american sociological review 2004 zajac 433 57
TRANSCRIPT
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
1/28
http://asr.sagepub.com/AmericanSociological Review
http://asr.sagepub.com/content/69/3/433Theonline version of this article can be foundat:
DOI: 10.1177/000312240406900306
2004 69: 433American Sociological ReviewEdward J. Zajac and James D. Westphal
Perspectives on Stock Market ReactionsThe Social Construction of Market Value: Institutionalization and Learning
Published by:
http://www.sagepublications.com
On behalf of:
American Sociological Association
can be found at:American Sociological ReviewAdditional services and information for
http://asr.sagepub.com/cgi/alertsEmail Alerts:
http://asr.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://asr.sagepub.com/content/69/3/433.refs.htmlCitations:
What is This?
- Jun 1, 2004Version of Record>>
by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from by Daniel Silva on November 1, 2013asr.sagepub.comDownloaded from
http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/content/69/3/433http://asr.sagepub.com/content/69/3/433http://asr.sagepub.com/content/69/3/433http://www.sagepublications.com/http://www.asanet.org/http://asr.sagepub.com/cgi/alertshttp://asr.sagepub.com/cgi/alertshttp://asr.sagepub.com/subscriptionshttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsPermissions.navhttp://asr.sagepub.com/content/69/3/433.refs.htmlhttp://asr.sagepub.com/content/69/3/433.refs.htmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://asr.sagepub.com/content/69/3/433.full.pdfhttp://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://asr.sagepub.com/http://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://asr.sagepub.com/content/69/3/433.full.pdfhttp://asr.sagepub.com/content/69/3/433.refs.htmlhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsReprints.navhttp://asr.sagepub.com/subscriptionshttp://asr.sagepub.com/cgi/alertshttp://www.asanet.org/http://www.sagepublications.com/http://asr.sagepub.com/content/69/3/433http://asr.sagepub.com/ -
8/13/2019 American Sociological Review 2004 Zajac 433 57
2/28
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
3/28
change in executive succession practices, orga-nizational structure, and acquisition targets inthe industry (Thornton 2001, 2002; Thorntonand Ocasio 1999). Moreover, Fligstein and col-leagues demonstrated how a shift in prevailing
conceptions of corporate control from a manu-facturing logic to a finance logic contributed tothe spread of multidivisional forms of organi-zation (Fligstein 1990; Fligstein and Brantley1992).
While organizational policies and practicescan acquire legitimacy and spread in a popula-tion to the extent that they are consonant withprevailing institutional logics, such policies alsoacquire further legitimacy by virtue of theirprevalence in a population. Fligstein (1990,
1991) found not only that firms were more like-ly to adopt the multidivisional form followingthe advent of a finance conception of control butalso that firms were most likely to adopt thisinnovation when many other firms in their orga-nizational field had already done so. This find-ing can be interpreted as suggesting that asmore firms adopted the multidivisional form itbecame progressively institutionalized, or taken-for-granted, as an appropriate means of organ-
izing production (cf., Chaves 1996; Davis andGreve 1997; Tolbert and Zucker 1983). Zajacand Westphal (1995) extended this literature byshowing that as prevailing beliefs about corpo-rate governance changed to suggest a more pos-itive view of certain corporate policies (e.g.,CEO incentive plans), and as such policiesbecame more prevalent among large firms, afocal firm was more likely to formally adopt theplans but less likely to implement them. Theysuggested that as executive incentive plans
became institutionalized, building symbolicvalue as normatively appropriate elements ofcorporate governance policy, f irms were morelikely to formally adopt the plans (to enhancetheir external legitimacy) while decoupling theplans from actual governance practices (to pre-serve informal routines and political interests ofexecutives) (Edelman et al. 1991; Meyer andRowan 1977).
This literature has thus demonstrated that
institutional processes can influence how orga-nizational actors perceive corporate policies,affecting firm-level adoption of policies andstructures and field-level change in prevailingorganizational practices. Yet, little research hasexamined whether and how these processes
influence the perceptions of actors in financial
markets and ultimately determine the market
value of corporate policies. Theories of capital
markets are dominated by a financial econom-
ics perspective in which the stock markets reac-
tion to a policy adoption is conceived as areliable, historically invariant indicator of the
efficiency benefits gained from adopting the
policy (David 1997; Fama 1970; Timmerman
1993). The market is seen as responding to the
value of corporate policies according to con-
sistent criteria of technical efficiency. This
response would only change over time to the
extent that market actors acquire new informa-
tion about the degree to which the policy satis-
fies those criteria and if they update theirassessment of the policys efficiency benefits
accordingly. This perspective does not recognize
the potential influence on market actors of his-
torical change in prevailing beliefs about the
sources or criteria of organizational efficiency.
Moreover, from this perspective, if there is evi-
dence that firms often adopt, but do not imple-
ment a corporate policy (i.e., efficiency benefits
are not realized), then the stock market should
incorporate this evidence and discount the value
of the policy (i.e., the policy would receive a lesspositive market reaction).
We develop a sociological perspective on the
process and outcomes associated with the stock
market valuation of corporate policies and prac-
tices, thus providing an alternative to the dom-
inant, financial economics perspective. Our
perspective highlights how investment behavior
is governed by social dynamics that can com-
promise market eff iciency. This theoretical
approach is consistent with recent sociologicalforays into the study of financial markets. For
example, studies of trader markets have shown
that investors base their investment decisions on
the actual or anticipated decisions of other mar-
ket players rather than on independent forecasts
of the assetseconomic performance (Abolafia
1996; Knorr Cetina and Bruegger 2002;
MacKenzie 2003). The importance of imita-
tion or social referencing in investment decisions
can, in fact, be traced to statements by Keynes(1936) and Merton (1948), as others have noted
(see Zuckerman 1999). We build on this litera-
ture by considering theoretically and empiri-
cally how different historical and institutional
contexts condition these social dynamics, and
434AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
4/28
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
5/28
ization of managers as self-interested grew, this
view displaced the pre-agency emphasis onvaluing managers for their professional expert-
ise. Whereas top executives had previously beenviewed as professionals who possessed the
unique strategic knowledge required for efficient
allocation of corporate resources, in the 1980sthey were increasingly characterized as rela-
tively fungible agents of shareholders. Useem(1993:21) quoted investor Carl Icahn as refer-
ring to executives as drones. Useems quali-tative research, together with Zajac and
Westphals (1995) archival analyses of proxystatements and Davis and colleagues (Davis,Diekmann, and Tinsley 1994; Davis and
Thompson 1994) analyses of Securities andExchange Commission (SEC) documents, court
decisions and publications of various investor
groups, all provided evidence that this shift inassumptions about top executives was prevalent
among the full spectrum of corporate stake-holders.
The agency logic also proceeds from differ-ent assumptions about the firm. In the agencyconception, there is nothing uniquely valuable
about the corporation itself. It is merely a nexusof contracts between individuals (Jensen and
Meckling 1976:311); the very notion of a cor-
poration is dismissed as a legal fiction (Davis
et al. 1994:548). As Davis and colleagues havesuggested, this conception of the firm eclipsed
the longstanding notion that corporations wereinstitutions with unique core competencies.
Agency assumptions led to a different model of
economic resource allocation, which severalauthors have characterized as investor capi-
talism, in contrast to managerial capitalism(Davis and Thompson 1994; Useem 1996:1;
Westphal and Zajac 1998). Specifically, if man-agers are merely fungible agents with no par-
ticular strategic expertise, and if firms aremerely nexus of contracts without unique corecompetencies, then resources can be allocated
by investors in capital markets rather than byexecutives in corporations. If executives are
also presumed to generate agency costs, then
capital allocation would be better left toinvestors.
Agency assumptions also led to differentinterpretations of specific governance practices.
In a study of verbal explanations for executiveincentive plans in proxy statements, Zajac andWestphal (1995) found that incentive plans in
the 1970s were typically justified as a mecha-nism to attract and retain scarce executive tal-
ent (consistent with the traditional corporate
436AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
Table 1. Agency Logic Versus Corporate Logic of Governance
X
Assumptions about:
Top managers
X
The firm
Concept of resource allocation
Links to high-order cultural frames
Links to theories of organization
Implications for governance
practices:
Compensation
Allocation of cash flow
Agency Logic
X
Relatively fungible agents of
shareholders
Pursue strategies that advancepersonal interests at expense of
shareholders.
A nexus of contracts; legal fiction.
Investor capitalism: shareholders
can diversify better and more
easily than firms.
Logic of capitalist markets
Agency theory (Jensen and
Meckling, 1976)X
Use incentives to align manage-
ment and shareholder interests.
Redistribute to shareholders.
Corporate Logic
X
Professionals with unique strategic
knowledge that is required for
efficient allocation of corporateresources.
Stewards of their organizations.
An institution with unique core
competencies.
Managerial capitalism: top man-
agers have primary responsibili-
ty for allocating resources to
different businesses in the corpo-
ration.
Norms of professional autonomy
Managerialist theory (Chandler,
1962)X
Use salary and other rewards to
attract and retain scarce manage-
ment talent.
Retain and reinvest.
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
6/28
conception of managers as professionals who
have unique strategic expertise). In the mid-to-
late 1980s, these plans were more frequently jus-
tified as a mechanism to align manager and
shareholder interests (consistent with the agency
logic).Governance researchers have generally attrib-
uted the rise of the agency logic to investor dis-
satisfaction with the performance of large U.S.
companies beginning in the mid- 1970s. Useem
(1993) provided qualitative evidence that when
corporate performance problems continued into
the early 1980s, despite large-scale deregulation
and a generally favorable political climate for
business, the search for causes veered in on
management itself (p. 21). In particular, per-formance problems of large companies were
widely attributed to excessive levels of corpo-
rate diversification, which in turn were increas-
ingly attributed to empire-building by top
executives (Useem 1993:21). According to
Useem (1993) and Davis and colleagues (Davis,
Diekmann, and Tinsley 1994; Davis and
Thompson 1994), this attribution of firm per-
formance problems to self-interested decision
making by managers was reinforced by (and to
some extent inspired by) the propagation ofagency theory by financial economists.
Ultimately, however, the popularity of agency
theory as a perspective on corporate governance
derives from its consonance with fundamental,
normative beliefs about the organization of eco-
nomic activity. Friedland and Alford (1991:250)
and Scott (2001) have suggested that institu-
tional logics draw legitimacy from higher-order
logics or cultural frames to which they are
connected. While the corporate logic of gover-nance draws legitimacy from norms of profes-
sional autonomy (Zajac and Westphal 1995),
the agency logic draws legitimacy from its con-
nection to the logic of capitalist markets
(Friedland and Alford 1991), given its empha-
sis on allocative efficiency through the invisi-
ble hand of the stock market rather than through
the visible hand of corporate managers
(Donaldson 1990). To the extent that such a
market logic of resource allocation becamemore prevalent in some sectors of the economy
in the 1980s (e.g., Scott 2000; Thornton 2001;
Thornton and Ocasio 1999), the legitimacy of
the agency logic may have been further
enhanced.
INSTITUTIONAL LOGICS OF GOVERNANCE ANDMARKET REACTIONS TO REPURCHASE PLANSWe now consider how the described shift ininstitutional logics of governance may havechanged the perceived value of a specific cor-
porate governance policy, namely stock repur-chase plans, resulting in changing stock marketreactions to this policy over time. A stock repur-chase or buyback plan is a formal (i.e., writ-ten) policy approved by the board of directorsto buy a portion of the firms outstanding sharesback from investors, typically in the open mar-ket, and retire them (Franz, Rao, and Tripathy1995). The plan typically specifies the numberof shares authorized by the board for repur-
chase and describes how the buybacks will befunded (typically with cash). Repurchase plansgrew more prevalent in the mid-1980s andremained quite common throughout the 1990samong mid- and large-sized companies (Grullonand Ikenberry 2000). From the mid-1980s to themid-1990s, U.S. firms announced plans to buyback more than $1 trillion of their stock.
Financial economists have developed a liter-ature on stock repurchase plans that consists pri-marily of event studies, which in this context
assess the value of a policy according to thestock markets reaction to its adoption. The find-ings of this literature are remarkably consis-tent. Published studies typically demonstratepositive, and often very large and persistentstock market reactions to the adoption of repur-chase plans (cf., Lee, Mikkelson, and Partch1992; Medury, Bowyer, and Srinivasan 1992;Raad and Wu 1995; Ratner, Szewczyk, andTsetsekos 1996). In explaining these findings,financial economists have typically invoked anagency logic. For instance, Jensen (1989)explained the markets enthusiastic response torepurchase plans by suggesting that when man-agers have free cash flow at their disposal theyare tempted to waste it on empire-buildingprojects or other perks that benefit themselvesat the expense of shareholders (p. 64); thus,shareholders benefit from policies such as repur-chase plans that return free cash flow toinvestors and allow them to invest it themselves
(Bagwell and Shoven 1988; Lazonick andOSullivan 2000:17; Medury et al. 1992).Financial gurus and members of the businesspress have offered similar interpretations. Forexample, the renowned investor Warren Buffettgave a widely quoted explanation for positive
SOCIALCONSTRUCTIONOFMARKETVALUE437
#1471-ASR 69:2 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
7/28
reactions to stock buybacks that invoked theagency logic:
By making repurchases . . . management clearlydemonstrates that it is given to actions that enhancethe wealth of shareholders, rather than to actions
that expand managements domain but do nothing(or even harm) shareholders. Seeing this, share-holders and potential shareholders increase theirestimates of future returns from the business,which produces higher market values. (SanFrancisco Chronicle, July 31, 1995, p. B1 1995)
From a financial economics perspective, eventstudies indicate the consequences of policy adop-tions for economic efficiency (David 1997; Fama1970). We suggest, however, that stock marketreactions are based on investorperceptionsabout
the sources of efficiency, rather than on anyunchanging standard of efficiency. More specif-ically, we contend that investor perceptions, likethe perceptions of other corporate stakeholders,are structured by historically variant institution-al logics. As Friedland and Alford (1991) suggest,any behavior/activity can carry with it alterna-tive meanings. . . . There is no one-to-one rela-tionship between an institution and the meaningcarried by the practices associated with it (p.
250).In this case, while the agency logic of gov-
ernance suggests a positive interpretation ofstock repurchase plans, the earlier corporatelogic suggests a different interpretation. Fromthe latter perspective, as discussed above, topmanagers are professionals with unique strate-gic knowledge that enables them to allocatecorporate resources efficiently. Accordingly,they should tend to allocate free cash flow onlyto profitable projects. If such projects are not
available, they would return the cash to share-holders to invest in other companies with moreattractive prospects. Thus, from the corporatelogic perspective, a stock repurchase plan couldbe perceived negatively as evidence that a firmhas only limited prospects for investing itsmoney.
We suggest, therefore, that stock market reac-tions to repurchase plan adoptions have changedover time, driven by changes in the prevailing
institutional logic of governance. Specifically,we predict that financial investorsperceptionsof stock repurchase plans starting in the mid-1980s (i.e., the period when the agency logic ofgovernance became a dominant institutionallogic) will be generally positive, resulting in
positive market reactions to the adoption ofrepurchase plans. We also expect, however, thatprior to this time period (i.e., when the corpo-rate logic of governance was dominant), finan-cial investors perceptions of stock buybacks
will be generally negative, resulting in negativemarket reactions to the adoption of these plans.Some financial economists have suggested thatinvestors react positively to repurchase planadoption in part because they view adoption asa signal that managers believe the firm is under-valued (Ikenberry, Lakonishok, and Vermaelen,2000). However, this perspective does notaddress the potential for market reactions tochange from negative to positive over time,which is the focus of our theory. This suggests
the following, initial hypothesis for the periodsof our study:
Hypothesis 1: Stock market reactions to repur-
chase plan adoptions shift from negative to
positive in the mid-1980s.
THE NON IMPLEMENTATION OF STOCKREPURCHASE PLANSOur discussion thus far has not distinguished
between the formal adoption of stock repur-chase plans and the actual implementation ofthose plans. From a neoinstitutional perspective,when firms adopt policies in an act of con-formity to prevailing institutional logics, theymay also decouple the formally adopted policyfrom informal routines in the organization.Decoupling reflects the overriding [impera-tive] of all systems .|.|. to maintain the integri-ty and continuity of the system itself .|.|.including the stability of informal [routines andpower relationships] within the organization(Scott 1998:117), while still conforming to insti-tutional environments (Selznick 1957). In a pre-vious study (Westphal and Zajac 2001), weposited that institutional decoupling can occurin the context of stock repurchase plans. Wedemonstrated that firms may formally announcea plan to repurchase stock and then subsequentlybuy back only a small fraction of the sharestargeted for repurchase in the formal plan; or
they may make no repurchases at all (see alsoStephens and Weisbach, 1998). We suggestedthat such nonimplementation constitutes aninstitutional decoupling, whereby managerssymbolically demonstrate commitment toagency values by formally adopting repurchase
438AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
8/28
plans, and then the managers neglect to redis-
tribute free cash to shareholders to preserveinformal routines and discretionary control ofcorporate resource allocation within the firm.
In our multivariate analyses, significant pre-
dictors of whether firms were likely to adopt butnot implement repurchase plans included CEO
power, network ties, and a firms history ofprior decoupling, but not post-announcement
stock returns or market-to-book value.Our present study includes data on repur-
chase plans from 1980 to 1994 among a sam-
ple of 463 Forbes 500 companies. Figure 1shows repurchase plan adoption and imple-mentation during this time period. The graph
shows a sharp increase in adoptions in the mid-1980s. It also suggests that the prevalence ofnonimplemented repurchase plans increased
during this period. Thus, firms were more like-ly to formally adopt repurchase plans during a
period in which the agency logic of governanceprevailed, but they were also less likely to actu-ally implement the formally adopted plans.
From a neoinstitutional perspective, it appearsthat repurchase plan adoptions became moresymbolic and less substantive over time
(Edelman et al. 1991; Pfeffer 1981). Whilefirms may gain legitimacy benefits in the stock
market from adopting policies that conform toprevailing institutional logics, we must alsoconsider whether such policies might lose their
symbolic value as evidence of nonimplemen-tation emerges. Would any positive market reac-tions to repurchase plans resulting from the rise
of an agency logic of governance in the mid-
1980s diminish in later time periods, or perhaps
disappear altogether, as evidence of nonimple-
mentation accumulates? We develop competing
hypotheses regarding the effects of prior
instances of adopting, but not implementing,stock repurchase plans on the market reactionto these plans.
MARKET LEARNING VERSUSINSTITUTIONALIZATION PERSPECTIVESFrom a financial economics perspective, stock
market reactions to repurchase plan adoptions
reflect the markets assessment of the efficien-
cy benefits of adopting and implementing the
plans (Bagwell and Shoven 1988; Raad and Wu1995; Ratner et al. 1996). That assessment, in
turn, is influenced by prior evidence regarding
the economic benefits of adoption. The larger
event study literature typically assumes semi-
strong capital market efficiency, wherein all
publicly available information about a givenpolicy at the time of adoption is reflected in the
stock markets response (Brealy and Myers
1991; Fama 1970). Although this literature has
not formally examined change in market reac-tions over time, it is assumed that the market has
responded less positively over time to corporate
strategies, such as unrelated acquisitions,
because of an absence of economic benefits
from the policy. In effect, the market reaction
to a particular policy adoption is assumed to
derive from a sophisticated and rapid statistical
SOCIALCONSTRUCTIONOFMARKETVALUE439
#1471-ASR 69:2 filename:69306-zajac
Figure 1. The Adoption and Decoupling of Stock Repurchase Programs: 19801994
0
10
20
30
40
50
60
70
80
90100
1980 1982 1984 1986 1988 1990 1992 1994
Not Implemented
Implemented
NumberofPlans
Year
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
9/28
analysis, based on all publicly available dataon prior adoptions and subsequent changes ineconomic performance. As more firms adopt thepolicy and their economic efficiency increasesor decreases, these cases are added to the
analysis, and the expected benefits from futureadoptions are modified based on the revisedresults. A corollary of this perspective is thatinvestors who are less informed about the effi-ciency benefits of particular policies (i.e.,investors who do not receive information aboutthe consequences of prior adoptions or whoreceive the information relatively late) are lesslikely to remain in the market over the longterm, because they earn a lower return. Thus,market knowledge about the efficiency benefits
that firms have enjoyed from adopting a par-ticular policy should tend to rise to the level ofthe most knowledgeable investor (Brealy andMyers 1991; Fama 1970).
In this way, markets should learn about thevalue of a particular policy through prior expe-rience with it. There is a growing theoretical lit-erature in information economics on stockmarket learning. However, theorists have typ-ically assumedthat market learning does occur.
The focus in this literature is on varying keyparameters of the learning process, such as thespeed with which investors respond to newinformation (e.g., David 1997; King et al. 1993;Timmermann 1993). Consequently, there is lit-tle empirical evidence for (or against) the under-lying assumption that the market learns(Bradley, Desai, and Kim 1983). Some researchhas examined whether market returns fromadopting a policy that is assumed to benefitshareholders, such as a related acquisition,
appears to dissipate if the policy is not subse-quently consummated (Bradley et al. 1983).However, as several studies in the event studyliterature have shown, the measurement of mar-ket reactions over extended time periods isfraught with difficulties (McWilliams and Siegel1997). Moreover, prior research has not direct-ly examined whether markets appear to learnabout particularpolicies, such that the marketsreaction to a policy is informed by prior expe-
rience with the same policy at other firms.Thus, we posit an initial hypothesis based on
the market-learning perspective. As discussedabove, in the mid-1980s, firms increasinglyadopted stock repurchase plans without actual-ly implementing the plans. When these repur-
chase plans were not implemented, they didnot, of course, enhance economic efficiency aswas expected by agency theorists. Consequently,from a market-learning perspective, as morefirms neglected to implement their repurchase
plans over time, the expected efficiency bene-fits from adoption should diminish, and themarkets reaction to the adoption of these plansshould diminish accordingly. Note that thiswould be true from a market-learning perspec-tive even if other motives for non-implementa-tion were sometimes also involved, sinceinvestors would, on average, come to expectfewer agency benefits from announced repur-chase plans, given growing accumulated evi-dence of non-implementation. Thus we have
the following hypothesis:
Hypothesis 2 (H2): The number of firms that
have adopted, but not implemented, stock
repurchase plans is negatively associated
with the stock market reaction to repur-
chase plan adoption at the focal firm.
A neoinstitutional perspective suggests acompeting prediction regarding the change inmarket reactions to repurchase plans as more
firms adopt (but do not implement) these plansover time. From this perspective, repurchaseplans may accumulate symbolic value over time,despite repeated instances of decoupling,through a process of institutionalization.Institutionalization refers to the social con-struction process by which organizational poli-cies become instilled with value and ultimatelytaken-for-granted among external constituentsas normatively appropriate (Selznick 1996;Tolbert and Zucker 1983:25). Institutionali-
zation is more likely in the context of a supra-organizational belief system that offers a positiveinterpretation of the policy (DiMaggio 1997;Scott 1994). In this case, the advent of an agencylogic of governance suggested a more positiveinterpretation of repurchase plans. Moreover, asmore firms adopt repurchase plans over time,the plans should become more strongly associ-ated with the agency values that they affirm.From a neoinstitutional perspective, policies
build symbolic value through reciprocatedinterpretations in which a policy is presentedto constituents as furthering their values and thatinterpretation is validated by the constituents(Berger and Luckmann 1967; Meyer and Rowan1977; Scott 1987:496, 1994). As this pattern is
440AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
10/28
repeated, the connection between the policy andthe belief system or institutional logic tightensand may ultimately become taken-for-granted.
Thus, stock repurchase plans may graduallyaccumulate symbolic value and thus become
institutionalized as more firms adopt them overtime. We consider how purely ceremonial actsof conformity to agency values, in the form ofdecoupled stock repurchase plans, could accu-mulate symbolic value in this way. Throughexamining change in market reactions to decou-pled repurchase plans as more firms adopt them,we are able to directly assess change in theirsymbolic value. As noted above, although mar-ket reactions are viewed in the financial eco-nomics literature as providing objective data
that reflect expected efficiency gains from pol-icy implementation, from an institutional per-spective, we view such reactions as subjectivedata that reflect the symbolic value of adoption,neatly quantified and aggregated, and we expectthat these numbers will increase as the policybecomes institutionalized.
How could repurchase plans accumulate sym-bolic value, as indicated by more positiveinvestor reactions, despite growing evidence
that they are often not implemented? We sug-gest that, since market reactions to announcedpolicy adoptions occur quickly and are based onimperfect communication among participants,investors are likely to predict the response ofother investors to a policy adoption by referringto prior market responses to similar events. Thissociohistorical estimation process serves to per-petuate market reactions. As more firms adopta policy and receive a favorable marketresponse, an individual investors uncertainty
about the likely response to the current adoptionis reduced, which should tend to result in amore positive reaction (Westphal and Zajac1998). Thus, the market value of repurchaseplans can build over time through a self-per-petuating, social construction process despite thepotential for decoupling. As more firms adoptthe plans and receive a positive reaction, thevalue of the plans becomes increasingly taken-for-granted among individual investors in that
their individual beliefs about the plans becomeless relevant to the evaluation process and areless likely to be referenced in any meaningfulway. In this way, the social referencing processthat underlies stock market reactions to policyadoptions can contribute to the institutional-
ization of corporate policies, and this occurs
even as evidence accumulates that the policies
do not enhance economic efficiency as origi-
nally promised (e.g., due to nonimplementa-
tion).
Moreover, from an institutional perspective,as a policy becomes institutionalized such that
it becomes closely identified with an institu-
tional logic, the collective reaction of external
constituents will increasingly be governed by a
logic of confidence and good faith, whereby
constituents (e.g., investors) accept it on faith
that organizational actors are behaving in accor-
dance with the legitimate goals of the policy
(Meyer and Rowan 1977:357). This shared
assumption that legitimate formal policies arefaithfully implemented not only simplifies the
decision-making process for investors, but also
has the ultimate function of maintaining con-
fidence in organizations and in the extent to
which they are efficiently governed (Meyer and
Rowan 1977:358). Thus, in opposition to
Hypothesis 2, the market-learning hypothesis (a
neo-institutional perspective on stock market
dynamics) leads to the following hypothesis:
Hypothesis 2a: The number of firms that haveadopted, but not implemented, stock repur-
chase plans is positively associated with
the stock market reaction to repurchase
plan adoption at the focal firm.
MARKET LEARNING VERSUS INSTITUTIONAL-
IZATION ACROSS RELATED POLICIES. Prior event
studies have treated different corporate gover-
nance policies and strategies as independent
events. Research on repurchase plans, forinstance, has examined the stock market reac-
tion to such plans independently of other, relat-
ed policy adoptions. However, the markets
reaction to a particular policy may be influ-
enced by the prior experience of investors with
other policies that had similar objectives. From
a market-learning perspective, stock repurchase
plans can be viewed as one element of a larger
set of policies that purport to control agency
costs in the firm. To the extent that marketactors recognize the common objectives that
underlie different policies, the potential for mar-
ket learning is broadened. Thus, market reac-
tions to repurchase plans could be influenced by
the failure to implement other governance
SOCIALCONSTRUCTIONOFMARKETVALUE441
#1471-ASR 69:2 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
11/28
policies that were ostensibly adopted to controlagency costs.
Prior to widespread adoption and decouplingof repurchase plans, many large firms hadadopted executive incentive plans known as
LTIPs (long-term incentive plans). These werealleged to reduce agency costs by aligning CEOinterests with the interests of shareholders(Kumar and Sopariwala 1992).1 Many firmsexplicitly invoked an agency logic in announc-ing LTIP adoption to shareholders in proxystatements, describing the plans as incentivealignment or control mechanisms that dis-courage executives from making decisions thatadvance their own interests at the expense ofshareholders (Wade, Porac, and Pollack 1997;
Zajac and Westphal 1995:283). Thus, the ration-ale for LTIP adoption, like the rationale forstock repurchase plans, was rooted in the agencylogic of governance. Whereas LTIPs partiallyresolve the agency problem by giving CEOs afinancial incentive to pursue shareholder objec-tives (i.e., rather than wasting corporate cash onempire-building strategies [Rajagopalan 1996]),repurchase plans prevent CEOs from wastingcash flow by taking it out of their hands entire-
ly and returning it to investors (Jensen andWarner 1988).Moreover, while implementation of repurchase
plans decreased in the mid-to-late 1980s, in theearly 1980s, many firms formally adopted anLTIP without making grants under the plan(Westphal and Zajac 1998). This represents an ear-lier instance of decoupling a corporate gover-nance policy purportedly designed to controlagency costs. Thus, the market-learning per-spective suggests that as more firms adopt LTIPs
with an agency rationale and then fail to imple-ment the plans (i.e., agency costs are not actual-ly reduced), the market may not only discount theexpected value of LTIPs but it may also lower itsestimation of other policies that purport to reduceagency costs, including stock repurchase plans.In effect, the market learns to recognize the poten-tial for nonimplementation of governance policiesin general, as investors learn from their experiencewith LTIPs to be skeptical about policies that
claim to control agency costs. This suggests the
following hypothesis:
Hypothesis 3: The number of firms that have
adopted LTIPs with an agency explanation
but have not implemented the plans is neg-
atively associated with the stock marketreaction to repurchase plan adoption at the
focal firm.
However, a neoinstitutional perspective sug-
gests an additional, competing prediction regard-
ing the effect of prior LTIP adoptions on the
stock markets reaction to repurchase plan adop-
tions. Just as repurchase plans can be seen as
acquiring greater symbolic value as more f irms
adopt them over time, the symbolic value that
accumulates to a policy from repeated instances
of adoption and endorsement by an external
audience may spread to other policies that can
be interpreted as having a similar objective. As
discussed earlier, institutionalization occurs
through a social construction process in which
a policy is presented to constituents as further-
ing a set of objectives, and that interpretation is
validated by constituents. As this pattern is
repeated over time, the benefits of the policy
become increasingly taken-for-granted. An insti-tutional perspective also suggests that as more
firms adopt a particular policy and invoke the
same or similar logic for adoption, not only the
policy but also the stated objective itself can
build institutional value and become taken-for-
granted as a legitimate rationale for action (i.e.,
the assumptions that underlie the alleged bene-
fits from adoption are less likely to be ques-
tioned) (Scott 2001; Zajac and Westphal 1995).
Thus, subsequent policies that appear to con-form to the same logic enjoy greater social
acceptance, and firms realize greater legitima-
cy benefits from adopting them (Meyer, Boli,
and Thomas 1987; Scott 1994, 2001; Suchman
1995). By virtue of this process, the prior dif-
fusion of executive incentive plans for top man-
agers may have enhanced the legitimacy of
stock repurchase plans. As more firms adopt-
ed LTIPs with an agency rationale, the agency
logic itself acquired greater legitimacy as asocially accepted rationale for corporate action.
Thus, as LTIPs became widely diffused in the
mid-1980s, firms realized greater legitimacy
benefits from adopting other policies that
appeared to control agency costs, including
stock repurchase plans.
442AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
1 LTIPs have been formally defined as incentive
programs that grant executives the right to receivecommon stock or cash on a particular date in thefuture to the extent that specific performance objec-
tives are met (Crystal 1991).
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
12/28
In this way, repurchase plans may have bor-
rowed symbolic value from the prior diffusion
of LTIPs. Moreover, this process could operate
despite prior decoupling of LTIPs. The social
estimation process that underlies the institu-
tionalization of stock market reactions, in whichinvestors reference prior reactions to similar
policies in determining their own response (i.e.,
regardless of their personal assessment of the
likelihood of implementation), should apply to
LTIPs as well as to repurchase plans. Thus,
market reactions to LTIPs should grow more
positive over time as more f irms adopt LTIPs
with an agency rationale, despite prior instances
of nonimplementation, and this increasingly
positive reaction should then generalize to stock
repurchase plans. This suggests the following
alternative to Hypothesis 3:
Hypothesis 3a: The number of firms that have
adopted LTIPs with an agency explanation
but have not implemented the plans is pos-
itively associated with the stock market
reaction to repurchase plan adoption at the
focal firm.
METHODDATAThe initial sample for this study included all
open-market stock repurchase plans adopted
from 1980 through 1994, among firms listed in
the 1980 Forbes 500 or Fortune 500 indexes. We
chose this time period because nonimplement-
ed repurchase plans became increasingly preva-
lent in the mid-1980s and previously were very
rare, as shown in Figure 1. We excluded casesfor which complete archival data were unavail-
able, leaving a final sample of 860 repurchase
plans adopted by 463 firms. These firms were
not significantly different in size (measured by
sales and assets) or performance (measured by
return on equity and return on assets in the pre-
vious year) from firms in the larger population
that adopted repurchase plans, as indicated by
Kolmogorov-Smirnov two-sample tests.2
We collected complete data for the period1980 to 1994. Later data were also collected tomeasure repurchase plan implementation, andearlier data were collected on LTIP adoption(1972 to 1980). Data on the adoption and imple-
mentation of repurchase plans were obtained pri-marily from an extensive database compiled bythe Securities Data Company. We carefullychecked the accuracy of these data using theWall Street Journal Index, Reuters, and TheInvestment Dealers Digest, and we obtainedadditional data on repurchase plan implementa-tion from COMPUSTAT (see Westphal and Zajac2001). We also obtained information on possibleconfounding events that were announced dur-ing the event window from the Wall Street Journal
Index (see discussion of control variables below).Data on stock market reactions and other finan-cial information were provided by COMPUSTATand the Center for Research in Security Prices(CRSP). We obtained data on board structureand ownership from Compact Disclosure,Standard and Poors Register of Corporations,Directors, and Executives, and corporate proxystatements. Data on CEO incentive plan adoptionand implementation were also collected direct-
ly from proxies.
INDEPENDENT VARIABLESPRIOR REPURCHASE PLAN ADOPTIONS. We creat-ed a dichotomous variable to indicate the adop-tion and nonimplementation of repurchaseplans, coded as 1 if firms adopted a plan and didnot subsequently repurchase any shares withinthree years. We also created a separate variableto indicate implemented repurchase plans, coded
as 1 if firms adopted a repurchase plan andactually repurchased shares within three years.In separate models we measured plan imple-mentation over different time periods (e.g., twoyears or five years following adoption), and theresults presented below were substantivelyunchanged. When firms adopted a new planduring the implementation window, the originalplan was coded as nonimplemented, becausefirms rarely had multiple plans operating simul-taneously.
For each case, we then developed count vari-ables to indicate the number of repurchase plans
SOCIALCONSTRUCTIONOFMARKETVALUE443
#1471-ASR 69:2 filename:69306-zajac
2 We followed prior studies in the financial eco-nomics literature in excluding repurchase plans thatwere adopted during the two-week period after the
1987 stock market crash (e.g., Raad and Wu 1995).These adoptions were made under extraordinary cir-
cumstances in which many firms were grossly under-valued by the market.
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
13/28
that were implemented or not implemented priorto the focal adoption (prior non-implementedrepurchase plans; prior implemented repur-chase plans). We also developed a continuousmeasure of prior repurchase plan implementa-
tion. For each prior adoption, we calculated thepercentage of shares available for repurchasethat were actually repurchased under the plan,and we then calculated the average level ofimplementation across all prior adoptions (aver-age implementation of prior adoptions).
PRIORLTIP ADOPTIONS WITH AGENCY EXPLA-NATIONS. LTIPs are routinely announced toshareholders in proxy statements. The
announcements often include an introductionthat summarizes the ostensible rationale foradopting the plan. We conducted a basic contentanalysis of these announcements (Holsti 1968;Zajac and Westphal 1995). Following Holsti(1968) we did not provide coders with a man-ual dictating how all possible phrases should beclassified; such a procedure can overstate reli-ability. Instead, we gave coders a capsuledescription of the agency perspective, includingkey concepts that characterize the theory (Zajac
and Westphal 1995). Three people coded theannouncements: two were doctoral candidatesin business, and the third was an undergraduatestudent who had taken no business coursework.Thus, the coders had different backgrounds anddifferent levels of exposure to organizationaltheory, which should provide a conservativetest of interrater reliability (Holsti 1968). Coderswere asked to indicate whether an agency expla-nation was present in relevant sections of theproxy. Interrater reliability was high (95 percentagreement rate), suggesting that coders facedminimal ambiguity in classifying the explana-tions. We analyzed proxies throughout the dif-fusion period of LTIPs to identify whetherlong-term incentive grants were made undereach formally adopted plan (see Westphal andZajac 1998). A count variable was created toindicate the number of LTIPs that were adopt-ed with an agency explanation but without actu-ally granting any incentives under the plans,
prior to the focal repurchase plan adoption(prior non-implemented LTIPs). A second vari-able represents the number of LTIPs that hadbeen adopted and implemented (i.e., with grantsmade under the plans) (prior implementedLTIPs).
CONTROL VARIABLES. We controlled for finan-cial and governance characteristics that mightinfluence the markets reaction to repurchaseplan adoption. The financial variables includecash flow per share (i.e., income before extraor-
dinary items, divided by total common shares),long-term debt to equity ratio, return on assets,and log of sales. The governance variablesincluded two indicators of apparent board inde-pendence from management (the ratio of out-side to inside directors, or outsider ratio, andseparation of the CEO and board chair positions,or CEO/chair separation), and the level of own-ership by institutional investors (institutionalownership). Moreover, we controlled for theprior adoption of repurchase plans or long-term
incentive plans at the focal firm (prior adoptionof non-implemented/implemented repurchase
plans; and prior adoption of non-implement-ed/implemented LTIPs). We controlled for majorpolicy announcements and other publicizedincidents that occurred during the event period(other events), including the possible con-founding events listed by McWilliams andSiegel (1997:640). Finally, we included dummyvariables for year and industry (using primary
two-digit SIC codes of adopting firms) (coef-ficients for these dummy variables are notreported and are available from the authors).
DEPENDENT VARIABLESWe measured stock market reactions to repur-chase plan adoption with excess stock returns,or the cumulative difference between a firmsobserved return and its expected return duringa specified time period surrounding adoption
(Brown and Warner 1985; Patell 1976; Gaver,Gaver, and Battistel 1992). It is assumed that inthe absence of stock price effects resulting fromrepurchase plan adoption, stock returns aredescribed by the following market model:
Rjt= j+ jRmt+ jt (1)
whereRjtis the return for firmjon day t,Rmtisthe market return on day t; jis the beta, or sys-tematic risk, of firmj(i.e., the market-adjust-
ed variance in stock returns for firmj), jis therate of return for firmjwhenRmtequals zero;and jt is a serially independent disturbanceterm (E(jt) = 0). The market model parametersfor each firm (jand j) are estimated over theperiod from day 259 to day 21 relative to
444AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
14/28
each adoption date t(Gaver et al. 1992). Theexcess daily return (ejt) for each firm is then esti-mated as follows:
ejt=Rjt aj bjRmt, (2)
where ajand bjare least squares estimates of jand j. Intuitively, this measure calculates thestock returns for a particular firm on a partic-ular day that exceed the returns that would havebeen expected based on the recent returns offirms with comparable betas (i.e., comparablevariance in their stock returns). To correct forheteroskedasticity, we estimated standardizedexcess returns using the Jaffe-Mandelker port-folio method (Binder 1998).
An important variable in event study design
is the event period, or the period over whichexcess returns are cumulated. Relatively longevent periods allow for the possibility of grad-ual diffusion of information about an event fol-lowing adoption. However, research hasgenerally shown that stock prices adjust veryquickly to the announcement of significant cor-porate events such as stock repurchase plans(e.g., within 15 minutes [Dann, Mayers, andRaab 1977; Ryngaert and Netter 1990]).
Moreover, longer event periods increase thelikelihood of contamination from extraneousorganizational events during the time period(we control for such events in this study)(McWilliams and Siegel 1997). Thus, we esti-mated excess returns over a 2-day period (t1 tot0) and an 11-day period (t5 to t+5). These eventperiods are commonly used in the event studyliterature (e.g., Franz et al. 1995; Raad and Wu1995). In the interest of thoroughness, we alsoran analyses of excess returns using a 31-dayevent period (t5 to t+25), and these results arepresented separately below.
ANALYSISEvent studies in the financial economics liter-ature often analyze the effects of independentvariables on excess returns using subgroupanalyses. Several authors have advocated the useof multiple regression analysis instead, in order
to control for possible third variables (e.g.,McWilliams and Siegel 1997). Given that ourdata set includes excess returns from differenttime periods, autocorrelation could occur in thedata (cf., Binder 1998). We use the Cochrane-Orcutt transformation to correct for first-order
autocorrelation. In separate analyses we use thePrais-Winsten method, and the results are verysimilar (Johnston and DiNardo 1997). Asidefrom the Cochrane-Orcutt transformation, weaddress the potential for spuriousness result-
ing from time trends in the data in two ways: (1)in separate models we control for higher-orderautocorrelation and find that the results are sub-stantively unchanged, and (2) we control fortime effects with robust standard errors clusteredby year or by longer time periods (e.g., 1980 to1984 vs. 1985 to 1994) given evidence that theagency logic emerged in the mid-1980s (Davisand Thompson 1994; Useem 1993; Zajac andWestphal 1995), and again we find that theresults are unchanged.
As discussed above, prior repurchase planadoptions were coded as nonimplemented if norepurchases had been made within three yearsof adoption. Given this implementation win-dow, the first three years of the sample areexcluded from the regression analysis. Thus,the sample for this analysis includes all repur-chase plans adopted from 1983 through 1994 (N= 778). To ensure that our results are not sensi-tive to this research design, we conducted two
sets of additional analyses. First, we ran analy-ses with different implementation windows,including windows of (1) one year and (2) sixmonths, and the results were consistent withthose we report below. A relatively short imple-mentation window seems justified, given thatfirms typically have no obvious economic rea-son for delaying implementation, and investorswould not expect such a delay. And from anempirical standpoint, firms that implement theirplans typically do so shortly after adoption (e.g.,the portion of plans that are not implementedincreases by less than 10 percent when imple-mentation is measured over a one-year win-dow).3
Finally, to ensure that sample selection bias-es would not affect our results, we estimate sep-
SOCIALCONSTRUCTIONOFMARKETVALUE445
#1471-ASR 69:2 filename:69306-zajac
3 We also conducted separate analyses that includ-ed the first three years of repurchase plan adoptions,
with additional data on prior adoptions in the late1970s provided by COMPUSTAT. Although complete
data were unavailable for 19 percent of the compa-nies in our sample during this earlier period, theoverall f indings were substantively unchanged from
the results of our primary analyses
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
15/28
arate Heckman selection models, in which one
equation estimates the likelihood of repur-
chase plan adoption, and parameters from that
model are included in a second equation that
estimates excess returns from adoption
(Heckman and Borjas 1980). The results aresimilar, and the hypothesized effects are
unchanged, suggesting that selection biases
are not affecting our results.
RESULTSHypothesis 1 is tested in an event study analy-
sis of repurchase plan adoptions. The results
are displayed in Figure 2. We assess the sig-
nificance of excess returns using the follow-
ing test statistic, which is commonly used in
the event study literature (Brown and Warner
1985):
At/S(At),
whereAtis the average cumulated excess return
over the relevant observation period, and S(At)
is the time-series standard deviation of excess
returns over a 238-day estimation period. Given
that average excess returns are independent,
identically distributed, and normal, this sta-
tistic has a student-tdistribution under the null
hypothesis. As shown in Figure 2, market reac-
tions were significantly positive in the late
1980s, consistent with results from the prior
event study literature, and this response con-
tinued through the 1990s. However, the results
also clearly show that market reactions became
more positive over time and were significant-
ly negative in the early years of the time peri-od. These negative responses are significantly
different from the positive responses in the
later period, consistent with Hypothesis 1 (the
difference in t-statistics between the begin-
ning and the end of the study period is signif-
icant at thep .01 level). This shift in market
reactions from negative to positive occurred
even as the cumulative evidence of nonimple-
mentation and the yearly rate of nonimple-
mentation were growing over the same period,as shown in Figure 1.
Descriptive statistics and correlation coeffi-cients for variables included in the regressionmodels are displayed in Table 2. Results of theCochrane-Orcutt regression analyses are pro-vided in Table 3. For this analysis, excess returnsare calculated using three different event peri-ods, and results are presented separately foreach. Results of models that estimate excess
returns over a 2-day period (t1 to t0) are shownin the first column, while models based on an11-day period (t5 to t+5) and a 31-day event peri-od (t5 to t+25) are shown in the second and thirdcolumns, respectively. These models control for
446AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
Figure 2. Excess Returns from Stock Repurchase Plan Adoptions: 19801994
Note: Average excess returns are presented on the vertical axis; tstatistics appear in parentheses. For years with
less than 50 observations, statistics are derived from a bootstrap distribution of excess returns (Dodd and Warner
1983; McWilliams and Siegel 1997). *p .05; **p .01; ***p .001 (one-tailed tests).
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
1980 1982 1984 1986 1988 1990 1992 1994
(-2.51**)
(-2.34*)
(-2.13*)
(-1.53)
(.97)
(1.72) (1.88) (2.08*)(2.25*)
(2.61**)(2.77**)
(2.83**) (2.78**)(2.71**)
(2.75**)
Year
Ave
rageExcessReturns
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
16/28
year effects and industry differences, as well as
the other control variables listed in the table. The
results consistently support Hypothesis 2a: Thenumber of firms that have previously adopted
but not implemented stock repurchase plans is
positively associated with the stock market reac-
tion to adoption of the focal plan, for all three
event periods. Conversely, Hypothesis 2, the
market-learning hypothesis, which argued that
markets would begin to discount repurchaseannouncements given accumulating evidence
of nonimplementation, is not supported. Note
that the number of prior implemented repur-chase plans is also associated with more posi-
tive market reactions. In effect, the number of
previously adopted plans predicts market reac-
tions regardless of whether those plans were
implemented.
The results in Table 3 also show support for
Hypothesis 3a. In particular, after controlling for
year effects, the number of firms that have pre-viously adopted LTIPs with an agency expla-
nation but not implemented the plans is
positively associated with the stock market reac-
tion to repurchase plan adoption at the focal
firm. This result, which holds for all three event
periods, is consistent with our contention that
the use and the spread of the agency logic of
governance as the interpretive lens for one cor-porate governance policy (i.e., top executive
incentive plans) can positively color investorperceptions of another corporate policy (stock
repurchase plans), even in the face of growing
evidence that the incentive plans were often not
implemented. Accordingly, the market-learn-
ing hypothesis (Hypothesis 3) is not supported.
SOCIALCONSTRUCTIONOFMARKETVALUE447
#1471-ASR 69:2 filename:69306-zajac
Table 2. Descriptive Statistics and Pearson Correlation Coefficients
Independent Variable Mean SD .(1) .(2) .(3) .(4) .(5) .(6) .(7) .(8)
0(1) Prior nonimplemented repurchase plans 59.20 63.22 . . . . . . . .
0(2) Prior implemented repurchase plans 210.7 135.7 .53 . . . . . . .
0(3) Prior nonimplemented LTIPs 98.47 45.13 .31 .24 . . . . . .
0(4) Prior implemented LTIPs 66.42 41.13 .20 .28 .41 . . . . .0(5) Cash flow per share 2.62 2.80 .03 .02 .07 .00 . . . .
0(6) Long-term debt/equity .73 1.21 .00 .03 .16 .06 .11 . . .
0(7) Outsider ratio .26 .16 .07 .02 .15 .00 .06 .11 . .
0(8) CEO/chair separation .23 .42 .19 .00 .01 .04 .04 .03 .06 .
0(9) Institutional ownership .31 .21 .07 .02 .03 .03 .00 .09 .04 .10
(10) Return on assets .06 .06 .17 .02 .03 .08 .09 .21 .02 .05
(11) Log of sales 7.92 1.25 .03 .07 .12 .04 .34 .01 .16 .05
(12) Other events .14 .35 .02 .01 .05 .01 .03 .08 .03 .21
(13) Prior adoptions:
(a) Nonimplemented repurchase program .12 .32 .21 .15 .14 .12 .03 .08 .03 .03
(b) Implemented repurchase program .30 .46 .13 .26 .10 .13 .03 .05 .05 .07
(c) Nonimplemented LTIP .14 .35 .16 .05 .30 .16 .01 .02 .02 .02(d) Implemented LTIP .09 .29 .08 .17 .17 .20 .01 .01 .00 .04
(14) Excess returns: 2 Days .71 .63 .36 .29 .18 .25 .23 .08 .07 .04
(15) Excess returns: 11 Days .84 .71 .31 .28 .13 .23 .19 .12 .09 .08
(16) Excess returns: 30 Days .91 .75 .28 .35 .14 .17 .20 .13 .09 .05
Independent Variable .(9) .(10) .(11) .(12) .(13a) .(13b) .(13c) .(13d).(14) .(15)
(10) Return on assets .03 . . . . . . . . .
(11) Log of sales .12 .01 . . . . . . . .
(12) Other events .01 .07 .05 . . . . . . .
(13) Prior adoptions:
(a) Nonimplemented repurchase program .05 .02 .02 .03 . . . . . .
(b) Implemented repurchase program .03 .05 .01 .06 .24 . . . . .(c) Nonimplemented LTIP .14 .03 .01 .00 .09 .04 . . . .
(d) Implemented LTIP .07 .01 .04 .01 .02 .05 .40 . . .
(14) Excess returns: 2 Days .10 .02 .02 .04 .02 .04 .03 .01 . .
(15) Excess returns: 11 Days .06 .00 .02 .05 .00 .03 .01 .05 .48 .
(16) Excess returns: 30 Days .10 .01 .01 .05 .01 .07 .02 .02 .31 .36
Note:N = 778
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
17/28
We also examined the robustness of these
findings in several ways. In supplementary
regression models, we used an alternative
approach to test Hypotheses 2 and 2a by exam-
ining the interaction between prior adoptions
and the average implementation of previously
adopted plans. These results were consistent
with the results in Table 3, showing that the
effect of prior adoptions is not contingent on
whether the plans were implemented (i.e., themarket learning argument is not supported).
Results of these analyses are provided in Table
1 of the Appendix. We also used a moving three-
year window to measure the number of prior
nonimplemented (and implemented) repurchase
plans, and we found results that essentially mir-
rored those reported in Table 3. Last, we rean-
alyzed our data using a sample that included all
repurchase plans announced during the excep-
tional two-week period following the market
crash of 1987 (these had been excluded to
ensure consistency with prior research, as noted
above); we found that the Table 3 results
remained robust.
It might be suggested that stock market reac-
tions to repurchase plans become more positive
over time despite a lower likelihood of imple-
mentation because (1) the efficiency benefits
from plans that were implemented increased
over time, or (2) the implementation of repur-
chase plans generally had a positive effect on
efficiency and investors learned about those
efficiency benefits over time. To examine these
possibilities, we conducted supplementaryanalyses of the performance effects of imple-
mented repurchase plans. We used data on stock
repurchases and performance for a random sam-
ple of 400 f irms from our sample frame (i.e.,
firms listed in the 1980 Forbes 500 or Fortune
500 indexes) from 1980 through 1994. We esti-
mated two measures of firm performance
return on assets and return on equityusing
cross-sectional time-series regression models
(Greene 1997). In the primary models we usedthe random-effects (GLS) estimator, but results
were robust to the fixed-effects estimator. As
shown in in Table 2 of the Appendix, the adop-
tion and implementation of repurchase plans
do not have a significant effect on either meas-
448AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
Table 3. Cochrane-Orcutt Regression Analyses of Excess Stock Returns for Firms Adopting Stock RepurchasePlans: 19801994
Independent Variable 2-Day Period 11-Day Period 30-Day Period
0(1) Prior nonimplemented repurchase Plans .0014*** (.0004) .0013*** (.0004) .0016*** (.0005)
0(2) Prior implemented repurchase Plans .0007*** (.0002) .0006*** (.0002) .0008*** (.0002)
0(3) Prior nonimplemented LTIPs .0016*** (.0005) .0015*** (.0006) .0019*** (.0007)0(4) Prior implemented LTIPs .0013*** (.0006) .0011*** (.0007) .0013*** (.0007)
0(5) Cash flow per share .020***0 (.007) .016***0 (.007) .022***0 (.008)
0(6) Long-term debt/equity .0004*** (.0002) .0004*** (.0002) .0004*** (.0002)
0(7) Outsider ratio .256***0 (.144) .270***0 (.162) .204***0 (.179)
0(8) CEO/chair separation .094***0 (.071) .176***0 (.079) .147***0 (.088)
0(9) Institutional ownership .261***0 (.101) .194***0 (.114) .290***0 (.126)
(10) Return on assets .0006*** (.0042) .0039*** (.0047) .0008*** (.0052)
(11) Log of sales .013***0 (.020) .015*** (.022) .008***0 (.025)
(12) Other events .154**0* (.094) .210***0 (.105) .215***0 (.117)
(13) Prior adoptions-focal firm:
(a) Nonimplemented repurchase plan .012***0 (.067) .001***0 (.075) .016***0 (.146)
(b) Implemented repurchase plan .058***0 (.051) .074***0 (.057) .095***0 (.064)(c) Nonimplemented LTIP .074***0 (.062) .016***0 (.131) .063***0 (.077)
(d) Implemented LTIP .002***0 (.001) .114***0 (.069) .088***0 (.146)
Constant .169***0 (.180) .360***0 (.201) .341***0 (.224)
Absorbed year effects (F) 3.83*** 3.90*** 3.68***
Model F 8.02*** 7.23*** 7.39***
R2 0.36*** 0.30*** 0.32***
Durbin-Watson statistic (D) 1.43*** 1.55*** 1.59***
Note: Coefficients shown with standard errors in parentheses. N = 778.
*p .05; **p .01; ***p .001 (one-tailed tests).
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
18/28
ure of firm profitability. The interaction ofrepurchase plan implementation and time ofadoption is also not significant: The effect ofrepurchase plans on profitability does notincrease or decrease over time. This result was
robust to the specification of alternative lagstructures: The table shows effects on prof-itability lagged by one to three years, and sim-ilar results were obtained with performancelagged by four or five years. These modelsincluded controls for industry effects as well asthe other control variables listed in the table.Thus, our findings regarding stock market reac-tions to repurchase plans are not an artifact ofthe efficiency benefits of implementing theplans or of the change in such benefits over
time.4
DISCUSSIONOverall, our results provide strong support forour sociological perspective on the market val-uation of corporate policies. The first set ofresults showed a dramatic change in stock mar-ket reactions to stock repurchase plan announce-ments over time, shifting markedly from anegative view in the late 1970s and early 1980sto a positive view from the mid-1980s onward.This finding supports our theoretical contentionthat, in the mid-1980s, as prevailing beliefsabout corporate governance shifted toward anagency conception of control (Davis andThompson 1994; Useem 1993, 1996; Zajac andWestphal 1995), repurchase plans were increas-ingly viewed positively as a means of prevent-ing managers from wasting free cash flow onempire-building projects. This resulted in more
positive market reactions to adopted plans. Incontrast, in earlier periods, when a corporatelogic prevailed, these repurchase plans were
seen as indications that managers lacked attrac-tive investment prospects, and this resulted innegative market reactions. Thus, our findingssuggest that changing institutional logics ofgovernance led to a shift over time in the finan-
cial markets assessment of stock repurchaseplans.
Of course, one might contend that the increas-ingly positive reaction to buyback announce-ments simply reflects the education of themarket as to the technical merits of buybacks forshareholder value. Therefore, it is particularlyrevealing that we find a growing positive reac-tion to repurchase plans despite a decreasingrate of implementation of adopted repurchase
plans over the same time period. In fact, our
multivariate analyses show that the number ofpreviously adoptedbut not implementedrepurchase plans is positively associated with themarket reaction to adoption at the focal firm.Prevailing assumptions in the financial eco-nomics literature are that market learning occursand that all publicly available information abouta corporate policy (including accumulated evi-dence about the efficiency benefits associatedwith adopting the policy) are reflected in the
stock markets response to adoption (David1997; Fama 1970; King et al. 1993;Timmermann 1993). While theorists haveexplored the implications of varying key param-eters of the market-learning process (e.g., thespeed with which investors respond to newinformation), the occurrence of learning itselfis generally taken-for-granted.
Our findings, however, show that as morefirms adopted but did not implement repur-chase plans over time (thus reducing the effi-
ciency benefits expected from adoption), themarket value of repurchase plans actuallyincreased. Additional analyses also ruled out thepossibility that our results might simply reflectincreasing efficiency benefits from implemen-tation over time or increasing investor learningabout efficiency benefits over time.
While our findings do not support a market-learning perspective, they are consistent with asociological perspective on stock market reac-
tions to corporate policy adoptions. From thisperspective, repurchase plans acquired greatersymbolic value over time, despite increaseddecoupling, through a process of institutional-ization. Institutionalization is more likely in thepresence of a prevailing belief system or insti-
SOCIALCONSTRUCTIONOFMARKETVALUE449
#1471-ASR 69:2 filename:69306-zajac
4 Additional analyses showed that repurchase planannouncements (including implemented and non-implemented plans) were also not associated with
subsequent performance, and this finding was sim-ilarly robust to different lag structures and both meas-ures of performance. Thus, our finding that the market
reacts more positively to repurchase plans even as therate of implementation decreases cannot be attributed
to any tendency for repurchase plan announcementsto signal that management is dedicated to increasingthe firms value to shareholders in some way (i.e.,
other than repurchase plan implementation).
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
19/28
tutional logic that suggests a sanguine inter-pretation of the policy (DiMaggio 1997; Scott1994). In this context, the advent of an agencylogic of governance provided a foundation forinstitutionalization by offering a positive inter-
pretation of repurchase plans. These findings areconsistent with our contention that as morefirms adopted repurchase plans over time, theplans became more strongly associated withthe agency values they affirm. In effect, repur-chase plans built symbolic value through recip-rocated interpretations, in which plans werepresented to constituents as furthering theagency logic, and that interpretation was vali-dated by constituents (Scott 1987:496). As thispattern was repeated, gradually the connection
between repurchase plans and agency valueswas strengthened and may have ultimatelybecome taken-for-granted as normatively appro-priate. This resulted in more positive investorreactions to plan adoptions over time, despite adecreasing rate of plan implementation (Scott1994; Selznick 1996; Zucker 1983).
We also argued theoretically, and foundempirically, that the earlier diffusion of execu-tive incentive plans (LTIPS), which had
advanced the agency logic of governance, con-tributed to this reinterpretation of stock repur-chase plans over time. This result is againconsistent with our institutional perspective,which suggests that the prior adoption of LTIPswith an agency explanation serves to reinforce,as well as reflect, the symbolic value of otherpolicies that appear to reduce agency costs,including repurchase plans (Meyer et al. 1987;Scott 1994). That is, it appears that the growingpredominance of an agency logic among mem-
bers of the financial community in the 1980s,which is reflected in (and reinforced by) theuse of agency explanations for executive incen-tive plans, conferred increased legitimacy onstock repurchase plans. It appears, in effect,that repurchase plans borrowed symbolic valuefrom the earlier diffusion of LTIPs.
More generally, our findings support ourinstitutional perspective on how the stock mar-ket values corporate policies and challenges
predominant perspectives rooted in financialeconomics. Whereas the dominant, financialeconomics perspective on capital markets con-ceives the markets reaction to a policy adoptionas a reliable, historically invariant indicator ofthe efficiency benefits from adoption, our the-
ory and findings demonstrate the influence on
market actors of historical change in prevailing
beliefs or institutional logics about the sources
of organizational efficiency. Moreover, assump-
tions of market learning in financial econom-
ics suggest that markets should discount thevalue of a policy given accumulated evidence
that efficiency benefits from the policy often are
not realized (i.e., due to nonimplementation), so
that the policy would effectively lose its legiti-
macy in the stock market. However, our theory
and consistent findings suggest how the market
value of a corporate policy can instead increase
as more firms adopt the policy over time, despite
a decreasing rate of implementation. Thus, our
results suggest how institutional processes thathave been shown to affect policy adoptions in
industrial markets can also influence policy
assessments in capital markets, a context that is
held up by financial economists as the closest
approximation of allocative efficiency. Our the-
ory and f indings ultimately suggest how stock
market reactions to corporate policiesand
thus the market value of those policiesare
socially constructed.
Our theoretical account of the social con-
struction of market value complements recentsociological research on the micro-social
dynamics of financial markets. This nascent
literature has provided evidence that in some
market contexts, trading and investment behav-
ior is governed by a social referencing process
in which market actors make investment deci-
sions by imitating or anticipating the decisions
of other market participants (i.e., rather than
independently forecasting assetseconomic per-
formance), such that the f inancial valuation ofeconomic assets may not reflect their true eco-
nomic value (Abolafia 1996; Westphal and
Zajac 1998; Zuckerman 1999). While this lit-
erature has yielded important insights, it begs
the question of how investors estimate the
beliefs and reactions of other market partici-
pants; that is, what are the inputs to the social
estimation process that drives stock market val-
uation? Our study addresses this question by
considering the macro-historical and institu-tional context in which the micro-social process-
es of stock market valuation occur. Specifically,
we suggest that investor perceptions of the value
of corporate policies are influenced by prevail-
ing institutional logics and prior market reac-
450AMERICANSOCIOLOGICALREVIEW
#1600-ASR 69:3 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
20/28
tions to the adoption of similar policies, and ourfindings support our theoretical claims.
Thus, our study advances the understandingof the social construction of financial marketsby considering the institutional determinants
of market value. Our theoretical perspectiveand findings suggest how the social referencingprocess that underlies stock market reactions topolicy adoptions can contribute to the institu-tionalization of corporate policies. We began bysuggesting that, given imperfect communicationamong market participants, investors are like-ly to reference prior market reactions to simi-lar events in estimating the reactions of otherparticipants to a focal policy adoption. As morefirms adopt a policy and receive a favorable
market response, each investors uncertaintyabout the likely response of other investors tothe current adoption decreases, which shouldtend to produce a more positive reaction. As aresult, the market value of corporate policies canincrease over time through a self-perpetuating,social construction process. As more firms adoptand receive positive reactions, the perceivedvalue of the policy becomes increasingly taken-for-granted by individual investors, and thus
the policy becomes institutionalized. Moreover,this social referencing process can contribute toinstitutionalization despite persistent or increas-ing rates of decoupling of formally adoptedpolicies from actual practices. Thus, our theo-ry and findings also advance neoinstitutionaltheory by suggesting how policies can becomeinstitutionalized, despite growing evidence ofnonimplementation, by virtue of the sociohis-torical estimation process that drives marketreactions. In this way, our perspective integrates
Meyer and Rowans (1977) decoupling thesiswith Zuckers (1983) thesis of institutionaliza-tion.
The theory and findings of our study also sug-gest new directions for research on corporategovernance. The governance literature has beendominated by financial economic perspectives,especially agency perspectives, and behavioralstudies have focused primarily on political andmicro-social factors that impede the imple-
mentation of agency prescriptions. Our studysuggests the need for further research that con-siders how governance practices, as well as thedeterminants and financial consequences ofthose practices, are conditioned by prevailinginstitutional logics and processes of institu-
tionalization. Future research could also exam-ine the key actors and actions involved in theestablishment of an institutitonal logic. Forexample, Zajac and Fiss (2003) perform a com-parative analysis of the adoption of a share-
holder value orientation among U.S. andGerman firms and describe how entire aca-demic and applied/practitioner journals (whosefounders were committed to the agency per-spective) helped push this change in the UnitedStates. They suggest that theJournal of AppliedCorporate Finance, founded in the mid-1980s,utilized the rapid rise of agency-based researchin corporate finance as an opportunity to trans-late agency theory into more accessible termsfor senior executives and policy makers. Much
of this research had appeared previously in theJournal of Financial Economics, a highlyregarded academic journal closely associatedwith agency theory (and with agency theorists).
Our study has implications for research onstock buybacks. Given that market reactions torepurchase plans are historically contingent,event studies should model the effects of timeand prior adoptions in estimating market reac-tions to adoption. Moreover, while some
researchers have suggested that investors reactpositively to repurchase plans in part becausethey interpret the plans as signals that man-agers believe their firm is undervalued, ourfinding that market reactions change from neg-ative to positive over time disconfirms theundervaluation interpretation, which cannotexplain negative reactions early in the time peri-od, nor a gradual increase over time.
Our study also addresses a centrally impor-tant question in the financial economics litera-
ture, albeit one that has curiously not attractedempirical attention: To what extent do financialmarkets learn? Perhaps strongly held beliefstructures in financial economics have ledresearchers to simply assume the answer is yes,but we suggest the answer is not obvious. Oneinterpretation of our findings is that, at thebroadest level, financial markets are teach-ablethey did reverse their earlier responsesto stock buybacks, viewing them much more
positively during the period in which a newdominant ideology (i.e., agency theory) tookroot. However, our central findings suggest thatfinancial markets only slowly incorporatedavailable information regarding the decouplingof stock buybacks and discounted the informa-
SOCIALCONSTRUCTIONOFMARKETVALUE451
#1471-ASR 69:2 filename:69306-zajac
-
8/13/2019 American Sociological Review 2004 Zajac 433 57
21/28
tional content of their announcements. Puttingthese two results together suggests that the effortexpended by financial economics researchers,business school professors, and the businessmedia to emphasize the rationality of stock buy-
backs may have been too successful, paradox-ically limiting the markets subsequentopportunity to learn about institutional decou-pling.
Future research could also examine the con-fluence of events or tipping points that lead oneinstitutional logic to begin to be supplanted byanother. In the context of corporate governance,for example, scandals involving Enron,Worldcom, and other major U.S. corporationshave led to numerous public and political dis-
cussions regarding the role of governance poli-cies such as the use of stock options for seniorexecutives. It is interesting that these discussionsreveal a reinterpretation of stock options.Specifically, while academic research and pub-lic discourse first viewed stock options throughthe lens of agency logic as a valued tool foraligning the incentives of top managers andowners, options are now viewed as making man-agers greedy and leading them to developschemes to inflate stock prices. The invest-ment community has begun to advocate exec-utive compensation vehicles, such as restrictedstock, that were previously criticized from anagency perspective as failing to adequately alignexecutive compensation with stock returns butthat are alleged to more effectively satisfy othercorporate goals, such as executive retention(Crystal 1991). The apparent shift in the debate
about executive incentive compensation (and
possibly future challenges to stock repurchase
plans) suggest that the predominance of agency
logic in the domain of corporate governance
may soon give way to a more pluralistic per-
spective on corporate governance in which mul-tiple institutional logics are represented.
Finally, while irrational stock market
behavior has increasingly been a focus of study
in the financial economics literature (under the
umbrella term of behavioral finance), we stress
that our sociological approach is quite different
from the typical psychological emphasis on
human decision-making limitations stemming
from individual decision-making biases.
Specifically, we seek to advance a logic of soci-
ological finance that emphasizes how macro-
level ideologies and institutionalization
processes affect financial markets and other
constituent groups in society. However, oppor-
tunities remain for a joint sociological and psy-
chological understanding of financial market
behavior. For example, the fact that psycholo-
gists have found that individuals prefer to seek
confirming, rather than disconfirming, evidence
when evaluating their cognitive schemas
(Nisbett and Ross, 1980) can add micro-leveldetail supporting the more sociological analy-
ses of market behavior we offer here. In con-
clusion, we hope our focus on the social
construction of financial markets contributes to
a more complete and accurate understanding of
the collective perceptions and behaviors of