• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site
Menu

Article

Концепция моделирования и прогнозирования вероятности отзыва лицензии российских банков

Биджоян Д. С., Богданова Т. К.

In this paper, we propose an approach to take into account the impact of daily changing macroeconomic variables in the development of a model for the probability of revoking a license from Russian banks on the basis of their annual financial indicators. The essence of this approach is that to take into account the influence of macroeconomic variables it is proposed to use not the average value for the sample, but the median and the volatility characteristics: standard deviation and variance. Based on the annual financial performance of banks for the period from 2004 to 2015, as well as the values of macroeconomic variables, a logistic regression model for estimating the likelihood of license revocation from Russian banks was built. The volatility of macroeconomic variables is characterized by the variables "Standard deviation" and "Variance". The problem of eliminating multicollinearity between these variables is proposed to be solved by including in the model standardized values of the variable "Standard deviation" and their squares. An approach is proposed for determining the cutoff threshold in binary selection models, according to which the I and II types of errors are assigned different weights. In the function of determining the cut-off value, an external parameter α is introduced that characterizes the investor's attitude to the I type error (the operating bank is classified as a bank with a revoked license). Classification of banks into operating banks and banks with a revoked license is based on the obtained of the cut-off value, taking into account the selected value of α, at which the function of determining the cut-off value reaches a minimum.

Thus, the approach of taking into account the volatility indicators of macroeconomic variables proposed in the study made it possible to improve the quality of the model for forecasting the revocation of a license from a Russian bank. The model has a stronger predictive ability than models that take into account only the average values of the exchange rate of currencies and other macroeconomic variables.