Impact of Microfinance on Food Security, Informal Credit, and Agricultural Wages: The Case of Bangladesh Open Access
Downloadable ContentDownload PDF
Since its beginning in Bangladesh in the 1970s, the spread of microfinance worldwide has been impressive, estimated at over 100 million members in 2009. While proponents argue that microfinance has improved the lives of millions of poor people, critics have raised doubts about its efficacy as a poverty alleviation strategy. Despite the doubts, few rigorous studies of impact have been completed. My dissertation is a contribution to the ongoing debate, and focuses on Bangladesh as a case study. It consists of three empirical research essays: (i) Chapter 2 estimates the impact of microfinance on seasonal food security of the ultrapoor in Northwestern Bangladesh; (ii) Chapter 3 assesses its impact on the informal credit market; and (iii) Chapter 4 looks at its impact on women's and men's agricultural wages. We take advantage of a unique dataset of 280,000 ultrapoor households generously shared with us by the Institute of Microfinance (InM) in Bangladesh and we apply novel econometric techniques to address the endogeneity of microfinance. Chapter 2 uses a subset of 150,000 ultrapoor households in Bangladesh to analyze whether microfinance helps a household cope with aggregate shocks such as seasonal famine, known as Monga in Bengali. To address selection on unobservables, we use the "Minimum-Bias Bias-Corrected" estimator due to Millimet and Tchernis (2012) that corrects for endogeneity bias without exclusion restrictions. The empirical results suggest that microfinance improves food security of poor households by increasing food consumption during the Monga season. Also evidence is strong that microfinance helps households avoid distress sale of labor, and reduces the probability of short-term migration in search of work during the seasonal famine. The expansion of microfinance programs through the developing world has raised some interesting questions regarding its impact on the existing informal credit market. Chapter 3 analyzes this question by looking at both the impact on the village moneylender interest rate and on informal borrowing by households. We use two data sets from Bangladesh: (i) a large cross-section that includes more than 800 villages (extracted from the above InM data set); and (ii) a nationally representative panel with household-level data from 62 villages from BRAC. Relying on heteroskedasticity and matching to achieve identification (Klein and Vella 2009a, 2010; Millimet and Tchernis 2012), our results suggest that the spread of microfinance has had a positive and statistically significant effect on the moneylender interest rate. Microfinance membership seems to decrease the probability of borrowing from informal lenders but increase the average amount borrowed. We consider three alternative hypotheses to explain these results: increased demand, cream skimming, or fixed costs. The theory of cream skimming is most consistent with our results: as MFIs reach villages, they attract the better borrowers, leaving behind a riskier pool, leading the moneylenders to raise their interest rates accordingly. Chapter 4 investigates the impact of microfinance coverage on women's agricultural wages, looking into both the impact on the general, "Normal", agricultural daily wage and the wage prevailing during the Monga. To address the endogeneity of microfinance coverage, we employ the Klein and Vella (2009a) instrumental variable and the Minimum Bias estimator of Millimet and Tchernis (2012). The results suggest that microfinance expansion tends to increase the average wage earned by women in agriculture, both during the Monga and Normal seasons, a significant poverty alleviation measure since only ultrapoor women work as agricultural laborers. The wages of men are found to increase as well. As microfinance expands, women tend to substitute their time away from the wage market towards their microenterprise. Consequently, the labor supply curve shifts to the left thus increasing the wage rate for those who continue to work in the labor market. Also by improving the situation during Monga, microfinance has an even greater impact during the non-Monga season, when more laborers--both men and women--now work at the going wage. In brief, the results of our research suggest that microfinance does have a significantly positive impact on the welfare of the ultrapoor. Microfinance membership reduces the likelihood that households will be forced to skip meals and practice distress sale of labor during the hungry season, while also reducing the poor's reliance on informal "exploitative" sources of finance, and increasing the agricultural wages earned by both men and women in the villages. At the same time, by capturing the less risky borrowers, MFIs do not bring down the interest rates for those fewer villagers who continue to borrow in the informal market. But higher average interest rates in the informal market do not necessarily worsen the situation of this group since they were risky borrowers in the first place, likely facing higher interest rates regardless. In fact they may even have easier access to informal credit as suggested by the larger average loan size.