Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Bank loans are more available and cheaper for new and small businesses in the U.S. in concentrated banking areas than in competitive banking areas. We explain this anomaly by analyzing banks' decisions to screen projects and their competition in loan provisions. It is shown that, by exacerbating the winner's curse, an increase in the number of banks can reduce banks' screening probability by so much that the number of banks that actively compete in loan provisions falls and the expected loan rate rises. This is the case when the screening cost is low, which induces all active bidders to be informed. The opposite outcome occurs when the screening cost is high, in which case there are sufficiently many uninformed banks in bidding to attenuate the winner's curse. We also examine the social optimum. JEL classification: G21, D44, L15