Empirical Modeling of International Banks’ Credit Risk: Assessment and Comparison of Credit Ratings

The work is devoted to credit risk modeling of international banks by constructing ordered logistic models of credit ratings assigned by the agencies: Moody’s, Standard & Poor’s, and Fitch ratings. It was demonstrated that mapping the credit ratings into a base scale helps to decrease the possible subjectivity of CRAs and increases models’ forecasting power. Using a random sample of 478 banks from more than 40 countries for the period of 2007–2019, a credit rating model was obtained that could be effectively used to assess credit risk using public information, which was demonstrated with a help of out-of-sample forecasts of the obtained regression. The quality of the model prediction was significantly improved by including interaction terms and applying the Principal Component Analysis. We observed that the transparent politics of the government can lead to increased credit ratings of international banks. This helps them in credit risk management and in the client’s base extension. Additionally, using marginal effects calculation, the empirical evidence of the importance of maintaining a high level of liquidity by banks in times of crisis was found. At the same time, the highest marginal effects of the whole sample time horizon were reached in factors of asset quality and bank’s size among all financial indicators included in the model.