Forecast of the credit risk of Moroccan companies

Abstract

Among the problems faced by banks is the assessment of the creditworthiness of borrowers. In this context, the credit rating method has been developed with increasingly sophisticated tools to address their concerns. The purpose of this article is to present and apply the logistic regression method as a means of classifying the business failure problem based on the accounting data of a sample of Moroccan companies. The results of our study are based on a database of 2,030 credit files granted to Moroccan companies (SMEs and GE) in 2015 and 2016. The resulting regression equation allowed us to select 12 variables (financial ratios). and identifiers) as important in analyzing the probability of non-repayment. After the necessary statistical tests and validation simulations, we recorded a prediction rate of over 93% for our model.