The Empirical Preconditions for Exploitation

In his treatise on the current state of the new economics of organizations, or transactions cost analysis, Oliver Williamson identifies two behavioral assumptions on which this research program rests and three dimensions along which transactions can vary: the two behavioral assumptions are bounded rationality and opportunism; the three dimensions are uncertainty, frequency, and asset specificity.

Bounded rationality. Standard neoclassical economics maintains that eco­nomic actors always maximize. This facilitates a formal treatment of a wide range of phenomena, but it is unrealistic as a description of existing human beings. (Whether or not that is a problem for neoclassical analysis is a con­troversial issue that will not be taken up here.) By contrast, Herbert Simon’s notion of bounded rationality comes closer to a description of what human beings are really like: it assumes that human beings are “intendedly rational but only Hmitedly so” (1961, xxiv).

What makes the assumption of bounded rationality (as well as the assump­tion of opportunism to be discussed shortly) so attractive is not simply its greater realism. It also suggests that human organizations and the rules or contracts by which they are defined can be understood as instrumentalities that are more or less useful for coping with these limitations of the human condition. This perspective on organizations is fundamental to transactions cost analysis and constitutes one of its great strengths. It invites the analyst to ask evaluative questions of a comparative nature, that is, how one type of organization compares to another as an instrumentality for dealing with a range of problems that would not exist but for various human limitations. This contrasts with the approach suggested by neoclassical economics, which holds up unrealizable ideals against which reality is to be measured.¹⁶

Opportunism. The other behavioral postulate of transactions cost analysis is opportunism. Specifically, the assumption is that most people are some­times given to opportunistic behavior. What is opportunistic behavior? Williamson’s definition—self-interest seeking with guile (1985, 47)—is sug­gestive but not very clear. To get a more helpful definition, consider the range of phenomena to which this term refers. Opportunism does not merely—or even primarily—consist in, lying, stealing, cheating, and bribing (though those are instances of the phenomenon par excellence).

As the term suggests, opportunism consists in the seizing of an opportunity. Specifically, it consists in seizing an opportunity for advancement (primarily for oneself but sometimes for others on whose behalf one acts as agent) at the expense of others not provided for in some contractual relation. This is what all of the listed behaviors have in common. Contracts (exchanges) are made against a background of legitimate expectations about how people will behave. The legitimacy of these expectations is partly normative and partly descriptive. In other words, the expectations derive, in part, from widely shared beliefs about how people ought to behave and in part from beliefs about how people do, in fact, behave. Consequently, opportunistic behavior includes not only vio­lations of implicit and explicit provisions of the contract but also the violation of any legitimate expectation surrounding the transaction.

Two factors that complicate the problem of dealing with opportunistic behavior are that the penchant for it varies from one person to another and that it is not “stamped on their foreheads”; that is, not only do people differ in their penchant for opportunism, but those differences are, for the most part, not public information. This is why the threat of shirking—or oppor­tunistic behavior generally—cannot be handled by a simple adjustment in the terms of the contract.

As indicated, the assumption that drives transactions cost analysis is that most people are sometimes given to opportunistic behavior. It would be implausible to assume that everyone is always on the lookout for a chance to act opportunistically, nor will everyone push to extract the maximum bene­fit from whatever chances are presented. This suggests that opportunistic behavior can vary along two dimensions: frequency and utilization. Clearly, some people are more inclined to seize opportunities than are others. And, people differ in the extent to which they will utilize, or take advantage of, the opportunities with which they are presented. Though perhaps most people will take small advantage of a large opportunity (the temptation being so great), some people will take maximum advantage of every opportunity, even the smallest. These are the quintessential parasites of human society.

There is, of course, a general moral prohibition against opportunistic behavior. How frequently people act opportunistically is a function of (1) their penchant for opportunism, (2) the power that moral prohibition has over them, (3) their capacity for clear-headed thinking about what they are doing, and (4) the opportunities to violate this prohibition that transactions present. The first two of these factors are self-evident, but the latter two require comment.

Presented with a chance to act opportunistically, people often interpret the situation so that it does not appear that way. This is illustrated in a com­mon saying among the workers in what used to be Yugoslavia, “They pre­tend to pay us, so we pretend to work.” In this way, people can rationalize opportunistic behavior as nonopportunistic. The more given someone is to such rationalizations, the more she will act opportunistically.

On the other point, transactions can be more or less vulnerable to oppor­tunistic behavior. The employment contracts and associated governance structures characteristic of universities and government bureaucracies gen­erally do a very poor job of monitoring workers. Part of the problem is that individual contributions (output) are inherently difficult to identify, but the problem is often compounded by institutional failure even to monitor inputs (e.g., professors “working at home”). These organizations are, therefore, highly vulnerable to various forms of worker opportunism (and would prob­ably be wiped out in a competitive struggle for survival if efficiency were the criterion for survival). The more vulnerable an organization is to this behav­ior, the more it will occur, all else equal.

This is not a simple consequence of people’s general penchant for oppor­tunism. For one thing, organizations known to be vulnerable to some form of opportunism tend to attract the type of people who are given to that form of opportunism. (This phenomenon is called adverse selection and will be dis­cussed in more detail shortly.) Second, there is something insidious about opportunistic behavior in that those who engage in it tend to prosper at the expense, in some way or other, of those who do not. When some form of opportunistic behavior reaches a certain threshold level, those who abstain begin to feel like “chumps” for their abstention. As a result, they begin to act opportunistically or to do so more frequently than before; in short, their pen­chant for opportunism changes. In this way, a self-reinforcing process gets under way that either destroys the organization or reaches a new equilibrium level—that is, a new low—of shirking and pay. For these two reasons, social organizations can be said to encourage opportunism to the extent that they do not embody structures and procedures to preclude or minimize it.

For example, the centrally planned economies of the former Soviet Union and communist China are so profoundly inefficient that it is extremely diffi­cult to survive by living according to the rules. Not coincidentally, most peo­ple do not; over the years, the distribution system in particular has become hopelessly corrupted. It has reached the point where it is only by participat­ing in the corruption that someone can move beyond bare survival. For this reason, these economic systems can be said to encourage corruption. It is in this way that opportunism is transformed from a human problem into a social problem.

People’s propensity for opportunistic behavior would be less of a problem if all contracts were unambiguous, completely determinate, and costlessly enforceable. Not only does none of these conditions hold, but each varies con­siderably from transaction to transaction. Much of transactions cost analysis consists of understanding how parties to various types of transactions deal with these sources of friction and the associated potential for exploitation. For example, since legal enforcement of contract provisions is often very costly, a contract might contain a clause mandating binding arbitration by a speci­fied arbiter for some range of disputes. Or one party might require some sort of bonding arrangement from the other party to ensure compliance with con­tract provisions.

What makes opportunistic behavior possible is almost always some infor­mational asymmetry between contracting parties—an asymmetry that one party deliberately creates or at least maintains. As Williamson says, “Oppor­tunism refers to the incomplete or distorted disclosure of information, espe­cially to calculated efforts to mislead, distort, disguise, obfuscate or otherwise confuse. It is responsible for real or contrived conditions of information asymmetry, which vastly complicate problems of economic organization” (1985, 47-48). Although Williamson maintains that the informational asym­metry is a form of opportunism, it is also very often a necessary condition for other forms of opportunistic behavior. In other words, the creation or main­tenance Ofinformational asymmetries makes opportunistic behavior possible. For example, the fact that one worker has a fairly good idea of what he is con­tributing to team production and others do not makes it possible for him to shirk. Here opportunism takes the form of shirking.

Two forms of informational asymmetries in particular have been system­atically investigated in transactions cost analysis: adverse selection and moral hazard. These concepts originally come from the insurance industry, but they have since been found to have much broader application. Adverse selection occurs when selection procedures systematically encourage people with undesirable characteristics to participate in a certain type of exchange with­out revealing that they have those undesirable characteristics. For example, suppose a retail firm hires security personnel based solely on an interview in which they are asked how they would deter and detect shoplifting. This firm might well fall victim to adverse selection because those who do best at this interview would probably include a disproportionate number of shoplifters. The general problem is that there are ex ante informational asymmetries that one party does nothing to dispel and that also work against the interests of the other party. Sometimes these problems can only be eliminated by proce­dures that are too costly to implement, but often they can be economically ameliorated. Efficient transactions are crafted to achieve just this result.

While adverse selection is an ex ante phenomenon, moral hazard is ex post. After a contract has been made, certain of its features can encourage— or not sufficiently discourage—violations of some of the terms of the contract or, at least, violations of legitimate expectations induced by the contract. This is a morally hazardous situation. In the insurance industry, for example, this problem arises when deductibles are too low or nonexistent so that the insured person is not sufficiently encouraged to exercise due care and cau-

tion. In the employment situation, the classic moral hazard problem is the incentive to shirk in unmonitored or loosely monitored team production. This, too, is a result of informational asymmetries. Other members of the team do not know what the shirker’s output is, allowing the latter to take advantage of that fact and thereby violate legitimate expectations about per­formance. Perfect monitoring is usually impossible, so the only recourse is to find economical means to discourage the shirking. This, of course, is a large part of the rationale for the classical capitalist firm (more on this in chapter 5)...

Inevitable informational asymmetries are responsible for the conditions of adverse selection and moral hazard. These problems are particularly important in principal-agent relationships, such as employer-employee, stockholder-manager, collective-individual worker. Much of transactions cost analysis seeks to understand how organizations and individuals craft trans­actions (e.g., the employment contract and even the job description) so as to avoid or mitigate these problems. We shall return to these issues later. Con­sider now the three key dimensions along which transactions vary.

Asset specificities. One of the most important features of transactions that transactions cost analysis considers is the specificity of the assets that support transactions. Physical and human assets have, or come to have, valuable but highly specific characteristics—characteristics that are most useful only in the context of a given contractual arrangement. Such assets are not easily rede­ployed once they have been committed.

There are three main types of asset specificities: (1) site specificities, (2) physical asset specificities, and (3) human asset specificities (Williamson 1985, 55). Site specificities would be illustrated by the example of an electric gen­erating plant built at the mouth of an already existing coal mine so as to reduce transportation costs for coal (Joskow and Schmalensee 1983). The assets represented by this plant have the value they do only because of their proximity to the coal mine. If coal had to be brought in by rail from another source, the value of the plant would fall accordingly. Physical asset specificity would be illustrated by the fact that the chemical composition of bauxite varies from one source to another and that idiosyncratic technologies are required for chemical processing, materials handling, and waste disposal in connection with aluminum smelting (Stuckey 1983). Wholly dedicated assets represent the most extreme form of physical asset specificity. A supplier fab­ricates a specialized piece of equipment to manufacture a product that only one customer can use. This piece of equipment has the highest possible degree of physical asset specificity. Human asset specificity means that employees build up firm-specific knowledge and skills that have little value elsewhere. For example, the parts manager in an auto dealership knows where every part in the warehouse is. That knowledge is extremely valuable but only in that particular dealership.

Uncertainty. ‘Uncertainty’ in this context refers to the fact that many of the contingencies that arise in the course of a transaction or contractual relation (especially a long-term one) cannot be foreseen and negotiated beforehand; often, even their probabilities cannot be known. For this reason and because of the inherent indeterminacy and ambiguity of language, complete contracts cannot be written to govern every contingency. Unanticipated disruptions in supply, transport, and so on can profoundly and adversely affect one of the parties to the exchange. Uncertainty is a serious problem in conjunction with asset specificities, bounded rationality, and opportunism. When highly spe­cific assets have been committed and the other party to an exchange is given to opportunism, what are claimed to be disruptions beyond that party’s con­trol can, in fact, be disruptions induced by opportunism or strategic behav­ior. Because of informational asymmetries, there is often no cost-effective way to know whether this is the problem or whether the problem is due to some genuinely exogenous disturbance. Transactions have to be crafted so as to minimize these problems; otherwise mutually beneficial exchanges will be for­gone or a premium will have to be paid to the party at risk.

Frequency. The frequency with which transactions take place is an impor­tant variable in determining the nature of the transaction, especially in con­junction with asset specificities. Frequent transactions involving nonspecific assets need only simple governance structures. For example, frequent pur­chases of wheat on the grain market require only ordinary contracts enforce­able under contract law. On the other hand, infrequent transactions involv­ing highly specialized assets require more complex governance structures. For example, the human capital that university professors bring to their jobs is highly specialized, and the market in this form of capital is fairly thin. At most universities, the terms and conditions of hiring and retention are gov­erned by an expensive structure of formidable complexity. Generally speak­ing, frequent transactions serve as an effective counterweight to bounded rationality and opportunism. One’s cognitive limitations and the potential rapacity of one’s green grocer are not serious handicaps to transactions in retail produce. Both are much more serious problems and require corre­spondingly complex governance structures when the transaction is the pur­chase of a mainframe computer or the labor of a university professor.