MICROFINANCE AND CONSUMPTION INEQUALITY: EVIDENCE FROM A RANDOMIZED CONTROLLED TRIAL IN ANDHRA PRADESH, INDIA

Existing evaluation studies of microfinance programs typically examine impact of microcredit on its borrowers in terms of certain broad indicators of development such as consumption, income, health, education, livelihoods, etc. Microfinance, a collateral free direct access to group-based credit, leads to financial deepening which has implications for consumption inequality. However, there is dearth of studies that systematically examine the relationship between microfinance and consumption inequality. Existing studies that examine such relationship lack methodological rigor. This paper fills this particular void in the literature. We examine the impact of access to microfinance on consumption inequality using a household-level panel dataset. The impact is analyzed in two ways viz. within slums and within treatment group. We compute standard measures of inequality such as coefficient of variation (CV), Gini coefficient, and Theil index and use regression technique to estimate the impact of microcredit on consumption inequality. We further test robustness of our results using bootstrap, a non-parametric technique used extensively in applied research. An inequality decomposition analysis following Shorrock's (1982) method is also conducted. The data is taken from three rounds of household surveys viz. baseline and two follow-up surveys conducted by Banerjee et al. (2013) as part of the randomized evaluation of microcredit programs carried out in 104 slums in Andhra Pradesh, India. We find that access to microcredit exacerbates consumption inequality both within slums and within experiment groups. A follow-up decomposition of inequality indices shows that this difference in consumption inequality is predominantly driven by expenditure on non-food items. However, once all households across treatment and control slums have equal access to microcredit in the long-run, the disparity in consumption inequality between treatment and control slums dissipates. Our results suggest that larger loan size and number of loan cycles do not matter for consumption inequality across treatment and control households. These results also suggest need for targeted livelihood support programs for those who cannot participate in microcredit programs.