Gradually Rebuilding a Relationship: The Emergence of Collusion in Retail Pharmacies in Chile (Job Talk)

How does collusion start? This paper studies the emergence of collusion in the presence of multimarket contacts. It analyzes price fixing among the three main retail drugstore chains in Chile. The pharmacies raised the prices of more than two hundred medicines, mostly best-selling brands, after months of low, even negative margins. The scope of collusion grew gradually as firms colluded on an increasingly larger number of drugs over a period of four months, raising the price of each product among themselves by means of staggered, sharp price increases. I use the large collusive price increases as supply-side shocks to estimate the demand for the drugs. My main result is that the pharmacies raised first and by a larger extent the prices of products in which firms are more differentiated. I claim that this behavior was due to the firms' concerns of not being followed by their competitors, which were stronger at the beginning of the coordination period. Collusion on differentiated products was safer due to smaller losses should the collusive scheme collapse, and thus gradualism allowed firms to learn their competitors' willingness to collude and build trust over time. In addition, as a simple model of collusion with two-sided uncertainty regarding the competitor's willingness to collude shows, firms would start colluding on differentiated products when they are either not too patient or pessimistic about the competitor's type. I support this argument with evidence from the depositions of the antitrust case and rich data on monitoring activity, which shows that price quotes on the brands that were added to the collusive scheme decreased over time.