Bank Credit Risk Modeling in Emerging Capital Markets

Abstract

Models for assessing the probability of default play an important role in the risk management systems of commercial banks, as they allow assessing the creditworthiness of various counterparties and transactions. Many Russian banks are trying to switch to an advanced approach based on internal ratings (IRB-approach) for evaluating regulatory capital. The main goals that banks pursue when switching to an advanced approach are: stability of credit risk assessment for the ability to carry out strategic planning; the validity of the credit risk assessment to simplify interaction with the regulator and external and internal audit; potential reduction of regulatory capital due to the high quality of the forecast capabilities of the developed models, which leads to a reduction in the regulatory capital of banks. To use internal rating models in the calculation of regulatory capital banks serve the petitions on them to the regulator, on basis of which external validation of the models is carried out and a decision about the possibility of using models for regulatory purposes is made. The main event of credit risk, the default event is determined by banks in the framework of credit policy, is consistent with the Central Bank and is predicted using models for assessing the probability of default. The PD models are the most popular in banking practice due to the fact that according to regulatory requirements, they are developed on the horizon of 1 year, and the minimum amount of statistical data for such models must be at least 5 years. The risk segments are identified using both economic and statistical evaluation criteria based on the banks available empirical data for each group of borrowers to build separate models (Allen, Financial risk management: a practioner’s guide to managing market and credit risk. Wiley, Hoboken, NJ, 288 p, 2003; Lobanov and Chugunov, Encyclopedia of financial risk management, 4th edn, Alpina Business books, 932 p, 2009; Rogov, Risk management, Finance and statistics, Moscow, 120 p, 2001). This paper will describe the specifics of developing models for low-default risk segments (bank assets), both low-default and high-default risk segments (corporate borrowers), and high-default risk segments, including taking into account the availability of a small amount of static data (residential real estate lending and project finance segments).