Accounting for the determinants of banks’ credit ratings

Abstract

The contribution of the banking industry to the recent financial crisis 2007/8 has raised public concerns about the excessive involvement of banks in risky activities. In addition there have been public concerns about the ability of credit rating agencies to evaluate these risks in advance. In this context, this study uses an ordered logit analysis to examine the determinants of banks’ credit ratings using a sample of US and UK banks’ accounting data from 1994 to 2009. Our intention is to examine to what extent banks’ ratings reflect banks’ risks. Our analysis shows that a small number of accounting variables, namely: bank size, liquidity, efficiency and profitability are able to correctly assign credit rating for approximately 74% to 78% the sample banks. Surprisingly, the association between banks’ credit ratings and each of leverage asset quality and capital is not robust, suggesting that the rating agency’s models did not pick them up despite their importance in the crisis. In addition, the relationship between banks’ credit ratings and liquidity is the reverse of that which an adequate early warning system would require. As banks benefit from higher credit ratings they will have addressed their determinants rather than taking care of systemic factors that affect underlying risk. Policy makers therefore need to intervene to address this market failure.