Tying Bundled Discounts and the Death of the Single Monopoly Profit Theory

Abstract

Chicago School theorists have argued that tying cannot create anticompetitive effects because there is only a single monopoly profit. Some Harvard School theorists have argued that tying doctrine’s quasi-per se rule is misguided because tying cannot create anticompetitive effects without foreclosing a substantial share of the tied market. This Article shows both positions are mistaken. Even without a substantial foreclosure share, tying by a firm with market power generally increases monopoly profits and harms consumer and total welfare, absent offsetting efficiencies. The quasi-per se rule is thus correct to require tying market power and a lack of offsetting efficiencies, but not a substantial tied foreclosure share. However, the quasi-per se rule should have an exception for products with a fixed ratio that lack separate utility, because those conditions generally negate anticompetitive effects absent a substantial foreclosure share. Cases meeting this exception should instead be governed by a traditional rule of reason that requires a substantial foreclosure share or effect. Bundled discounts can produce the same anticompetitive effects as tying without substantial tied foreclosure, but only when the unbundled price exceeds the but-for price. Thus, when the unbundled price exceeds the but-for price, bundled discounts should be condemned based on market power absent offsetting efficiencies, with the same exception for products with a fixed ratio that lack separate utility. When the unbundled price does not exceed the but-for price, bundled discounts should be condemned only when there is substantial foreclosure or direct proof of anticompetitive effects. Alternative tests for judging bundled discounts, such as comparing the effective price to cost, are not only underinclusive, but perversely exempt the bundled discounts with the worst anticompetitive effects.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 68 citations exist upstream.