What do we know about the impact of microfinance? The problems of statistical power and precision

• Published in: World development 2020-04, Vol.128, p.104773, Article 104773
• Main Authors: Dahal, Mahesh, Fiala, Nathan
• Format: Article
• Language: English
• Subjects: Clients; Clinical trials; Microfinance; Policy making; Power calculations; Profits; RCTs; Replication; Statistical power
• ISSN: 0305-750X; 1873-5991
• DOI: 10.1016/j.worlddev.2019.104773

•Studies of microfinance generally show no or minimal impact from providing microloans.•We replicate all RCTs on microfinance and find that results are insignificant, though most coefficients are large.•Every one of the studies is significantly underpowered to detect reasonable effect sizes.•Analysis on a pooled sample shows treatment effects of 28% to 40% increase in profits, significant at the 1% level.•This suggests that we still know very little about the impact of microfinance. We review all eight randomized control trial studies of microfinance published in peer-reviewed journals. The studies generally show no or minimal impact from providing microloans to clients and have led many researchers and policy makers to conclude that microfinance has been proven to have little or no positive impacts on people’s lives. We review these studies in detail and find four main results. First, we are able to replicate the results using the researcher’s original data. Second, we observe that while the results are generally insignificant at traditional levels, most estimated coefficients are large. Third, every one of the studies is underpowered to detect reasonable effect sizes, often due to low take-up of the financial product offered. Pooling the data from the studies together improves power for most outcomes, but minimum detectable effect sizes are still large. Finally, when we run analysis on a pooled sample, we find a treatment effect on business profits, business revenue and household assets, significant at the 1% level. We argue that existing research on the impact of microfinance is generally underpowered to identify impacts reliably and suggests that we still know very little about the impact of microfinance. We end by discussing ways to improve future research.