Microcredit: Rhetoric vs Reality

Many people have inferred from the results of the first wave of microcredit randomised controlled trials that microcredit has had a negligible average impact on business performance and household welfare. In this post I am going to argue that, firstly, these people are correct in that microcredit has had a negligible impact on certain measures of business performance, and indicators of welfare, across the microfinance population in the randomised studies. However, I will then argue that this does not mean that it has had a negligible overall impact on peoples’ lives. Firstly, I will draw attention to the fact that the microcredit studies only analysed a specific subpopulation – compliers in expansion of credit. Secondly, the muted average impacts mask important heterogeneity in business performance, with some businesses benefitting immensely. Lastly, just because microcredit has not transformed many businesses does not mean that it has not been welfare enhancing; its role as a money-management tool is vital.

The initial vision of microcredit, championed by Grameen Bank founder Muhammad Yunus, revolved around the idea that the poor were ‘frustrated entrepreneurs’. They had the ideas, knowledge and skills to run their own highly-profitable businesses – all they lacked was the start-up capital. Thus, the initial rhetoric was very much a story of the poor facing a unidimensional poverty trap, whereby there is a threshold level of income needed to invest in businesses and prosper. Microcredit, by providing this credit, would transform businesses and help families escape the poverty trap.

Contrary to expectations, this is not what the initial evidence showed. Banerjee et al. (2015) evaluated six randomised microcredit expansions in six diverse countries and failed to find transformative effects on business or welfare indicators in any of them. Notably, they found that microcredit had very modest take-up. This provides clear evidence against the vision that the majority of the poor are frustrated entrepreneurs who are yearning for provision of capital to transform their businesses. In terms of impact on welfare indicators, none of the six studies found a statistically significant increase in total household income, and none of the four studies with household consumption found any positive effect. Additionally, the authors found a lack of transformative effects on business activity. Businesses were started and sustained, however, they failed to show the rapid growth that the initial visionaries of microcredit expected. The authors say that all studies did find some evidence of increased activity; however, businesses simply did not grow on average.

To get a more holistic picture of the impact of microcredit expansion, and to overcome some econometric issues, Meager (2019) performed a meta-analysis of the microcredit expansions using a partial pooling method that accounts for heterogeneity in the different contexts. She found that, overall, the average treatment effects on household business performance and consumption were typically small and uncertain. They seemed to be around 5% of the control mean – certainly not transformative.

Therefore, when analysing the overview of the initial microcredit randomised controlled trials, it is easy to understand why some say that microcredit has had negligible effects on both business performance and welfare. However, I do not believe that this is the case. Firstly, as highlighted by Morduch (2020), the experiments discussed above focus on a very particular aspect of microcredit – expansion to marginal customers. The studies focus on customers who only took up the product when new institutions expanded into their areas. However, these impacts could be very different from the those of the ‘infra-marginal customers’ – the customers who were the first to be targeted by microcredit institutions. Wydick (2016) argues that, in the case of microcredit, we have good reason to expect that the impact on the early-adopters exceeds those of the later marginal adopters. This corresponds with economic theory; those early adopters may have been desperate to obtain capital as they were the ones with the highest return to capital. This implies that, while the available evidence may yield unbiased estimates of the impact of microcredit on marginal customers, it is likely to underestimate the impact of microcredit as a concept.

Beyond this distinction, however, the average effects of the expansion of microcredit mask important heterogeneity. Banerjee et al. (2019) highlight that individuals are likely to benefit from microcredit in different ways; some may borrow purely for consumption, others may start a business but lack the ambition or skills to grow it, and only a small subpopulation have the willingness and potential to use the credit to transform their businesses. The authors call this last group ‘gung-ho entrepreneurs’ (GE). In her meta-analysis, Meager (2019) found that, on average, and in five of the six studies, individuals who owned a business prior to microcredit expansion had more positive treatment effects than the average. In the Hyderabad study, Banerjee et al. (2015) found very small average effects; however, for businesses that existed before microcredit was introduced, they found a 100% increase in profits. In light of this, Banerjee et al. (2019) follow up on the Hyderabad results in a new wave collected six years after the treatment had access to microcredit, and four years after control gained access. In this time, there was a severe microcredit crisis in the state and all institutions discontinued operations two years before the new wave of data was collected. Despite this, they still found positive effects that were entirely being driven by the GE. These firms’ assets, investment, employment hours, revenues and profits were all significantly higher among treatment firms than control, despite the control firms gaining access to microcredit, and both losing access completely. This thus suggests that, while the expansion of microcredit may have had negligible impacts on business performance on average, for the subpopulation of owners who had existing businesses, it had large, important effects.

Lastly, even if microcredit has negligible effects on both business performance and consumption levels, this does not mean that it does not have important welfare impacts. By focusing only on business owners, there may be a fundamental gap between microcredit’s vision and what it actually is. The evidence suggests that individuals are taking out loans that they may not have been able to obtain otherwise; thus, perhaps, what microcredit offers the most is a set of money-management tools. It allows poor families to obtain the right amount of money at the right time to manage adverse shocks and fluctuations to income. This is seen vividly in the Netflix documentary Living On One Dollar, where families in Nicaragua are shown to be using their loans to start small businesses and manage shocks to the household. In fact, Banerjee et al. (2015) said that the lack of transformative effects should not obscure other important effects that microcredit has. The studies find evidence that access to credit improved occupational freedom, consumption choices, female decision-making and risk management. Thus, if development is freedom (as in Amartya Sen’s capability approach), and freedom means having a larger choice set, then microcredit is certainly successful development. In light of this, Morduch (2013) argues that if microcredit deserves support, it’s because it provides means for poor families to reduce the stresses of everyday life; families will never get ahead if they fail to address emergencies.

In sum, the preliminary evidence of the microcredit randomised controlled trials suggested that microcredit had negligible impacts on business performance and welfare indicators such as income and consumption. Thus, no matter what, microcredit certainly did not live up to Yunus’ initial vision. However, I argue that this does not mean that it has had negligible effects. First, the studies only focused on one sub-sample of the population. Second, the average effects hide important, and large, heterogeneity on business performance, and third, improved financial management services offered by microcredit led to important welfare gains by allowing poor households greater agency.

Note: This blog post is based on an essay that I wrote for my master's at The University of Oxford.

References:

Banerjee, A., Karlan, D. and Zinman, J. (2015). Six Randomized Evaluations of Microcredit: Introduction and Further Steps. American Economic Journal: Applied Economics, 7(1): 1–21.

Banerjee, A., Breza, E., Duflo, E. and Kinnan, C. (2019). CAN MICROFINANCE UNLOCK A POVERTY TRAP FOR SOME ENTREPRENEURS? NBER WORKING PAPER SERIES.

Banerjee, A., Dluflo, E., Glennerster, R. and Kinnan, C. (2015). The miracle of microfinance? Evidence from a randomized evaluation.

Meager, R. (2019). Understanding the Average Impact of Microcredit Expansions: A Bayesian Hierarchical Analysis of Seven Randomised Experiments. American Economic Journal: Applied Economics, 11(1): 57-91.

Morduch, J. (2020). Why RCTs failed to answer the biggest questions about microcredit impact. World Development.

Morduch, J. (2013). How Microfinance Really Works. The Milken Institute Review.

Wydick, B. (2016). Microfinance on the margin: why recent impact studies may understate average treatment effects. Journal of Development Effectiveness, vol. 8(2), pages 257-265.