Modeling and Empirical Validation of Revised Altman's Credit Risk Model for Indian Banks

Abstract

The major cause of serious banking problems over the years continues to be directly related to lax credit standards for borrowers and counterparties, poor portfolio risk management, or a lack of attention to deterioration in the credit standing of a bank's counterparties. According to BIS, credit risk is most simply defined as the potential that a bank borrower or counterparty will fail to meet its obligations in accordance with agreed terms. The effective management of credit risk is a critical component of a comprehensive approach to risk management and essential to the long-term success of any banking organization. Since exposure to credit risk continues to be the leading source of problems in banks world-wide, banks and their supervisors should be able to draw useful lessons from past experiences. Hence, in this paper, we try to address how banks assess the creditworthiness of borrowers. We understand that banks consider, among other factors, the current and prospect patrimonialisation and profitability, the borrower's history, as well as its industrial sector and how the borrower is positioned in it. The objective of the paper is to develop an internal credit rating model for banks which will improve their current predictive power of financial risk factors. The study aims at validating the efficiency of Altman's Z-Score model for credit risk evaluation through empirical data. Further a new revised model is developed for Indian banks. We try to fit-in credit risk models like the Altman's Z-Score Model (with modifications of some variables) to the Indian Banks, and try to arrive at an equation of the Z-Score, which will enable the banks to predict future defaulters, and hence take the necessary actions accordingly. Both the models are applied and tested on a sample of 60 default and non default companies. The revised Altman's model is able to predict default one year prior with an accuracy of over 80%. This is further verified on a separate holdout sample.