Collusion, Efficiency, and Antitrust

THE structuralist approach to antitrust rests upon two fundamental propositions. One is that high seller concentration facilitates explicit and/or tacit collusion. The other is that leading firms in concentrated industries tend to be larger than the minimum optimal size dictated by scale economies. Together the propositions imply that a policy of deconcentration will promote lower prices without sacrificing production efficiency. These propositions are not without substantial empirical support. Empirical studies consistently reveal a positive if weak relationship between industry concentration and profitability.' These results are interpreted as evidence that concentration induces monopoly pricing through collusion. Leading engineering cost studies indicate a strong tendency in concentrated industries for firm sizes to exceed estimated minimum optimal scales.2 This evidence is generally corroborated by statistical cost studies and survivor tests.3 A number of economists, most notably Yale Brozen, John McGee, and Harold Demsetz, have challenged the thesis that deconcentration will increase economic welfare.4 Expressing some skepticism regarding the economic meaningfulness of the concentration-profitability correlation noted above, these economists go on to argue that the concentration itself is explained by the superior efficiency of large firms. Hence, while deconcentra-