The Effect of the Current Expected Credit Loss Model on Conditional Conservatism of Banks and Its Spillover Effect on Borrower Conservatism

Abstract

Under the Current Expected Credit Loss (CECL) model, banks should fully recognize expected lifetime credit losses upon loan origination while gradually recognizing interest revenues. This timelier recognition of losses versus gains (i.e., conditional conservatism) makes banks more capital constrained. To mitigate this, banks may (1) offset timelier credit losses by lowering conservatism in other earnings components and (2) reduce credit losses by demanding greater borrower conservatism. We find that, under CECL, banks increase conservatism in loan losses but decrease conservatism in other earnings components, making overall conservatism only marginally increase. In sharp contrast, their borrowers increase conservatism by 40 percent, and borrowers' increase is twice that of banks. This substantial spillover effect suggests that, by greatly increasing borrowers' conservatism, CECL may strengthen debt governance of a broad scope of firms in the economy, thereby having economy-wide consequences beyond the banking industry and potentially enhancing the stability of the entire economy.

Keywords: Current Expected Credit Loss Model, accounting conservatism, bank regulation, debt governance. JEL Classifications: G21, M41, M48