Does Parallel Behavior Provide Some Evidence of Collusion?

Antitrust authorities often consider parallel pricing and market share stability to be clues of illegal collusion. To analyze whether this inference is correct, I develop a model of price competition with differentiated products in which demand and costs vary over time. In many cases parallel pricing does not distinguish between a competitive and a collusive outcome. However, in some cases perfect parallel pricing is compatible only with a competitive equilibrium, and therefore provides some evidence that firms did not collude. I also show that the competitive equilibrium is characterized by a higher market share stability than a collusive equilibrium.