Collection Score and the Opportunities for Non-Performing Loans Market

• Autor(a) principal: Gonçalves, Eric Bacconi
• Data de Publicação: 2021
• Outros Autores: Santos, Francisco Carlos Barbosa dos, Bontempo, Paulo César
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista JRG de Estudos Acadêmicos
• Texto Completo: http://revistajrg.com/index.php/jrg/article/view/503
• topic: Non-performing loans; Collection Scoring; Regressão Logística; Modelos Estatísticos; Cobrança; Non-performing loans; Collection Scoring; Logistic Regression; Statistical Models

In the academic literature, credit scoring models are widely studied, while collection scoring models are less explored; likewise, there are few articles dealing with the Brazilian non-performing-loans market. This work has as main contributions: the use of scoring models in the area of collection and working with non-performing-loans data. The objective of this paper is to develop a collection scoring model through Logistic Regression to identify, in a portfolio of clients with non-performing-loans, to verify if it is possible to adjust a good model and to indicate which clients are more likely to pay the debts nonperforming credits The results show that the model worked well for the database, obtaining an excellent fit (accuracy of classification greater than 83% for the two samples and KS=68), pointing the viability of this methodology.