Prediction of a financial crisis in Latin American companies using the mixed logistic regression model

Abstract

The development of statistical methods for predicting the financial crisis of a company is a real contribution to scientific research. These methods identify possible adverse financial situations of the companies, through the behavior of their financial indicators. The contribution of this work is to compare the binary classification by different prediction methods of mixed logistic models to predict a future financial crisis in new companies. The results based on an application involving companies from the Argentina, Peru and Chile Stock Exchange showed that all prediction methods were able to predict with high accuracy the financial crisis of the next year.