BAYESIAN LEARNING IN AN INFANT INDUSTRY MODEL

Most analyses of the infant industry argument assume that production e arly on benefits subsequent production in deterministic fashion. This paper presents an alternative framework for evaluating the infant indu stry argument that focuses on imperfect information as a barrier to en try. Initial entrants provide information of social value by reducing uncertainty for potential followers about the suitability of local con ditions. In contrast to earlier work, I show that factors that increas e the informational barrier to entry may lower the optimal subsidy to the infant industry, and that the standard test of the success of infa nt industry policy is not always valid. (C) 1997 Elsevier Science B.V.