Mutual Forbearance under Experimental Conditions

Since Corwin Edwards first raised it 30 years ago [4], the question whether conglomerate mergers and diversification will affect competition within markets has remained unanswered. Diversification or conglomeration has no direct impact on entry conditions or market shares or concentration ratios. Aside from claims for advantages from size per se (and not from conglomeration itself), the most promising theoretical rationale to explain potential anticompetitive effects has been the forbearance, or spheres of influence, theory. By forbearance, we mean the possible tendency for firms meeting in more than one market to account for rivals' reactions across market boundaries by adopting less competitive price and output strategies. Possibly because real-world oligopolists have a large set of decision variables available to act on (price, output, capacity, advertising, research and development, etc.), the effects of interrelationships among firms such as mutual forbearance have been difficult to determine by subjecting industrial data to statistical testing.' On the other hand, in a controlled laboratory setting the validity of particular behavioral assumptions can be examined with some precision. Sherman has discussed results of previous oligopoly experiments [19], which generally have supported Cournot (noncooperative) solutions rather than cooperative (collusive) ones, and Plott has surveyed experiments related to antitrust and industrial organization issues [17]. Under various information and