Bust, Trickle or Scare? How Enforcing Competition Policy Affects Competition

How does the enforcement of competition policy lead to greater competition? While competitive markets are often touted as having wide-ranging economic benefits, we know surprisingly little about whether and how public policies designed to promote competition -- in particular competition law and policy -- actually promote competition. In this paper, we identify and test three possible mechanisms through which the enforcement of competition law may lead to increased competition. First, the enforcement of competition law seeks to put an end to a particular firm’s anti-competitive behavior, which should correspondingly increase competition in the wider industry of which the firm is a part. Second, the enforcement of competition policy in one industry may have spill-over effects to competition in related industries. For example, where industries serve as perfect complements to each other, (e.g., the automobile and gasoline industries), greater competitiveness in one industry should lead to greater competitiveness in its complement. Third, the enforcement of competition law in any given industry may lead to greater avoiding of anti-competitive behavior in all industries, especially when high-profile examples of competition enforcement send a signal to business leaders across all industries the need to avoid anti-competitive conduct and to engage in more competition-compatible behavior to avoid similar prosecution. While these mechanisms are not mutually exclusive, knowing their relative importance would greatly enhance our understanding of what makes competition law and policy effective. We test the above hypotheses using panel analyses of an original dataset of some four thousand U.S. Department of Justice antitrust enforcement actions from 1958 to 2010, coded by the authors with a large team of research assistants. The dataset codes, inter alia, the type of enforcement action taken, the outcome of the action, and the industry, which allows us to conduct analyses with industry-year as our unit of observation. We measure our dependent variable, competition, using industry-level Herfindahl-Hirschman indices from the U.S. Census and the CRSP Compustat database, as well as industry-level price data from the U.S. Department of Labor Bureau of Labor Statistics. To test our first hypothesis, we create a binary indicator of whether an antitrust case was successfully prosecuted in a particular industry. To test our second hypotheses, we create an indicator of whether an antitrust case was successfully prosecuted in a linked industry, weighting by the degree of complementarity between the two industries. Finally, to test the third hypotheses, we collect data on the media coverage of different antitrust enforcement actions using the LexisNexis Newspapers & Wires Database (using the names of the plaintiffs and the term ‘antitrust’ or the legal basis under which the case was prosecuted). The number of newspaper articles for each case serves as our measure of the degree of media scrutiny for a particular case.