Examining economic organization through the lens of contract

Transcription

1 Industrial and Corporate Change, Volume 12, Number 4, pp Examining economic organization through the lens of contract Oliver E. Williamson The lens of contract approach to the study of economic organization is partly complementary but also partly rival to the orthodox lens of choice. Specifically, whereas the latter focuses on simple market exchange, the lens of contract is predominantly concerned with the complex contracts. Among the major differences is that non-standard and unfamiliar contractual practices and organizational structures that orthodoxy interprets as manifestations of monopoly are often perceived to serve economizing purposes under the lens of contract. A major reason for these and other differences is that orthodoxy is dismissive of organization theory whereas organization theory provides conceptual foundations for the lens of contract. 1. Introduction As discussed elsewhere (Williamson, 2002a,b), the lens of contract approach to economic organization differs from orthodoxy in three significant respects. First and foremost, the lens of contract focuses predominantly on gains from the trade whereas orthodoxy is focused on resource allocation (prices and output). Second, and related, orthodoxy describes the firm as a production function while the lens of contract describes the firm as a governance structure. Third, the lens of contract is a more microanalytic construction and is more amenable to the lessons of organization theory. This paper adopts the lens of contract perspective. I begin with a sketch of organization theory, both curriculum and research, at the Graduate School of Industrial Administration (GSIA) at Carnegie in the 1960s. I then examine some of the leading lessons of organization theory for the study of economic organization, as viewed through the lens of contract. Consequential differences that accrue upon applying the lens of contract to economic organization are illustrated by three examples in Section 4 and in my discussion of full formalism in Section 5. Concluding remarks follow. 2. Organization theory Like many others, my graduate education involved a discovery of interests. It began with my first year in the MBA program at Stanford (in , before the Stanford Business School underwent a massive curriculum change), where I discovered that Industrial and Corporate Change 12/4 ICC Association 2003 All rights reserved.

2 918 O. E. Williamson economics was a fascinating subject that used much of the same analytical apparatus as engineering. At Jim Howell s suggestion, I took most of my second year courses in the economics and statistics departments at Stanford. Following up on discussions with another member of the Stanford Business School faculty who was a recent graduate of GSIA, Charles Bonini, I moved to Carnegie the following year. As I have indicated elsewhere (Williamson, 1996: ch. 1, 2002c), being a student in the Ph.D. program at GSIA in the late 1950s early 1960s was exhilarating. Not only had Carnegie assembled an extraordinary faculty, but the faculty and students were engaged in interdisciplinary social science research that surpassed anything that I have witnessed before or since. The literature in organization theory is not only huge and diffuse, but some of it is muddled. Having already taken a muddled version of organization theory (in which proverbs of administration were the vehicle for transmitting knowledge) 1 before I came to GSIA, I was not at all sure what to expect when I first sat in on Jim March s organization theory course. Before the end of that first week (and against Jim s advice), I had signed up to take his course for credit. What was especially distinctive about Carnegie is the way that it combined organization theory with the contiguous social sciences. My understanding of and approach to the study of economic organization underwent a vast and permanent change in the process. Although organization theory is a separate field, it is not a separate social science. Instead, it works off of and joins economics, sociology, political science, and aspects of cognitive and social psychology. I tell students of mine who have interests in organizations and institutions that organization theory is foundational, although the benefits vary with what version of organization theory is presented and with who teaches the course. As between rational, natural, and open systems treatments (Scott, 1987), the Carnegie version was mainly a (boundedly) rational and natural systems mix, with emphasis on the former. 2 As for who teaches the course, I would like to have my students take organization theory with James March or his equivalent. Inasmuch, however, as Jim March is one of those almost unique, irreplaceable research workers, teachers, administrators to whom Jacob Marschak (1968: 14) made reference, equivalency is a demanding standard. Be that as it may, students of mine who take organization theory normally come away with lasting benefits. The overarching benefit is their realization that economic 1 Herbert Simon took exception with this tradition in Administrative Behavior (1947: ch. 2). Also see James March and Herbert Simon, Organizations (1958: ch. 2). 2 Both rational and natural systems approaches trace their origins to Chester Barnard (1938). Rational systems theory views administration as a conscious, deliberate, purposeful process (Barnard, 1938: 4). Organization is subject to bounds on rationality, is given to subgoal pursuit, and is influenced by informal organization (Simon, 1947). Natural systems theory makes greater provision for the idea that organization, like the law, has a life of its own. Spontaneous intertemporal transformations thereby arise (Selznick, 1948, 1949, 1950). It is elementary that all significant economic regularities need to be uncovered and the ramifications worked out.

3 Examining economic organization through the lens of contract 919 organization is more complicated, less orderly, and more interesting than they had previously imagined. Rather than confidently pronounce that This is the law here (as revealed, for example, by the disciplined application of price theory), they are much more apt to ask the question What s going on here? when confronted with non-standard or puzzling forms of contract or organization. Such an orientation takes them into the microanalytic mechanisms and intertemporal process transformations of economic organization, both of which are familiar terrain to organization theory. Robert Michels s famous study of Political Parties (1911) led him to discern an intertemporal regularity that had previously gone unnoticed: initial democratic purposes notwithstanding, democratic organizations were often subverted by oligarchy. Such transformations appeared gradually and were attributable to positional advantages that accrued to the leadership: it is organization that gives rise to the dominion of the elected over the elector, of the mandatories over the mandator, of the delegates over the delegatees. Who says organization, says oligarchy (Michels, 1962: 365). Michels did not, however, conclude that democracy should be abandoned for this reason. Rather, upon uncovering the propensity for oligarchy and discerning the mechanisms that were responsible for it, the organizational design ramifications needed to be wrung out: nothing but a serene and frank examination of the oligarchical dangers of democracy will enable us to minimize these dangers (Michels, 1962: 370). Although Ronald Coase was not an organization theorist, he had a natural curiosity about organization and took exception with the propensity of economists to confidently pronounce that puzzling contractual practices and organizational structures had monopoly purpose and effect. Maybe, but maybe not. Indeed, he ascribed the backward state of industrial organization in the 1960s precisely to excessive reliance on monopoly/price theoretic reasoning (Coase, 1972: 67): One important result of this preoccupation with the monopoly problem is that if an economist finds something a business practice of one sort or another that he does not understand, he looks for a monopoly explanation. And as in this field we are very ignorant, the number of ununderstandable practices tends to be very large, and the reliance on a monopoly explanation, frequent. Organization theorists had no such monopoly predilections. Thus although some organization theory interpretations of non-standard practices and structures were whimsical and others tautological (e.g. power ), organization theorists were not predisposed to impose an all-purpose monopoly rationale. It was my good fortune to have my understanding of and curiosity with economic organization vastly transformed by being a part of the GSIA program where organization was viewed as both important and susceptible to analysis. Indeed it could be argued that, in combination

4 920 O. E. Williamson with an economic lens (of a contractual, or game theoretic, or price theoretic kind), organization theory confers an unfair advantage The lens of contract 3.1 General Hal Varian has recently distinguished between important ideas and big ideas and describes Ronald Coase s classic paper on The nature of the firm (1937) as a big idea (Varian, 2002: C2). Although there is widespread agreement on this, the nature of the big idea took a long time to register. Thus, as of 1972, 35 years after the publication of The nature of the firm, Coase described his 1937 article as much cited and little used (1972: 63). It was much cited because it was onto something important, perhaps even big. But it was little used because the big idea was only dimly perceived and/or lacked operationalization (Coase, 1992: ). The essence of the Coasian contribution has been variously described (Williamson, 1994: 202; North, 2000: 37; Werin, 2000: 45). On reflection, and with the benefit of James Buchanan s declaration that mutuality of advantage from voluntary exchange is... the most fundamental of all understandings in economics (Buchanan, 2001: 29), I would say that the overarching big idea is to move from the orthodox lens of choice to bring the lens of contract systematically to bear on economic phenomena of all kinds. For many transactions, of which the make-or-buy decision is one (Coase, 1937), the contractual structure is easily recognized. Other transactions, such as the externality problem (Coase, 1960), needed to be reformulated to bring out their latent contractual features. The orthodox resource allocation approach to economics was famously described by Lionel Robbins (1932: 16) as follows: Economics is the science which studies human behavior as a relationship between ends and scarce means which have alternative uses. Choice is based on two parallel constructions: the theory of consumer behavior, in which consumers maximize utility; and the theory of the firm as a production function, in which firms maximize profit. Economists who work out of such setups emphasize how quantities are influenced by changes in relative prices and available resources. The resulting resource allocation paradigm became dominant for economics throughout the twentieth century (Reder, 1999: 48). Buchanan (1975: 255), however, avers that economics as a discipline went wrong in its preoccupation with the science of choice and the optimization apparatus associated therewith. What was needed was the parallel development of a science of contract. Awaiting this, some phenomena would go unnoticed, others would be poorly understood, and public policy error would result. As perceived by Buchanan, the principal needs for a science of contract were in the 3 David Kreps (1992: 1) contends that: almost any theory of organization which is addressed by game theory will do more for game theory than game theory will do for it.

5 Examining economic organization through the lens of contract 921 field of public finance and took the form of public ordering: Politics is a structure of complex exchange among individuals, a structure within which persons seek to secure collectively their own privately defined objectives that cannot be efficiently secured through simple market exchanges (Buchanan, 1987: 296; emphasis added). For Buchanan and others, thinking contractually about issues where exchange was beset with problems of collective organization led into an examination of the rules of the game. Issues of public choice and constitutional economics are posed (Buchanan and Tullock, 1962; Brennan and Buchanan, 1985). Whatever the rules of the game, the lens of contract is also usefully brought to bear on the play of the game. Much of the latter works through private ordering, which entails self-help efforts by the immediate parties to a transaction to align incentives and craft governance structures that are better attuned to their exchange needs. John R. Commons s prescient statement of the economic problem can be viewed as a precursor to a lens of contract construction: the ultimate unit of activity... must contain in itself the three principles of conflict, mutuality, and order. This unit is the transaction (Commons, 1932: 4). Transaction cost economics not only takes the transaction to be the basic unit of analysis but views governance as the means by which to infuse order, thereby to mitigate conflict and realize mutual gains. As against orthodoxy, issues of organization, rather than technology, become salient. Also, as against faceless transactions and spontaneous order in the market, conflict between bilaterally dependent parties can now arise for which intentional (sometimes hierarchical) support is warranted. Kenneth Arrow s interpretation of the lessons of the market failure literature provides an additional bridge from orthodoxy into comparative contractual analysis. Thus he proposes that market failures be given an organizational interpretation: market failure is a more general category than externality; and both differ from increasing returns in a basic sense, since market failures in general and externalities in particular are relative to the mode of organization, while increasing returns are essentially a technological phenomenon (Arrow, 1969: 48; emphasis added). Accordingly, to ascribe market failures and externalities to technology is unduly restrictive, even wrong-headed, in that many of the problems owe their origins to organization. Furthermore, the importance of transaction costs comes to the fore when market failure is reconceptualized in this (lens of contract) way: market failure is not absolute; it is better to consider a broader category, that of transaction costs, which in general impede, and in particular cases completely block the formation of markets (Arrow, 1969: 48; emphasis added). The incidence and mitigation of transaction costs thus become the objects of analysis. 3.2 Operationalization Examining economic organization through the lens of contract has massive ramifications for industrial organization and more generally. 4 Firms are no longer 4 As Avinash Dixit puts it in his monograph on The Making of Economic Policy (1996: 9): The standard

6 922 O. E. Williamson production functions (a technological phenomenon) but are a mode of organization (a governance structure, to be examined in relation to other feasible modes of governance). Also, absent pre-existing monopoly power (which is the exception rather than the rule), non-standard and unfamiliar forms of contract are presumed to have efficiency rather than monopoly purposes. 5 Transaction cost economizing becomes the main case. This comparative contractual approach to economic organization is implemented through a series of six supporting moves, many of which have organization theory origins. Human actors. If Nothing is more fundamental in setting our research agenda and informing our research methods than our view of the nature of the human beings whose behavior we are studying (Simon, 1985: 303), then economists and other social scientists are well-advised to describe the attributes of the human actors on which their analyses rest. Transaction cost economics names three: cognition, self-interest, and foresight. Bounded rationality is the cognitive assumption, the main economic import of which is that all complex contracts are unavoidably incomplete. The self-interestedness assumption is that of opportunism, on which account parties to a long-term contract will contemplate defection from the spirit of a contract and revert to self-interested bargaining when a contract is pushed out of alignment by significant disturbances. Issues of strategizing that orthodoxy had suppressed arise when opportunism is introduced (Makowski and Ostroy, 2001). Feasible foresight assumes that parties to a contract have the capacity to look ahead, uncover the salient hazards, ascertain the mechanisms through which they work, and fold these back into the ex ante design of governance. 6 Ludwig von Mises s description of entrepreneurs as those who have more initiative, more venturesomeness, and a quicker eye than the crowd normative approach to policy analysis views... [the] process as a social-welfare-maximizing black box, exactly as the neoclassical theory of production and supply viewed the firm as a profit-maximizing black box. While some useful insights follow from this, it leaves some very important gaps in our understanding and gives us some very misleading ideas about the possibilities of beneficial policy intervention. Economists studying business and industrial organization have long recognized the inadequacy of the neoclassical view of the firm and have developed richer paradigms and models based on the concepts of various kinds of transactions costs. Policy analysis also stands to benefit from such an approach, opening the black box and examining the actual workings of the mechanism inside. 5 Efficiency is always examined in a comparative institutional way in relation to alternative feasible forms of organization, all of which are flawed. The remediableness criterion applies (Williamson, 1996). 6 As Richard Dawkins (1976: 20) put it, the capacity to simulate the future in imagination... [saves] us from the worst consequences of the blind replicators. Foresight is nonetheless limited and is often combined with perspicacious hindsight (learning). If things went wrong, what were the reasons and how can the problems be mitigated in the future? But this is also symmetrical. If things went right, what were the reasons and how can the benefits be enhanced?

7 Examining economic organization through the lens of contract 923 (von Mises, 1949: 255) is in this spirit, as is Rudolf Spreckels remark that When I see something badly done or not done at all, I see an opportunity to make a fortune. Unit of analysis. The natural unit of analysis for lens of contract purposes is the transaction. Naming a unit of analysis is easier, however, than identifying the critical dimensions for describing the unit of analysis as witness the fact that the key attributes for so many would-be units of analysis are never identified. 7 Awaiting dimensionalization, transaction cost economics remained a largely tautological construction. To be sure, transactions can be variously described depending on the purpose. Three dimensions that have pervasive ramifications for governance are: asset specificity (which is a measure of non-redeployability and takes a variety of forms physical, human, site, dedicated, brand name); the unanticipated disturbances to which transactions are subject (which pose the need for adaptations, 8 whereupon maladaptation costs are incurred should bilaterally dependent parties decline to adapt cooperatively unless the contract is renegotiated ); and the frequency with which transactions recur (which bears both on the efficacy of reputation effects in the market and the incentive to incur the cost of specialized governance). The absence of asset specificity describes the ideal transaction in law and economics for which competition works well: sharp in by clear agreement; sharp out by clear performance (Macneil, 1974: 734). As asset specificity builds up, bilateral dependency sets in and, in combination with uncertainty (which pushes incomplete contracts out of alignment), the aforementioned contractual complications appear. Intertemporal transformations. The idea that bilateral dependency sets in as asset specificity builds up is due to the fundamental transformation, to wit: although there may be a large number of qualified suppliers at the outset, before a contract is awarded and transaction-specific investments are made, parity between the winning bidder and all others can no longer be presumed once investments in durable-specific assets are in place. That is because the buyer and the initial winning bidder now have a continuity interest. Not only would the supplier experience a loss of productive value upon redeploying these assets to best alternative uses, but the buyer could not procure the good or service from generic sources except at a greater cost. A large numbers bidding 7 Examples of would-be units of analysis for which operational content is missing include the role (see Simon s critique, 1957a: xxx), the decision-premise (which is Simon s candidate, but which has found little application outside of cognitive psychology (Newell and Simon, 1972), and the routine, which has been worked up for special purposes (Cyert and March, 1963) but not more generally (Nelson and Winter, 1982). 8 As March and Simon (1958: 159) observe: Interdependence by itself does not cause difficulty if the pattern of interdependence is stable and fixed. For, in this case, each subprogram can be designed to take account of all the subprograms with which it interacts. Difficulties arise only if program execution rests on contingencies that cannot be predicted perfectly in advance. The need for adaptive, sequential decision making is posed by the latter.

8 924 O. E. Williamson competition at the outset thus gets transformed into a bilateral exchange relation during contract implementation and at the contract renewal interval. The efficacy of simple market contracting is compromised as a consequence. Internal organization also undergoes intertemporal transformations. The aforementioned condition of oligarchy to which Michels called attention is one example. This is only one manifestation, however, of the proposition that organization has a life of its own (Selznick, 1949, 1950). Identifying the key intertemporal transformations and working out their comparative institutional significance are vital to an understanding of the governance of contractual relations. Bureaucratization (Williamson, 1985: ch. 6) and informal organization (Barnard, 1938; Simon, 1991) are especially important intertemporal phenomena. Learning is yet another intertemporal process that has important performance ramifications. One of the issues here is whether learning in organizations is subject to biases, such as myopia (Levinthal and March, 1993). As with bureaucratization, the challenge is to discover the biases, uncover the mechanisms, and work out the comparative institutional ramifications. Note, moreover, that the opportunities for learning and the ramifications that accrue thereto are conditional, among other things, on the attributes of transactions and pre-existing conditions of governance. The organization theory literature often describes intertemporal transformations under the rubric of path dependency, especially as it relates to a comparison of first movers with later entrants. First movers often enjoy an advantage in such comparisons because later entrants need to incur start-up costs of an out-of-pocket kind whereas such costs for first movers are historical. A common error is to infer inefficiency from the fact that a later entrant with a superior technology (start-up costs aside) cannot profitably enter. Transaction cost economics advises that the efficiency consequences of intertemporal transformations/path dependencies be judged with reference to the remediableness criterion. 9 Operational purpose: adaptation. Interestingly, both the economist Friedrich Hayek (1945) and the organization theorist Chester Barnard (1938) were in agreement that adaptation is the central problem of economic organization. Hayek focused on the adaptations of economic actors who adjust spontaneously to changes in the market, mainly as signaled by changes in relative prices: Upon looking at the price system as... a mechanism for communicating information, the marvel of the market resides in how little the individual participants need to know to be able to take the right action (Hayek, 1945: ). By contrast, Barnard featured coordinated adaptation among economic actors working through administration (hierarchy). The latter is accomplished not spontaneously but in a conscious, deliberate, purposeful way (Barnard, 1938: 9). Prices no longer serve as sufficient statistics. Instead, a wider range of 9 The remediableness criterion holds that an extant mode of organization for which no feasible superior alternative can be described and implemented with expected net gains is presumed to be efficient. Such efficiency is, however, a rebuttable presumption (Williamson, 1996).

9 Examining economic organization through the lens of contract 925 information is gathered and shared, otherwise divergent expectations are supplanted by a common information base from which common projections are made (Malmgren, 1961), coordinated investments and operating responses are also reached, contingent plans for revisiting the issues are worked up, and disputes are settled not in the courts but internally. The adaptive properties of markets and hierarchies differ for each of these reasons. The upshot is that the autonomous adaptations to which Hayek refers, whereby individual parties respond to market opportunities as signaled by changes in relative prices, need to be distinguished from the cooperative adaptations of concern to Barnard, accomplished through administration within the firm. Because a high performance economic system will display adaptive capacities of both kinds, an understanding and appreciation for both markets and hierarchies (which differs from the ideological dichotomy of markets or hierarchies) is needed. There is no ideological predilection for either firm or market. Both are described as modes of governance. The lens of contract, as against the lens of choice, becomes the cutting edge. One of the advantages of focusing on adaptation is that it brings added meaning to the idea of mutual gain. It is elementary that gains from trade will always be realized by moving onto the contract curve. But how is this to be accomplished in a world where complex contracts are incomplete and are implemented over time in the face of disturbances for which contingent provisions either have not been made or, if made, are often in error? More attention to the choice of governance structures that have good adaptive properties (and less to concentrating all of the action in the ex ante incentive alignment stage) is one of the central lessons of viewing economic organization through the lens of incomplete contracting. Governance structures. Examining economic organization through the lens of contract not only places the spotlight on ex post adaptation but, in the process, gives prominence to the role of governance. Specifically, transaction cost economics holds that each generic mode of governance is defined by a syndrome of internally consistent attributes to which different adaptive strengths and weaknesses accrue. Such a description is broadly consonant with Simon s view that organizations differ in discrete structural ways (1978: 8). Discrete structural analysis differs from marginal analysis, in that the former recognizes differences in kind whereas the latter involves differences in degree. But there is more. Not only is mode A governance unable to replicate a particular attribute (e.g. incentive intensity) of mode B, but mindless efforts at piecemeal replication can compromise the integrity of mode A. An examination of the three governance structures shown in Table 1 spot market, hybrid and hierarchy is instructive. As shown, the rudimentary attributes for describing a governance mode (for transaction cost economics comparisons) are incentive intensity, administrative control and contract law regime, where each can take on any of three values: much (++), some (+) and little (0). Given that each attribute can take on any of three values and that there are three attributes to consider, there are 27

10 926 O. E. Williamson Table 1 Attributes that define three viable modes of governance Governance structures Governance attributes Incentive intensity Administrative control Contract law regime Spot market Hybrid Hierarchy possible combinations. What is distinctive about the three combinations shown in Table 1 is that these are complementary configurations (Williamson, 1991a). Market and hierarchy, it will be noted, are polar opposites. Thus whereas markets engage high-powered incentives, have little administrative apparatus, and settle disputes in a legalistic way (in the courts), hierarchy works out of low-powered incentives, has considerable administrative apparatus, and settles disputes internally (the firm is its own court of ultimate appeal). Given these governance differences, markets enjoy the advantage in autonomous adaptation respects while the advantage accrues to hierarchy for cooperative adaptation. The hybrid is a compromise mode of governance for managing bilateral dependency. Discriminating alignment. Firms, markets, hybrids, bureaux, etc. are all alternative modes of governance. What distinguishes these governance structures is that each deploys a different syndrome of attributes. As described in Unit of analysis above, transactions also differ in their attributes. The discriminating alignment hypothesis out of which transaction cost economics works is this: transactions, which differ in their attributes, are aligned with governance structures, which vary in their cost and competence, so as to effect a (mainly) transaction cost economizing result. 10 The basic regularity is that as asset specificity (which gives rise to bilateral dependency) increases, and as disturbances (which push the parties off of the contract curve) become more consequential, the needs for cooperative adaptation increase. The efficient governance response to added needs for cooperative adaptation is to first move transactions from spot markets to hybrid contracting and, if unmet needs for added coordination persist, to hierarchies. Generic transactions are thus those for which markets are well suited; complex transactions are managed by hierarchy; and hybrid modes of governance are employed 10 In effect, transaction cost economics assumes that the attributes of any particular transaction are held constant in deciding upon which is the least cost mode of governance. That is a simplifying assumption and can be thought of as a means by which to focus on first-order effects. In fact, transaction attributes and governance structures are interactive, but this is ordinarily a second-order effect (Riordan and Williamson, 1985).

11 Examining economic organization through the lens of contract 927 for those in between. This pattern applies, moreover, not merely to transactions in intermediate product markets but to any issue that arises or can be reconceptualized as a contracting problem. This is gratifying, in that a fundamental hypothesis of science is that appearances are deceptive and that there is a way of looking at or interpreting or organizing the evidence that will reveal superficially disconnected and diverse phenomena to be manifestations of a more fundamental and relatively simple structure (Friedman, 1953: 33). 4. Some comparisons Because many issues arise as and many others can be reconceptualized in contracting terms, the lens of contract approach to economic organization has broad application. Issues to which the lens of contract has been brought to bear (and contrasted with rival interpretations) are set out in Table 2. These are merely suggestive, moreover. 11 What follows is a sketch of four applications: franchise bidding, game theory, public bureaux, and incomplete contracting. The first three of these are examined in this section. Incomplete contracting is examined in Section 5. The recurrent argument throughout is this: because the purported efficacy of all consequential mechanisms is everywhere regarded as problematic, the microanalytics of the relevant mechanisms of governance need to be named and explicated, whereupon the plausibility of each is examined in a hard-headed way. 4.1 Franchise bidding for natural monopoly Franchise bidding for natural monopoly is an imaginative application of contractual reasoning to the problem of natural monopoly that Milton Friedman once described as a choice among three evils: there is unfortunately no good solution to technical monopoly. There is only a choice between three evils: private monopoly, private monopoly regulated by the state, and government operation (Friedman, 1962: 128). Harold Demsetz (1968) was not persuaded and advanced a fourth possibility: use the state to organize an ex ante competition for the market. Demsetz introduced an important intertemporal distinction between an ex ante condition of competition and an ex post condition of monopoly. Ex ante, there are large numbers of qualified bidders. Ex post, it is efficient for one firm to supply the market by reason of economies of scale. Upon making this distinction, the new solution to the problem of natural monopoly is to run an ex ante bidding competition and award the monopoly franchise to the firm that offers to supply on the best (lowest price) terms. Demsetz illustrates the argument with reference to bidding for automobile license plate manufacture and suggests that the proposal has general application. The Demsetz argument for using ex ante competition for the market as a solution to 11 Still other issues to which the lens of contract has been applied include agricultural contracting practices, the organization of health care, business strategy, the elusive concept of trust, and the list goes on.

12 928 O. E. Williamson Table 2 Rival interpretations Phenomenon or condition Alternative interpretations Lens of contract Rival lens Firm Governance structure (organizational) Production function (technological) Contractual completeness Incomplete Complete Vertical integration Organizational response to contractual hazards Correct against inefficient factor proportions Vertical market restrictions To mitigate subgoal pursuit Monopoly purpose or muscle (power) Labor organization Governance arises to infuse Monopoly unionism credibility Organization of work Economizing purposes Power or abuse Finance Asset based financing Modigliani Miller Natural monopoly No one all-purpose best response All-purpose franchise bidding Public bureaux As efficient response to some Intrinsically inefficient transactions Contract law Varies among governance structures One, single all-purpose legalistic form Economic reform Gradualist Big bang Reputation effects Problematic efficacy Reliably efficacious natural monopoly operates at a high level of generality. The automobile license plate example, moreover, is merely a sketch. As Lester Telser puts it (1971: 364), the proposal works out of a vaguely described bidding process, the specifics of which would need to be addressed if actual applications were to be attempted. The timely issue of cable television was very shortly thereafter addressed by Richard Posner (1972), who proposed to implement franchise bidding for the right to serve the cable television market through the use of three mechanisms: (i) a mechanism for ascertaining the preferences of customers (thereby to define the optimal package); (ii) a mechanism to value non-redeployable assets (thereby to facilitate the objective and undisputed transfer of ownership among parties); and (iii) a mechanism to conduct periodic rebidding at fixed intervals. Rather, however, than engage the microanalytics of any of these three mechanisms, Posner took the position that to expound the details of particular regulations and proposals... would serve only to obscure the basic issues (Posner, 1972: 98). Perhaps that is because the relevant details are not self-evident without the assistance of a focused lens and Posner s favored lens (textbook orthodoxy) is poorly suited to

13 Examining economic organization through the lens of contract 929 examine contractual issues. Be that as it may, Posner blithely asserts that his wellintentioned but scantily described bidding framework suffices. Alas, the cost of good intentions are often great (Morris, 1980). Upon examining Posner s sparingly described mechanisms through the lens of transaction cost economics by making note of the attributes of the cable television transaction (with special reference to asset specificity and uncertainty), ascertaining potential contracting problems that accrue thereto, and uncovering the weaknesses of the customer solicitation, asset valuation, and periodic rebidding mechanisms described by Posner (to say nothing of his neglect of politicization hazards) I concluded that serious problems would be encountered if franchise bidding for cable television were implemented in this way in the 1970s (Williamson, 1976). (An accompanying case study of franchise bidding in Oakland, moreover, disclosed that my concerns were not merely hypothetical.) This is not to say that franchise bidding will never work well, hence is wholly lacking in merit. On the contrary, it is an imaginative proposal. Those, however, who would prescribe franchise bidding for real world (as against toy) problems need to engage the microanalytics, thereby to mitigate the hazards and delimit the circumstances. The recent award of a long-term electricity supply contract by London Underground Ltd (Littlechild, 2002) is a constructive real world example. 4.2 Game theory Not only is there broad agreement that game theory is a powerful mode of analysis, but some view it as the theory with which to study the economics of organization (Gibbons, 2000) and strategic behavior (Shapiro, 1989). My arguments are that (i) the mechanisms through which game theory works are sometimes invoked uncritically and (ii) game theoretic appeals to organization, as the means by which to reshape otherwise bad games, are frequently truncated. The efficacy of reputation effects. What is commonly referred to as the trust game is a one-sided version of the prisoner s dilemma in which party A moves first and decides whether to put himself at hazard ( trust B ) or not ( do not trust B ). If party A accepts the hazard, then party B decides whether to take advantage of A ( abuse A s trust ) or not ( honor A s trust ). The payoffs are such that the joint gain is maximized by the trust/honor outcome. But since B s immediate gains are maximized if he abuses A s trust, the no trust/no trust result will obtain if played as a one-shot game. Consider, however, a variant in which there is a high probability that each round of play will be followed by another. If, in these circumstances, A says to B, I will begin by trusting you, hoping that you will honor that trust. Indeed, I will continue to trust you as long as you do not abuse that trust. But if ever you abuse that trust, I will never again trust you, if B hears and believes that statement, and if the game is played repeatedly (with high probability), then the honor-trust arrangement is self-enforcing (Kreps, 1990a: ). The argument, moreover, can be extended to a sequence of As who successively deal

14 930 O. E. Williamson with the same B and a further extension is possible by creating a succession of B s in which each successor B buys the reputation of his predecessor. Repeated play in combination with use of a simple trigger strategy if ever you abuse trust, I will never trust you again 12 thus appears to salvage what would otherwise be another tragic failure of opportunism. But while simplicity is a virtue, that which is simple may not always be plausible. The fundamental question to be asked is this: When do simple trigger strategies work well and when poorly? Examining this brings in the microanalytics. For what types of transactions? For what types of populations? For what types of information assumptions? For what institutional environments? David Kreps (1990a) notes that there are inference problems if A observes only his own payoff, rather than B s actions, and if payoffs are probabilistic. Unforseeability, moreover, compounds the difficulties: When is B judged to be out of compliance if a contingency arises that has not been foreseen? Bernard Williams (1988) also expresses precautions over game theoretic treatments of trust; and I name 10 additional factors that put the efficacy of trigger strategies in question (Williamson, 1991b). Ignoring these considerations leads to undue reliance on analytically convenient but sometimes implausible reputation effect mechanisms. Bad games and strategizing. The prisoners dilemma is much the most familiar example of the coercive logic of game theory. The game involves two criminals who have been apprehended and face jail time of, say, three years if both persist with claims of innocence. Being smart, the police confront each criminal separately with an inducement to confess: if you confess and the other does not, you will receive a light sentence (say, 1 year) while the other will get a draconian sentence of 15 years. But should both confess, each is sentenced to 10 years. Unable to coordinate, the dominant strategy for each criminal is to confess. The coercive logic of the game prevails. The cops win. This, however, as many others have observed, is a truncated presentation. Once again, playing the prisoners dilemma as a repeated game introduces new options, of which organization is one. For example, the Mafia is commonly interpreted as a means by which to introduce contingent payoffs that defeat, as it were, the dilemma. Thus suppose that criminals who are admitted to membership in the Mafia are not only advised never to confess but are informed that the Mafia will impose penalties of its own on those who confess. Specifically, should both confess, both criminals are considered unreliable and are thereafter denied membership in the Mafia. If only one confesses, and the other holds out, the defector is punished by the Mafia as soon as he is released from his light sentence whereas the holdout is promised aid for his family and 12 Not only is a trigger strategy easy to describe but it is also easy to understand. If, therefore, A is ever placed in the position of being the first to move in a trust game, A s decision to trust B will be accompanied by an unambiguous declaration that A is a trigger strategy player. This being heard by B, and the outcome of the game being known to all interested outsiders (who are potential traders with B and are also trigger strategists in relation to previous plays, with them or other As), B has more powerful incentives to honor trust.

15 Examining economic organization through the lens of contract 931 legal assistance to get his sentence reduced. The cooperative strategy (neither confesses) can thus be induced by thinking ahead and embedding the game in a larger set of contingent payoffs, whereupon the prisoners dilemma vanishes. Accordingly, just as parties to a complex contract can often look ahead, identify potential hazards, uncover the mechanisms, and fold these back into the ex ante design of governance, so too can some bad games be reshaped. 13 Game theory, so construed, is much more than a sophisticated tool for working out the strategic interactions of the game as initially presented. Once the mechanisms that generate (possibly unwanted) outcomes are displayed, participants who now understand the strategic situation of which they are a part will be better able to reshape bad games than will those who are not so informed. These parallels between reshaping games and contracts notwithstanding, the contractual approach usually pushes beyond the game theory approach in its uses of organization theory reasoning. In contrast with many game theory accounts which treat the Mafia as an efficacious solution to the prisoners dilemma, the governance approach maintains that all feasible forms of organization are flawed, hierarchy (the Mafia) included. Accordingly, governance draws on organization theory to uncover added significant regularities to include weaknesses that accrue to organization. Questions as to where and why the Mafia will work well and poorly thus arise under the governance approach that are ordinarily ignored by game theory. (Recall in this connection that governance views hierarchy as the organization form of last resort. Rather, therefore, than jump immediately to the Mafia, the governance approach also examines the comparative efficacy of intermediate forms.) Here as elsewhere, the action resides in the details of transactions on one hand and governance structures on the other. Note, moreover, that the (usual) mind-set of strategizing differs from that of contracting. Strategizing commonly involves clever ploys and positioning, whereby one strategic actor imposes its will upon another. Contract, by contrast, places emphasis on voluntary exchange with the prospect of mutual gains. Although some might construe 13 Sometimes this may be unrealistic, which is Kreps s response to a query that is posed in the final pages of his Clarendon lectures (1990b: 182): Whenever a game theoretic analysis of some situation is presented to the economics department at Tel Aviv University, one of the faculty there... can be counted on to object, If the players of this game are so smart, why are they playing this silly game? Why don t they change the rules and play a game where the can do better? One explanation is that people play silly games because they are not quite so smart as we typically assume in our analyses and because rules that are socially embedded experience inertia (Kreps, 1990b: ). My contention is that some silly games actually are reshaped (enlarged), whereupon the players commonly do better. Thus consider the dating game where X and Y plan to have dinner and agree that X will buy the wine while Y will buy the main course, each on the way back from work. Both prefer red wine with steak and white wine with chicken, but X prefers the first combination, and Y prefers the latter. Failing to coordinate in advance, they face a dilemma. New communication technologies (cell phones) could relieve the dilemma. A pre-cell phone solution would be for X to think ahead and buy both red and white wines, opening the one that is appropriate, and reminding Y that the backup wine will be opened at the next dinner.

16 932 O. E. Williamson this as taking contract out of the strategic arena, that would be too hasty. Consider predation. The strategic view is that incumbents can use predation to deter entry (if the requisite structural preconditions are satisfied). The contractual view is that inefficiency invites its own demise provided that a cost-effective response can be devised. For example, faced with the prospect that established firms will contingently cut prices to deter entry, only to restore prices after an entrant has been vanquished, disadvantaged customers and would-be entrants appear to be helpless pawns. Not necessarily. Sometimes customers and entrants can commit to pre-entry contracts, thereby better to assure the viability of new entry [the airlines industry being a recent illustration (Brannigan, 2002)]. The uses of contractual reasoning to better understand complex economic organization are endless. 4.3 Public bureaux The public sector counterpart for describing firms as profit-maximizing black boxes, which transform inputs into outputs according to the laws of technology, is to describe public bureaux as omniscient, omnipotent, and benevolent actors (Dixit, 1996: 8). Omniscience annihilates the information asymmetries that might otherwise plague public sector transactions; omnipotence annihilates implementation problems; and benevolence annihilates subgoal pursuit. If, however, public bureaux are described in this way, why is not the government bureau a superior, all-purpose form of organization? That will be recognized as a variant of the Coasian question, Why is not all production organized in one large firm? (Coase, 1937: 394). Transaction cost economics responds to both of these questions identically: because hypothetical (ideal) forms of organization are fanciful, open the black box and examine the actual workings of the mechanisms inside. Specifically, examine these mechanisms in a comparative institutional way. As it turns out, the public bureau, like other modes of governance has (comparative) strengths and weaknesses whereupon it is well suited to some transactions and poorly suited to others. Thus, although the public bureau displays very low-powered incentives and experiences added bureaucratic burdens in comparison with private bureaux, there are some especially complex transactions (of which foreign policy administration is an example) for which the public bureau is the least cost mode (Williamson, 1999). The common error to be avoided is to pronounce that governance structures are efficient or inefficient without reference to the transactions which is responsible for the mistaken mantra that spot markets are superior to firms which are superior to public bureaux. The logic of discriminating alignment reveals, however, that there are some transactions for which the firm is superior to the market, its bureaucratic disabilities notwithstanding. Similarly, its even greater bureaucratic disabilities notwithstanding, there is a place for the public bureau within the logic of organization. More generally,