A logical analysis of banks’ financial strength ratings

We evaluate the creditworthiness of banks using statistical, as well as combinatorics-, optimization-, and logic-based methodologies. We reverse-engineer the Fitch risk ratings of banks using ordered logistic regression, support vector machine, and Logical Analysis of Data (LAD). The LAD ratings are shown to be the most accurate and most successfully cross-validated. The study shows that the LAD rating approach is (i) objective, (ii) transparent, and (iii) generalizable. It can be used to build internal rating systems that (iv) have varying levels of granularity, and (v) are Basel compliant, allowing for their use in the decisions pertaining to the determination of the amount of regulatory capital.

► Reverse-engineer the risk ratings of banks using the Logical Analysis of Data (LAD) method.
► Develop an LAD rating approach that is objective, transparent and generalizable.
► Show that the LAD ratings are the most accurate and cross-validate best.
► Can be used to build internal rating systems with varying levels of granularity.
► Provide Basel compliant credit risk ratings for banks.