The Paradox of Vertical Integration

Abstract

Antitrust policy and commentary include a persistent suspicion that vertical integration may increase market power in the integrated industries. The facet of this suspicion that economists have analyzed most is conveyed as follows. Suppose that a pure monopoly sells an intermediate good to the firms in a final product industry and that the monopolist acquires all of those firms. Since the firms in the final product industry have been combined, vertical integration has extended monopoly to a second industry. The integration does not improve production technologies or product traits in either industry nor reduce transaction costs, thus abstracting from any economies of integration. This paper concerns the paradox that economic models usually show vertical integration increasing economic welfare, in spite of the extension of monopoly, and even in the absence of any economies. Since economies would also enhance welfare, these models contradict the suspicion that vertical integration is harmful. In a recent survey of these models, Blair and Kaserman have commented that [o]n the basis of these results, the economics profession generally argued that the appropriate antitrust policy regarding vertical integration ... was conciliation at best or neutrality at worst [2, 2]. The only contrary result is in models where inputs are substitutable in final production: sometimes vertical integration reduces welfare.' However, in order to focus sharply on the paradoxical welfare result, this paper assumes fixed input proportions and ignores any economies of integration. Section I has a summary model that sets the stage for the succeeding analysis. The principal discovery of this paper is that the paradox is not in general valid, even for a simplified model. This discovery appears when the assumption of pure input monopoly, which is universal in the existing literature, is relaxed in section II. The new model includes situations in which integration reduces welfare, contrary to the usual result. These contradictory situations have strong implications for antitrust policy, because the analysis also shows that they ate the only kinds of situations that antitrust enforcers are likely to confront: