Credit risk analysis in microcredit: How does gender matter?

This paper is the first to analyze the credit risk of a microfinance institution based on the loan portfolio of a leading Maghrebian microfinance institution, both in terms of number of clients served and of portfolio size. This allows us to work with a proprietary data set of 1,144,770 contracts issued between 1997 and 2007. Using a resampling technique, we estimate the probability density function of losses and value-at-risk measures for a portfolio of loans granted to female and male microfinance clients separately. Results show that loss rates are higher for a male client population than for a female client population, both on average as for percentiles 95 to 99.99. We find that this difference is due to lower default probabilities for female clients, while recovery rates for male and female clients are similar. We also analyze diversification effects, where we find that the proportion of diversifiable risk in total risk is bigger for portfolios of loans granted to female clients than for portfolios of loans granted to male clients. Finally we show that capital requirements determined by the 99.9 percentile remain below those required by the Basel 2 Accords, which opens perspectives for a specific treatment of microfinance if financial regulation becomes applicable to the sector.