In order to focus attention on what I believe to be the critical issues. I will assume, initially, that oligopolistic agreements are lawful, so that there is no legal bar to collusion, but that oligopolists cannot appeal to the courts for assistance in enforcing the terms of an oligopolistic agreement. The oligopolists can, however, themselves take punitive actions to bring deviant members into line, provided only that laws prohibiting the destruction of property, libel, and so forth are respected. Entry is assumed to be difficult; also, I assume that profit pooling is permitted but that horizontal mergers between the firms are disallowed.
I will argue that oligopolists will commonly have difficulty in reaching, implementing, and enforcing agreements under these circumstances, but this argument does not mean that laws regarding oligopoly are of no account. The stipulations that horizontal mergers are disallowed and that collusive agreements are unenforceable in the courts are both important in this connection. If, however, it can be shown that monopolistic outcomes are difficult to effectuate even when the laws permit collusion, then the performance differences between dominant firms (monopoly) and oligo- polistic firms are not principally to be attributed to the unlawfulness of collusion among the latter.8 It follows, of course, that if express and lawful agreements (contracts) are difficult for oligopolists to reach and implement, tacit agreements are even less reliable instruments for achieving collusion.
A contract between two or more parties will be attractive in the degree to which: (1) the good, service, or behavior in question is amenable to unambiguous written specification; (2) joint gains from collective action (agreement) are potentially available; (3) implementation in the face of uncertainty does not occasion costly haggling; (4) monitoring the agreement is not costly; and ( 5) detected noncompliance carries commensurate penalties at low enforcement expense. Consider the oligopoly problem with respect to each of these.
1. Contract Specification
Recall that express oligopolistic collusion is assumed to be lawful. The parties to the collusive arrangement can therefore negotiate openly and express the details of the agreement in writing without exposing themselves to prosecution. The question to be assessed here is whether the latitude thus afforded will permit a comprehensive collusive agreement to be specified.
I submit that, except in rather special and unlikely circumstances, a comprehensive joint-profit maximizing statement (which does not entail merger) will usually be infeasible. The reason for this is that a comprehensive statement of such a kind requires an inordinate amount of knowledge about the cost and product characteristics of each firm, the interaction effects between the decision variables within each firm, and the interaction effects of decision variables between firms. Not only is the relevant information costly to come by, to say nothing of digesting it and devising the appropriate adaptation for each of the firms to make, but this needs to be done ex ante, for a whole series of contingent future events, most of which will never materialize, if anything approximating a complete contract is to be written.
The point is that joint-profit maximization, even as an abstract exercise, is very difficult to accomplish once one departs from the simplest sort of textbook exercise. Homogeneous products, identical linear and horizontal cost curves, and static markets constitute the “ideal.” Maintaining these product and cost assumptions in the face of changing demand does not greatly complicate the abstract analysis, in that the conditions of joint- profit maximization are easv to display, but the operational problems become somewhat more difficult in the face of uncertainty, for the reasons given in Section 2.3, below.
In circumstances, moreover, where differentiated products are involved, product and process innovations are occurring, organization form changes are taking place, selling expense and financial strategies are open to determination, and the like, the resulting complexity becomes impossibly great in relation to the bounded rationality capacities of planners. When, in addition, the optimization problem is cast in a multiperiod framework under conditions of uncertainty, abstract analysis breaks down.
The recent study by Hamilton and Moses (1973) of intrafirm profit maximization in a multidivisional enterprise is of some interest in this con- nection. They attempt operationally to address the problem of joint-profit maximization in a multiproduct firm where (1) product lines are independent., (2) only heuristic rather than full-blown optimization methods are attempted, and (3) only the financial decision is considered. Their model, which has since been implemented by the International Utilities Corporation, contains approximately 1,000 variables and 750 constraints and tests not one but various configurations of the strategic variables. Replicating such an arrangement by interfirm agreement boggles the mind. Complicating the analysis further to include interdependent products (which, of course, is the case in oligopoly) and the full range of decision variables referred to above reveals the manifest impossibility of attempting to comprehensively joint- profit maximize — even by heuristic simulation methods, much less by determinant contractual agreements. One concludes, accordingly, that the absence of legal prohibitions to collusive contracting is not what prevents comprehensive collusive contracts from being reached.9 Rather, elementary bounded rationality considerations explain this condition.
2. Joint Gain Agreement
Suppose, arguendo. that it were possible to specify the joint-profii maximizing strategy contractually. Would the parties then be prepared to make such an agreement? I submit that, but for the simple textbook cases referred to above, they would commonly decline to accept comprehensive joint-profit maximization of the profit pooling kind.
Partly disagreement might arise, as Fellner suggests, on account of differences between the parties concerning the appropriate discount rates to be used in evaluating future prospects. Surely more fundamental, however, are the risks and monitoring expenses that profit pooling entails. As Fellner notes, some of the parties must accede to reductions in relative output and to contractions in relative firm size if the joint-profit maximizing result is to be realized. This, however, is hazardous. Firms which are authorized to expand relatively as a result of the agreement will be powerfully situated to demand a renegotiated settlement at a later date. Wary of such opportunism, firms for which retrenchment is indicated will decline from the outset to accept a full- blown profit pooling arrangement. Moreover, even setting such concerns aside, monitoring the profit pooling agreement will be costly because of the pairing of opportunism with information impactedness. These issues are best addressed in the context of the moral hazard discussion in Section 2.4, below.
3. Implementation Under Uncertainty
Implementing an agreement under conditions of uncertainty requires the parties to agree, when changes in the environment occur, on what new state of the world obtains. Problems can arise if, for any true state of the world description, (1) some parties would realize benefits if a false state were to be declared, and either (2a) information regarding the state of the world is dispersed among the parties and must be pooled, or (2b) despite the possession of identical information by all the parties, definitive agreement must still be reached. The issues are those discussed in Section 3 of Chapter 2 and need not be repeated here. I merely point out that implementing an oligopolistic agreement under conditions of uncertainty can occasion costly haggling if the parties are given to making opportunistic representations with respect to the data and an outside arbiter can be apprised of true conditions only at great cost.
4. Monitoring Contract Execution
As Stigler points out, and as is widely recognized, oligopolists have an incentive to cheat on price fixing agreements if they believe that cheating will go, for a time at least, undetected. Given that information about individual sales is impacted, in that the seller knows exactly what the terms were but, given uncertainty, his rivals do not and can establish the terms only at some cost, the individual seller can often cut prices below the agreed level to the disadvantage of the other parties to the conspiracy. The pairing of opportunism (which is here manifested as cheating) with information impactedness explains this condition.
The argument, moreover, applies to oligopolistic collusion with respect to nonprice decision variables as well. If anything, agreeing to collude with respect to marketing expense, R & D efforts, and the like is even more hazardous for nonopportunistic parties, who are prepared to abide by the agreements, than is price collusion. Although it is easy to establish after the fact that a rival has made significant design changes or introduced a new product in violation of the agreement, such information may come too late. If recovery from a large shift in market share, attributable, say, to an “illicit” innovation is inordinately expensive, the detection of such a violation is to little avail. Although firms can mitigate these risks by maintaining a defensive posture against such contingencies, comprehensive collusion in nonprice respects can scarcely be said, in that event, any longer to be operative.
The matter of profit pooling, referred to in Section 2.2, above, needs also to be examined in monitoring respects. Even if firms were prepared to enter into profit pooling agreements, in which all profits are pooled and each participant is assigned his share of the total, there is still the problem of determining what the contribution of each firm to the pool should be. Individual firms have an incentive to understate true profits in these circum- stances.
Moreover, merelv to audit the earnings of each firm, even to the extent that all sources of revenue and cost are fully disclosed, is not sufficient to avoid distortion. An assessment of individual expense items must also be made. The problems facing the auditor here are akin to those that face the defense agencies in monitoring cost-plus (or, more generally, cost-sharing) defense contracts.170 Unless it can be established that certain types or amounts of actual costs are unwarranted, and hence will be disallowed, each firm has an incentive to incur excessive costs.
Expense excesses can take any of several forms. Perhaps the simplest is to allow some operations to run slack, which is to say that the management and workers in the firm take part of their rewards as on-the-job leisure. A second way is to allow emoluments to escalate, in which case corporate personal consumption expenditures exceed levels which, from a profit maximizing standpoint, would be incurred. Third, and most important, firms may incur current costs which place them at a strategic advantage in future periods. Developing new and improved technology, training the work force, and so forth are examples of the latter. Evaluating individual firm performance in these several respects is at least an order of magnitude more difficult than simple audits of revenue and cost streams. Conformably, profit pooling, even were it legal, poses severe contract enforcement problems.
5. Penalizing Contract Violations
Recall that it was assumed that while collusive agreements are not unlawful, the participants in such agreements cannot call upon the courts to help enforce the agreement. Instead, violators must be determined and penalties must be administered by the parties to the contract. Problems of two types arise in this latter connection. First, there is the problem of whether the penalties imposed will be efficacious. Put differently, do the penalties, if implemented, constitute an effective deterrent to the would-be violator? Second, even if penalties can be devised that would be efficacious, will the parties to the conspiracy be prepared to impose them?
Because the conspirators lack legal standing, conventional penalties such as fines and jail sentences are presumably unavailable. Rather, penalties are exacted in the marketplace bv confronting the violator with unusually adverse circumstances. Price reductions are matched and perhaps even undercut. Normal types of interfirm cooperation (for example, supply of components) is suspended. Key employees may be raided, and the like. Except, however, as deviant firms are highly dependent on rivals for vital supplies, such market reactions may well be ones that the deviant is prepared to risk.
For one thing, the contract violator is not the only firm to be adversely affected by exacting these penalties in the marketplace. The firms meting out the penalties also incur costs. Second, and related, securing the collective action needed to punish the violator may be difficult. Thus, although all firms may agree both that a violation has taken place and that the violator deserves to be punished, not all may be prepared to participate in administering it. Defectors (for example, those willing to supply the deviant with the essential component, perhaps at a premium price), which is to say opportunists, who refuse to incur the costs of punishing the violator, naturally reduce the prospective costs of being detected in violation of the agreement. Where such defection is deemed likely, collusive contracts are all the less likely.
Source: Williamson Oliver E. (1975), Markets and hierarchies: Analysis and antitrust implications, A Study in the Economics of Internal Organization, The Free Press.