Monday, April 13, 2009
Posted by D. Daniel Sokol
Herb Hovenkamp of the University of Iowa Law School and Erik N. Hovenkamp, University of Iowa - Department of Economics, B.S. Candidate 2009 (Herb's son?) have written an article on Complex Bundled Discounts and Antitrust Policy.
ABSTRACT: A bundled discount occurs when a seller conditions a discount or rebate on the buyer's purchaser or two or more different products. Firms that produce fewer than all the good in the bundle find it difficult to compete because they must amortize the discount across a smaller range of goods. For example, if the dominant firm offers a 10% discount for purchase of both good A and good B, but the rival makes only good B, it will have to offer a discount that is large enough to match the dominant firm's B discount as well as the foregone discount on A. The Antitrust Modernization Commission and several courts have adopted an "attribution" test for assessing the antitrust legality of bundled discounts. The test attributes the full discount to the product(s) for which rivals are claiming exclusion, and asks whether the resulting price is below cost. This test contains some features of the cost-based rule for single product predatory pricing, but it also differs in important respects. Both tests query whether an equally efficient rival can match the dominant firm's price.
Most models of bundled discounting consider two goods that are purchased in a one-to-one ratio. None of the judicial decisions involve such simplicity. In most the bundle consists of more than two goods, and different rivals may produce differing subsets of the dominant firm's bundle. Further, in nearly all of the cases the proportion of goods in the bundle can be varied at the will of the customer. We show that in such situations antitrust analysis of the bundle is significantly more complex and anti-competitive exclusion must typically be assessed on a rival-by-rival and customer-by-customer basis. This has important implications for the certification of class actions in bundled discount cases. We also provide some apparatus for assessing bundled discounts in these situations.