While it is traditionally assumed that inter-firm collusion creates deadweight losses that harm society, over the past quarter century several studies have suggested that collusion can result in increased economic efficiency and hence benefit society. Webb (1980), Bittlingmayer (1982), Troesken (1989), Sjostrom (1989, 2004), Pirrong (1992), and Kinghorn (1996) have highlighted individual industries such as coal, steel, cast-iron pipe, and ocean shipping in which cartelization appears to have facilitated welfare-enhancing outcomes. How widespread such efficiency gains from collusion are, however, is unknown since scholars have primarily relied upon case studies rather than a broad sample of industry cartels to examine this "efficient-cartel" hypothesis.
This project employs a monthly panel of industries covered by the National Industrial Recovery Act (NIRA) of 1933 to perform a broad empirical test of the efficient-cartel hypothesis. The NIRA required firms in every U.S. manufacturing industry-over 500 industries in all-to write up a specific cartel "code" of rules. These rules were legally binding from the date of each code's passage until the legislation was ruled unconstitutional in May of 1935. Thus, the NIRA creates a natural testing ground for analyzing the welfare effects of collusion via a broad sample of industries covered by overt cartels for as many as 23 months.
Findings from a 64-industry sample suggest that cartelization under the NIRA, as a whole, reduced output. This suggests that cartels are harmful to society as is consistent with orthodox theory. However, when the 64 industry regressions are run individually, around one-quarter of the industries experienced output expansion under collusion, other factors held constant, suggesting that the welfare effects of collusion vary by industry as is suggested by the efficient-cartel theory. The core of this project is to analyze the impact industry-level attributes such as the intensity of fixed costs, degree of demand uncertainty, number of firms, industry concentration ratio, cost and product heterogeneity, etc. have upon economic welfare under collusion.
By departing from the case-study approach generally employed in efficient-cartel research, this project makes a substantive intellectual contribution to the theory of collusion. The research provides empirical insight into (1) how widespread efficient-cartel dynamics are and, perhaps more importantly, (2) the welfare impact cartels are likely to produce under the presence of various industry characteristics.
With respect to the broader policy impact of this study, it is well known that legal institutions