How Does Recognition of Forward-Looking Estimates Affect Learning about the Macroeconomy? Evidence from CECL

Abstract

We use the adoption of current expected credit loss reporting for banks as a setting to examine how recognition of managers’ forward-looking estimates affects their learning from stock prices. We posit that recognition of banks’ expected credit losses can reduce managerial learning in two ways. First, managers are forced to invest in information systems that allow them to generate decision-useful information and to rely less on alternative sources, such as stock price. Second, the disclosure of expected credit losses reduces incentives for investors to privately collect and trade based on this information, thereby reducing the informativeness of stock prices for banks’ lending decisions. We find robust evidence that bank managers learn from stock prices under the incurred loss model and that this learning is attenuated under the expected credit loss model. The results vary with banks’ information advantage over equity market participants and are robust to alternative treatment and control groups and measurement approaches. We present evidence that investors incorporate less macroeconomic information in banks’ stock prices and that the reduction in learning hurts banks’ lending efficiency.

Keywords: Forward-Looking Estimates, Managerial Learning, Expected Credit Losses

JEL Classification: M41, D83, G21