Application of Artificial Neural Network to Predict Wilful Default for Commercial Banks in India

Abstract

Since 2014, the problem of the rise of Non-Performing Assets (NPA) in the Indian banking system has been a subject of investigation. A major impact of mounting NPA has tested the bank’s ability to recover bad loans and its capacity to lend in the short run. Among others, wilful default has been a significant category of NPA; it is detrimental to the financial health of the banking system. The share of wilful default in the total NPA for the year 2018 stands at 44%. By and large, wilful default indicates ‘intend of fraud’. As declared by various banks, around 106 companies are identified as wilful default companies from those listed between 2000 and 2018. The research paper constructs a model to predict the wilful default using an artificial neural network. The model is based on 106 wilful default companies and 106 non-default companies. The model predicts an accuracy rate of 92.2% and the variables with the highest degree of importance are found to be Profit before Interest and Taxes/Total Assets, followed by Enterprise Value/Total Assets, Operating Profit Margin, Cash Flow Financing/Cash Flow Investing, Total Debt/Total Asset, Sales/Capital Employed, Retained Earnings/Total Assets, Return on Shareholders’ Funds, PBIT/Sales, and others.

Keywords: Artificial Neural Network, Wilful Default, Non-Performing Assets, Bankruptcy Prediction Model

View PDF