Um modelo de duration para a estimativa da duração efetiva de contratos do agroamigo do Banco do Nordeste entre 2009 e 2011

• Autor(a) principal: Rousselet, Tiago
• Data de Publicação: 2018
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Ceará (UFC)
• Texto Completo: http://www.repositorio.ufc.br/handle/riufc/39796
• topic: Microcrédito; Crediamigo; Banco do Nordeste

The main objective of this study is to estimate the effective duration of microcredit contracts using data from contracts and the profile of the borrowers of Banco do Nordeste’s Agroamigo between 2009 and 2011. Regularization and multivariate regression techniques were compared, as well as a treatment for data (the first one with the removal of observation units with missing data and the other with the creation of dummies called "Missing" in the missing cases). The regularization model using the dummy base "Missing" presented the least prediction error for the effective duration. The total difference between the actual duration of the contracts and the estimated duration was 0.1057 years, which made it possible to reduce the error in the present value of the simulated contracts of R $ 14.49 between the actual value and the contract value for R $ 0.77 between the actual and the estimated value, obtaining a reduction in the projection error of -94.7%. It was also estimated the change in the present value resulting from the application of fines of 1% per year against late borrowers and a reduction of -0.15% in the average loss was evidenced with the consequence of an increase in deviation of 0.86%.