Year of Graduation
Modelling of Defaulting Companies in the Financial Sector
Konstantin Lvovich Polyakov
This study is devoted to mathematical and statistical modeling of Russian banks defaults. The peculiarity of the work is the automatic determination of the most adequate specification of the model and the nature of the entry of variables. The aim is to identify the statistical relationship of financial indicators of banks with the probability of default. The simulation uses logistic regression with panel data, the functional form is determined by the apparatus of Multivariable Fractional Polynomials. It is shown that the inclusion of polynomials significantly improves the quality of the model.