Modelagem de risco de crédito para o Programa Agroamigo do Banco do Nordeste do Brasil

Abstract

As the operating costs of microcredit are high, the institutions that operate it need to pay attention to credit market risks, in this work the concept of default risk was used, a broader term than default commonly used in financial institutions. In addition, microcredit programs have been widely accepted as a policy option to alleviate poverty among family farmers. The objective of this paper is to analyse the main factors that affect the default risk of financing for clients of the Agroamigo program, with micro-information from the Program, in the period from 2014 to 2019, and to propose a default risk model or a driver for monitoring and customer tracking. As local beneficiaries are directly or indirectly linked to their credit agents in the program, default risk probability models are estimated with random effects for them. The results confirm the need to use this functional form, suggesting the importance of the credit agent for the program and especially in the default risk modelling. Lower risk factors were also observed for female beneficiaries, with better educational levels, married, with up to 5 years of participation in the program and who had amounts above R$ 5 thousand.