The effect of corruption on microfinance loan portfolio: A semiparametric analysis

In this paper, we examine the extent to which corruption affects the loan portfolio of microfinance institutions (MFIs). We employ robust econometric estimation on a sample of 507 MFIs across 63 countries from 2005 to 2018. Our results show that corruption is negatively associated with the loan portfolio. However, in semiparametric analysis, we find that lower-level corruption is beneficial to increase the loan portfolio while higher-level corruption is detrimental. The results imply that it is not just corruption that matters as far as its effect on MFIs' loan portfolio is concerned; what matters is the degree of corruption. In further analyses, we find that corruption reduces both the number of active borrowers and average loan per borrower indicating that corruption reduces both coverage and amount of credit extension. The results suggest that the effect of corruption on the loan portfolio is gender-sensitive. Corruption facilitates an increase in loans to female borrowers. Our results are robust to alternative variable measurements and different identification strategies, including two-stage least square.