Optimal industrial targeting with unknown learning-by-doing

Abstract

We examine a government's optimal targeting policy when it has limited information about the learning curves of domestic producers. Popular arguments suggest that in order to promote learning-by-doing, the government might want to protect domestic producers from foreign competition by temporarily closing the domestic market to foreign producers. We identify a set of conditions under which such trade intervention is not optimal. Instead, domestic welfare is better fostered either by no government intervention, or by providing subsidies to the most capable domestic producers who are willing to set a particularly low domestic price for their product. © 1995.