9410711 Kennan This project aims to provide an evaluation of the empirical relevance of strategic bargaining models. The ultimate goal is to develop theoretically coherent empirical analyses of important policy issues in collective bargaining. The strategic importance of private information has been emphasized in many recent theoretical models, with applications in many fields. It remains to be seen whether this large body of theory has much empirical substance. The analysis of labor contracts is a good test-bed, since interesting theories have been proposed, and good data are available. A serious limitation of the recent literature is that each contract is treated in isolation. In practice, labor negotiations involve contracts covering several years, and the parties seem mindful of how the outcome of the current contract will affect negotiations on the next contract. There is strong evidence that the likelihood of a strike is influenced by what happened before the previous contract was signed: a short strike last time means that the probability of a strike now is about 40%, while if there was either a long strike or no strike last time the current strike probability is below 20%. To analyze whether private information might explain this pattern, a model of negotiations between a buyer and a seller will be developed where the "pie" to be divided in each contract follows a Markov chain, with transitions that are observed privately by the buyer. Each negotiation consists of a sequence of offers made by the seller; the buyer can only accept or reject each offer, and a contract is signed as soon as the buyer accepts.
