A model for risk classification of banks

This paper deals with designing a bank risk classification scheme based on readily available performance data. This risk rating is referred to as ‘Risk rating’. Due to non-availability of data on CAMEL rating (C rating), R risk rating has potential for studying risk-based premiums insurance policy and for determining optimal frequencies for variable frequency on-site examination policy. A composite non-performance measure is developed to estimate probability of failure of a bank based on performance data available in bank call reports by fitting a Logit curve and estimating its parameters using maximum likelihood method. Division of banks into healthy and watchful types is based on critical dividing value of probability of failure.