Year of Graduation
Statistical Analysis and Valuation of the Credit Risk of Commercial Banks in the Russian Federation
School of Statistics, Data Analysis and Demography
STATISTICAL ANALYSIS AND VALUATION OF THE CREDIT RISK OF COMMERCIAL BANKS IN THE RUSSIAN FEDERATION Egor SafarbakovMoscow, NRU HSE,scientific directorprof. Arkhipova M.With the development and expansion of the credit market in Russia the task of predicting the customer creditworthiness and reliability is more important for banks. The scoring system of classification became the most widespread in the world. Scoring models are an important and integral part of each forecasting system of credit risk. There are used to solve the three major tasks: evaluation of the borrower under a loan (application scoring), monitoring the behavior of the borrower over the life cycle of the loan product (behavioral scoring) and evaluation of the borrower's debt collection (collection scoring). Building of such models requires information about borrowers in terms of their ability to pay, and experience in dealing with other banks. In foreign practice for a long time this information is provided by specialized organizations - credit bureaus. After the adoption of the relevant federal law in 2005, the credit bureaus operate in Russia. Currently, their number exceeds 30. Largest of them: "National Credit Bureau", "Equifax Credit Services" and "United Credit Bureau".Creating of quality scoring model which can effectively divide borrowers into "good" and "bad", requires not only the availability of representative data, but also a deep understanding of the market lending, its trends and challenges prevailing in the banking practice.Despite considerable scientific interest in this subject the domestic scientific literature almost has no works on the statistical analysis of the consumer credit market in the Russian Federation. The structure and trends of modern consumer credit market in Russia were analyzed in the work. Five stages of formation of retail lending were highlighted. Development of the credit market, which peaks at 2012, contributes to the rapid increase of the delay. The banks’ portfolios should go the way of recovery, which would be a more careful selection of borrowers and the introduction of new developments for risk management.To build the Application scoring model there were taken the data from the commercial bank in Russia and three largest credit bureaus.The developed mathematical model let to identify patterns between individual characteristics of the borrower and its creditworthiness. Also it gives recommendations to banks in order to determine better the degree of risk in lending. The existing quality characteristics for scoring models were described in the work, marking advantages and disadvantages of each. All characteristics showed excellent efficacy and significant predictive power of the model.