Dr. David Rapson
This paper revisits the question of whether automobile manufacturers engaged in tacit collusion in the 1950s. In his famous paper, Bresnahan (1987) develops a model of strate- gic interaction between multi-product firms to explain a brief spike in quantity in 1955, and concludes that the episode was caused by a temporary breakdown in tacit collusion. However, his model imposes strong restrictions on the nature of consumer preferences and intra-firm, multi-product pricing strategies. I reconstruct his original dataset and solve a random coefficients logit model that allows for a broader range of substitution patterns and pricing strategies. I estimate the model under multiple firm behavioral hypotheses: Bertrand competition between firms and between brands, and perfect collusion. For no year in 1954-1956 can either form of Bertrand competition be rejected in favor of tacit collusion.