In this paper, we explore the consequences of low nominal interest rates for credit supply, macroeconomic outcomes and policy. Using the protracted period of negative rates and detailed micro data from Japan, we show that banks with higher ex-ante market power face relatively higher funding costs, have lower profitability, and decrease loan supply in a low nominal rate environment. We build a macroeconomic model that rationalizes our key empirical findings and characterizes optimal long-run rates. Market power in deposits helps mitigate lending frictions, but is sensitive to nominal rates due to the existence of cash. Under such lending frictions, the optimal nominal rate is higher than suggested by the Friedman rule. Calibrating the model with cross-sectional panel evidence, we find that low rates resulted in significantly lower aggregate lending, negatively affecting output. Tiering bank reserves only marginally alleviates the negative effects of low rates on credit supply, while taxing cash is more effective.
