Almost all industries have a vertical dimension where the good or service moves from upstream units (e.g. manufacturing) to downstream units (e.g. wholesalers). The effects of different organizational formats of this vertical structure have been the focus of much theoretical work in industrial organization. This research project studies from an empirical perspective the ambiguous theoretical predictions of the effects of two central topics in this literature: vertical integration (VI) and vertical restraints (VR). Vertical integration refers to a decision by a firm to perform both upstream and downstream tasks, whereas vertical restraints refer to contract clauses between the (separated) upstream and downstream units whereby the downstream unit is forced to engage in particular sales formats. The data allows this project to focus on exclusive dealing and exclusive territories as VR.
A brand level dataset from supermarket scanners on beer and softdrink sales in the US has been made available for this research. A reduced form approach is used to study the effects of VI/VR on equilibrium price and quantity. Price and quantity are regressed on the VI/VR variable and, after controlling for endogeneity and other factors, unambiguous conclusions can usually be reached; for example, if VI produces a larger quantity and a lower price, then this market structure can be interpreted as procompetitive with respect to vertical separation. While in some cases this reduced form approach allows inference about the possible direction of change in welfare, a monetary estimate of such change is not possible. Moreover, a reduced form approach can not determine who (consumers or producers) benefits or suffers more, and by how much. Thus, to complement the reduced form approach, this research also adopts a structural approach where supply and demand parameters are estimated and welfare estimates are computed.
Intellectual Merit: The main contribution of this research is based on the fact that empirical work has been limited relative to the quite large amount of theoretical results on the consequences of VI and VR. In addition, choosing to study these topics in these industries is particularly important. The industries exhibit a unique variation across: a) time and space on the decisions of firms regarding VI (softdrinks) and exclusive dealing (beer and softdrinks), and b) across states regarding the legal mandates on exclusive territories (beer). Also, the available dataset consists of a three dimensional panel on dozens of brands, which spans across the main 63 metropolitan areas of the US over a period of five years; this richness allows the project to address the critical issue of endogeneity of the VI/VR decision that has frequently been neglected in prior work.
Broader Impact: The proposed research has key implications from a policy angle since US antitrust activity, like theory, has lacked of consistency in determining the (anti) competitive nature of VI/VR. For example, antitrust policy regarding VR has fluctuated over the last century starting off as lawful up to the 1940s, turning to largely illegal by the 1960s and then back to lenient in the late 1970s after the Sylvania case in 1977. Finally, the Vertical Guidelines of 1985 reflecting this lenient view were eliminated in the 1990s, removing any formal antitrust policy guidance for VR (Lafontaine and Slade, 2006). Antitrust policy towards VI (vertical mergers) has also shown significant fluctuations, as reflected in the differential treatment of vertical mergers by the three different merger guidelines that the Department of Justice has had over time. The proposed research will thus provide useful evidence to guide future antitrust policy towards vertical integration and vertical restraints.
