ICYMI: A Deeper Look at Microcredit Research (Part 3)

The next installment of our series on microfinance innovation research brings us to Ulaanbaatar. The motivating question of most microcredit evaluations is the impact on poverty for the “average” microcredit borrower. But “average” microcredit doesn’t typically serve ultra-poor or other high risk populations.  So one important dimension of innovation in microfinance is products or interventions specifically targeting these marginal groups, rather than marginal typical borrowers.

Interventions targeting highly vulnerable groups, in this case sex workers in Mongolia, can be thought of as an alternative or as a precursor – establishing more stability (financial or otherwise) and creating a foundation to allow an individual to pursue sustainable income-generating activities (and become an “average” microcredit client).

 In Mongolia, half of all HIV infections occur among female sex workers.  Authors Susan Witte et al. investigate whether a matched savings product could have an impact on health, namely in reducing HIV infection rates in this population.  The researchers conducted an RCT with a treatment offering the women an intensive financial literacy program and a savings account with a matching component of up to $160 over four months of deposits.  To maintain the savings program, the women needed to attend 9 out of 12 financial literacy and business development training sessions and 6 out of 10 mentoring sessions also focused on business development.  The sessions included topics such as banking services, savings, budgeting, debt management, and financial negotiation.  The control group received training and mentoring but not the savings account.

On the one hand, the women in the treatment group demonstrated a significant reduction in high-risk sexual behavior with fewer reported partners and incidences of unprotected sex during six-month follow up surveys.  However, understandably this was self-reported data, and therefore it’s possible that the respondents were reporting what they thought the surveyors wanted to hear (it’s unclear, based on the current version of the paper, whether both treatment and control groups received the exact same information on avoiding high-risk behavior).  The matched savings were disbursed to the women to purchase materials to start a small business or to pay for vocational training (again,  there’s lack of clarity on whether  there were enforcement mechanisms for these restrictions).  The authors note anecdotally that some women started small business or expressed interest in a vocational training program to improve their business skills.  Additionally, the (likely necessarily) small sample size of 107 and the lack of reporting on non-health outcomes limits the extrapolation of impact. 

This research underscores the difficulty of assessing financial intervention impacts among highly vulnerable populations.  It can be very hard to identify and recruit a large enough sample, and even more difficult to gather objective data on risks and behaviors.

Overall research on  the impact of grants or other treatments on the ultra-poor has had some success.  Chris Blattman’s paper on providing women in post-conflict Uganda with cash grants earmarked for business development reports an increase in entrepreneurship, income, and consumption.  However, there was no impact on female empowerment and the post-conflict setting may contribute to the high rate of returns to capital observed.

Jonathan Bauchet, Jonathan Morduch, and Shamika Ravi’s recently published paper (paywall) on an ultra-poor programs in South India reports no lasting net impacts for treatment groups.  In fact, many individuals dropped out of the programs providing livestock, training, and other interventions to pursue wage work as wages for unskilled labor were rising in the area.  Conversely, evaluations of similar ultra-poor programs in different contexts seem to have had quite positive results but have not been published yet.

Witte et al. note, “until and unless there is increased regulation or oversight of microlending to build in more protections for more vulnerable or stigmatized populations, there should be more attention paid to income enhancement approaches that do not put individuals at more potential adversity and risk.”  While the research on such approaches (like ultra-poor programs) is mixed, further studies may show where (and what design of) microfinance interventions could have the greatest impact for at-risk populations.