Expected loss recognition and banks’ management forecasts

Abstract

Accounting rules for credit impairment recognition have been shifting to a more forward-looking approach based on expected losses. We examine how the adoption of an expected loss model (ELM) influences banks’ management forecasts, which also are forward-looking. In a difference-in-differences setting of gradual implementations of the ELM worldwide, we find that banks enhance management forecasts after adopting the future-oriented provisioning model, as manifested in higher likelihood of forecast issuance, higher frequency of forecasts, more precise forecasts, and higher overall forecast quality. This forecast-enhancing effect is more prominent when accounting standards are more strictly enforced, when banks experience larger changes in loss allowance after ELM implementation, and when forecasting is more challenging such as during the onset of the COVID-19 pandemic. Moreover, banks’ post-ELM forecasting performance also improves in terms of greater forecast accuracy and persistency. Overall, our results suggest a complementary relation between expected loss recognition and banks’ management forecasts.