Developing a Credit Risk Model for Mauritian Bankers

Abstract

The objective of this chapter is to develop a credit risk model which gauges the repayment capacity of household borrowers in Mauritius. The research is innovative as it differs from prior credit risk models by moving away from default modelling to repayment capacity modelling through the use of the Debt Service Coverage Ratio (DSCR). The model was developed by the author while he was working for an international bank. The chief benefit of our modelling approach is that it is not only practical, but also induces a sense of proactivity as the lender often has knowledge pertaining to the underlying forces which either boost or undermine DSCR. To trigger a holistic assessment, a whole set of models is employed such as continuous (Ordinary Least Squares), binary (Probit and Logit) and a ‘discretisized’ (Ordered logit) metric for DSCR in the study. To unleash richer results, our model provides due consideration to the interaction variables. Findings show that DSCR is positively influenced by margin cover and the loan tenor. Negative forces are conspicuously noted in the case of the cost of the loan and dummy variables such as loan for construction and borrowers who are employed by the public sector. Some evidence prevails as to the interaction between town dummy and margin cover, both of which have an effect on DSCR. Poor evidence is found in the case of age, arrears and town dummy. In general, findings advocate that a careful approach should be taken by banks whenever they compute the margin cover as it is found to constitute a particularly vital risk-mitigating mechanism when it comes to alleviating the moral hazard problem that is inherent in any lending activity.