Ex-post credit risk and the cyclicality of banks' self-discipline: empirical evidence from US banks

PurposeThe aim of this research is to empirically examine if both credit cycle of large lending and business cycle affect the ex-post credit risk (i.e. non-performing loans (NPLs)) in the banking system of USA. A unique data set is created by using the Statistics on Depository Institutions (SDI) report compiled by the Federal Deposit Insurance Corporation (FDIC) covering the period between 2010Q1 and 2021Q4.Design/methodology/approachThe differenced generalised method of moments (GMM) is used as econometric methodology to deal with endogeneity issues that may arise when we regress non-perfoming loans (NPLs) on the credit cycle of large lending (proxy of banks' self-discipline) and business cycle (proxy of macroeconomic conditions).FindingsWhat we found is that both current credit cycle of large lending and one period lag business cycle can influence the US NPLs negatively due to the self-discipline role of large lending and the adverse macroeconomic conditions respectively. The excess credit influence of credit cycle of large lending on the US NPLs emerges after a period lag. Similarly, unexpectedly, current business cycle affects the US NPLs positively. Moreover, we noticed that the US NPLs have not a symmetric sensitivity between both business cycle and credit cycle of large lending.Originality/valueThe authors offer crucial evidence that credit cycle of large lending can also be associated negatively with NPLs enriching the relevant literature, which supports that credit cycle influences positively NPLs due to loan credit expanding, implying that NPLs can be reduced by large lending through the improvement of banks' cost efficiency.