O uso de indicadores não contábeis para a redução do risco de concessão de crédito comercial: modelo de apoio à decisão baseado na técnica estatística de regressão logística

• Autor(a) principal: Jesus, Edmilson Luiz de
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações do Mackenzie
• Texto Completo: http://tede.mackenzie.br/jspui/handle/tede/4482
• topic: risco de crédito; risco de crédito; regressão logística; inadimplência; CNPQ::CIENCIAS SOCIAIS APLICADAS::ADMINISTRACAO

Given the relevance of applying statistical models to support the credit decision – making process, this dissertation proposes the adoption of a commercial credit decision support model based on the application of the statistical technique of logistic regression. The model sought to identify the variables chosen by the credit analyst at V.I Industria e Comercio Ltda. To base the concession decisions. With a client base composed of companies of various segments and sizes,the studied company uses a judgmental decision model based on information from various sources and on the behavioral history of its customers. Although the system involves significant subjectivity in the decision-making process, the non-accounting indicators represented by the registration and historical variables used as parameters for deciding on the concession have been reasonably effective, since historical data revealed that in the period from 2014 to 2018, default rates observed in the company were below 1%, well below the rates published by public institutions and companies specializing in credit risk.