STUDY OF THE DISCRIMINATORY ABILITY OF ROA AND ROE FINANCIAL RATIOS TO IDENTIFY PROBLEM BANKS (RUSSIAN EXPERIENCE)

Abstract

The approach is to construct a generalized linear regression model with a dependent variable of the indicator of default of the Bank. The explanatory variables in this model are the financial coefficients ROA and ROE. According to the constructed model, the ROC curve is reconstructed and the Gini coefficient is calculated, which serves as a quantitative measure of the discriminative ability of the exogenous variables ROA and ROE to explain the bank default.