See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/31462357

Transaction Cost Economics and Organization Theory

Article  in  Industrial and Corporate Change · January 1993

DOI: 10.1093/icc/2.1.107 · Source: OAI

CITATIONS READS

1,130 19,002

1 author:

Óliver Williamson
University of California, Berkeley
167 PUBLICATIONS   93,453 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Dolores and Oliver View project

Oliver Williamson View project

All content following this page was uploaded by Óliver Williamson on 25 November 2015.

The user has requested enhancement of the downloaded file.

 4 Transaction Cost Economics and
Organization Theory
Oliver E. Williamson

A NEW institutional economics has been taking in which law, economics, and organization are
pe over the past twenty-five years, of which j~ined.~
saction cost economics is a part. This new in- Joint ventures sometimes evolve into mergers
tional economics turns on two propositions: and sometimes unravel. I t is unlikely that either
(1) institutions matter, and (2) institutions are will happen here. That merger is not in prospect
susceptible t o analysis (Matthews 1986, p. 903). is because economics, organization theory, and
Tell economists something that they did not pre- law have separate as well as combined agendas. A
viously know about phenomena of interest to full-blown merger, moreover, would impoverish
them, display the logic, and demonstrate that the the evolving science of organization-which has
; data line up: that will get their attention. By con- benefited from the variety of insights that are re-
trast, mere critiques of orthodoxy and mere asser- vealed by the use of different lenses. Most likely
tions that institutions matter would assuredly the joint venture will hold until one of the parties
spelled the demise of the new institutional has learned enough from the others t o go it alone.
economics-which was the experience of its pred- Progress attended by controversy is to be ex-
pected for the remainder of the decade.
This chapter develops the argument that eco- This chapter focuses on the emerging relation-
d sociological approaches to economic ship between transaction cost economics and
on have reached a state of healthy ten- organization theory and argues that this relation-
also Sabel, chap. 6 in this Handbook). ship has three main aspects. First and most impor-
be contrasted with an earlier state of tant, transaction cost economics has been (and
hich the two approaches were largely will continue t o be) massively influenced by con-
disjunct, hence ignored one another, or described cepts and empirical regularities that have their or-
s research agendas and research ac- igins in organization theory. Second, there are
wedberg 1990, p. many transaction cost economics concepts t o
genuine give-and- which organization theorists can (and many do)
er the obsolescence of organization productively relate (see in this connection the
ich Charles Perrow has recently al- glossary of key terms in transaction cost econom-
. 162), nor the capitulation of eco- ics at the end of this chapter). But third, healthy
hich James March (tongue-in- tension survives-as revealed by an examination
of phenomena for which rival interpretations have
A more respectful relation, perhaps even a been advanced that remain unresolved and pro-
nization theory are voke controversy.
s evident in W. Rich- The chapter begins with some background o n
le important areas of institutional economics, both old and new. In
. disagreement remain, more consensus exists than subsequent sections, a three-level schema for
'., is at first apparent" (1992, p. 3), in game theorist studying economic organization is proposed;
t "almost any theory some of the more important ways in which trans-
ressed by game the- action cost economics has benefited from organi-
ory will do more for game theory than game the- zation theory are examined; the key concepts in
;?;.,ory will d o for it" (1992, p. l ) , and in my argu- transaction cost economics are sketched; empiri-
ment that a science of organization is in progress cal regularities, as discerned through the lens of
78 Williamson
transaction cost economics, that are pemnent to institutional change on the other, and (d) historical
organization theory are discussed; and finally, appreciation o f how customs, legal precedents, and
contested terrain between transaction cost eco- laws of a society evolve to construct a collective stan-
nomics and organization theory is surveyed. dard of prudent reasonable behavior for resolving
disputes between conflicting parties in pragmatic and
ethical ways.
Albeit in varying degree, transaction cost eco-
Older Traditions nomics is responsive to Commons in all four of
these r e ~ p e c t s . ~
Leading figures in the older institutional eco- Commons and his colleagues and students
nomics movement in the United States were were very influential in politics during and after
Wesley Mitchell, Thorstein Veblen, and John R the Great Depression-in shaping social security,
Commons. Although many sociologists appear to labor legislation, public utility regulation, and,
be sympathetic with the older tradition, there more generally, public policy toward business.
is growing agreement that the approach was Possibly because of its public policy successes, the
"largely descriptive and historically specific" Wisconsin School was remiss in developing its
(DiMaggio and Powell 1991, p. 2) and was not intellectual foundations. The successive opera-
cumulative (Granovetter 1988, p. 8). tionalization-from informal into preformal,
Criticisms of the old institutional economics by semiformal, and fully formal modes of analysis-
economists have been scathing. Thus George associated with transacti.on cost economics (Wil-
Stigler remarks that "the school failed in America liamson 1993a) never materialized. Instead, the
for a very simple reason. It had nothing in it ex- institutional economics of Commons progressed
cept a stance of hostility to the standard theoreti- very little beyond the informal stage.
cal tradition. There was no positive agenda of re- There is also an older institutional economics
search" (Stigler 1983, p. 170). Similar views are tradition in Europe. Of special importance was
expressed by R C. 0. Matthews (1986, p. 903). the German Historical School. (Interested read-
Ronald Coase concurs: the work of American in- ers are advised to consult Terrence Hutchison
stitutionalists "led to nothing. . . . Without a the- [1984] and Richard Swedberg [1991] for assess-
ory, they had nothing to pass on except a mass of ments.) And, of course, there were the great
descriptive material waiting for a theory or a fire. works of Karl Marx.
So if modern institutionalists have antecedents, it A later German school, the Ordoliberal or
is not what went immediately before" (Coase Freiburg School, also warrants remark. As dis-
1984, p. 230). cussed by Heinz Grossekettler (1989), this school
The general accuracy of these assessments not- was inspired by the work of Walter Eucken, whose
withstanding, an exception should be made for student Ludwig Erhard was the German minister
John R Commons. Not only is the institutional of economics from 1949 to 1963 and chancellor
economics tradition at Wisconsin still very much from' 1963 to 1966, and is widely credited with
alive (Brornley 1989), but the enormous public being the political father of the "economic mira-
policy influence of Commons and his students cle" in West Germany. Grossekettler describes
and colleagues deserves to be credited. Andrew numerous parallels between the Ordoliberal pro-
Van de Ven's summary of Commons's intellectual gram and those of property rights theory, transac-
contributions is pertinent to the first of these tion cost economics, and especially constitutional
points: economics (ibid., pp. 39, 64-67).
The Ordoliberal program proceeded at a very
Especially worthy of emphasis [about Commons] are
high level of generality (ibid., p. 47) and featured
his (a) dynamic views of institutions as a response to
scarcity and conflicts o f interest, (b) original formu-
the application of lawful principles to the entire
lation o f the transaction as the basic unit o f analysis, economy (ibid., pp. 46-57). Its great impact on
(c) part-whole analysis of how collective action con- postwar German economic policy nonvithstand-
strains, liberates, and expands individual action in ing, the influence of the school declined after
countless numbers o f routine and complementary the mid-1960s. Although Grossekettler attributes
transactions on the one hand, and how individual the decline to the "wide scale of acceptance of the
wills and power to gain control over limiting or con- Keynesian theory . . . [among] young German
tested factors provide the generative mechanisms for intellectuals" (ibid., pp. 69-70), an additional
 Transaction Cost Economics 79
problem is that the principles of Ordoliberal eco- proach appear to be rival theories of organization,
nomics were never given operational content. much of that tension is relieved by recognizing
Specific models were never developed; key trade- that the New Institutional Economics has actually
of's were never identified; the mechanisms re- developed in two complementary parts. One of
mained very abstract. The parallels with the these parts deals predominantly with background
Wisconsin School-great public policy impact, conditions (expanded beyond property rights to
underdeveloped conceptual framework, loss of include contract laws, norms, customs, conven-
intellectual influence-are suiking. tions, and the like) while the second branch deals
with the mechanisms of governance. The two-
The New Institutional Economics part definition proposed by Lance Davis and
Douglass North (1971, pp. 5-6; emphasis added)
The New Institutional Economics comes in a is pertinent:
variety of flavors and has been variously defined
(compare Hodgson, chap. 3 in this Handb~ok).~ The innitutional environment is the set o f funda-
mental political, social and legal ground rules that es-
The economics of property rights-as developed
tablishes the basis for production, exchange and dis-
especially by Coase (1959; 1960), Armen Alchian
tribution. Rules governing elections, property rights,
(1961), and Harold Demsetz (1967)-was an
and the right o f contract are examples. . . .
early and influential dissent from orthodoxy. An
An institutional arrangement is an arrangement
evolutionary, as opposed to a technological, ap-
between economic units that governs the ways in
proach to economic organization was advanced,
which these units can cooperate and/or compete. It
according to which new property rights were cre-
ated and enforced as the economic needs arose, if
. . . [can] provide a structure within which its mem-
bers can cooperate . . . or [it can] provide a mecha-
and as these were cost-effective.
nism that can effect a change in laws or property
The definition of ownership rights advanced by
rights.
Eirik Furuboul and Svetozq Pejovich is broadly
pertinent: "By general agreement, the right of Interestingly, these two parts correspond very
ownership of an asset consists of three elements: closely with the much earlier division of effort be-
(a) the right to use the asset. . . ,(b) the right to tween "economic sociology" and "economic the-
appropriate the returns from the asset. . . ,and (c) ory" described by Joseph Schumpeter-where
the right to change the asset's form and/or sub- economic sociology was expected to study the in-
stance" (1974, p. 4). Strong claims on behalf of stitutional environment and economic theory was
the property rights approach to economic organi- concerned principally with the mechanisms of
zation were set out by Coase as follows (1959, p. governance (Schumpeter [1951] 1989, p. 293).
14): "A private enterprise system cannot hnction As it turns out, a large number of economists
unless property rights are created in resources, have productively worked on issues relating to the
and when this is done, someone wishing to use a institutional environment. These include a prodi-
resource has to pay the owner to obtain it. Chaos gious amount of research by North, who defines
disappears; and so does the government except institutions as "the humanly devised constraints
that a legal system to define property rights and to that structure political, economic, and social in-
arbitrate disputes is, of course, necessary." As it teractions. They consist of both informal con-
turns out, these claims overstate the case for the straints (sanctions, taboos, customs, traditions,
property rights approach. Not only is the defini- and codes of conduct), and formal rules (consti-
tion of property rights sometimes costly-con- tutions, laws, property rights)" (1991, p. 97).
sider the difficult problems of defining intellectual Elsewhere he argues that "institutions consist of a
property rights-but court ordering can be a costly set of constraints on behavior in the form of rules
way to proceed. A comparative contractual ap- and regulations; a set of procedures to detect de-
proach-according to which court ordering is often viations from the rules and regulations; and, fi-
(but selectively) supplanted by private ordering nally, a set of moral, ethical behavioral norms
for purposes of governing contractual relations which define the contours and that constrain the
(Macneil1974; 1'978; Williamson 1979; 1991a)- way in which the rules and regulations are speci-
rather than a pure property rights approach, fied and enforcement is carried out" (North
therefore has a great deal to recommend it. 1986, p. 233). These definitions, as well as related
Although the earlier property rights approach definitions by Allan Schmid (1972, p. 893),
and the more recent comparative contractual ap- Daniel Bromley (1989, p. 41), Andrew Schotter
80 Williamson
(1981, p. 9), and Eirik Furubotn and Rudolf individual and argue that the three levels of analy-
Richter (1991, p. 3), are akin t o the definition of sis are "nested, where organization and institu-
institutions offered by the sociologists Talcott tion specify progressively higher levels of con-
Parsons and Neil Smelser, who observe that insti- straint and opportunity for individual action"
tutions are "ways in which the value patterns of (1991, p. 242).
the common culture of a social system are inte- The causal model proposed here is akin to and
grated in the concrete action of its units in their was suggested by, but is different from, the causal
interaction with each other through the defini- model recently proposed by W. Richard Scott
tion of role expectations and the organization of (1992, p. 45), who is also predominantly con-
motivation" (1956, p. 102). cerned with governance. There are three main ef-
This emphasis on property rights, customs, fects in the present schema (see fig. 1). These are
norms, conventions, and the like is especially per- shown by the solid arrows. Secondary effects are
tinent for purposes of doing intertemporal, inter- drawn as dashed arrows. As indicated, the institu-
national, or cross-cultural comparisons. What the tional environment defines the rules of the game.
economics of organization is predominantly con- If changes in property rights, contract laws,
cerned with, however, is this: holding these back- norms, customs, and the like induce changes in
ground conditions constant, why organize eco- the comparative costs of governance, then a re-
nomic activity one way (e.g., procure from the configuration of economic organization is usually
market) rather than another (e.g., produce t o implied.
your own needs: hierarchy)? That is the Coasian
question (Coase [1937] 1988). I t is the focus of
transaction cost economics, and explains much of
the interest of organization theorists and the soci-
ology of organization in the new institutional ec-
onomics. Not only does the study of governance
raise different issues, but much of the predictive s
Institutional
Environment

h T
t
I Strategic
------ - 1
I
I

content and most of the empirical research in in- Parameters

stitutional economics has been at the governance
level (Matthews 1986, p. 907).
1 Governance I I
I
I
I
I
I
I
Transaction cost economics is mainly con- Behavioral -Endogenous I
cerned with the governance of contractual rela-
Attributes Preferences + II
tions. Governance does not, however, operate in
isolation. The comparative efficacy of alternative
modes of governance varies with the institutional
environment on the one hand and the attributes
1 Individual ------A I

of economic actors on the other. A three-level

schema is therefore proposed, according to which FIGURE1. A Layer Schema
the object of analysis, governance, is bracketed by
more macro features (the institutional environ- The solid arrow from the individual to govern-
ment) and more micro features (the individual). ance carries the behavioral assumptions within
Feedbacks aside (which are underdeveloped in which transaction cost economics operates, and
the transaction cost economics setup), the institu- the circular arrow within the governance sector
tional environment is treated as the locus of shift reflects the proposition' that organization, like the
parameters, changes in which shift the compara- law, has a life of its own. This proposition is dis-
tive costs of governance, and the individual is cussed in the next section.
where the behavioral assumptions originate. Although behavioral assumptions are fre-
Roger Friedland and Robert Alford also pro- quently scanted in economics, transaction cost
pose a three-level schema in which environment, economics subscribes t o the proposition that eco-
governance, and individual are distinguished, but nomic actors should be described in workably re-
their emphasis is very different. They focus on the alistic terms (Simon 1978; Coase 1984). Interest-
 Transaction Cost Economics 81

ingly, "outsiders," especially physicists, have long tunistic, the Prince was advised to engage in re-
been insistent that a better understanding of the ciprocal and even preemptive opportunism-to
actions of human agents requires more self-con- breach contracts with impugnity whenever "the
scious attention to the study of how people's reasons which made him bind himself no longer
minds work (Bridgeman 1955, p. 450; Waldrop exist" (Machiavelli [1513] 1952, p. 92). The
1992, p. 142). Herbert Simon concurs: subtle form is strategic and has been described
Nothing is more hndamental in setting our research elsewhere as "self-interest seeking with guile"
agenda and informing our research methods than (Williamson 1975, pp. 26-37; 1985, pp. 46-52,
our view of the nature of the human beings whose 64-67). The natural form involves tilting the sys-
behavior we are studying. . . . It makes a difference tem at the margin. The so-called "dollar-a-year"
to research, but it also makes a difference for the men in the Office of Production Management, of
proper design of political institutions. James Madi- which there were 250 at the beginning of World
son was well aware of that, and in the pages of the War 11, were of concern to the Senate Special
Federalist Papers he opted for this view of the human Committee to Investigate the National Defense
condition (Federalist, No. 55): Program because "such corporate executives in
As there is a degree of depravity in mankind high official roles were too inclined to make deci-
which requires a certain degree of circumspection sions for the benefit of their corporations. 'They
and distrust, so there are other qualities in human have their own business at heart,' [Senator] Tru-
nature which justify a certain portion of esteem man remarked. The report called them lobbyists
'in a very real sense,' because their presence inevi-
and confidence.
tably meant favoritism, 'human nature being what
-a balanced and realistic view, we may concede, of
it is' " (McCullough 1992, p. 265). Michel Cro-
bounded human rationality and its accompanying
zier's treatment of bureaucracy makes prominent
frailties of motive and reason. (1985, p. 303)
provision for all forms of opportunism, which he
Transaction cost economics expressly adopts the describes as "the active tendency of the human
proposition that human cognition is subject to agent to take advantage, in any circumstances, of
bounded rationality-where this is defined as be- all available means to further his own privileges"
havior that is "intendedly rational, but only limit- (Crozier 1964, p. 194).
edly so" (Simon 1957a, p.'xxiv)-but differs from Feedback effects from governance to the insti-
Simon in its interpretation of the "degree of de- tutional environment can be either instrumental
pravity" to which Madison refers. or strategic. An example of the former would be
Whereas Simon regards the depravity in ques- an improvement in contract law, brought about
tion as "frailties of motive and reason," transac- at the request of parties who find that extant law
tion cost economics describes it instead as oppor- is poorly suited to support the integrity of con-
tunism-to include the use of guile in pursuit of tract. Strategic changes could take the form of
one's own interests. The former is a much more protectionist trade barriers against domestic and/
benign interpretation, and many social scientists or foreign competition. Feedback from govern-
understandably prefer it. Consider, however, ance to the level of the individual can be inter-
Robert Michels's concluding remarks about oli- preted as "endogenous preference" formation
garchy: "Nothing but a serene and frank examina- (Bowles and Gintis 1993), due to advertising or
tion of the oligarchical dangers of democracy will other forms of "education." The individual is also
enable us to minimize these dangers" ([1911] influenced by the environment, in that en-
1962, p. 370). If a serene and frank reference to dogenous preferences are the product of social
opportunism alerts us to avoidable dangers which conditioning. Although transaction cost econom-
the more benign reference to frailties of motive ics can often relate to these secondary effects,
and reason would not, then there are real hazards other modes of analysis are often more pertinent.
in adopting the more benevolent construction. More generally, the Friedland and Alford
As discussed below, the mitigation of opportun- scheme, the Scott scheme, and the variant that I
ism plays a central role in transaction cost eco- offer are not mutually exclusive. Which to use
when depends on the questions being asked. To
pportunism can take blatant, subtle, and nat- repeat, the main case approach to economic or-
forms. The blatant form is associated with ganization that I have proposed works out of the
achiavelli. Because he perceived that the agents heavy line causal relations shown in figure 1, to
th whom the Prince was dealing were oppor- which the dashed lines represent refinements.
82 Williamson
and the ramifications can be folded back into the
organizational design. Unwanted costs will then
be mitigated and unanticipated benefits will be
enhanced. Better economic performance will or-
Richard Swedberg (1987; 1990), Robert Frank dinarily result.
(1992), and others have described numerous re- Unintended effects are frequently delayed and
spects in which economics has been influenced by are often subtle. Deep knowledge of the details
sociology and organization theory. The value and intertemporal process transformations that
added which is referred to here deals only with attend organization is therefore needed. Because
those aspects where transaction cost economics organization theorists have wider and deeper
has been a direct and significant beneficiary. knowledge of these conditions, economists have
The behavioral assumptions discussed above- much to learn and ought to be deferential. Four
bounded rationality and opportunism-are per- specific illustrations are sketched here.
haps the most obvious examples of how trans- Demands for control. A natural response to
action cost economics has been shaped by perceived failures of performance is to introduce
organization theory. But the proposition that or- added controls. Such efforts can have both in-
ganization has a life of its own (the circular arrow tended and unintended consequences (Merton
in the governance box in figure 1)is also impor- 1936; Gouldner 1954).
tant. And' there are yet additional influences as One illustration is the employment relation,
well. where an increased emphasis on the reliability of
behavior gives rise to added rules (March and
Intertemporal Process Transformations Simon 1958, pp. 3 8 4 0 ) . Rules, however, serve
not merely as controls but also define minimally
Describing the firm as a production function acceptable behavior (Cyert and March 1963).
invites an engineering approach to organization. Managers who apply rules to subordinates in a le-
The resulting "machine model" of organization galistic and mechanical way invite "working to
emphasizes intended effects to the neglect of un- rules," which frustrates effective performance.
intended effects (March and Simon 1958, chap. These unintended consequences are picked up
3). But if organizations have a life of their own, by the wider peripheral vision of organization
and if the usual economic approach is unable to theorists. In the spirit of farsighted contracting,
relate to the intertemporal realities of organiza- however, the argument can be taken yet a step
tion, then-for some purposes at least-an extra- further. Once apprised of the added conse-
economic approach may be needed. quences, the farsighted economist will make al-
Note that this does not amount to a proposal lowance for them by factoring these into the orig-
that the economic approach be abandoned. inal organizational design. (Some organization
Rather, the "usualn or orthodox economic ap- theorists might respond that this last is fanciful
proach gives way to an augmented or extended and unrealistic. That can be decided by examining
economic approach. That is very different from the data.)
adopting an altogether different approach-as, Oligarchy. The "iron law of oligarchy" holds
for example, that of neural networks. that "it is organization which gives birth to the
As it turns out, the economic approach is both dominion of the elected over the electors, of the
very elastic and very powerful. Because it is elastic mandatories over the mandators, of the delegates
and because increasing numbers of economists over the delegators. Who says organization, says
have become persuaded of the need t o deal with oligarchy" (Michels [1911] 1962, p. 365). AC-
economic organization "as it is," warts and all, all cordingly, good intentions notwithstanding, the
significant regularities whatsoever-intended and initial leadership (or its successors) will inevitably
unintended alike-come within the ambit. Be- develop attachments for the office. Being strategi-
cause it is very powerful, economics brings added cally situated, the leadership will predictably en-
value. Specifically, the "farsighted propensity" or trench itself-by controlling information, manip-
"rational spirit" that economics ascribes to eco- ulating rewards and punishments, and mobilizing
nomic actors permits the analysis of previously ne- resources to defeat rivals. Even worse, the en-
glected regularities to be taken a step h t h e r . trenched leadership will use the organization to
Once the unanticipated consequences are under- promote its own agenda at the expense of the
stood, those effects will thereafter be anticipated membership.
 Transaction Cost Economics 83

One response would be to eschew organization well-considered theory of organization will make
in favor of anarchy, but that is extreme. The bet- provision for failures of all kinds.
ter and deeper lesson is to take all predictable reg- Albeit underdeveloped, the bureaucratic failure
ularities into account at the outset, whereupon it literature is vast, partly because purported failures
may be possible t o mitigate foreseeable oligarchi- are described in absolute rather than comparative
cal excesses at the initial design stage.6 terms. Unless, however, a superior and feasible
Identity/capability. The proposition that iden- form of organization to which to assign a trans-
tity matters has been featured in transaction cost action (or related set of transactions) can be
economics from the outset. As developed below, identified, the failure in question is effectively ir-
identity is usually explained by some form of remediable. One of the tasks of transaction cost
"asset specificity." The "capabilities" view of the economics is t o assess purported bureaucratic fail-
firm (Selznick 1957; Penrose 1959; Wernerfelt ures in comparative institutional terms.
1984; Teece, Pisano, and Shuen 1990) raises re- The basic argument is this: it is easy to show
lated but additional issues. that a particular hierarchical structure is beset
One way t o unpack the "capabilities" view of with costs, but that is inconsequential if all feasi-
the firm is to ask what-in addition t o an inven- ble forms of organization are beset with the same
tory of its physical assets, an accounting for its fi- or equivalent costs. Efforts to ascertain bureau-
nancial assets, and a census of its workforce-is cratic costs that survive comparative institutional
needed to describe the capabilities of a firm. Fea- scrutiny are reported elsewhere (Williamson
tures of organization that are arguably important 1975, chap. 7; 1985, chap. 6), but these are very
include the following: (1) the communication provisional and preliminary. Although intertem-
codes that the firm has developed (Arrow 1974); poral transformations and complexity are recur-
(2) the routines that it employs (Cyert and March rent themes in the study of bureaucratic failure,
1963; Nelson and Winter 1982); and ( 3 ) the cor- much more concerted attention to these matters
porate culture that has taken shape (Kreps is needed.
1990b). What do we make of these?
One response is to regard these as spontaneous Adaptation
features of economic organization. As interpreted
by institutional theory in sociology, "organiza- The economist Friedrich Hayek maintained
tional structures, procedures, and decisions are that the main problem of economic organization
largely ritualistic a n d symbolic, especially so when was that of adaptation and argued that this was
it is difficult or impossible to assess the efficacy of realized spontaneously through the price system.
organizational decisions on the basis of their tan- Changes in the demand or supply of a commodity
gible outcomes" (Baron and Hannan 1992, p. give rise to price changes, whereupon "individual
57; emphasis added). participants . . . [are] able to take the right ac-
If, of course, efficiency consequences are im- tion" (Hayek 1945, p. 527; emphasis added).
possible to ascertain, then intentionality has noth- Such price-induced adaptations by individual ac-
ing to add. Increasingly, however, some of the tors will be referred to as autonomous adapta-
subtle efficiency consequences of organization are tions.
coming to be better understood, whereupon they The organization theorist Chester Barnard also
are (at least partly) subject to strategic determina- held that adaptation was the central problem of
tion. If the benefits of capabilities vary with the organization. But whereas Hayek emphasized au-
attributes of transactions, which arguably they tonomous adaptation of a spontaneous kind, Bar-
do, then the cost effective thing to do. i u nard was concerned with cooperative adaptation
shape c i i T t u r ~ d i f ~ c Z m m u n i c a f i ocodes,
n v and of an intentional kind. Formal organization, espe-
m b a e . E u ~ n _ _ e _ S ~ ~ - ~ - 4 ~ ~ b , e _ cially ~ ~ a hierarchy,
rcCtl~p was the instrument through
s p ~ ~ w i ! YImplementing
i the intentionality which the "conscious, deliberate, p ~ r p o s e h l co-
'~
view will require that the microanalytic attributes operation to which Barnard called attention was
that define culture, communication codes, and accomplished (Barnard 1938, p. 4). Barnard's in-
routines be uncovered, which is an ambitious sights, which have had a lasting effect on organi-
exercise. zation theory, should have a lasting effect on eco-
Bureaucratization. As compared with the nomics as well.
study of market failure, the study of bureaucratic Transaction cost economics (1) concurs that
failure is underd&eloped. It is elementary that a adaptation is the central problem of economic
84 Williamson
organization, (2) regards adaptations of both au- and the "purchase" of votes is much more cum-
tonomous and cooperative kinds as important, bersome. Because, moreover, the gains that result
( 3 ) maintains that whether adaptations to distur- from improved efficiency accrue (in the first in-
bances ought to be predominantly autonomous, stance, at least) to private sector owners in
cooperative, or a mixture thereof varies with the portion to their ownership, private incentives to
attributes of the transactions (especially on the concentrate ownership and remove inefficiency
degree to which the investments associated with are greater.
successive stages of activity are bilaterally or mul- Even setting voting considerations aside, how- -
tilaterally dependent), and (4) argues that each ever, there is another factor that induces politi-
generic form of governance-market, hybrid, and cians to design agencies inefficiently. Incumbent
hierarchy--differs systematically in its capacity to politicians who create and design bureaus are
adapt in autonomous and cooperative ways. A se- aware that the opposition can be expected to win
ries of predicted (transaction cost economizing) a majority and take control in the future. Agencies
alignments between transactions and governance will therefore be designed with reference to both
structures thereby obtain (Williamson 1991a)- immediate benefits (which favors responsive
which predictions invite and have been subjected mechanisms) and possible future losses (which
to empirical testing (Joskow 1988; Shelanski often favors crafting inertia into the system). A
1991; Masten 1992). farsighted majority party will therefore design
some degree of (apparent) inefficiency into the
Politics agency at the outset-the effect of which will be
to frustrate the efforts of successor administra-
Terry Moe (1990) makes a compelling case for tions to reshape the purposes served by an
the proposition that public bureaucracies are dif- agency.
ferent. Partly that is because the transactions that
are assigned to the public sector are different, but
Embeddedness and Networks
Moe argues additionally that public sector bu-
reaucracies are shaped by politics. Democratic poli- Gary Hamilton and Nicole Biggart (1988) take
tics requires compromises that are different in exception to the transaction cost economics inter-
kind from those posed in the private sector and pretation of economic organization because it im-
presents novel expropriation hazards. Added "in- plicitly assumes that the institutional environ-
efficiencies" arise in the design of public agencies ment is everywhere the same-namely, that of "

on both accounts. Western democracies, and most especially that of

The inefficiencies that result from compromise the United States. They observe that large firms in
are illustrated by the design of the Occupational East Asia differ from U.S. corporations in sig-
Safety and Health Administration (OSHA). "If nificant respects and explain that "organizational
business firms were allowed to help design practices . . . are fashioned out of preexisting i
OSHA, they would structure it in a way that it teractional patterns, which in many cases date to
could not do its job. They would try to cripple it. preindustrial times. Hence, industrial enterprise is
This is not a hypothetical case. Interest groups a complex modern adaptation of preexisting pat-
representing business actually did participate in terns of domination to economic situations in
the design of OSHA, . . . [and] OSHA is an ad- which profit, efficiency, and control usually form
ministrative nightmare, in large measure because the very conditions of existence" (Hamilton and
some of its influential designers fully intended t o Biggart 1988, p. S54).
endow it with structures that would not work" The evidence that East Asian corporations dif-
(Moe 1990, p. 126). fer is compelling. The argument, however, that
To be sure, private sector organization is also transaction cost economics does not have applica-
the product of compromise. Egregious ineffi- tion to East Asian economies goes too far.
ciency in the private sector is checked, however, The correct argument is that the institutional
by competition in both product and capital mar- environment matters and that transaction cost ec-
kets. Note with reference to the latter that the onomics, in its preoccupation with governance,
voting rules in the private and public sectors are has been neglectful of this. Treating the institu-
very different. The private rule is one share-one tional environment as a set of shift parameters-
vote, and shares may be concentrated through changes in which induce shifts in the comparative
purchase. The public rule is one person-one vote, costs of governance-is, to a first approximation
 Transaction Cost Economics 85

at least, the obvious response (Williamson closely to the transaction cost economics treat-
1991a). That is the interpretation advanced above ment of the hybrid form of economic organiza-
and shown in figure 1 . tion (Williamson 1983, 1991a). For another, as
The objection could nevertheless be made that the discussion of Japanese economic organization
this is fine as far as it goes, but that comparative below reveals, transaction cost economics can be
statics-which is a once-for-all exercise-does not and has been extended t o deal with a richer set of
go far enough. As Mark Granovetter observes, network effects.
"more sophisticated . . . analyses of cultural in-
fluences . . . make it clear that culture is not a
once-for-all influence but an ongoing process, con- Discrete Structural Analysis
tinuously constructed and reconstructed during
interaction. It not only shapes its members but is Capitalism and socialism can be compared in
also shaped by them, in part for their own strate- both discrete structural (bureaucratization) and
gic reasons" (1985, p. 486; emphasis in original). marginal analysis (efficient resource allocation)
This objection is not without merit, but it respects. Interestingly, Oskar Lange (19 38, p.
should be pointed out that "more sophisticated 109) conjectured that, as between the two, bu-
analyses" must be judged by their value added. reaucratization posed a much more severe danger
What are the deeper insights? What are the added to socialism than did inefficient resource allo-
implications? Are the effects in question really be- cation.
~ o n the
d reach of economizing reasoning? That he was sanguine with respect to the latter
Consider, with reference to this last, the em- was because he had derived the rules for efficient
beddedness argument that "concrete relations resource allocation (mainly of a marginal cost
and structures" generate trust and discourage pricing kind) and was confident that socialist
malfeasance of noneconomic or extraeconomic planners and managers could implement them.
kinds: Joseph Schumpeter (1942) and Abram Bergson
(1948) concurred. The study of comparative eco-
Better than a statement that someone is known to be nomic systems over the next fifty years was pre-
reliable is information from a trusted informant that dominantly an allocative efficiency exercise.
he has dealt with that individual and found him so. Bureaucracy, by contrast, was mainly ignored.
Even better is information from one's own past deal- Partly that is because the study of bureaucracy was
ings with that person. This is better information for believed to be beyond the purview of economics
four reasons: (1) it is cheap; (2) one trusts one's own and belonged to sociology (Lange 1938, p. 109).
information best-it is richer, more detailed, and Also, Lange held that "monopolistic capitalism"
known to be accurate; (3) individuals with whom was beset by even more serious bureaucracy prob-
one has a continuing relation have an economic mo- lems (ibid., p. 110). If, however, the recent col-
tivation to be trustworthy, so as not to discourage lapse of the Soviet Union is attributable more to
future transactions; and (4) departing from pure eco- conditions of waste (operating inside the frontier)
nomic motives, continuing economic relations often than to inefficient resource allocation (operating
become overlaid with social content that carries at the wrong place on the frontier), then it was
strong expectations of trust and abstention from op- cumulative burdens of bureaucracy-goal distor-
portunism. (Granovetter 1985, p. 490) tions, slack, maladaptation, technological stagna-
tion-that spelled its demise.
This last point aside, the entire argument is con-. The lesson here is this: always study first-order
sistent with, and much of it has been anticipated (discrete structural) effects before examining sec-
by, transaction cost reasoning. Transaction cost ond-order (marginalist) refinements. Arguably,
economics and embeddedness reasoning are evi- moreover, that should be obvious: waste is easily
dently complementary in many respects. a more serious source of welfare losses than are
A related argument is that transaction cost eco- price-induced distortions (compare Harberger
nomics is preoccupied with dyadic relations, so [1954] with Williamson [1968]).
that network relations are given short shrift. The Simon advises similarly. Thus he contends that
premise is correct,* but the suggestion that net- the main questions are
work analysis is beyond the reach of transaction
cost economics is too strong. For one thing, Not "how much flood insurance will a man buy?"
many of the network effects described by Ray but "what are the structural conditions that make
Miles and Charles Snow (1992) correspond very buying insurance rational or attractive?"
 86 Williamson
Not "at what levels will wages be fixed" but "when ences that distinguish alternative modes of gov-
will work be performed under an employment con- ernance explains earlier claims that firms and
tract rather than a sales contract?" (1978, p. 6) markets are indistinguishable in fiat and control
Friedland and Alford's recent treatment of the respects.
logic of institutions is also of a discrete structural
kind (1991, p. 248). Transaction cost economics
proceeds similarly, but whereas Friedland and Al- TRANSACTION
COSTECONOMICS,
ford are concerned with discrete structural logics THESTRATEGY
between institutional orders-apitalism, the
state, democracy, the family, etc.-transaction The transaction cost economics program for
cost economics maintains that distinctive logics studying economic organization has been de-
within institutional orders also need t o be distin- scribed elsewhere (Williamson 1975,198 1,1985,
guished. Within the institutional order of capital- 1988a, 1991a; Klein, Crawford, and Alchian
ism, for example, each generic mode of govern- 1978; Alchian and Woodward 1987; Davis and
ance-market, hybrid, and hierarchy-possesses Powell 1992). The aim here is t o sketch the gen-
its own logic and distinctive cluster of attributes. eral strategy that is employed by transaction cost
, Of special importance is the proposition that each
economics, with the suggestion that organization
generic mode of governance is supported by a dis- theorists could adopt (some already have
tinctive form of c adopted) parts of it.
As developedW k-welsew ere (Williamson 1991a),
The five-part strategy proposed here entails (1)
transaction cost economics holds that classical a main case orientation (transaction cost econo-
contract law applies to markets, neoclassical con- mizing), (2) choice and explication of the unit of
tract law applies t o hybrids, and forbearance law is analysis, (3) a systems view of contracting, (4) ru-
the contract law of hierarchy. As between these dimentary trade-off apparatus, and (5) a remedia-
three concepts of contract, classical contract law is bleness test for assessing "failures."
the most legalistic, neoclassical contract law is
somewhat more elastic (Macneil 1974; 1978), The Main Case
and forbearance law has the property that hierar-
Economic organization being very complex
chy is its own court of ultimate appeal. But for
and our understanding being primitive, there is a
these contract law differences, markets and hierar-
need t o sort the wheat from the chaff. I propose
chies would be indistinguishable in fiat respects.
for this purpose that each rival theory of organiza-
Recall in this connection that Alchian and
tion should declare the main case out of which it
Demsetz introduced their analysis of the "classical
works and develop the refutable implications that
capitalist firm" with the argument that "It is com-
accrue thereto.
mon t o see the firm characterized by the power to
Transaction cost economics holds that econo-
settle issues by fiat. . . . This is delusion. The firm
mizing on transaction costs is mainly responsible
. . . has no power of fiat, no authority, no discipli- for the choice of one form of capitalist organiza-
nary action any different in the slightest degree
tion over another. It thereupon applies this hy-
from ordinary market contracting" (1972, p.
pothesis to a wide range of phenomena-vertical
777). That is a provocative formulation and integration, vertical market restrictions, labor or-
places the burden on those who hold that firm ganization, corporate governance, finance, regu-
and market differ in fiat respects t o show wherein lation (and deregulation), conglomerate organi-
those differences originate.
- zation, technology transfer, and, more generally,
The transaction cost economics response is t o any issue that can be posed directly or indi-
that courts treat interfirm and intrafirm disputes rectly as a contracting problem. As it turns out,
differently-serving as the forum of ultimate ap- large numbers of problems that on first examina-
peal for interfirm disputes while refusing t o hear tion d o not appear t o be of a contracting kind
identical technical disputes that arise between di- turn out t o have an underlying contracting struc-
visions (regarding transfer prices, delays, quality, ture-the oligopoly problem (Williamson 1975,
3 and the like). Because&crg,rchy-&i,t~ y n court of chap. 12) and the organization of the company
ultimate appeal (Williamson 1991a), fir@&can, town (Williamson 1985, pp. 35-38) being exam-
------ -
fiat that markets
.--- cannot. Prior ne- ples. 'comparisons with other-rival or comple-
mentary-main case alternatives are invited.
glect of the discrete structural contract law differ-
 Transaction Cost Economics 87 j
Three of the older main case alternatives are on out there?-rather well, but public policy needs
that economic organization is mainly explained to be more circumspect. As discussed below, the
by (1) technology, (2) monopolization, and (3) relevant test of whether public policy intervention
efficient risk bearing. More recent main case can- is warranted is that of remediableness.
didates are (4) contested exchange between labor These important qualifications notwithstand-
and capital, (5) other types of power arguments ing, transaction cost economics maintains that
(e.g., resource dependency), and (6) path depen- economizing is mainly determinative of private
dency. The first three alternatives may be dis- sector economic organization and, as indicated,
posed of as follows. (1) Technological nonsepara- invites comparison with rival main case hypothe-
bilities and indivisibilities explain only small ses. Nicholas Georgescu-Roegen's views on the
groups and, at most, large plants, but explain nei- purpose of science and the role of prediction are
ther multiplant organization nor the organization pertinent: "The purpose of science in general is
of technologically separable groups/activities not prediction, but knowledge for its own sake,"
(which should remain autonomous and which yet prediction is "the touchstone of scientific
should be joined?). (2) Monopoly explanations knowledge" (1971, p. 37). There being many
require that monopoly preconditions be satisfied, plausible accounts from which to choose, it is vital
but most markets are competitively organized. that each be prepared to show its hand (offer its
(3) Although differential risk aversion may apply predictions).
to many employment relationships, it has much
less applicability to trade between firms, where
Unit of Analysis
portfolio diversification is more easily accom-
plished and where smaller firms (for incentive in- A variety of units of analysis have been pro-
tensity and economizing, but not risk-bearing, posed to study economic organization. Simon has
reasons) are often observed to bear inordinate proposed that the decision premise is the appropri-
risk. Responses to the last three are developed ate unit of analysis (1957, pp. xxx-xxxii). Owner-
more hlly below. In brief, they are that (4) the ship is the unit of analysis for the economics of
failures to which contested exchange refers are property rights. The industry is the unit of analy-
often irremediable; (5) resource dependency is a sis in the structure-conduct-performance ap-
truncated theory of contract; and (6) although proach to industrial organization (Bain 1956;
path dependency is an important phenomenon, Scherer 1970). The individual has been nomi-
remediable inefficiency is rarely established. nated as the unit of analysis by positive agency
To be sure, transaction cost economizing does theory (Jensen 1983). Transaction cost econom-
not always operate smoothly or quickly. Thus we ics follows John R. Commons (1924; 1934)
should "expect [transaction cost economizing] to and takes the transaction to be the basic unit of
be most clearly exhibited in industries where analysis.
entry is [easy] and where the struggle for survival Whatever unit of analysis is selected, the critical
is [keen]" (Koopmans 1957, p. 141).9 Transac- dimensions with respect to which that unit of
tion cost economics nevertheless maintains that analysis differs need to be identified. Otherwise
later, if not sooner, inefficiency in the commercial the unit will remain nonoperational. Also, a para-
sector invites its own demise-all the more so as digm problem to which the unit of analysis ap-
international competition has become more vig- plies needs to be described. Table 1 sets out the
orous. Politically imposed impediments (tariffs, relevant comparisons.
quotas, subsidies, rules) can, however, delay the As shown, the representative problem with
reckoning;1° and disadvantaged parties (railroad which transaction cost economics deals is that of
workers, longshoremen, managers) may also be vertical integration-when should a firm make
able to delay changes unless compensated by rather than buy a good or service? The focal di-
buyouts. mension on which much of the predictive content
The economizing to which I refer operates of transaction cost economics relies, moreover, is
through weak-form selection-according to which asset specificity-which (as discussed below) is a
the fitter, but not necessarily the fittest, in some measure of bilateral dependency. More generally,
absolute sense, are selected (Simon 1983, p. transaction cost economics is concerned with the
69).11 Also, the economizing in question works governance of contractual relations (which bears
through a private net benefit calculus. That suits a resemblance to the "going concerns" to which
the needs of positive economics-what's going Commons referred). As it turns out, economic
88 Williamson
TABLE1. Com~arisonof Units of Analvsis
Unit of Analysis Critical Dimensions Focal Problem
Decision premise Role; information; Human problem solvingb
idiosyncratica
Ownership "Eleven characteristics"' Externality
Industry Concentration; Price-cost margins
barriers to entry
Individual Undeclared Incentive alignment
Transaction Frequency; uncertainty; Vertical integration
asset specificity
" Simon (1957a, pp. xxx-xxxi).
Newell and Simon (1972).
' Bromley (1989, pp. 187-90).
organization-in intermediate products markets, Lest claims of farsightedness be taken to hyper-
labor markets, capital markets, regulation, and rationality extremes, transaction cost economics
even the family-involves variations on a few key concedes that all complex contracts are unavoid-
transaction cost economizing themes. The pre- ably incomplete. That has both practical and the-
dictive action turns on the following proposition: oretical significance. The practical lesson is this:
transactions, which differ in their attributes, are all of the relevant contracting action cannot be
aligned with governance structures, which differ concentrated in the ex ante incentive alignment
in their costs and competence, in a discriminat- but some spills over into ex post governance. The
ing-mainly, transaction cost economizing-way. theoretical lesson is that differences among orga-
The arguments are familiar and are developed nization forms lose economic significance under
elsewhere. Suffice it to observe here that empirical comprehensive contracting setup-because an
research in organization theory has long suffered form of organization can then replicate any other
from the lack of an appropriate unit of analysis (Hart 1990).
and the operationalization-which is to say, di- Transaction cost economics combines incom-
mensionalization-thereof. pleteness with farsighted contracting by describ-
ing the contracting process as one of "incomplete
Farsighted Contracting contracting in its entirety." But for incomplete-
ness, the above-described significance of ex p
The preoccupation of economists with direct governance would vanish. But for farsightedness,
A d intended effects to the neglect of indirect and transaction cost economics would be denied ac-
(often delayed) unintended effects is widely inter- cess to one of the most important "tricks" in the
preted as a condition of myopia. In fact, however, economist's bag-namely, the assumption that
most economists are actually farsighted. The economic actors have the ability to look ahead,
problem is one of limited peripheral vision. discern problems and prospects, and factor these
Tunnel vision is both a strength and a weak- back into the organizational/contractual design.
ness. The strength is that a focused lens-pro- "l'lausible farsightedness," as against hyperration-
vided that it focuses on core issues-can be very ality, will often suffice.
powerful. The limitation is that irregularities that Consider, for example, the issue of threats.
are nonetheless important will be missed and/or, Threats are easy to make, but which threats are to
even worse, dismissed. be believed? If A says that it will do X if B does Y7
Transaction cost economics relates to these but if after B does Y, A's best response is to do 2, :
limitations by drawing on organization theory. then the threat will not be perceived t o be credi- ,G.

Because organization has a life of its own, transac- ble to a farsighted B. Credible threats are thus
tion cost economics (1) asks to be apprised of the those for which a farsighted B perceives that A's
more important indirect effects, whereupon (2) it ex post incentives comport with its claims-be-
asks what, given these prospective effects, are the cause, for example, A has made the requisite kind .
ramifications for efficient governance. A joinder and amount of investment t o support its threats '
of unanticipated effects (from organization the- (Dixit 1980).
ory) with farsighted contracting (from econom- Or consider the matter of opportunism. As
ics) thereby obtains. described above, Machiavelli worked out of a
 Transaction Cost Economics 89 I 1

P k s
Node A PI 0 0
- -
Node B P k 0
A - A
FIGURE2. Simple Contractual
Node C P k s
Schema

myopic logic, whereupon he advised his Prince to even condition. The break-even price that is
reply to opportunism in kind (get them before associated with Node A is p, . There being no haz-
they get you). By contrast, the farsighted Prince ards, k = 0. And since safeguards are unneeded,
is advised to look ahead and, if he discerns poten- s = 0.13
tial hazards, to take the hazards into account Node B is more interesting. The contractual
by redesigning the contractual relation-often by hazard here is 6. If the buyer is unable or unwill-
devising ex ante safeguards that will deter ex post ing to provide a safeguard, then s = 0. The corre-
opportunism. Accordingly, the wise prince is sponding break-even price is P .
advised to give and receive "credible commit- Node C poses the same contractual hazard,
namely 6. In this case, however, a safeguard in
To be sure, it is more complicated to think amount s^ is provided. The break-even price that is
about contract as a triple (p, k, 5)-where p refers projected under these conditions is j. It is ele-
to the price at which the trade takes place, k refers mentary that j < j.
to the hazards that are associated with the ex- Note that Jeffrey Bradach and Robert Eccles
change, sdenotes the safeguards within which the contend that "mutual dependence [i.e., k > 0]
exchange is embedded, and price, hazards, and between exchange partners . . . [promotes] trust,
safeguards are determined simultaneously-than [which] contrasts sharply with the argument cen-
as a scalar, where price alone is determinative. The tral to transaction cost economics that . . . depen-
simple schema shown in figure 2 nevertheless cap- .
dence . . fosters opportunistic behavior" (1989,
tures much of the relevant action.'' p. 111). What transaction cost economics says,
It will facilitate comparisons to assume that however, is that because opportunistic agents will
suppliers are competitively organized and are not self-enforce open-ended promises to behave
risk neutral. The prices at which product will be responsibly, efficient exchange will be realized
supplied therefore reflect an expected break- only if dependencies are supported by credible
90 Williamson
commitments. Wherein is trust implicated if par- courts, and the like are operationally irrelevant.
ties t o an exchange are farsighted and reflect the That does not deny that hypothetical ideals can ,
relevant hazards in terms of the exchange? (A bet- be useful as a reference standard, but standards
ter price [ j < @ ] will be offered if the hazards are often arbitrary. Is unbounded rationality the
[k > 01 are mitigated by cost-effective contractual relevant standard? How about perfect steward-
safeguards [j> 01.) ship, in which event opportunism vanishes?
As it turns out, the farsighted approach t o con- Lapses into ideal but operationally irrelevant
tracting has pervasive ramifications, some of reasoning will be avoided by (1) recognizing that
which are developed below. it is impossible t o d o better than one's best; (2)
insisting that all of the finalists in an organiza-
tional form of competition meet the test of feasi-
bility; (3) symmetrically exposing the weaknesses
The ideal organization adapts quickly and ef- as well as the strengths of all proposed feasible
ficaciously to disturbances of all kinds, but ac- forms; and (4) describing and costing out the
tual organizations experience trade-offs. Thus, mechanisms of any proposed reorganization. Such P'
whereas more decentralized forms of organiza- precautions seem to be reasonable, transparent,
tion (e.g., markets) support high-powered incen- even beyond dispute; yet all are frequently violated.
tives and display outstanding adaptive properties Note in this connection that "inefficiencym is
t o disturbances of an autonomous kind, they are unavoidably associated with contractual hazards.
poorly suited in cooperative adaptation respects. The basic market and hierarchy trade-off that is
Hierarchy, by contrast, has weaker incentives and
is comparatively worse at autonomous adaptation
but is comparatively better in cooperative adapta-
incurred upon taking transactions out of markets
and organizing them internally substitutes one
form of inefficiency (bureaucracy) for another
-
tion respects. (maladaptation). Other examples where one form
Simple transactions (for which k = 0)-in in- of inefficiency is used t o patch up another are (1)
termediate product markets, labor, finance, regu- decisions by firms t o integrate into adjacent stages
lation, and the like-are easy t o organize. The of production (or distribution) in a weak intellec-
requisite adaptations here are preponderantly of tual property rights regime, so as thereby t o miti-
an autonomous kind and the market-like option is gate the leakage of valued know-how (Teece
efficacious (so firms buy rather than make, use 1986); (2) decisions by manufacturers' agents to
spot contracts for labor, use debt rather than eq- incur added expenses, over and above those
uity, eschew regulation, etc.). Problems with needed t o develop the market, if these added ex-
markets arise as bilateral dependencies, and the penses strengthen customer bonds in a cost-effec- *
need for cooperative adaptations, build up. Mar- tive way, so as thereby to deter manufacturers
kets thus give way to hybrids which in turn give from entering and expropriating market develop-
way t o hierarchies (which is the organization ment investments (Heide and John 1988); and
form of last resort) as the needs for cooperative (3) the use of costly bonding t o deter franchisees
adaptations (k > 0 ) increase. from violating quality norms (Klein and Leffler
More generally, the point is this: informed 1981). Organization also has a bearing on the
choice among alternative forms of organization distribution of rents as well as asset protection.
entails trade-offs. Identifying and explicating Concern over rent dissipation influenced the de-
trade-offs is key t o the study of comparative eco- cision by U.S. automobile industry firms to inte-
nomic organization. Social scientists-econo- grate into parts (Helper and Levine 1992) and
mists and organization theorists alike-as well as also helps t o explain the resistance by oligopolies'
legal specialists need t o come t o terms with that to industrial unions.
proposition. To be sure, any sacrifice of organizational effi-
ciency, for oligopolistic rent protection reasons 0
otherwise, poses troublesome public policy is
Remediableness
sues.14 A remediableness test is nonetheless re-
Related t o this last is the concept of remedia- quired t o ascertain whether public policy should
bleness. If all feasible forms of organization are attempt t o upset the oligopoly power in question.
flawed (Coase 1964), then references t o be- The issues are discussed further later on, in rela-
nign government, costless regulation, omniscient tion to path dependency.
 Transaction Cost Economics 91

meeting competitive bids from qualified rivals.

A ~ D E DR ~ U L A R I T I B S Rivals cannot be presumed to operate on a parity,
however, once substantial investments in transac-
It is evident from the foregoing that the com- tion-specific assets are put in place. Winners in
parative contractual approach out of which trans- these circumstances enjoy advantages over non-
action cost economics works can be and needs to winners, which is to say that parity at the renewal
be informed by organization theory. Transaction interval is upset. Accordingly, what was a large
cost economics, however, is more than a mere numbers bidding condition at the outset is effec-
user. It pushes the logic of self-interest seeking to tively transformed into one of bilateral supply
deeper levels, of which the concept of credible thereafter. The reason why significant reliance in-
commitment is one example. More generally, it vestments in durable, transaction-specific assets
responds to prospective dysfunctional conse- introduce contractual asymmetry between the
quences by proposing improved ex ante designs winning bidder on the one hand and nonwinners
and/or alternative forms of governance. Also, on the other is because economic values would be
and what concerns us here, transaction cost eco- sacrificed if the ongoing supply relation were to
nomics has helped to discover added regularities be terminated.
that are pertinent to the study of organization. Faceless contracting is thereby supplanted by
These include ( I ) the fundamental transforma- contracting in which the pairwise identity of the
tion, (2) the impossibility of selective interven- parties matters. Not only would the supplier be
tion, ( 3 ) the economics of atmosphere, and (4) an unable to realize equivalent value were the spe-
interpretation of Japanese economic organization. cialized assets to be redeployed to other uses, but
a buyer would need to induce potential suppliers
to make similar specialized investments were it to
The Fundamental Transformation
seek least-cost supply from an outsider. Such par-
The Fundamental Transformation is the princi- ties therefore have strong incentives to work
pal transaction cost economics way of dem- things out rather than terminate. More generally,
onstrating that "identity matters." It helps to ex- farsighted agents will attempt to craft Node C
plain how firms take on distinctive identities and safeguards ex ante. As previously indicated, that
why identity matters.15 entails a progression from markets to hybrids and,
Economists of all persuasions recognize that if that does not suffice, to hierarchies. Given its
the terms upon which an initial bargain will be bureaucratic disabilities, hierarchy is the organiz-
struck depend on whether noncollusive bids can ational form of last resort.
be elicited from more than one qualified supplier.
Monopolistic terms will obtain if there is only a The Impossibility of Selective Intervention
single highly qualified supplier, while competitive
terms will result if there are many. Transaction Large established firms purportedly have ad-
cost economics fully accepts this description of ex vantages over smaller potential entrants because
ante bidding competition but insists that the "the leader can at least use [inputs] exactly as the
study of contracting be extended to include ex entrant would have . . . ,and earn the same profit
port features. as the entrant. But typically, the leader can im-
Contrary to earlier practice, transaction cost prove on this by coordinating production from
economics holds that a condition of large num- his new and existing inputs. Hence [inputs] will
bers bidding at the outset does not necessarily be valued more by the dominant firm" (Lewis
imply that a large numbers bidding condition will 1983, p. 1092).
obtain thereafter. Whether ex post competition is That argument has the following implication:
fully efficacious or not depends on whether the if large firms can everywhere do as well as a collec-
good or service in question is supported by dura- tion of smaller firms, through replication, and can
ble investments in transaction-specific human or sometimes do better, through selective interven-
physical assets. Where no such specialized invest- tion, then large firms ought to grow without
ments are incurred, the initial winning bidder re- limit. That is a variant of the Coasian puzzle
alizes no advantage over nonwinners. Although it "Why is not all production carried on in one big
may continue to supply for a long period of time, firm?" (1937, p. 340).
this is only because, in effect, it is continuously The simple answer to that query is that replica-
92 Williamson
tion and/or selective intervention are impossible. for compensation made accordingly, if what had
But that merely moves the argument back one hitherto been considered to be harmless by-prod-
stage. What explains these impossibilities? ucts of normal social intercourse are suddenly de-
The underlying difficulty is this: the integrity ofclared to be compensable injuries. The transfor-
rule governance is unavoidably compromised by mation of relationships that will ensue can easily
allowing discretion (Williamson 1985, chap. 6). lead to a lower level of felt satisfaction among the
Accordingly, any effort to combine rule govern- parties than prevailed previously-at least transi-
ance (as in markets) with discretionary govern- tionally and possibly permanently.
ance (hierarchy) experiences trade-offs. The pro- Part of the explanation is that filing claims for
posal to "implement the rules with discretion" is petty injuries influences attitudes toward other
simply too facile. transactions. My insistence on compensation for
That comes as no surprise to those who ap- A leads you to file claims for B, C, and D, which
proach the study of governance in discrete struc- induces me to seek compensation for E and F, etc.
tural terms-whereupon each generic form of Although an efficiency gain might be realized
governance possesses distinctive strengths and were it possible to isolate transaction A, the over-
weaknesses and movements between them entail all impact can easily be negative. Realizing this to
trade-offs. The puzzle of limits to firm size none- be the case, some individuals will be prepared to
theless eluded an answer for fifty years and more overlook such injuries. But everyone is not simi-
(Williamson 1985, pp. 132-35) and still occa- larly constituted. Society is rearranged to the
sions confusion. advantage of those who demand more exacting
correspondences between rewards and deeds if
Atmosphere metering at the margin is everywhere attempted.
Were the issue of compensation to be taken up as
The unintended effects described earlier are of a constitutional matter, rather than on a case-by-
a more local kind than the atmospheric effects ex- case basis, a greater tolerance for spillover would
amined here. "Atmosphere" refers to interactions commonly obtain (Schelling 1978).
between transactions that are technologically sep- Also pertinent is that individuals keep informal
arable but are joined attitudinally and have sys- social accounts and find the exchange of recipro-
tems consequences.I6 cal favors among parties with whom uncompen-
Thus, suppose that a job can be split into a se- sated spillovers exist to be satisfying (Gouldner
ries of separable functions. Suppose further that 1954). Transforming these casual social accounts
differential metering at the margin is attempted into exact and legal obligations may well be de-
with reference to each. What are the conse- structive of atmosphere and lead to a net loss of
quences? satisfaction between the parties. Put differently,
If functional separability does not imply attitu- pervasive pecuniary relations impair the quality of
dinal separability, then piecemeal calculativeness "contracting"--even if the metering of the trans-
can easily be dysfunctional. The risk is that push- actions in question were costless.17
ing metering at the margin everywhere to the The argument that emerges from the above is
limit will have spillover effects from easy-to-meter not that metering ought to be prohibited but that
onto hard-to-meter activities. If cooperative atti- the calculative approach to organization that is
tudes are impaired, then transactions that can be associated with economics can be taken to ex-
metered only with difficulty, but for which con- tremes. An awareness of attitudinal spillovers and
summate cooperation is important, will be dis- nonpecuniary satisfactions serves to check such
charged in a more perfunctory manner. The ne- excesses of calculativeness.
glect of such interaction effects is encouraged by
piecemeal calculativeness, which is to say by an in- Japanese Economic Organization
sensitivity to atmosphere.
A related issue is the matter of externalities. Transaction cost economics deals predomi-
The question may be put as follows: Ought all ex- nantly with dyadic contractual relations. Viewing
ternalities to be metered which, taken separately, the firm as a nexus of contracts, the object is to
can be metered with net gains? Presumably this prescribe the best transaction/governance struc-
turns partly on whether secondary effects obtain ture between the firm and its intermediate prod-
when an externality is accorded legitimacy. All uct market suppliers, between the firm and its
kinds of grievances may be "felt," and demands workers, between the firm and finance, etc. Japa-
 ~1

' I
Transaction Cost Economics 93 I

Employment

Personnel and
Enterprise Unions

FIGURE3. Supports for Lifetime Employment against the Hazards of (1) Adversity, (2) Shirking,
(3) Breach, (4) Equalitarianism. Subcontracting reduces (1) and (4); the personnel office and enter-
prise unions reduce (2) and (4); and banking reduces (3).

nese economic organization appears to be more tems solution shown by figure 3 is (arguably)
complicated. Employment, banking, and sub- more effective still.
contracting relations need to be examined simul- Figure 4 is somewhat more complicated and
interested readers are referred to discussions else-
ent, and subcontracting where (Williamson 1991b; Aoki 1992; and Sabel,
ese and U.S. economic chap. 6 in this Handbook). Suffice it to observe
xplicated by Masahiko here that banking and subcontracting (1) are not
Asanuma (1989), Erik only supports for the employment relation in the
nald Dore (1983), Michael core firm, but (2) are supported by the employ-
Gerlach (1992), James Lincoln (1990), Paul ment relation, and (3) are supports for each other.
Sheard (1989), and others. I am not only per-
suaded that these three areas of divergence are
believe that transaction cost economics UNRESOLVED
TENSIONS
the complementarities (Wil-
The healthy tension referred to at the outset
4 display the nature of the corn- has contributed to better and deeper understand-
e 3 depicts the contractual ings of a variety of phenomena. The matters that
d by lifetime employment. concern us here-power, path dependence, trust,
, These are (1) economic adversity--due, say, to and "toshn-are ones where differences between
, periodic decreases in demand-mkes it costly to transaction cost economics and organization the-
offer lifetime employment; (2) workers who enjoy ory are great.
ay treat it as a sinecure and
s who are induced by promises of Power~esourceDependence
ent to specialize their assets to a
to a breach-of-contract hazard; That efficiency plays such a large role in the
rian pressures develop within economic analysis of organization is because par-
the offer of lifetime employ- ties are assumed to consent to a contract and do
ment to key workers (where the justification is this in a relatively farsighted way. Such volun-
'Strong) spreads to all workers (to include those tarism is widely disputed by sociologists, who
for whom the justification is w e d ) . Although "tend to regard systems of exchange as embedded
each of these can be addressed separately, the sys- within systems of power and domination (usually
94 Williamson
Employment

FIGURE4 . Japanese Corpo-

rate Connectedness through
Contracting. Heavy arrows
indicate strong support; solid
arrows, average support;
dashed arrows, weaker sup-
port. Benefits are (1) greater
homogeneity, (2) greater
contract stability, (3) feed-
back stability, ( 4 ) reliable
Banking responses to adversity,
(5) financial planning (con-
vergent expectations), and
(6) no surprises.

regarded as grounded in a class structure in the sumes myopia. The argument here is that myopic
Marxian tradition) or systems of norms and val- parties to contracts are victims of unanticipated
ues" (Baron and Hannan 1992, p. 14). and unwanted dependency. Because myopic par-
The concept of power is very diffuse. Unable to ties do not perceive the hazards, safeguards will
define power, some specialists report that they not be provided and the hazards will not be
know it when they see it. That has led others t o priced out.
conclude that power is a "disappointing concept. Evidence pertinent to the myopic versus far-
It tends to become a tautological label for the un- sighted view of contract includes the following.
explained variance" (March 1988, p. 6). (1) Are suppliers indifferent between two tech-
Among the ways in which the term "power" is nologies that involve identical investments and
used are the following: the power of capital over have identical (steady-state) operating costs, but
labor (Bowles and Gintis 1993); strategic power one of which is much less redeployable than the
exercised by established firms in relation to extant other? (2) Is the degree of nonredeployability evi-
and prospective rivals (Shapiro 1989); special- dent ex ante or is it revealed only after an adverse
interest power over the political process (Moe state realization (which induces defection from
1990); and resource dependency. Although all the spirit of the agreement) has materialized? (3)
are relevant to economic organization, the last is Do added ex ante safeguards appear as added spe-
distinctive to organization theory,18 and conse- cificity builds up? And (4) Does contract law doc-
quently is examined here. uine and enforcement reflect one or the other of
Two versions of resource dependency can be these concepts of contract? Transaction cost eco-
distinguished. The weak version is that parties nomics answers these queries as follows: (1) the
who are subject to dependency will try to mitigate more generic (redeployable) technology will al-
it. That is unexceptionable and is akin to the safe- ways be used whenever the cetera are paria; (2)
guard argument advanced above. There are two nonredeployability can be discerned ex ante and is
significant differences, however: (1) resource de- recognized as such (Palay 1984; 1985; Masten
pendency nowhere recognizes that price, hazards, 1984; Shelanski 1993); (3) added ex ante safe-
and safeguards are determined simultaneously; guards do appear as asset specificity builds up
and (2) resource dependency nowhere remarks (Joskow 1985; 1988); and (4) because truly unu-
that asset specificity (which is the source of con- sual events are unforeseeable and can have puni-
tractual hazard) is intentionally chosen because it tive consequences if contracts are enforced liter-
is the source of productive benefits. ally, various forms of "excuse" are recbgnized by
The strong version of resource dependency as- the law, but excuse is granted sparingly.19
 I
Transaction Cost Economics 95 I

if introduced, would yield netgains. That is to be

Path Dependency
distinguished from hypothetical net gains, where
Transaction cost economics not only sub- the inefficiency in question is judged by compar-
scribes to the proposition that history matters but ing an actual alternative with a hypothetical ideal.
relies on that proposition to explain the differen- To be sure, big disparities between actual and
tial strengths and weaknesses of alternative forms hypothetical sometimes signal opportunities for
of governance. The Fundamental Transforma- net gains. But a preoccupation with hypotheticals
tion, for example, is a specific manifestation of the comes at a cost: "It has directed economists' at-
proposition. (Transactions that are not subject to tention away from the main question, which is
the Fundamental Transformation are much easier how alternative arrangements will actually work
to manage contractually.) The bureaucracy prob- in practice" (Coase 1964, p. 195; emphasis
lems that afflict internal organization (entrench- added).
ment; coalitions) are also the product of experi- Consider Brian Arthur's (1989) numerical ex-
ence and illustrate the same proposition. Were it ample of path dependency in which the payoffs to
not that systems drifted away from their initial individual firms upon adopting either of two
conditions, efforts to replicate markets within hi- technologies (A or B) depend on the number of
erarchies (or the reverse) and selectively intervene prior adoptions of each. Technology A has a
would be much easier-in which event differences higher payoff than B if there are few prior adop-
between organization forms would diminish. tions, but the advantage switches to technology B
The benefits that accrue to experience are also if there have been many prior adoptions. The
testimony to the proposition that history matters. "problem" is that if each potential adopter con-
Tacit knowledge and its consequences (Polanyi sults only its own immediate net gain, then each
1962; Marschak 1968; Arrow 1974) attest to this. will select A and there will be lock-in to an infe-
More generally, firm-specific human assets--of rior technology. A tyranny of micromotives
both spontaneous (e.g., coding economies) and thereby obtains (Schelling 1978).
intentional (e.g., learning) kinds-are the prod- As S. J. Liebowitz and Stephen Margolis ob-
uct of idiosyncratic experience. The entire insti- serve of this argument, however, whether choice
tutional environment (laws, rules, conventions, of technology A is inefficient or not depends on
norms, etc.) within which the institutions of gov- what assumptions are made about the state of
ernance are embedded is the product of history. knowledge (1992, p. 15). Also, even if individual
And although the social conditioning that oper- parties could be assumed to know that technol-
ates within governance structures (e.g., corporate ogy B would become the more efficient choice
culture [Kreps 1990bl) is reflexive and often in- after thirty or fifty adoptions, the added costs of
tentional, this too has accidental and temporal collective actionTto deter individuals from
features. choosing technology A-would need to be taken
That history matters does not, however, imply into account. If it is unrealistic to assume that in-
that only history matters. Intentionality and dividuals possess the relevant knowledge that a
economizing explain a lot of what is going on out switchover (fiom A to B) will occur upon thirty or
there. Also, most of the path dependency litera- fifty adoptions, or if, given that knowledge, the
ture emphasizes technology (e.g., the QWERTY costs of orchestrating collective action are prohib-
typewriter keyboard, discussed below) rather itive, then the inefficiency in question is effec-
than the organizational consequences referred to tively irremediable through private ordering.
above, Paul David's recent paper (1992) being an Sometimes, however, public ordering can do
exception. I am not persuaded that technological, better. The issues here are whether (1) the public
as against organizational, path dependency is as sector is better informed about network externali-
important as much of that literature suggests. ties; (2) the requisite collective action is easier to
Many of the "ineffi~iencies'~ to which the techno- orchestrate through the public sector (possibly by
logical path dependency literature refers are of an fiat); and/or ( 3 ) the social net benefit calculus
irremediable kind. differs from the private in sufficient degree to war-
Remediable inefficiencies. As described above, rant a different result. Absent plausible assump-
transaction cost economics emphasizes remedia- tions that would support a prospective net gain
ble inefficiencies-that is, those conditions for (in either private or social respects), the pur-
which afeasible alternative can be described which, ported inefficiency is effectively irremediable.
96 Williamson
That is regrettable, in that society would have sign meet essentially the same resistance . . . 2"
done better if it had had better knowledge or if a (ibid., 1986, p. 34).
reorganization could have been accomplished There are several possibilities. These include
more easily. Hypothetical regrets-the "nirvana nonrational behavior, conspiracy among type-
economics" t o which E. A. G. Robinson (1934) writer firms, and path dependency (ibid., pp. 34-
and Harold Demsetz (1969) refer-are neither 46). David makes a strong case for the last, but
here nor there. Real costs in relation t o real there is a fourth possibility, subsequently raised
choices is what comparative institutional eco- and examined by Liebowitz and Margolis (1990):
nomics is all about. neither the Navy study nor Apple advertising copy
Quantitative signi$cance. Path dependency, can support the astonishing claims made on their
remediable or not, poses a greater challenge if the behalf. Upon going back t o the archives and ex-
effects in question are large and lasting rather amining the data, Liebowitz and Margolis con-
than small and temporary. I t is not easy t o docu- clude that "the standard history of QWERTY ver-
ment the quantitative significance of path depen- sus Dvorak is flawed and incomplete. . . . [The]
dency. Arthur provides a series of examples and claims of superiority of the Dvorak keyboard are
emphasizes especially the videocassette recorder, suspect. The most dramatic claims are traceable t o
where VHS prevailed over the Beta technology Dvorak himself, and the best documented experi-
(1990, p. 92), and nuclear power, where light ments, as well as recent ergonomic studies, sug-
water reactors prevailed over high-temperature gest little or no advantage for the Dvorak key-
gas-cooled reactors (ibid., p. 99). But while both board" (ibid., p. 2 1 ). If that assessment stands up,
are interesting examples of path dependency, it is then path dependence has had only modest effi-
not obvious that the "winning" technology is sig- ciency effects in the QWERTY keyboard case.
nificantly inferior t o the loser,-or even, for that Such effects could easily fall below the threshold
matter, whether the winner is inferior at all. of remediable inefficiency.
Much the most widely cited case study is that Recent studies by sociologists of the evolution
of the typewriter keyboard. The QWERTY key- of particular industries also stress path depend-
board story has been set out by Paul David (1985; ency. Population ecologists have used the ecolog-
1986). It illustrates "why the study of economic ical model of density-dependent legitimation and
history is a necessity in the making of good econ- competition t o examine the evolutionary pro-
omists" (David 1986, p. 30). cess-both in particular industries (e.g., the tele-
QWERTY refers t o the first six letters on the phone industry [Barnett and Carroll 19931) and
top row of the standard typewriter keyboard. in computer simulations. Glenn Carroll and Rich-
Today's keyboard layout is the same as that which ard Harrison conclude from the latter that
was devised when the typewriter was first invented "chance can play a major role in organizational
in 1870. The early mechanical technology was evolution" (1992, p. 26).
beset by typebar clashes, which were mitigated by Although their simulations d o suggest that
the QWERTY keyboard design. Subsequent de- path dependency has large and lasting effects,
velopments in typewriter technology relieved Carroll and Harrison d o not address the matter of
these problems, but the QWERTY keyboard per- remediability. Until a feasible reorganization of
sisted in the face of large (reported) discrepancies the decision process for choosing technologies
in typing speed between it and later keyboard de- can be described, the effect of which is t o yield
signs. Thus the Dvorak Simplified Keyboard expected net private or social gains, it seems pre-
(DSK), which was patented in 1932, was so much mature t o describe their experiments as a test of
faster than the standard keyboard that, according the "relative roles of chance and rationality"
t o U.S. Navy experiments, the "increased (ibid., p. 13). Large but irremediable inefficien-
efficiency obtained with DSK would amortize the cies nevertheless d o raise serious issues for model-
cost of retraining a group of typists within the first ing economic ~rganization.~'
ten days of their subsequent full-time employ- Perspectives. David persuasively contends and
ment" (ibid., p. 33). More recently, the Apple I am persuaded that "there are many more
IIC computer comes with a built-in switch which QWERTY worlds lying out there" (1986, p. 47).
instantly converts its keyboard from QWERTY t o An unchanged keyboard layout does not, how-
DSK: "If as Apple advertising copy says, DSK 'lets ever, strike me as the most important economic
you type 2 0 4 0 % faster,' why did this superior de- attribute of typewriter development from 1870 t o
 I
I
Transaction Cost Economics 97
8
I
the present. What about improvements in the me- commitments (through the use of bonds, hos- I
chanical technology? What about the electric tages, information disclosure rules, specialized !
typewriter? What about personal computers and dispute settlement mechanisms, and the like) is to
laser printers? Why did these prevail in the face of create functional substitutes for trust (Grano-
path dependency? Were other "structurally supe- vetter 1985, p. 487). Albeit vitally important to
1
rior" technologies (as defined by Carroll and Har- economic organization, such substitutes should
rison) bypassed? If, with lags and hitches, the not be confused with (real) I
i
more efficient technologies have regularly sup- That calculativeness plays a larger role in eco-
planted less efficient technologies, shouldn't that nomics than in the other social sciences is evident
be featured? Possibly the response is that from the discussion of farsighted contracting. But
"everyone knows" that economizing is the main calculativeness can also be taken to excesses,
case: I t goes without saying that economizing is which is the main point in the discussion of atmo-
II
the main case to which path dependency, mono- sphere. Sometimes, however, an altogether differ-
polizing, efficient risk bearing, etc. are qualifica- ent orientation is needed. Thus whereas the
tions. response to excesses of monitoring is to be more
The persistent neglect of economizing reason- sophisticatedly calculative (take the dysfunctional
ing suggests otherwise. Thus the "inhospitality effects into account), there are other circum-
tradition" in antitrust proceeded with sublime stances where the response is to avoid being calcu-
confidence that nonstandard and unfamiliar busi- lative.
ness practices had little or no efficiency rationale Thus relations that are subject to continuous
but mainly had monopoly purpose and effect. Bayesian updating of probabilities based on expe-
Similarly, the vast inefficiencies that brought rience are thoroughly calculative (Williamson
down the economies of the Soviet Union and 1993b). And because commercial relations are in-
Eastern Europe may now be obvious, but that variably calculative, the concept of calculated risk
could never have been gleaned from the postwar (rather than calculated trust) should be used to
literature on comparative economic systems or describe commercial transactions.
from CIA intelligence estimates. The preoccupa- Continuous experience rating need not obtain
tion in the area of business strategy with clever everywhere, however. Indeed, because some per-
"plans, ploys, and positioning" to the neglect of sonal relations are unique and because continu-
economizing is likewise testimony to the wide- ous updating, even if only of a low-grade kind,
spread tendency to disregard efficiency (William- can have corrosive effects,24certain personal rela-
son 1991b). And the view that the "effective tions are treated in a nearly noncalculative way.
organization is (1) garrulous, (2) clumy, (3) su- That is accomplished by a discrete structural re-
perstitious, (4) hypocritical, (5) monstrous, (6) oc- classification, according to which personal rela-
I

~
~
I 1,
topoid, (7) wandering, and (8) grouchy" (Weick tions are dealt with on an all-or-none, rather than
1977, pp. 193-94; emphasis in original) is recon- a continuous updating, basis. '
ciled with economizing only with effort. More re- The upshot is that personal/trust relations and I
cent "social construction of industry" arguments commercial/calculative risk relations differ in
reduce economizing to insignifi~ance.~~ kind. Commercial relations are in no way deni-
If economizing really does get at the funda- grated as a result (Robbins, 1933, pp. 179-80).
mentals, then that condition ought to be continu-
ously featured. Some progress has been made
(Zald 1987), but there is little reason to be
complacent .22 The legal philosopher Lon Fuller distinguished
between "essentials" and "tosh," where the for-
mer involves an examination of the "rational
Trust
core" (Fuller 1978, pp. 359-62) and "tosh" is
There is a growing tendency, among econo- preoccupied with "superfluous rituals, rules of
mists and sociologists alike, to describe trust in procedure without clear purpose, [and] needless
calculative terms: both rational choice sociolo- precautions preserved through habit" (ibid., p.
gists (Coleman 1990) and game theorists (Das- 356). According to Fuller, to focus on the latter
gupta 1988) treat trust as a subclass of risk. I would "abandon any hope of fruitful analysis"
concur with Granovetter that to craft credible (ibid., p. 360).
 98 Williamson
This last goes too far: a place should be made might be offered as a contradiction. Hamilton
for "tosh," but "tosh7' should be kept in its and Biggart, however, go well beyond "tosh" (as
place.25Consider in this connection the Friedland described by Fuller) to implicate the institutional
and Alford interpretation of Clifford Geertz's de- environment-to include property rights, con-
scription of Balinese cockfights: "Enormous sums tract law, politics, and the like.
of money can change hands at each match, sums Thus, although both "tosh" and the institu-
/ that are irrational fiom an individualistic, utili- tional environment refer to background condi-
tarian perspective. The higher the sums, the more tions, the one should not be confused with the
evenly matched the cocks are arranged to be, and other. "Tosh" is a source of interesting variety
the more likely the odds on which the bet is made and adds spice to life. Core features of the institu-
are even. The greater the sum of money at stake, tional environment-as defined by North (1991)
the more the decision to bet is not individualistic and others (Sundaram and Black 1992)-are ar-
and utilitarian, but collective--one bets with guably more important, however, to the study of
'
one's kin or village-and status-oriented" (Fried- comparative economic ~ r g a n i z a t i o n . ~ ~
land and Alford 1991, pp. 247-48; emphasis
added).
That there are social pressures to support one's
kin or village is a sociological argument. Absent
these pressures, the concentration of bets on The science of organization to which Barnard
evenly matched cocks would be difficult to ex- made reference (1938, p. 290) over fifty years ago
plain. It does not, however, follow that it is irra- has made major strides in recent decades. All of
tional to bet enormous sums on evenly matched the social sciences have a stake in this, but none
cocks. Given the social context, it has become more than economics and organization theory.
nonviable, as a betting matter, to fight unevenly If the schematic set out in figure 1 is an accu-
matched cocks. rate way to characterize much of what is going
Thus suppose that an "objective matchmaker" on, then the economics of governance needs to be
would set the odds for a proposed match at four informed both from the level of the institutional
to one. In the absence, however, of such a match- environment (where sociology has much to con-
maker-and there is none in this instance-on- tribute) and from the level of the individual
siderations of local pride may preclude effective (where psychology is implicated). The intertem-
odds greater than three to two. Such a match will poral process transformations that take place
not attract much betting-because those from the within the institutions of governance (with re-
village with the lesser cock who view it fiom an spect to which organization theory has a lot to
individualistic, acquisitive perspective will make say) are also pertinent. The overall schema works
only perfunctory bets. Accordingly, the only in- out of the rational spirit approach that is associ-
teresting matches are those where social pressares ated with economics.28
are relieved by the even odds.26The "symbolic con- This multilevel approach relieves some-per-
struction of reality" to which Friedland and Al- haps much-of the strain to which Baron and
ford refer thus has real consequences. It delimits Hannan refer: "We think it important to under-
the feasible set within which rationality operates; stand the different assumptions and forms of rea-
but rationality is fully operative thereafter. soning used in contemporary sociology versus ec-
One interpretation of this is that "tosh" has onomics. . . . These disciplinary differences . . .
discrete structural effects and that rationality, op- represent major barriers to intellectual trade be-
erating through the marginal calculus, applies tween economics and sociology" (1992, p. 13).
thereafter. Indeed, that seems to fit the Balinese If, however, deep knowledge at several levels is
cockfight rather well. Whether the social con- needed and is beyond the competence of any one
struction of reality has such important conse- discipline, and if a systems conception can be de-
quences more generally is then the question. vised in which intellectual trade among levels can
Most probably it varies with the circumstances. be accomplished, then some of the worst misun-
"Tosh" is arguably more important in non- derstandings of the past can be put behind us.
commercial circumstances-state, family, reli- The following are some of the principal re-
gion-than in the commercial sector, although spects in which the healthy tension referred to at
the Hamilton and Biggart (1988) examination of the outset has supported intellectual trade-f
differences in corporate forms in the Far East which more is in prospect.
 Transaction Cost Economics 99

Trade-offs. Because each mode of governance

Organization Theory Supports for
is a syndrome of attributes, the move from one
Transaction Cost Economics
mode to another involves trade-offs. The key
Behavioral assumptions. Organization the- trade-offs need to be stated and explicated.
ory's insistence on workably realistic-as opposed Remediability. Relevant choices among feasi-
to analytically convenient-behavioral assump- ble forms of organization are what the analysis of
tions is a healthy antidote to artificiality. Transac- comparative economic organization is all about.
tion cost economics responds by describing eco-
nomic actors in terms of bounded rationality and
opportunism.
Adaptation. The cooperative adaptation em-
phasized by Barnard is joined with the autono- This paper has benefited from oral presentations to the
mous adaptation of Hayek, with the result that Macro Organization Behavior Society at the October 1992
transaction cost economics makes an appropriate meeting at Northwestern University, the Stanford Center
place for both market and hierarchy. for organizational Research, the Institutional Analysis
Workshop at the University of California, Berkeley, and the
Unanticipated consequences. The subtle and "Handbook of Economic Sociology Conference" at the
unintended consequences of control and organi- Russell Sage Foundation in February 1993. Helpful com-
zation need to be uncovered, whereupon provi- ments by James Baron, Paul DiMaggio, David Levine, Neil
sion can be made for these in the ex ante organiz- Smelser, and Richard Swedberg are gratefully acknowl-
ational design. edged.
1. A chronic defect of many critiques of neoclassical ec-
Politics. Because property rights in the public onomics is that these do not develop rival research agendas.
arena are shaped by democratic politics, provision Useful as it is to point out that neoclassical economics
needs to be made for these in the ex ante organiz- works out of strong and perhaps implausible assumptions
ational design of public sector bureaus. and/or is inattentive to important purposes, a reigning or-
thodoxy is never defeated by mere critique.
Embeddedness. The first-order response to 2. James March advised the Fourth International Con-
the proposition that embeddedness matters is to ference of the Society for the Advancement of Socio-Eco-
regard the institutional environment as a locus of nomics that economics had been so m y reformed that the
shift parameters, changes in which change the audience should "declare victory and go homen (Coughlin
comparative costs of governance. 1992, p. 23).
3. Richard Posner comes out differently. He argues that
Discrete structural analysis. Each generic form "organization-theory . . . [adds] nothing to economics that
of organization is described as a syndrome of at- the literature on information economics had not added
tributes and possesses its own logic. These dis- years earlier" (1993, p. 28).
creteness features need to be discovered and ex- 4. Briefly, the transaction cost economics responses are:
(1) institutions respond to scarcity as economizing devices;
plicated-both within and between sectors. (2) the transaction is expressly adopted as the basic unit of
analysis; (3) conflicts are recognized and relieved by the
creation of credible commitments/ex post governance ap-
Transaction Cost Economics Supports for paratus; and (4) the institutional environment is treated as
Organization Theory a set of shift parameters that change the comparative costs
of governance. Although these may be incomplete re-
Unit of analysis. Any theory of organization sponses, the spirit of the transaction cost economics enter-
that fails to name the unit of analysis out of which prise nevertheless makes serious contract with Commons's
it works and thereafter identify the critical dimen- prescription.
sions with respect to which that unit of analysis 5. Whereas Hodgson mainly describes the relation be-
tween institutional economics and neoclassical economics
varies is nonoperational at best and could be as one of opposition, I view the two in more complemen-
bankrupt. tary terms. Partly that is because the two some$mes deal
The main case. All rival theories of organiza- with different issues, but it is also because (1) transaction
tion are asked to nominate the main case, develop cost economics works out of the "rational spiritn of ortho-
the rehtable implications that accrue thereto, and doxy (Arrow 1974) and (2) neoclassical economics is elastic
and has made concessions to institutional economics.
examine the data. Economizing on transaction Consider, with respect to this last, the three defining at-
costs is the transaction cost economics candidate. tributes that Hodgson associates with neoclassical econom-
Farsighted contracting. Looking ahead, rec- ics-the assumption of rational, maximizing behavior; the
ognizing hazards, and folding these back into the preoccupation with equilibrium analysis; and the absence of/
chronic information problems. Although most of the cur-
design of governance is often feasible and explains rent microeconomic textbooks display all three of these fea-
a very considerable amount of organizational tures, leading-edge economic theory has begun to take
variety. bounded rationality very seriously (Kreps 1990a, pp. 151-
i oo Williamson
56), and information asymmetries and disparities of cogni- which the typewriter keyboard (discussed later in this chap-
tive competence are being recognized and dealt with as ter) is purportedly an example. Assuming, however, that
well (Milgrom and Roberts 1992). Equilibrium analysis has the appropriate criterion for judging superiority is that of
been more resistant t o assault, but even that is changing remediability, there are grave doubts that significant tech-
(see Summers and Shleifer [1988] and chapters in this nological or organizational efficiencies can be delayed in-
handbook. bv Richard Nelson).
> , ~ ~ definitely.
But for the fact that the new institutional economics was 11. The Schumpeterian process of "handing onn-
making headway by introducing bounded rationality, infor- which entails "a fall in the price of the product to the new
mation impactedness, and disequilibrium contracting, neo- level of costs" (Schumpeter 1947, p. 155) and purportedly
classical analysis would have been more apt t o remain works whenever rivals are alert t o new opportunities and
within the "protective belt" that is so often ascribed to it. are not prevented by purposive restrictions from adopting
Presented with challenges, however, it has found ways to them-is pertinent. The efficacy of handing o n varies with
respond. the circumstances. When are rivals more alert?What are the
6. Oligarchy is usually applied to composite organiza- underlying information assumptions? Are there other capi-
tion, but it applies t o subdivisions as well. Whether a firm tal market and/or organizational concerns?
should make or buy is thus a matter for which oligarchy has 12. The remainder of this subsection is based on Wil-
a bearing. If the decision t o take a transaction out of the liamson (1993a).
market and organize it internally is attended by subsequent 13. Another way of putting it is that (transition prob-
information distortions and subgoal pursuit, then that lems aside), each party can go its own way without cost to
should be taken into account at the outset (Williamson the other. Competition provides a safeguard.
1975, chap. 7; 1985, chap. 6). Not only d o operating costs 14. This has public policy ramifications. As between two
rise but a constituency develops that favors the renewal of oligopolies, one-of which engages in rent-protective meas-
internal facilities. An obviousresponse is to demand high ures while the other does not, and assuming that they are
hurdle rates for new projects, thereby to protect against the identical in other respects, the dissolution of the rent-pro-
unremarked but predictable distortions (added costs; advo- tective oligopoly will yield larger welfare gains.
cacy efforts) to which internal (as compared with market) 15. This subsection is based on Williamson (1985, pp.
procurement is differentially subject. 61-63).
The argument applies t o public sector projects as well. 16. This subsection is based on Williamson (1993a).
Because of the deferred and undisclosed but nevertheless 17. The buying of "rounds" in English pubs is an exam-
predictable distortions to which "organizationn is subject, ple. Would a costless meter lead t o a superior result? Sup-
new projects and regulatory propos~lsshould be req;ired pose that everyone privately disclosed a willingness t o pay
to display large (apparent) net gains. and that successive bids were solicited until a break-even
7. This is an interesting and important argument. Poli- result was projected. Suppose that the results of the final
tics really is different. But it is not as though there is no solicitation either are kept secret or posted, depending on
private sector counterpart. The more general argument is preferences, and that rounds are thereafter delivered to the
this: weak property rights regimes-both public and pri- table on request. Monthly bills are sent out in accordance
vate-invite farsighted parties to provide added protec- with the break-even condition. How is camaraderie af-
tions. The issues are discussed hrther in conjunction with fected?
remediableness. 18. Friedland and Alford (1991, p. 235) identify re-
Note,, as a comparative institutional matter, that secure source de~endencvas one of the two dominant theories of
totalitarian regimes can, according t o this logic, be ex- organization (the other being population ecology).
pected t o design more efficient public agencies. That is 19. Because contracts are incomplete and contain gaps,
neither here nor there if democratic values are held to be errors, omissions, and the like, and because the immediate
paramount-in which event the apparent inefficiencies of ~ a r t i e smav not be able t o reconcile their differences when
agencies under a democracy are simply a cost of this form of an unanticipated disturbance arises, parties t o a contract
governance. will sometimes ask courts t o be excused from performance.
8. Interdependencies among dyadic contracting rela- Because, moreover, literal enforcement can pose unaccept-
tions and the possible manipulation thereof have, however, ably severe contractual hazards-the effects of which are to
been examined (Williamson 1985, pp. 318-19). Also see discourage contracting (in favor of ertical integration)
the discussion of appropriability later in this chapter. and/or t o discourage potentially cost effective investments
9. The statement is a weakened variant on Tjalling in specialized assets-some relief from strict enforcement
Koopmans. Where he refers to "profit maximization," recommends itself. How much relief is then the question.
"easiest," and "keenest," I have substituted "transaction Were excuse to be granted routinely whenever adversity OC-
cost economizing," "easy," and "keen." curred, then incentives t o think through contracts, choose
10. Joel Mokyr observes that resistance t o innovation technologies judiciously, share risks efficiently, and avert
"occurred in many periods and places but seems t o have adversity would be impaired. Accordingly, transaction
been neglected by most historians" (1990, p. 178). He cost economics recommends that (1) provision be made
nevertheless gives a number of examples in which estab- for excuse but (2) excuse should be awarded sparingly--
lished interests, often with the use of the political process, which it evidently is (Farnsworth 1968, p. 885; Buxbaum
set out t o defeat new technologies. In the end, however, 1985).
the effect was not t o defeat but to delay machines that 20. I have argued that dominant firm industries in
pressed pinheads, an improved slide rest lathe, the ribbon which chance plays a role do warrant public policy interven-
loom, the flying shuttle, the use of Arabic numerals, and tion (Williamson 1975, chap. l l ) , but whether net gains
the use of the printing press (ibid., pp. 178-79). That, of would really be realized by implementing that proposal (es-
course, is not dispositive. There may be many cases in pecially as international competition becomes more inten-
which superior technologies were in fact defeated--of sive) is problematic.
 Transaction Cost Economics i oi

21. The "new economic sociology" holds that "even in is that "human behavior must be understood in terms of
identical economic and technical conditions, outcomes the fusion of individually-based and communally-based
may differ dramatically if social structures are different" forces, which Etzioni labels the I and We. The I represents
(Granovetter 1992, p. 9). The "social construction of in- the individual acting in pursuit of his or her own pleasure;
dustry" argument is developed in a forthcoming book by the We stands for the obligations and restraints imposed by
Patrick McGuire, Mark Granovetter, and Michael Schwartz the collectivity" (Coughlin 1992, p. 3). That is close to the
on the origins of the American electric power industry. The interpretation of the Balinese cockiights advanced here.
authors work out of a sociological perspective and maintain 27. This is pertinent, among other things, to the study
that the reason why central station generation and distribu- of the multinational enterprise. As Anant Sundaram and
tion ofelectricity was chosen over home generation was not J. Stewart Black observe, MNEs "pursue different entry/
for reasons of economy. Rather, the preferences of "power- involvement strategies in different markets and for different
ful actors" prevailed. products at any given time" (1992, p. 740). Their argu-
Granovetter challenges economic sociology to improve ment, that transaction cost economics "is inadequate for
upon the explanatory and predictive ability of rival theories explaining simultaneously different entry modes because
(1992, p. 5) and mentions residential heating (ibid., p. 8), . . . asset specificity . . . [is] largely the same the world
where the experience reversed that of electricity-in that over" (ibid., p. 740) assumes that the governance level op-
individual home heating units won out over central station erates independently of the institutional environment
steam generation (a few communities, which tried and under a transaction cost setup. This assumption is mis-
abandoned the latter, excepted). The parallel argument, taken.
presumably, is that the households had more power than 28. I borrow the term "rational spirit" from Kenneth
industry in the contest between these two technologies. Arrow (1974, p. 16). The rational spirit approach holds
ath her than appeal to power, I would argue that the two that there is a logic to organization and that this logic is
technologies were not on a parity in economizing respects mainly discerned by the relentless application of economic
and that the winner in each case was the more economical reasoning (subject, however, to cognitive constraints). The
mode. A more contemporary question has been posed by rational spirit approach is akin to but somewhat weaker (in
James Robinson, who asks, "Which theory of professional that it eschews stronger forms of utility maximization) than
dominance would have predicted the stupendous decline in the rational choice approach associated with James Cole-
power and authority suffered in recent years by organized man (1990).
medicine?", whereupon Robinson argues that comparative
transaction cost efficiencies are implicated (1992, pp. 13-
14).
- -

22. An earlier version of this chapter discussed the ten-

sion between transaction cost economics and the worker-
managed enterprise. For an examination of these issues, see Italicized terms in the dejnitions serve as cross-refer-
Williamson (1989, pp. 22-24). My argument is that fi-
nance will be made available on very adverse terms if (1) ences to other entries in theglossary.
finance is denied a major role in decision-making and (2) asset specificity: A characteristic of a specialized invest-
the firm is highly invested in specific assets. John Bonin and ment, whereby it cannot be redeployed t o alterna-
Louis Putterman (1987) ignore the complications that tive uses o r by alternative users except at a loss of
arise because of asset s~ecificitvand advise that finance will productive value. Asset specificity can take several
accept the denial of a decision-making role in a worker-
managed firm without adjusting the cost of capital ad- forms, of which human, physical, site, and dedi-
versely. cated assets are the most common. Specific assets
23. Note that the trust that Granovetter ascribes to give rise t o bilateral dependency, which complicates
ongoing relations can go either way. That is because experi- contractual relations. Accordingly, such invest-
ence can be either good (more confidence) or bad (less ments would never be made except as these con-
confidence)-which, if contracts of both kinds are re- tribute to prospective reductions in production
newed, will show up in differential contracting (Crocker costs o r additions t o revenue.
and Reynolds 1993). bilateral dependency: An ongoing dependency relation
24. Not only can intendedly noncalculative relations be obtains between a buyer and a supplier when one
upset by Type I error, according to which a true relation is or both have made durable specialized investments
incorrectly classified as false, but calculativeness may be
subject to (involuntary) positive feedback. Intendedly non- in support of the other. Although sometimes this
calculative relations that are continuously subject to being condition exists from the outset (the familiar bilat-
reclassified as calculative are, in effect, calculative. eral monopoly condition), often it evolves during
25. The evolution of cooperation between opposed ar- the course of an ongoing contractual relation. Bi-
mies or gangs that are purportedly engaged in "deadly lateral dependency, in which one o r both parties
combat" is illustrated by Robert Axelrod's examination of specialize t o the other, is a much more widespread
"The Live-and-Let-Live System in Trench Warfare in condition than preexisting bilateral monopoly.
World War I" (1984, pp. 73-87). Interesting and impor- Such dependency poses contractual hazards in the
tant as the live-and-let-live rituals were, these nonviolent face of incomplete contracting and opport~nism,in
practices should not be mistaken for the main case. Rather,
these rituals were the exception to the main case-which response t o which contractual safeg~ardsare com-
was that British and German troops were at war. monly provided.
26. Richard M. Coughlin contends that the 'essence" bounded rationality: Refers t o behavior that is in-
of the socioeconomic approach proposed by Amitai Etzioni tendedly rational but only limitedly so; a condition
 10 2 Williamson
/

of limited cognitive competence t o receive, store, mon understanding of the nature of the future
retrieve, and process information. All complex contingencies cannot be reached; (4) a common
contracts are unavoidably incomplete because of and complete understanding of the appropriate
bounds o n rationality. adaptations to future contingencies cannot be
bureaucracy: The support staff that is responsible for reached; (5) parties are unable t o agree on what
developing plans, collecting and processing infor- contingent event has materialized; (6) parties are
mation, operationalizing and implementing execu- unable to agree on whether actual adaptations to
tive decisions, auditing performance, and, more realized contingencies correspond with those spec-
generally, providing direction t o the operating ified in the contract; and (7) even though the par-
parts of a hierarchical enterprise. Bureaucracy is at- ties may both be fully apprised of realized contin-
tended by low-powered incentives (due to the im- gency and the actual adaptations that have been
possibility of selective intervention) and is given to made, third parties (e.g., courtsj can be fully ap-
subgoal pursuit (which is a manifestation of oppor- prised of neither, in which event costly haggling
tunism). between bilaterally dependent parties may ensue.
contract: An agreement between a buyer and a supplier institutional arrangement: The contractual relation o r
in which the terms of exchange are defined by a tri- governance structure between economic entities
ple: price, asset specificity, and safeguards. (This as- that defines the way in which they cooperate and/
sumes that quantity, quality, and duration are all or compete.
specified.) institutional environment: The rules of the game that
credible commitment: A contract in which a promisee is define the context within which economic activity
reliably compensated should the promisor prema- takes place. The political, social, and legal ground
turely terminate or otherwise alter the agreement. rules establish the basis for production, exchange,
This is to be contrasted with noncredible commit- and distribution.
ments, which are empty promises, and semicredi- market: The arena in which autonomous parties engage
ble commitments, in which there is a residual in exchange. Markets can either be thick o r thin.
hazard. Credible commitments are pertinent t o Classical markets are thick, in which case there are
contracts where one or both parties invest in spe- large numbers of buyers and sellers o n each side of
cific assets. the transaction and identity is unimportant, be-
discriminating alignment: The assignment of least-cost cause each can go its own way at negligible cost to
governance structures to manage transactions. the other. Thin markets are characterized by
governance structure: The institutional matrix within fewness, which is mainly due to asset specificity.
which the integrity of a transaction is decided. Hybrid contracts and hierarchy emerge as asset spe-
Within the commercial sector, three discrete struc- cificity builds up and identity matters.
tural governance alternatives are commonly recog- opportunism: Self-interest seeking with guile, t o in-
nized: classical market, hybrid contracting, and hi- clude calculated efforts t o mislead, deceive, obfus-
erarchy. cate, and otherwise confuse. Opportunism is to be
hierarchy: Transactions that are placed under unified distinguished from simple self-interest seeking, ac-
ownership (buyer and supplier are within the same cording t o which individuals play a game with fixed
enterprise) and subject to administrative controls rules that they reliably obey.
(an authority relation, to include fiat) are managed private ordering: The self-created mechanisms t o ac-
by hierarchy. The contract law of hierarchy is that complish adaptive, sequential decision-making be-
of forebearance, according t o which internal orga- tween autonomous parties to a contract, including
nization is its own court of ultimate appeal. information disclosure, dispute settlement, and
hybrid: Long-term contractual relations that preserve distributional mechanisms to deal with gaps, er-
autonomy but provide added transaction-specific rors, omissions, and inequities. (Court-ordering,
safeguards as compared with the market. however, is normally available for purposes of ulti-
incentive intensity: A measure of the degree to which a mate appeal.)
party reliably appropriates the net receipts (which safeguard: The added security features, if any, that are
could be negative) associated with its efforts and introduced into a contract, thereby to reduce haz-
decisions. High-powered incentives obtain if a ards (due mainly to asset specificity) and infuse con-
party has a clear entitlement to and can establish fidence. Safeguards can take the form of penalties,
the magnitude of its net receipts easily. Lower- a reduction in incentive intensity, and/or more
powered incentives obtain if the net receipts are fully developed private ordering apparatus t o deal
pooled and/or if the magnitude is difficult t o as- with contingencies.
certain. selective intervention: Would obtain if bureaucratic in-
incomplete contracting: Contracts are effectively in- tervention between the semiautonomous parts of a
complete if (1)not all of the relevant future contin- hierarchical enterprise occurred only but always
gencies can be imagined; (2) the details of some of when there is a prospect of expected net gain. Be-
the future contingencies are obscure; ( 3 ) a com- cause promises to intervene selectively lack credi-
 Transaction Cost Economics 1o3

bility, selective intervention is impossible. Were it ship-Specific Skill." Journal of Japanese and Inter-
otherwise, everything would be organized in one national Economies 3: 1-30.
large firm. Because, however, selective intervention Axelrod, Robert. 1984. The Evolution of Cooperation.
is impossible, hierarchies are unable t o replicate New York: Basic Books.
market incentives. Bain, Joe. 1956. Barriers t o New Competition. New
subgoal: A local goal, the pursuit of which ofien impairs York: John Wiley & Sons.
the performance of the system. Barnard, Chester. 1938. The Functions of the Executive.
transaction: T h e microanalytic unit of analysis in trans- Cambridge, M A : Harvard University Press (fif-
action cost economics. A transaction occurs when teenth printing, 1962).
a good o r service is transferred across a technologi- Barnett, William, and Glenn Carroll. 1993. "How In-
cally separable interface. Transactions are mediated stitutional Constraints Affected the Organization
by governance structures (markets, hybrids, hierar- of the Early American Telephone Industry." Jour-
chies). nal of Law, Economics, and Organization 9:99-
transaction cost: The ex ante costs of drafting, negotiat- 126.
ing, and safeguarding an agreement and, more Baron, James, and Michael Hannan. 1992. "The Im-
especially, the ex post costs of maladaptation and pact of Economics o n Contemporary Sociology."
adjustment that arise when contract execution is Stanford University, Stanford, CA. Unpublished
misaligned as a result of gaps, errors, omissions, manuscript.
and unanticipated disturbances; the costs of run- Becker, Gary. 1976. The Economic Approach t o Human
ning the economic system. Behavior. Chicago: University of Chicago Press.
weak form selection: Selection from among the better Berglof, Erik. 1989. "Capital Structure as a Mechanism
of the feasible alternatives, as contrasted with selec- of Control-A Comparison of Financial Systems."
tion of the best from among all possible-includ- Pp. 237-62 in The Firm as a Nexus of Treaties, ed-
ing hypothetical alternatives. In a relative sense, ited by Masahiko Aoki, Bo Gustafsson, and Oliver
the fitter survive, but these may not be the fittest in Williamson. London: Sage.
any absolute sense. Bergson, Abram. 1948. "Socialist Economies," Pp.
430-58 in Survey of Contemporary Economies, ed-
ited by Howard Ellis. Philadelphia: Blakiston.
Bonin, John, and Louis Putterman. 1987. Economics of
Cooperation and Labor Managed Economies. New
York: Cambridge University Press.
Alchian, Armen. 1961. Some Economics of Property. Bowles, Samuel, and Herbert Gintis. 1993. "The Re-
RAND D-2316. Santa Monica, CA: RAND Cor- venge of H o m o Economicus: Contested Exchange
poration. and the Revival of Political Economy." Journal of
Alchian, Armen, and Harold Demsetz. 1972. "Produc- Economic Perspectives 7 3 3 - 1 00.
tion, Information Costs, and Economic Organiza- Bradach, Jeffrey, and Robert Eccles. 1989. "Price, Au-
tion." American Economic Review 62:777-95. thority, and Trust." American Review of Sociology
Alchaian, Armen, and Susan Woodward. 1987. "Reflec- 15:97-118.
tions o n the Theory of the Firm." Journal of Insti- Bridgeman, Percy. 1955. Rejections of a Physicist. 2d
tutional and Theoretical Economics 143:llO-36. ed. New York: Philosophical Library.
Aoki, Masahiko. 1988. Information, Incentives, and Bromley, Daniel. 1989. Economic Interests and Institu-
Bargaining in the Japanese Economy. New York: tions. New York: Basil Blackwell.
Cambridge University Press. Buxbaum, Richard. 1985. "Modification and Adapta-
. 1990. "Toward an Economic Model of the tion of Contracts: American Legal Developments."
Japanese Firm." Journal of Economic Literature Studies in Transnational Law 3:31-54.
28:l-27. Carroll, Glenn, and J. Richard Harrison. 1992.
-. 1992. "The Japanese Firm as a System of "Chance and Rationality in Organizational Evolu-
Attributes: A Survey and Research Agenda." Stan- tion." University of California, Berkeley. Unpub-
ford University, Stanford, CA. Unpublished man- lished manuscript.
uscript. Coase, R. H . [1937] 1988. The Firm, the Market, and
Arrow, Kenneth J. 1974. The Limits of Organization. the Law. Chicago: University of Chicago Press.
1st ed. New York: W. W. Norton. . 1959. "The Federal Communications Com-
Arthur, Brian. 1989. "Competing Technologies, In- mission." Journal of Law and Economics 2:1-40.
creasing Returns, and Lock-In by Historical . 1960. "The Problem of Social Cost." Journal
Events." Economic Journal 99:116-31. of Law and Economics 3:1-44.
- 1990. "Positive Feedbacks in the Economy." . 1964. "The Regulated Industries: Discussion."
Scientific American 262 :80-8 5. American Economic Review 54:194-97.
Asanuma, Banri. 1989. "Manufacturer-Supplier Rela- . 1972. "Industrial Organization: A Proposal for
tionships in Japan and the Concept of Relation- Research." Pp. 59-73 in Policy Issues and Research
 1 04 Williamson
Opportunities in Industrial Organization, edited Dore, Ronald. 1983. "Goodwill and the Spirit of Market
by V . R. Fuchs. New York: National Bureau of
Economic Research.
Capitalism." British Journal of Sociology 34:459-82.
Farnsworth, Edward Allan. 1968. "Disputes over i
1
. 1984. "The New Institutional Economics." Omissions in Contracts." Columbia Law Review 1
Journal of Institutional and Theoretical Economics 68:860-91.
140:229-31. Frank, Robert. 1992. "Melding Sociology and Eco-
Coleman, James. 1982. The Asymmetric Society. nomics." Journal of Economic Literature 30:147-
Syracuse, NY: Syracuse University Press. 70.
. 1990. The Foundations of Social Theory. Cam- Friedland, Roger, and Robert Alford. 1991. "Bringing
bridge, MA: Harvard University Press. Society Back In: Symbols, Practices, and Institu-
Commons, John R. 1924. Legal Foundations of Capi- i
tiopal Contradictions." Pp. 232-66 in The New In- I
talism. New York: Macmillan. stitutionalism in Organizational Analysis, edited
. 1934. Institutional Economics. Madison: Uni- by Walter Powell and Paul DiMaggio. Chicago: i
versity of Wisconsin Press. University of Chicago Press. i
Coughlin, Richard. 1992. "Interdisciplinary Nature of Fuller, Lon L. 1978. "The Forms and Limits ofAdjudi-
Socio-Economics." Unpublished manuscript. cation." Harvard Law Review 92:353409.
Crocker, Keith, and Kenneth Reynolds. 1993. "The Ef- . 1981. "Human Interaction and the Law." Pp.
ficiency of Incomplete Contracts: An Empirical 2 1 2 4 6 in The Principles of Social Order: Selected
Analysis of Air Force Engine Procurement." Rand Essays of Lon L. Fuller, edited by Kenneth I. Win-
Journal of Economics 1 2 6 4 6 . 1
Crozier, Michel. 1964. The Bureaucratic Phenomenon.
Chicago: University of Chicago Press.
ston. Durham, NC: Duke University Press.
Furubotn, Eirik, and Svetozar Pejovich. 1974. The Eco-
nomics of Property Rights. Cambridge, MA: Ballin-
1
Cyert, Richard M., and James G. March. 1963. A Be- ger.
i
havioral Theory of the Firm. Englewood Cliffs, NJ:
Prentice-Hall.
Dasgupta, Partha. 1988. "Trust as a Commodity." Pp.
Furubotn, Eirik, and Rudolf Richter. 1991. The New
Institutional Economics. College Station: Texas !
A&M University Press.
I
49-72 in Trust: Making and Breaking Cooperative Georgescu-Roegen, Nicholas. 1971. The Entropy Law 1
Relations, edited by Diego Gambetta. Oxford: and Economic Process. Cambridge, MA: Harvard '
Basil Blackwell. University Press. I
David, Paul. 1985. "Clio in the Economics of Gerlach, Michael. 1992. Alliance Capitalism: The So- 1
QWERTY." American Economic Review 75:332- cial Ovganization of Japanese Business. Berkeley: j
37. University of California Press. I
-. 1986. "Understanding the Economics of Gouldner, Alvin W. 1954. Industrial Bureaucracy.
I
1
QWERTY: The Necessity of History." Pp. 3 0 4 9 Glencoe, IL: Free Press.
in Economic History and the Modem Economist, ed- Granovetter, Mark. 1985. "Economic Action and SO- I
ited by W. N. Parker. New York: Basil Blackwell. cial Structure: The Problem of Embeddedness."
-. 1992. "Heroes, Herds, and Hypteresis in American Journal of Sociology 9 1 :48 1-501.
Technological History." Industrial and Corporate . 1988. "The Sociological and Economic Ap- j
Change 1:129-80. proaches to Labor Market Analysis." Pp. 187-218
Davis, Gerald F., and Walter W. Powell. 1992. "Organ- in Industries, Firms, and Jobs, edited by George i
ization-Environment Relations." Pp. 315-75 in Farkas and Paula England. New York: Plenum. :
Handbook of Industrial and Organizational Psy- . 1990. "The Old and the New Economic Soci- :
chology, vol. 3, edited by M. Dunnette. 2d ed. ology: A History and an Agenda." Pp. 89-1 12 in
New York: Consulting Psychologists Press. Beyond the Marketplace, edited by Roger Friedland
Davis, Lance E., and Douglass C. North. 1971. Institu- and A. F. Robertson. New York: Aldine.
tional Change and American Economic Growth. . 1992. "Economic Institutions as Social Con-
Cambridge: Cambridge University Press. structions: A Framework for Analysis." Acta Social.
Demsetz, Harold. 1967. "Toward a Theory of Property ogica 35:3-11.
Rights." American Economic Review 57:347-59. Grossekettler, Heinz. 1989. "On Designing an ECO-
. 1969. "Information and Efficiency: Another nomic Order: The Contributions of the Freiburg
Viewpoint." Journal of Law and Economics 12:l- School." Pp. 38-84 in Perspectives on the History 0,
22. Economic Thought, vol. 2, edited by ~ o n a l c
DiMaggio, Paul, and Walter Powell. 1991. "Introduc- Walker. Aldershot: Edward Elgar.
tion." Pp. 1-38 in The New Institutionalism in Or- Hamilton, Gary, and Nicole Biggart. 1988. "Market
ganizational Analysis, edited by Walter Powell and Culture, and Authority." American Journal of So
Paul DiMaggio. Chicago: University of Chicago ciology (Supplement) 94:S52-S94.
Press. Harberger, Arnold. 1954. "Monopoly and Resourct
Dixit, A. 1980. "The Role of Investment in Entry De- Allocation." American Economic Review 44:77-
terrence." Economic Journal 90:95-106. 87.
 Transaction Cost Economics 1o 5

Hart, Oliver. 1990. "An Economist's Perspective on cialism." Pp. 55-143 in On the Economic Theory of
the Theory of the Firm." Pp. 154-71 in Organiza- Socialism, edited by Benjamin Lippincott. Minnea-
tion Theory, edited by Oliver Williamson. New polis: University of Minnesota Press.
York: Oxford University Press. Lewis, Tracy. 1983. "Preemption, Divestiture, and For-
Hayek, Friedrich. 1945. "The Use of Knowledge in So- ward Contracting in a Market Dominated by a Sin-
ckty." American Economic Review 35:5 19-30. gle Firm." American Economic Review 73:1092-
Hechter, Michael. 1987. Principles of Group Solidarity. 1101.
Berkeley: University of California Press. Liebowitz, Stanley J., and Stephen Margolis. 1990.
Heide, Jan, and George John. 1988. "The Role of De- "The Fable of the Keys." Journal of Law and Eco-
pendence Balancing in Safeguarding Transaction- nomics 33:l-26.
Specific Assets in Conventional Channels." Jour- -. 1992. "Path Dependency, Lock-In, and His-
nal of Marketing 52:20-35. tory." College Station: Texas A&M University.
Helper, Susan, and David Levine. 1992. "Long-Term Unpublished manuscript.
Supplier Relations and Product-Market Struc- Lincoln, James. 1990. "Japanese Organization and Or-
ture." Journal of Law, Economics, and Organiza- ganization Theory." Research in Organizational
tion 8:561-81. Behavior 12:255-94.
Hinds, Manuel. 1990. Issues in the Introduction ofMar- Llewellyn, Ibrl N. 1931. "What Price Contract? An
ket Forces in Eastern European Socialist Economies. Essay in Perspective." Yale Law Journal 40:704-
The World Bank, Washington DC. Report No. 5 1.
IDP-0057. McCain, Roger. 1977. "On the Optimal Financial En-
Horvat, Branko. 1991. Review of Janos Kornai, The vironment for Worker Cooperatives." Zeitschrift
Road to a Free Economy. Journal of Economic Be- fur Nationalokonomie 37:355-84.
havior and Organization 1 5:408-10. McCullough, David. 1992. Truman. New York: Simon
Hutchison, Terrence. 1984. "Institutional Economics & Schuster.
Old and New." Journal of Institutional and Theo- McGuire, Patrick, Mark Granovetter, and Michael
retical Economics 140:20-29. Schwartz. 1992. "The Social Construction of In-
Jensen, Michael. 1983. "Organization Theory and dustry." Unpublished book prospectus.
Methodology." Accounting Review 50:319-39. Machiavelli, Niccolb. [1513] 1952. The Prince. Trans-
Joskow, Paul L. 1985. "Vertical Integration and Long- lated by Christian Gauss. New York: New Ameri-
Term Contracts." Journal of Law, Economics, and can Library.
Orgnnization 1:33-80. Macneil, Ian R. 1974. "The Many Futures of Con-
. 1988. "Asset Specificity and the Structure of tracts." Southern California Law Review 47:691-
Vertical Relationships: Empirical Evidence." Jour- 816.
nal of Law, Economics, and Organization 4:95- -. 1978. "Contracts: Adjustments of Long-Term
117. Economic Relations under Classical, Neoclassical,
IUein, Benjamin, R. A. Crawford, and A. A. Alchian. and Relational Contract Law." Northwestern Uni-
1978. "Vertical Integration, Appropriable Rents, versity Law Review 72:854-906.
and the Competitive Contracting Process." Jour- March, James G. 1988. Decisions and Organizations.
nal of Law and Economics 2 1:297-326. Oxford: Basil Blackwell.
Klein, Benjamin, and K. B. Lemer. 1981. "The Role of March, James G., and Herbert A. Simon. 1958. Organ-
Market Forces in Assuring Contractual Perfor- izations. New York: John Wiley & Sons.
mance." Journal of Political Economy 89:615-41. Marschak, Jacob. 1968. "Economics of Inquiring,
Koopmans, Tjalling. 1957. Three Essays on the State of Communicating, Deciding." American Economic
Economic Science. New York: McGraw-Hill Book Review 58:l-18.
Company. Masten, Scott. 1984. "The Organization of Produc-
Kornai, janos. 1990. "The Affinity between Ownership tion: Evidence from the Aerospace Industry."
Forms and Coordination Mechanisms: The Com- Journal of Law and Economics 27:403-18.
mon Experience of Reform in Socialist Countries." . 1993. "Transaction Costs, Mistakes, and Perfor-
Journal of Economic Perspectives 4:131-47. mance: Assessing the Importance of Governance."
Kreps, David M. 1990a. Game Theory and Economic Management and Decision Sciences 14:119-29.
Modelling. New York: Oxford University Press. Matthews, R. C. 0. 1986. "The Economics of Institu-
-. 1990b. "Corporate Culture and Economic tions and the Sources of Economic Growth." Eco-
Theory." Pp. 90-143 in Perspectives on Positive Po- nomic Journal 96:903-18.
litical Economy, edited by James Alt h d Kenneth Merton, Robert K. 1936. "The Unanticipated Conse-
Shepsle. New York: Cambridge University Press. quences of Purposive Social Action." American So-
-. 1992. "(How) Can Game Theory Lead to a ciological Review l 394-904.
Unified Theory of Organization?" Stanford Uni- Michels, Robert. [ 1 911] 1962. Political Parties. Trans-
versity, Stanford, CA. Unpublished manuscript. lated by Edan and Cedar ~ a u l Glencoe,
: IL: Free
Lange, Oskar. 1938. "On the Theory of Economic SO- Press.
1 06 Williamson
Miles, Ray, and Charles Snow. 1992. "Causes of Failure Scherer, F. M. 1970. Indust~pialMarket Structure and
in Network Organizations." California Manage- Economic Performance. 1st ed. Chicago: Rand Mc-
ment Review 34:5 3-72. Nally.
Milgrom, Paul, and John Roberts. 1992. Economics, Or- Schmid, Allan. 1972. "Analytical Institutional Eco-
ganization and Management. Englewood Cliffs, nomics." American Journal of Agricultural Eco-
NJ: Prentice-Hall. nomics 54:893-901.
Moe, Terry. 1990. "Political Institutions: The Ne- Schotter, Andrew. 1981. The Economic Theory of Social
glected Side of the Story: Comment." Journal of Institutions. New York: Cambridge University
Law, Economics, and Ovganization 6:213-54. Press.
Mokyr, Joel. 1990. The Lever of Riches. New York: Ox- Schumpeter, Joseph A. 1942. Capitalism, Socialism,
ford University Press. and Democracy. New York: Harper & Brothers.
Nelson, Richard R., and S. G. Winter. 1982. A n Evolu- -. 1947. "The Creative Response in Economic
tionary Theory of Economic Change. Cambridge, History." Journal of Economic History 7:149-59.
MA: Harvard University Press. -. [1951] 1989. Essays on Entrepreneurs, Innova-
Newell, Alan, and Herbert Simon. 1972. Human Prob- tions, Business Cycles, and the Evolution of Capital-
lem Solving. Englewood Cliffs, NJ: Prentice-Hall. ism. New Brunswick, NJ: Transaction Publishers.
North, Douglass. 1986. "The New Institutional Eco- Scott, W. Richard. 1992. "Institutions and Organiza-
nomics." Journal of Institutional and Theoretical tions: Toward a Theoretical Synthesis." Stanford
Economics 142:230-37. University, Stanford, CA. Unpublished manuscript.
-. 1991. "Institutions." Journal of Economic Per- Selznick, Philip. 1949. TKA and the Grass Roots. Berke-
spectives 5:97-112. ley: University of California Press.
North, Douglass, and Robert Thomas. 1981. The Rise -. 1957. Leadership in Administration. New
of the Western World. Cambridge: Cambridge Uni- York: Harper & Row.
versity Press. Shapiro, Carl. 1989. "The Theory of Business Strat-
Palay, Thomas. 1984. "Comparative Institutional Eco- egy." Rand Journal of Economics 20:125-37.
nomics: The Governance of Rail Freight Contract- Sheard, Paul. 1989. "The Main Bank System and Cor-
ing." Journal of Legal Studies 13:265-88. porate Monitoring in Japan." Journal of Economic
-. 1985. "The Avoidance of Regulatory Con- Behavior and Organization 11: 3 9 9 4 2 2 .
straints: The Use of Informal Contracts." Journal Shelanski, Howard. 1991. "Empirical Research in
of Law, Economics, and Organization 1:155-75. Transaction Cost Economics: A Survey and Assess-
Parsons, Talcott, and Neil Smelser. 1956. Economy and ment." University of California, Berkeley. Unpub-
Society. New York: Free Press. lished manuscript.
Penrose, Edith. 1959. The Theory of Growth of the Firm. . 1993. "Transfer Pricing." University of Cali-
New York: John Wiley & Sons. fornia, Berkeley. Unpublished Ph.D. dissertation.
Perrow, Charles. 1992. "Review of the New Competi- Simon, Herbert. 1957a. Admintstrative Behavior. 2d
tion." Administrative Science Quarterly 37:162-66. ed. New York: Macmillan.
Pfeffer, Jeffrey. 1981. Power in Organizations. Marsh- . 1957b. Models of Man. New York: John Wiley
field, MA: Pitman Publishing. & Sons.
Polanyi, Michael. 1962. Personal Knoudedge: Towards a . 1978. "Rationality as Process and as Product of
Post-Critical Philoso~hy. New York: Harper & Thought." American Economic Review 68: 1-16.
Row. . 1983. Reason in Human Affairs. Stanford,
Posner, Richard. 1993. "The New Institutional ECO- CA: Stanford University Press.
nomics Meets Law and Economics." Journal of In- . 1985. "Human Nature in Politics: The Dia-
stitutional and Theoretical Economics 149:73-87. logue of Psychology with Political Science." Amer-
Putterman, Louis. 1984. "On Some Recent Explana- ican Polttical Science Review 79:293-304.
tions of Why Capital Hires Labor." Economic In- . 1991. "Organizations and Markets." ~ournal
quiry 22:171-87. of Economtc Perspectives 5 : 2 5 4 4 .
Robbins, Lionel, ed. 1933. The Common Sense of Politi- Stigler, George J. 1968. The Ovganization of Industry.
cal Economy, and Selected Papers on Economic The- Homewood, IL: Richard D. Irwin.
ory, by Philip Wicksteed. London: G. Routledge . 1983. Comments on "The Fire of Truth: A Re-
and Sons, Ltd. membrance of Law and Economics at Chicago,
Robinson, E. A. G. 1934. "The Problem of Manage- 1932-1970," edited by Edmund W. Kitch. Jour-
ment and the Size of Firms." Economic Journal nal of Law and Economics 26:163-234.
44:240-54. Summers, Lawrence, and Andrei Shleifer. 1988.
Robinson, James C. 1992. "A New Institutional Eco- "Breach of Trust in Hostile Takeovers." Pp. 38-55
nomics of Health Care." University of California, in Corporate Takeovers, edited by Alan ~ u e r b a c h .
Berkeley. Unp.ublished manuscript. Chicago: University of Chicago Press.
Schelling, Thomas C. 1978. Micromotives and Mac- Sundaram, Anant, and J. Stewart Black. 1992. ''The
robehavior. New York: Norton. Environment and Internal Organization of ~ u l t i -
 Transaction Cost Economics 107

national Enterprise." Academy of Management Re- . 1983. "Credible Commitments: Using Hos-
view 17:729-57. tages to Support Exchange." American Economic
Swedberg, Richard. 1987. "Economic Sociology: Past Review 73:519-40.
and Present." Current Sociology 35 :1-22 1. . 1985. The Economic Institutions of Capitalism.
. 1990. Economics and Sociology. Redefining New York: Free Press.
Their Boundaries: Conversations with Economists -. 1988a. "The Logic of Economic Organiza-
and Sociologists. Princeton, N J: Princeton Univer- tion." Journal of Law, Economics, and Organiza-
sity Press. tion 4:65-93.
-. 1991. "Major Traditions of Economic Sociol- -. 1988b. "The Economics and Sociology of Or-
ogy." Annual Review of Sociology 17:251-76. ganization: Promoting a Dialogue." Pp. 159-85 in
Teece, David J. 1986. "Profiting from Technological Industries, Firms, and Jobs, edited by George Far-
Innovation." Research Policy 15:285-3 05. kas and Paula England. New York: Plenum.
Teece, David J., Gary Pisano, and Amy Shuen. 1990. -. 1989. "Internal Economic Organization." Pp.
"Firm Capabilities, Resources, and the Concept of 7-48 in Perspectives on the Economics of Organiza-
Strategy." University of California, Berkeley. Un- tion, edited by Oliver E. Williamson, Sven-Erik
published manuscript. Sjostrand, and Jan Johanson. Lund, Sweden: Lund
Van de Ven, Andrew. 1993. "The Institutional Theory University Press.
of John R. Commons: A Review and Commen- -. 1991a. "Comparative Economic Organization:
tary." Academy of Management Review 18:139-52. The Analysis of Discrete Structural Alternatives."
Waldrop, M. Mitchell. 1992. Complexity. New York: Administrative Science Quarterly 36:269-96.
Simon & Schuster. -. 1991b. "Economic Institutions: Spontaneous
Weick, Karl E. 1977. "Re-Punctuating the Problem." and Intentional Governance." Journal of Law, Eco-
Pp. 193-225 in New Perspectives on Organiza- nomics, and Organization 7:159-87.
tional Effectiveness, edited by Paul S. Goodman and -. 1991c. "Strategizing, Economizing, and Eco-
Johannes M. Pennings. San Francisco: Jossey-Bass. nomic Organization." Strategic Management Jour-
Wernerfelt, Birger. 1984. "A Resource-Based View of nal 12:75-94.
the Firm." Strategic Management Journal 5:171-80. -. 1993a. "The Evolving Science of Organiza-
Williamson, Oliver E. 1968. "Economies as an Anti- tion." Journal of Institutional and Theoretical Eco-
trust Defense: The Welfare Tradeoffs." American nomics 149:36-63.
Economic Review 58:18-35. . 1993b. "Calculativeness, Trust, and Economic
. 1975. Markets and Hierarchies: Analysis and Organization." Journal of Law and Economics
Antitrust Implications. New York: Free Press. 36:453-86.
. 1979. "Transaction-Cost Economics: The Williamson, Oliver, and Sidney Winter, eds. 1991. The
Governance of Contractual Relations." Journal of Nature of the Firm. New York: Oxford University
Law and Economics 22:233-61. Press.
. 1981. "The Economics of Organization: The Zald, Meyer. 1987. "Review Essay: The New Institu-
Transaction Cost Approach." American Journal of tional Economics." American Journal of Sociology
Sociology 87:548-77. 93:701-8.

View publication stats