Loan Loss Measurement and Bank Lending

I develop and estimate a model bank lending under regulatory capital requirements that rely on accounting measurement. I use the model to assess how alternative measurement approaches to banks’ loan loss allowances affect lending. Through counterfactuals, I find that: (1) the Current Expected Credit Loss (CECL) method increases loan loss reserves on average by 14.7% relative to the Incurred Credit Loss (ICL) method; (2) the difference between CECL loan loss allowances and ICL loan loss allowances is larger in economic downturns than upswings; (3) banks reduce lending on average by 5.16% (71 basis points) when CECL had been imposed instead of ICL; and (4) CECL results in less procyclical lending than ICL, specifically, the sensitivity of lending to GDP growth decreases by 1.3% when moving from ICL to CECL