The effect of the covid-19 pandemic on different types of microfinance borrowers - Empirical evidence from Uganda

Abstract

Microfinance has been growing in significance over the past decades, with considerable increases in the number of microfinance loans taking place especially across the developing world during this period. It has even been considered as one of the most effective ways to combat poverty, though it has also been the target of some criticism. At the onset of the COVID-19 pandemic, these microfinance borrowers were likely rather exposed to both the health- and economic effects caused by it and the various government responses to it.

In this master’s thesis, I study the effects of the COVID-19 pandemic and the related lockdowns on particularly affected borrowers, as well as how microfinance borrowers differentiated by certain borrower level attributes fared during the same period. For this, I utilize various data that originate mainly from BRAC Uganda, a microfinance institution that operates in Uganda. The data includes both loan performance data and survey data collected from the borrowers.

I split the empirical analysis in to two sections, with the first of which focuses on the differences experienced by those borrowers who were employed in a field affected by the lockdowns. I use the difference-in-differences method, supplemented by a two-way fixed effects regression to infer a treatment effect based on being one of the borrowers affected by the lockdowns. I find additional effects ranging from 0.421% to 0.718% for being in the affected group for the general population with the difference-in-differences framework, when measuring the fraction of their loans disbursement the borrowers had incurred in arrears. The two-way fixed effects results support the difference-in-difference results, though neither effect persists for all the alternative specifications.

The second section of the empirical analysis applies the logistic regression framework to study how borrower level attributes can be used to predict a borrower’s performance in loan repayment. I find that how the borrowers report using their loans, and how they perceive their COs interactions with themselves and the group both may be good indicators for this purpose.