We are grateful to Ilya Segal and Michael Whinston for improving our analysis. We are pleased they confirm our two main conclusions. The first is that normally a firm cannot use contracts with its customers or suppliers inefficiently to exclude a rival from competition, because the high price of these contracts will make this strategy unprofitable. This is an old point, well summarized in Robert Bork's 1978 book. Second, and in contrast, exclusionary contracts can be profitable, effective, and socially inefficient -- under certain limited conditions. One condition is that firms in the industry must be able to operate only at or above some minimum efficient scale. Another condition is that the victims -- customers or suppliers -- must expect that the exclusionary tactic will succeed, and must be unable to coordinate their actions to defeat the tactic. An excluding firm in this situation can buy naked exclusion affordably because it can scare victims into selling cheaply; no single victim can stop the exclusion by itself, so no single victim has any bargaining power. At a theoretical limit, the excluding firm can gain the exclusionary rights for free. This striking result has implications for antitrust policy by suggesting that naked exclusion is, in theory, a potentially viable threat to efficient competition. Also striking from an antitrust perspective, however, is the lack of fit between this theory and the cases in which the United States Supreme Court has forged the law most relevant to exclusionary conduct. A simple legal label for contracts of naked exclusion is "exclusive dealing": "You agree to deal only with me, and not with my competitors." The facts of the three relevant Supreme Court cases, however, clearly violate the assumptions of the naked exclusion theory, as we explain in Rasmusen, Ramseyer and Wiley (1998). Two important conclusions follow. We cannot establish whether this kind of naked exclusion ever really happens by looking at the three legally most relevant cases. The theory awaits other empirical testing. And second, naked exclusion - if it ever really occurs - cannot be the only explanation for exclusive dealing. Rather, exclusive dealing "often" serves legitimate business purposes, as Judge (now Justice) Stephen Breyer wrote in his 1987 opinion in Interface Group, Inc v. Massachusetts Port Authority. The theory at hand thus does not support outlawing exclusive dealing on a per se or summary basis. If a legal prohibition is justified at all, any sensible legal test would have to be far more discriminating. Lawyers and judges who might seek to translate this theory into practice, please take note
