MACROECONOMIC SCENARIOS BUILDING AND BANK CREDIT RISK ASSESSMENT, BY THE EXAMPLE OF PSC PROMINVESTBANK

Based on foreign research, the article developed a methodology for stress testing of the bank and carried out its approbation in Prominvestbank. Dependence of the level of non-payments on macroeconomic indicators is determined by the logistic function, the parameters of which are determined based on the maximum likelihood method on historical data. The macroeconomic indicators that are most important for constructing a qualitative model for forecasting the level of non-payments are determined. A forecast of the future level of non-payments for a basic and stressful scenario is made. For this purpose, based on ARIMA models 1000 different macroeconomic scenarios that cover each quarter from 2017 to 2019 inclusive were built; 1000 different forecasts of the level of non-payments were received. For each quarter, the median of this set is chosen as the baseline level of non-payments, and 99% quantile is used as an adverse one. The corresponding sets of macroeconomic data are the baseline and adverse scenario. As the next steps for each scenario, the projected level of losses of the bank, the level of its capital, as well as the capital adequacy ratio are determined. Based on the latter, one can draw conclusions about the potential stability of the bank in the event of economic shocks. The stress testing technique allows, based on autoregressive models such as ARIMA, to generate a baseline and adverse macroeconomic scenarios, as well as analyze the losses, capital and capital adequacy ratio of the bank for each of these scenarios.