1. General
Reciprocity transforms a unilateral supply relation where A sells X to B into a bilateral one, whereby A agrees to buy Y from B as a condition for making the sale of X and both parties understand that the transaction will be continued only if reciprocity is observed. The resulting contractual relation is thereby expanded. Although, reciprocalselling is widely held to be anticompetitive (Stocking and Mueller, 1957; We, 1973), others regard it more favorably. George Stigler offers the following affirmative rationale for reciprocity:
The case for reciprocity arises when prices cannot be freely varied to meet supply and demand conditions. Suppose that a firm is dealing with a colluding industry which is fixing prices. A firm in this collusive industry would be willing to sell at less than the cartel price if it can escape detection. Its price can be reduced in effect by buying from the customer-seller at an inflated price. Here reciprocity restores flexibility of prices.1
Inasmuch, however, as many industries do not satisfy the prerequisites for oligopolistic price collusion (Posner, 1969; Williamson, 1975, chap. 12), and reciprocity is sometimes observed among them, reciprocity presumably has other origins as well. Tiebreaking is one. A second is the possible advantageous governance strucure benefits of reciprocity. The two can be distinguished by the type of product being sold.
The tiebreaker explanation applies where firm B, which is buying spe- cialized product from A, asks that A buy standardized product from B on the condition that B meets market terms. Other things being equal, procurement agents at A are apt to accede. F. M. Scherer notes, “Most of the 163 corporation executives’ responding to a 1963 survey stated that their firms’ purchases were awarded on the basis of reciprocity only when the price, quality, and delivery conditions were equal” (1980, p. 344).
The more interesting case is where reciprocity involves the sale of spe- cialized product to B conditioned on the procurement of specialized product from B. The argument here is that reciprocity can serve to equalize the exposure of the parties, thereby reducing the incentive of the buyer to defect from the exchange—leaving the supplier to redeploy specialized assets at greatly reduced alternative value. Absent a hostage (or other assurance that the buyer will not defect), the sale by A of specialized product to B may never materalize. The buyer’s commitment to the exchange is more assuredly signaled by his willingness to accept reciprocal exposure of specialized assets. Defection hazards are thereby mitigated.
The original (1968) Merger Guidelines of the U.S. Department of Justice took a wholly different approach. Although the subject was conglomerate mergers, the concern with reciprocity as a contracting practice was general. The language is instructive:
(a) Since reciprocal buying (i.e., favoring one’s customer when making purchases of a product which is sold by the customer) is an economically unjustified business practice which confers a competitive advantage on the favored firm unrelated to the merits of the product, the Department will ordinarily challenge any merger which creates a significant danger of reciprocal buying . . .
(c) Unless there are exceptional circumstances, the Department will not accept as a justification for a merger creating a significant danger of reciprocal buying the claim that the merger will produce economies, because, among other reasons, the Department believes that in general equivalent economies can be achieved by the firms involved through other mergers.
The Guidelines are noteworthy in several respects. For one thing, they plainly have origins in the inhospitality tradition: Reciprocity, like other non- standard contracting practices, is an “economically unjustified business prac- tice.” Second, and related, the Guidelines are informed by the technological tradition: Reference is made to the “merits of the product” (a technological standard), but the “merits of the transaction” (a governance concern) go unmentioned. Third, the Guidelines assert that the relevant economies can be realized without posing reciprocity hazards: Socially valued (technological) economies can ordinarily be realized by finding substitute mergers for which reciprocity hazards are absent. The possibility that nonstandard contracting, of which reciprocity is a member, can yield valued economies is simply ignored. Such economies are evidently so implausible that potential contractual benefits do not even have to be admitted in principle before being dismissed.
Lon Fuller’s interesting discussion of reciprocity, although posed in much more general terms than those set out here, is plainly apposite:
I think we may discern three conditions for the optimum efficacy of the notion of duty. First, the relation of reciprocity out of which the duty arises must result from a voluntary agreement between the parties immediately affected; they themselves “create” the duty. Second, the reciprocal performances of the parties must in some sense be equal in value. Though the notion of voluntary assumption itself makes a strong appeal to the sense of justice, the appeal is reinforced when the element of equivalence is added to it . . .Third the relationship of duty must in theory and in practice be reversible…
When we ask, “In what kind of society are these conditions most apt to be met?” The answer is a surprising one: in a society of economic traders. [Fuller, 1964, pp. 22-23]
Lest the hostage argument be uncritically considered a defense for re- ciprocal trading quite generally, note that it applies only where specialized assets are placed at hazard by both parties. Where only one or neither invests in specialized assets, the practice of reciprocity plainly has other origins.2
2. Exchanges
Although reciprocal trading among nonrivals may occasionally be justified, the exchange of product among nominal rivals is surely more puzzling and troublesome. Firms that are presumed to be in head-to-head competition ought to be selling product against one another rather than to one another. Inasmuch as neoclassical benefits are not plausibly imputed to the continuing exchange of product between rivals, public policy toward exchanges has been mainly negative and even hostile.
Several distinctions are useful in considering exchanges. First, trade among rivals—short-term or long-term, unilateral or bilateral—is feasible only if product is fungible. That is not true for many differentiated goods and services, so the issue of trade among rivals never arises for those. Second, short- term supply agreements are usefully distinguished from long-term. The former may be explained as an “occasional exception,” whereby one rival will sell product to another on a short-term, gap-fdling basis so as to provide temporary relief against unanticipated product shortfalls (occasioned by either demand or supply changes). Recognizing that the shoe may be on the other foot next time, otherwise rivalrous firms may assist one another for stopgap purposes. Public policy can presumably recognize merit in such trades and, so long as they lack a pattern, hence do not give rise to a “web of interdependence,” will regard them as unobjectionable. Long-term trading among rivals is, however, much less consistent with the notion of effective head-to-head rivalry. At the very least, such arrangements warrant scrutiny.
Whether there are efficiency incentives for rivals to supply product to one another on a long-term basis turns initially on prospective realization of production cost savings. The realization of production cost savings through long-term trade between rivals requires that economies of scale be-large in relation to the size of geographic markets and, if they are, that firm-specific reputation effects extend across geographic market boundaries. The former is obvious since, absent economies of scale, every firm would presumably supply everywhere to its own long-term needs. Where scale economies are significant, however, each market will support only a limited number of plants of minimum efficient size.
But fungibility and scale economies do not establish that gains from trade will be realized from such sales. That will occur only if the value of (identical) product sold by rivals exceeds that sold by the local supplier. The issue here is whether valued reputation effects will go unrealized if rivals are unable to secure local product on favorable terms. Firms that possess valued reputations extending beyond their local market to include distant markets are thus the ones for which long-term supply by rivals will be attractive.
Even supposing that fungibility, scale economy, and reputation effect conditions are satisfied, that merely establishes that unilateral long-term trade among rivals can yield economies. A justification for bilateral (exchange) agreements is not reached by the same arguments. Indeed, the usual defense for exchanges—that inefficient cross-hauling will occur if every firm is required to supply everywhere to its own needs—conveniently suppresses the obvious alternative, which is not the absence of trade at all but unilateral longterm trade. Failure to address such matters directly and demonstrate wherein exchanges enjoy comparative institutional advantages over more standard and familiar forms of unilateral trade presumably explains the suspect or hostile attitude with which exchanges are typically regarded. The argument that emerges from this chapter is that bilateral exchanges offer prospective advantages over unilateral trade if the resulting exposure of transaction-specific assets effects a credible commitment without simultaneously posing expropriation hazards.
The type of specific asset that is placed at hazard by unilateral long-term trade, but which a reciprocal long-term exchange agreement serves to protect, is that of a dedicated asset. Recall that dedicated assets were described as discrete additions to generalized capacity that would not be put in place but for the prospect of selling a large amount of product to a particular customer. Premature termination of the contract by the buyer would leave the supplier with a large overhang of capacity that could be disposed of only at distress prices. Requiring buyers to post a bond would check that hazard, but only by posing another: The supplier may contrive to expropriate the bond. More generally, the interests of the supplier in adapting efficiently to new circum-stances are not fully engaged. Reciprocal trading supported by separate but concurrent investments in specific assets provides a mutual safeguard against this second class of hazards. The hostages thereby created have the interesting property, moreover, that they are never exchanged. Instead, each party retains possession of its dedicated assets should the contract be prematurely terminated.
The usual argument that exchanges are justified because they avoid costly cross-hauling does not get to the issues described above and, by itself, is not arTadequate justification for widespread use of exchanges. Were it only that transportation cost savings were realized, unilateral trading would suffice. Indeed, firms that buy from and sell product to rivals should be expected to create a central exchange in which supplies and demands were brought into correspondence by an auctioneer. Firms would end up selling to each other only by accident in such circumstances. Where dedicated assets are exposed, however, the identity of the parties clearly matters. Trades of that kind will not go through an auction market but will be carefully negotiated between the parties. Reciprocity in those circumstances is thus a device by which the continuity of a specific trading relation is promoted with risk attenuation effects.
Source: Williamson Oliver E. (1998), The Economic Institutions of Capitalism, Free Press; Illustrated edition.