We investigate the effect of an exogenous change in loan loss provisioning rules on bank risk taking. To identify the effect we exploit that only banks that revealed a high conditional accounting conservatism (CAC) in the pre-adoption period should respond to the change. We conduct a difference-in-differences analysis using a large sample of matched bank-firm data around the introduction of the dynamic loan loss provisions in Spain in 2000. The main result is that banks with a high CAC in the pre-adoption period significantly increased their risk taking in the post-adoption period. These banks lend significantly more to ex ante riskier borrowers, accept more borrowers with lower accounting quality, and they display a higher loan growth. Our findings on bank risk taking are consistent with reduced screening and monitoring incentives and highlight unintended side effects of the change in the loan loss provisioning rules for banks. (Coauthors Manuel Illueca and Gregory F. Udell.)