Prediction of Wilful Defaults: An Empirical Study from Indian Corporate Loans

Lakshmi Karthik ； M. Subramanyam ； Arvind Shrivastava ； A. R. Joshi

《International Journal of Intelligent Technologies and Applied Statistics》 11卷1期 (2018/03) Pp. 15-41

https://doi.org/10.6148/IJITAS.201803_11(1).0002

摘要

Using a sample of 558 Indian companies’ financial indicators, during the period 2002–2016, the study investigates empirically the utility of key financial variables that could mitigate the credit risk by improving information on the probability of wilful defaults. The paper has attempted to build a credit risk model using logistic regression to predict the probability of wilful defaulters and classify them with that of the non-defaulting companies. This study has employed an ex-ante approach to pre-empt identification of wilful defaults at an early stage. This could avert/minimize the high costs associated with identification, declaration and legal recourse undertaken by filing a suit. The result of this model with introduction of three cash flow variables has contributed positively to the existing financial distress prediction model. Further, it is empirically tested that the high outstanding debt increases the probability of wilful defaults since the benefits exceeds its implied costs. The resulting estimate from logistic regressions demonstrates that they are statistically significant and the coefficient estimates’ signs also possess the predicted sign. The model developed is able to classify with an accuracy of more than 70%, the likelihood of wilful default in the next 2–3 years and the non-defaulting companies with an accuracy of around 98.0% on the above data-set.

關鍵字

Credit risk ； Corporate default ； Predicting wilful defaulters ； Logistic model