How communication makes the difference between a cartel and tacit collusion: A machine learning approach

This paper sheds new light on the role of communication for cartel formation. Using machine learning to evaluate free-form chat among firms in a laboratory experiment, we identify typical patterns both explicit formation and indirect attempts collude tacitly. We document that are less likely communicate explicitly about price fixing more use messages when sanctioning institutions present. effect sanctions reinforces direct cartel-deterring as collusion is difficult reach sustain without an agreement. Indirect have no, or even negative, prices.