A new theory of change is emerging for microfinance. People from poor households tap microfinance services to smooth consumption and build assets to protect against risks ahead of time and cope with shocks and economic stress events after they occur—leading to widespread poverty alleviation but not widespread poverty reduction.
This is the narrative coming out of the financial diaries reviewed by Portfolios of the Poor and the research summarized in Poor Economics and Due Diligence. More or less independently, perhaps in spite of that research, the microfinance industry has been adjusting toward supporting resilience strategies of the poor as it has become more sensitive to client demand – by moving toward a mix of loan, saving and other services and greater flexibility and choice to accommodate the use of microfinance for supporting diverse household needs rather than focusing just on the needs of microentrepreneurs.
What is the evidence so far to support the emerging theory of change? What is the evidence that participation in microfinance actually generates these “resilience” benefits? Focusing on increase of consumption and reduction of poverty, researchers have treated consumption smoothing as a secondary interest. Yet evidence of smoothing now seems the most logical justification of public and private social investment in microfinance. Consider how evidence that microfinance doesn’t improve consumption smoothing would be an embarrassment equal to the lack of evidence that microfinance increases incomes and reduces poverty.
Many consider consumption smoothing, or even the resilience it reflects, a consolation prize at best in the quest to eliminate world poverty. But what could be more fundamental to poverty reduction than helping the poor make sure they have enough to eat and other basic necessities throughout the week, the month and the year? From that stable ground, the poor are in a much better position to seize whatever opportunities are provided by health and education services and a decent economy. If financial services by themselves cannot provide these opportunities, they do seem quite capable of helping the poor provide the stable ground to stand and build upon. That is the significance of consumption smoothing, risk mitigation and resilience in poor households.
But there is a nagging issue: Poor households are obviously not fully successful in smoothing consumption. Does use of microfinance services improve consumption smoothing over and above the effects of other, mostly informal types of financial instruments the poor already use?
Answering that question won’t be easy. Measuring consumption smoothing means we have to measure variation over many months, if not years.
My Freedom from Hunger colleague Megan Gash has drawn my attention to the 1995 paper by Jonathan Morduch on “Income Smoothing and Consumption Smoothing.” He defines both consumption and income smoothing in terms of variance measures (“average coefficient of variation”), which seems obvious. What is not so obvious is how feasible are the research designs that can get the data needed to construct a variance measure. Jonathan included this telling footnote:
By focusing on the smoothness of consumption over time, the tests take into account the net contributions of all risk-mitigating mechanisms employed by households, no matter how subtle and difficult they may be to observe in isolation. While the data requirements are steep [emphasis added] (the tests require information on consumption and income data for the same households over time), several new data sets have been collected in low-income economies, some of which cover five or more years.
The existence of these data sets is encouraging. Perhaps they can be mined for correlation between use of microfinance and reduced consumption variability. However, I anticipate all sorts of problems in trying to wring meaningful conclusions from data sets collected without this specific question in mind.
I suspect we have to acquire the needed time-series data with the time-intensive financial diaries method. To tease out the intra-year smoothing effects of microfinance use vs. the effects of informal methods already in use, we need to compare financial diaries from treatment households with those from control households. I am hopeful that such comparison will be made in new studies that are being designed, are under way, or are in final analysis.
While waiting on such definitive studies to remove doubt about the impact of microfinance on consumption smoothing, there is another approach we can take: look at shock-coping behavior in treatment and control groups.
Shocks are the more extreme tests of the consumption smoothing capabilities, or resilience, of households. Seasonal shocks are predictable, like scarcity of food in the months before the next harvest or downturn in business activities when local households are focusing on cultivating, planting and weeding the fields. Unexpected shocks, of course, are unpredictable, like illness or death of household income-earners or natural calamity, like drought, storms or floods.
Certainly we should not contrive shocks to test the resilience of households in a field experimental design. But we can (and do) set up treatment and control comparisons for periods of time in which we can expect with fair confidence that the study households will be subject to one or more relatively severe shocks, both seasonal and unexpected. In such studies, I suspect there is good evidence of shock-coping, resilience-enhancing effects of microfinance access, but this evidence has not been the center of attention.
I challenge the academic and practitioner research communities to catalog this available evidence and get more of it, including a few really good, long-term financial diary studies of both treatment and control groups. Let’s really nail down this basic benefit of microfinance, which by itself could justify the building of a whole new industry and investment asset class. Then we can turn our attention to discovering what else microfinance does or could do for the poor.
Ed. Note: This post was adapted from an earlier post on Chris Dunford’s blog, The Evidence Project