We empirically quantify the welfare implications of bank entry in the United States between 2000 and 2008. We use a fully structural framework that combines a differentiated demand model with an endogenous product model to investigate the market outcomes. We find no evidence for under- or over-entry. Compared with the socially efficient outcome, there is a mild welfare loss resulting from banks entering wrong locations in product space. Compared with the observed outcome, consumer surplus drops by 20-38% and bank profits decline by 48-59% when banks are homogeneous. Therefore product differentiation significantly improves welfare under free entry. (C) 2013 Elsevier B.V. All rights reserved.