The Impact of Microcredit in Rural Bangladesh : A Fuzzy Regression Discontinuity Design Approach

Microcredit has expanded rapidly in recent years but its effect at the household level is still controversial and there is little empirical evidence on its externalities to the non-borrowers. This paper estimates the effect of participation in microcredit programs on expenditure and savings both in micro and macro level using a unique dataset collected from 69 villages in Bangladesh during the period 2006-2007. A regression discontinuity (RD) design is used to identify the credit effect. To justify the validity of RD design we test the discontinuities in the conditional density of the forcing variable (in our case land), as suggested by McCrary. We did not find any substantial evidence of manipulation of forcing variable at the threshold to invalidate the RD design. The results show that access to microcredit increase household saving, possibly for start a new business or expending the existing enterprise, instead of spending on durable or nondurable goods. We also find that 10% increase in the average credit would yield a 23% increase in expenditure for per school going child. At the village level microcredit generates positive externalities to non-borrowers on their expenditure