Análise de concessão de crédito à pessoa física em uma empresa comercial com a utilização da técnica de análise discriminante

Abstract

This paper is intended to propose a probability model of credit risk to individual entities, in a commercial company, with the use of the statistical method called Discriminant Analysis. In relation to methodology, this research is classified in quantitative, review of literature and descriptive, and the method chosen for it was the case study. For this research it was used a sample of 95 customers who purchased in installments between 1/8/2012 and 31/1/2013. The customers‟ information was obtained from the database registration provided by the company. First of all, the profile of such customers were identified through descriptive statistics method, and in order to draw up a model of credit risk, it was used – as a dependent variable – the default; and four independent variables were also tested (gender, age, income and marital status). The findings demonstrated that the profile of the individual entity customer – who purchased in installments – are mostly women (51,58%), married (54,74%), average age 51 and average monthly income R$1.561,31. Among the explanatory variables analyzed, the only one statistically significant – with a level of significance of 10% - was the Variable Income (p-value = 0,05).