Cartels and Collusion - Empirical Evidence
Margaret C. Levenstein
University of Michigan at Ann Arbor - Survey Research Center; The Stephen M. Ross School of Business at the University of Michigan, Business Economics and Public Policy
Valerie Y. Suslow
Johns Hopkins Carey Business School
Ross School of Business Paper No. 1182
Chapter prepared for publication in Oxford Handbook on International Antitrust Economics, Roger D. Blair and D. Daniel Sokol, editors. Cartels occur in a wide range of industries and engage in a wide range of behaviors in their efforts to increase profits. In this chapter, we discuss the wide variety of techniques that cartels use to increase prices and profits. Studies of national and international markets across the twentieth century find cartels in a wide variety of products and services, and these cartels typically last between five and eight years. The most important determinant of cartel breakup is effective antitrust policy. While it has often been presumed that cartels’ demise results from cheating by member firms tempted by short term profits, empirical analysis suggests that cheating rarely destroys cartels. The potential profits from collusion provide sufficient incentives for cartels to develop creative ways to limit the temptations that inevitably arise. While scholars and policy makers have often been concerned that business cycle downturns are associated with cartel formation, the evidence we review here does not suggest strong cyclical effects. There is evidence that cartels are formed during periods of falling prices, but these are more likely to be the result of market integration or an increase in competitive intensity than macroeconomic fluctuations. Similarly, cartel breakup does not evidence strong cyclicality.
Number of Pages in PDF File: 32
Keywords: cartels, collusion, antitrust, competition
JEL Classification: L13, L16, L41
Date posted: November 30, 2012
© 2016 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollobot1 in 1.687 seconds