Impact on Household Poverty—the Poverty Gap and the Poverty Bubble

In Theme Three, I found that the more credible impact studies of access to microfinance had uncovered little statistically significant positive impact on household income and consumption. This general conclusion is affirmed by the just-released results of the impact study of Compartamos in the state of Sonora, Mexico, led by Manuela Angelucci, Dean Karlan and Jonathan Zinman. The logical conclusion is that poverty reduction has not occurred. However, poverty “reduction” usually means a reduction in the percentage of people who live below a national or international poverty “line.” None of these studies actually focuses on what happens to borrowing households that start below a given poverty line (in the Compartamos study, hardly any households were considered poor—without reference to a poverty line). Even if these several studies had demonstrated increased household income and consumption, we would not know if poverty reduction had occurred.

Poverty reduction starts with poverty outreach—the percentage of incoming clients who live in households living below a poverty line. Before we ask whether microfinance reduces poverty, we should ask whether microfinance providers are even serving households living in poverty. That is, does microfinance even have the potential to reduce poverty? The answer appears to “yes, there is potential,” but not nearly as much as we might have believed.

It is emblematic of the discomfort in microfinance circles with poverty outreach data that the best documentation to date of poverty outreach by microfinance institutions (MFIs) worldwide is hard to find on an obscure website. In June 2012, I listened in on a webinar produced by Local Voice 4 Development, featuring Lucia Spaggiari of MicroFinanza Rating and Sanjay Sinha of M-CRIL. I devote this post to repeating the substance of Lucia’s “Poverty outreach: big expectations, mixed results, transparency opportunity,” which of the two presentations is global in scope and more completely documented on the website. To see her nifty charts, you should check out the original. She starts off with the most compact yet comprehensive summary ever:

  • Does microfinance reach the poor?
  • Sometimes yes, sometimes in part;
  • Sometimes it does not really want to;
  • Generally we say it does;
  • In most of the cases we do not know.

Source of information

    • Source: comprehensive social ratings carried out by MicroFinanza Rating, including a field survey on recent clients of MFIs.
    • Household poverty level estimated using the Progress out of Poverty Index (PPI) or consumption data.
    • Poverty line used in this study: US2$PPP/day, for comparability reasons.
    • MFI’s mission orientation towards the poor analyzed in social ratings.
    • Sample: 65 MFIs, 30 Countries, 12,000 new clients (~180/MFI).


    • Focus on one monetary dimension of poverty, within the multidimensional definition of poverty adopted in social rating analysis (see Social Rating Methodology, section 6).
    • Focus on the gap between the poverty rate among clients and the poverty rate in the country, rather than on the poverty rate value; the gap is confirmed using other poverty lines. Different poverty lines are relevant in each specific context for management purposes.
    • Sample not meant to be representative of the entire microfinance industry, due to the still limited size, and the possible better performance of MFIs undertaking a social rating.
    • More research needed: larger sample size and deeper analysis will shed more light on the topic.

Evidence from social ratings

Globally, only 19 percent of the MFI clients in the sample are poor (daily per capita consumption of US$2.00 (adjusted for purchasing power parity) or less.

There are differences by type of institutional charter: 28 percent for NGOs, 21 percent for credit unions, 13 percent for non-bank financial institutions (NBFIs), and 3 percent for banks. However, banks and NBFIs may reach a larger number of absolute poor clients when they have much greater overall outreach than NGOs and credit unions.

There are differences in poverty outreach across regions, reflecting the overall poverty levels in these regions:

    • Africa: Poor people are 41% of the clients of the average MFI vs. 65% of total population of the average country
    • Asia: 41% vs. 55%
    • Latin America and the Caribbean: 10% vs. 28%
    • Europe and Central Asia: 9% vs. 19%

Most important in these data is the gap shown between MFI poverty outreach and the poverty level of the country in which the MFI operates—the poverty gap. To match the generic rhetoric of microfinance, we might expect at minimum that the percentage of poor people among the average MFI’s clientele would match the percentage of poor people in the population at large. This is a pretty low bar to hurdle; we would expect MFIs to have a higher percentage of poor people in their clientele. Yet we see just the opposite on average—a lower percentage; in fact, a much lower percentage. Asian MFIs do the best, but they still fall short by 14 percentage points! As Lucia gently put it: the poverty outreach message communicated by the industry may not always be reflected in reality.

To be fair, Lucia points out that not all MFIs have a poverty alleviation mission: some of them do, others may have different development objectives. The breakdown of the MFIs by mission (as determined by the social rating) shows that MFIs with the intention of reaching the poor tend to achieve a higher poverty outreach (36% of clientele vs. 47% of the population) compared to the MFIs with no such specific intention (11% of clientele vs. 31% of the population). This is excellent corroboration (albeit only a correlation) of what makes intuitive sense. Yet the poverty gap even for the “poor oriented” MFIs is 11 percentage points below the poverty level of the country’s population. To be more precise, Lucia shows the data for individual poor-oriented MFIs and finds that 45 percent of them have a higher percentage of poor clients than the country’s total population, most of them just barely higher. Of the 55 percent of the poor-oriented MFIs that have a lower percentage of poor clients, most of them fall well below the country’s total poverty rate.

Here is Lucia’s insightful and cautionary conclusion:

Every poverty alleviation mission raises expectations. To alleviate poverty, we need to reach the poor first. Yet, very little is known about the actual poverty outreach results, as few MFIs have measurement systems in place. The research suggests that poverty outreach should not be taken for granted: not every MFI has a poverty-oriented mission; among the MFIs with a poverty-oriented mission, some are doing a great job at reaching the poor, while others may be exposed to the risk of mission drift. The investors’ portfolios may reflect the investees varied poverty outreach results.

Keeping the expectations artificially high without sufficient measurement and transparency systems—the poverty outreach bubble—may be a risky strategy for an industry depending on its poverty alleviating reputation. A new reputation risk crisis may be prevented with enhanced MFI internal measurement systems, independent transparency tools such as social ratings, and investors’ commitment to focus on MFIs that achieve the results they support.