Method of indirect estimation of default probability dynamics for industry-target segments according to the data of Bank of Russia

TL;DR

This paper introduces an indirect modeling approach to estimate default probability dynamics for industry-specific segments using Bank of Russia data, validated through retail lending statistics, aiding macroeconomic credit risk analysis.

Contribution

It develops a novel filtering model based on debt balance equations and Hodrick-Prescott filtering to estimate default probabilities without direct statistical data.

Findings

Model effectively estimates default probabilities for corporate segments.

Validation confirms the model's reliability over historical data.

Provides exogenous variables for macroeconomic credit risk modeling.

Abstract

A direct method for calculating default rates by industry and target corporate segments is not possible given the lack of statistical data. The proposed paper considers a model for filtering the dynamics of the probability of default of corporate companies and other borrowers based on indirect data on the dynamics of overdue debt supplied by the Bank of Russia. The model is based on the equation of the balance of total and overdue debts, the missing links of the corresponding time series are built using the Hodrick_Prescott filtering method. In retail lending segments (mortgage, consumer lending), default statistics are available and supplied by Credit Bureaus. The presented method is validated on this statistic. Over a historical limited period, validation has shown that the result is trustworthy. The resulting default probability series are exogenous variables for macro_economic modelling of sectoral credit risks.