for adoption were 1982 and 1983. By the end
butepoli sahgorewemtheanttsaspeaermacshpuatertiacgurleaerlmyeinmtpisorctaonmtpbaercaabuleset,oasotohneer of 1982, the variations implemented by early adopters
rin contained all of the major contract provisions we iden-
sto acquirer stated, “Some shareholders will be gunning for
dissi directors in every case” (Proeksch 1986, p. F1). How- tified as relevant for our analysis of practice variation.
dmi In 1984, agreements dropped to a lower but mostly
aner ever, the board’s ability to negotiate a standard package steady level through 1986. Paradoxically, 1984’s Con-
p is likely to be countered by a CEO’s ability to demand
e gressional action to tax golden parachute payments sig-
rticland athecuosutotsmidizeesdhopuacldkafgaec.eIfnewpearrticcounlastrr,aCinEtsOisnhaisrkedingfrofomr naled that the idea of a golden parachute had truly
as arrived. Then, 1987 marked the beginning of a mostly
thisright pwahriamchsu,treescuoltnintrgacintsptaarialocrheudtetopatchkeaigressptehcaitfiacreneleesdsssaimnd- dooffwsnhwaraprldytraetntdh—e d1e9c8a9dwe’asseanbdl.ipGeoxincgepbtieoyno—ndthoautrtadilaetda
og ilar to the average contract. This suggests the following
tn analysiswindow,otherstudiesofgoldenparachutessug-
htudi hypothesis. gest that adoptions remained extremely low in the early
rigncl Hypothesis 6B (H6B). HavinganoutsiderCEOwill 1990s (Narayanan and Sundaram 2000).
yi
pn, haveanegativeeffectonthesimilarityofafirm’sgolden Thus, we select 1980–1990 as the period in which
coatio parachute plan to those of prior adopters. golden parachutes diffused and became standard cor-
sm porate practice (Davis and Greve 1997). For purposes
d
olor of our analysis, we define the early versus middle ver-
hnf Data and Methods sus late years of the period as 1980–1983, 1984–1986,
i
MSal To test our hypotheses about extensiveness and similar- and 1987–1990, respectively. These roughly equal peri-
n
Ro ity of golden parachute agreements, we collected details ods line up well with three different levels and trends
FOditi on all golden parachute contracts adopted by publicly of adoption activity, but the growth and decline at the
Nd
IA listed U.S. firms included in either the 1980 or the 1986 beginning and end of the decade are steep enough that
Fortune500lists.FollowingDavisandGreve(1997),we expanding the middle period a year either way has little
identified parachute adoption as a formal obligation to effect on our analyses.
provide compensation to an executive contingent upon
a “change in control” in the corporation. If a severance Defining Features of Golden Parachutes:
agreementdidnotexplicitlymentionachange-in-control A Concept Network Approach
clause, it was not counted as a golden parachute. In To analyze changes in the features of golden parachute
contrast to Davis and Greve, however, we do include contracts, we identified the range of distinct provisions
packages that consisted primarily of compensation other they included. From a close reading of all the agree-
than cash (e.g., accelerated stock options), benefits (e.g., ments,weidentified13suchprovisions,whicharelisted
Fiss,Kennedy,andDavis: DiffusionandVariationofaControversialPractice
1086 OrganizationScience23(4),pp.1077–1099,©2012INFORMS
Table1 GoldenParachuteProvisions
Normaleigenvectorcentrality
Figure2
Description abbreviation 1980–1983 1984–1986 1987–1990
1. Voluntaryterminationoftheofficerasatrigger Voluntary 71(cid:48)4 71(cid:48)9 71(cid:48)0
2. Involuntaryterminationoftheofficerasatrigger Involuntary 56(cid:48)2 61(cid:48)2 62(cid:48)0
3. Continuedemploymentoftheofficeraftera Continuation 55(cid:48)1 53(cid:48)0 60(cid:48)4
changeincontrol
s).g/. 4. Lump-sumpaymentoptionforcoveredofficers Lumpsum 51(cid:48)6 51(cid:48)7 50(cid:48)6
r(or 5. Bonusawardedtotheofficer(s) Bonus 49(cid:48)5 43(cid:48)6 45(cid:48)6
hos. 6. Healthinsurancecontinuationforofficer(s)after Insurance 35(cid:48)8 34(cid:48)7 30(cid:48)1
utm achangeincontrol
aor
enf 7. Retirementbenefitscontinuationforofficer(s) Retirement 33(cid:48)5 33(cid:48)4 25(cid:48)8
ths.i afterachangeincontrol
oal 8. Otherbenefitscontinuationforofficer(s)aftera Benefits 18(cid:48)3 25(cid:48)3 29(cid:48)6
tn changeincontrol
r
syou 9. Stockoptionpaymentstoofficers Stockoptions 26(cid:48)0 19(cid:48)3 12(cid:48)3
urtep://j 1101.. SRteolcokcagtiroanntfepeasymfoernotfsfictoerosfficcoevresred GReralonctsation 99(cid:48)(cid:48)44 137(cid:48)(cid:48)59 104(cid:48)(cid:48)94
cohtt 12. Legalfeesforofficerscovered Legalfees 2(cid:48)7 10(cid:48)9 9(cid:48)2
aat 13. Noncompetitionclauseincluded Noncompete 3(cid:48)0 3(cid:48)6 2(cid:48)0
asble
a
pyail in Table 1. Of these 13 distinct provisions, 11 had the takeover. For several years, penalties were signifi-
coav appeared in at least one agreement by the end of 1981, cant for both employers and departing executives. For
sis and all 13 had appeared by the end of 1982. These the former, they included a loss of the company’s tax
thies, 13 provisions of golden parachutes represent the range deduction on the extra compensation; for the latter, a
dci of options adopting boards of directors might choose to surcharge of 20% on the executive’s personal income
butepoli include when entering into change-of-ownership sever- tax. Between limits to these payoffs and our interest in
rin ance arrangements with their officers. We coded every variation rather than excess, we use salary percentage as
distsio agreement for the presence or absence of each of the a control variable.
s
dmi 13 provisions. To understand the importance of provisions for the
aner Table 1 shows these 13 features of golden parachutes overallpractice,wedrawonnetworkmethodstoanalyze
p
ticleand iwneirgahntkedorbdyerthoeffcreoq-oucecnucryreonfcecow-oitchcuorrthinegr pprroovviissiioonnss,. tthateiopnrowviisthionosthienrTparbolveis1ioinns.teTrmhes oinftuthiteiiorncohiemrepliesmtehna-t
r
sahts Fpaorracehxuatmesplbee,cathmeefikrnsotwfinveasitceomnstratcetlsl tuhsatthgautargaonltdeeend provisionsthatco-occurwithothersfrequentlyoralmost
thirig officers a bonus for selling the company, usually as a always are most central to the concept, whereas those
og that rarely co-occur are more peripheral. To assemble
tn lump-sum payment even though the promise of con-
htudi tinued employment was typically part of the deal as such a view of which contract features were most essen-
gcl tial to golden parachute implementations, we created for
yriin well, regardless of whether covered executives decided eachperiodatwo-modeaffiliationmatrixrelatingagree-
copation, toTlehaeve1v3olfuenattaurrielsy oinr bTeacbaluese1threeyflewcterteheletvgaori.ety and meaesnytsvitsouacloiznatrtiaocnt,pwroevicsoionnvse.rtFedurtthheartmtworoe-,mtoodaellmowatrainx
sm rangeofconcernscoveredinourearlierdiscussionofthe
d into a single-mode adjacency matrix showing how the
olor history of golden parachutes. As explained previously,
hnf golden parachutes departed from prior executive sever- contract provisions co-occurred in agreements. In such
i
MSal ance agreements by offering compensation and benefits a model of relations among contract provisions, those
n
Ro in the case of ownership changes—and doing so with thatco-occurmostoftencanbeviewedastheprovisions
FOditi generosity designed to overcome executive resistance to most essential to the emerging standard for what ought
INAd such changes. Another commonly discussed feature of to be implemented in golden parachute agreements.
goldenparachutesisthepercentageofsalaryincludedin Furthermore, core versus peripheral conditions can be
such plans. Although this feature of contracts has been tangibly measured using a variety of measures of net-
examined previously (Singh and Harianto 1989), the work centrality. For our study, we chose eigenvector
U.S. Congress imposed limits on it in the Tax Reform centrality for two benefits that meet our needs as sum-
Act of 1984. Specifically, that act imposed increased tax marized by Bonacich and colleagues (Bonacich 2007,
liabilitiesonexecutivesfor“excessive”goldenparachute Bonacich et al. 1998). First, eigenvector centrality offers
payments, usually defined as payments whose present the advantage of paying attention to the weight or
value equaled or exceeded three times the executive’s frequency of links between items, so more frequent
average total compensation in the five years preceding co-occurrences between items are taken into account.