An Empirical Analysis of Loan Loss Provisioning Behavior: The Case of U.S. Banking Sector

This paper is an examination of loan loss provisioning behaviour in the American banking sector spanning from 2011 to 2018, using the Generalized Method of Moments (GMM) to jointly test three traditional hypotheses. These hypotheses consist of income smoothing hypothesis, business cycle hypothesis and capital management hypothesis. In addition, X-efficiency is also incorporated into the GMM model as one independent variable affecting provisioning decision, as well as it firstly estimated by the Stochastic Frontier Analysis. From the empirical results, there is no obvious evidence to confirm that U.S. banks use their loan loss provisions for income smoothing, as well as for the management of regulatory capital. Whereas the business cycle has an impact on banks' LLP decision-making. Furthermore, it is notable that large commercial banks in America have a relatively high level of cost efficiency, however, its efficiency scores do not influence loan loss provisioning behaviour across sample bank