First Randomized Trial of Microsavings

July 3, 2009

I fear I got swept up in the disproportionate focus on credit in pouncing to blog the first randomized trial of microcredit while neglecting its counterpart for microsavings. I learned about it from Jonathan Morduch’s blog in February just as I was launching this one. (Hat tip to Daniel Radcliffe for nudging it back up my reading list.)

Pascaline Dupas and Jonathan Robinson, both rising stars in economics, performed the study. They did so outside the big evaluation houses, J-PAL and IPA, and have the (small) budget and sample to prove it. Their sample contains about 200 people, compared to nearly 7,000 households in J-PAL’s microcredit evaluation. Nevertheless, the paper strikes this amateur economist (have I mentioned I practice without a license?) as creative, careful, and seminal.

In the rural market town of Bumala on the road between Nairobi and Kampala, Dupas and Robinson worked with a local “village bank” to offer free savings accounts to randomly selected “market vendors, bicycle taxi drivers, hawkers, barbers, carpenters, and other artisans”—existing microentrepreneurs. The accounts paid no interest and in fact charged for withdrawals, which may have been a good thing for people looking for discipline to help them save. (The village bank here is a Financial Services Association, a type of institution in Kenya that resembles the 19th century “village banks” more than the “village banks” you usually hear about today, in that members buy shares.) Uniquely among the randomized impact studies I have seen, the research team followed up with the subjects not once, through the usual door-to-door survey, but daily over several months, through logbooks not unlike the financial diaries at the heart of Portfolios of the Poor.

Of the 122 people offered an account, some 67 opened one and actually used it. That’s a small group with which to study impacts, yet the authors find statistically significant differences. Women invested more in their businesses and increased personal spending within 6 months. The savings accounts appeared to help them accumulate money for major purchases for their businesses, such as stock for their stores. That may have increased profits. The pattern did not hold for men. As the authors discuss, it is unclear whether the accounts helped primarily by giving women more control over their own impulses to spend rather than saving to invest, or by giving them a way to deflect family requests for money. Especially if the latter, the women’s gains may have come at the expense of relatives outside the study group.

Dupas and Robinson’s numbers also suggest that the accounts helped women maintain financial stability when they or a family member contracted malaria; this pattern is mostly insignificant statistically, however, meaning that it might have showed up by chance.

The study seems important in two respects. First, it is a gold-standard randomized trial showing that in one context, availability of savings cut poverty. That contrasts with the story so far for classical microcredit. Second, in the focus on savings and the use of logs, it sets an example for studies that I’m sure will follow. It cries out for replications with larger samples and in more places. I bet Daryl Collins of the Financial Diaries could help improve the daily logging.

The paper also reminds us about the diversity of poor people. Because impact studies look for the average effect they can lead us to imagine that everyone who uses the service experiences that same, average effect. But people differ in how much they need any given financial service, in how much they perceive that they need it, in how they use it, and in how well their investments pay off.

In particular, it is interesting that usage and impacts of the savings accounts were concentrated among a self-selected minority. Of those offered an account, 34% declined, 11% accepted but never used it, and another 12% made only one deposit in six months. Among those using it more, a relatively small group, mainly female, made most of the deposits and withdrawals. The reluctant embrace of savings accounts is disappointing since savings can help almost everyone, whereas not everyone should borrow. But we should not read too much into it. Perhaps some people mistrusted the village bank with good reason. Perhaps universal adoption just takes a few years.

Note also that the study deliberately targeted people who are already microentrepreneurs. But Dupas and Robinson also cite a finding that only a quarter of households living on less than $2/day/person include a microentrepreneur. In this sense, the study population does not represent poor people in general. Interestingly, actual and likely entrepreneurs were the ones whose businesses appeared to benefit in the J-PAL microcredit study too (even if that did not affect bottom-line poverty indicators). I wonder if microcredit and microsavings have their biggest positive impacts on the same people—while debt, more than savings, leaves some in the less-entrepreneurial majority worse off.

You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

1. Ron Says:
   July 4th, 2009 at 3:30 am

   David:

   Dupas’ budget just got bigger.

   http://3ieimpact.ning.com/profiles/blogs/round-1-of-3ie-funded-grants

2. Tim Ogden Says:
   July 14th, 2009 at 10:42 am

   David,

   two of the more interesting things to me about the study:

   * those who took up and used the accounts achieved pretty startling returns, which really makes you wonder why take up was so low.

   * Men who took up the accounts didn’t see a statistically significant gain from using the savings accounts, only women did. Is this because the men have some other savings mechanism not available to the women? If not, why do the gains accrue strongly on gender lines?

3. Pascaline Dupas & Jon Robinson Says:
   July 15th, 2009 at 5:43 am

   Tim,

   Thank you for your insights, these are great points. Some brief responses:

   * There might be heterogeneity in returns to capital across individuals, so it is not clear that the increase in business size would have been as large for those who didn’t use the account as it was for those who did. It is also important to keep in mind that the savings accounts did not pay any interest and charged substantial withdrawal fees, which means that people who have no difficulty saving at home would not use the account.

   It’s also debatable whether the take-up we observe (60% of women used the account at least once) should be considered high or low. Take-up of this savings program was higher than the take-up that has been observed for other financial services such as microfinance (around 20% in Morocco and India), or rainfall insurance for farmers (under 25% in recent studies in India and Malawi).

   Finally, the bank we worked with was very rural and most people in the sample had no experience with financial services of any kind prior to the program. It is likely that some of those that didn’t use the accounts were not familiar or did not feel comfortable with banks in the first place, and that take-up will be higher as people get more exposure to banks.

   * Indeed the gender differences we find are surprising and need to be more fully explored in future work. One hypothesis is that since most men in the study were bicycle-taxi drivers, their “production function” is different from that of women – besides their bike, which they already owned at baseline, they don’t need capital to invest in inventory every day. Thus the returns to saving could be lower for men, possibly explaining the low take-up. Another is that men face less severe savings problems than women so that the accounts aren’t as valuable for them.

   We are currently exploring this issue further. Among other things, we are currently working to expand our sample to men with market businesses. We are also randomizing whether the husband or the wife receives an account, to test whether the total impact of access to savings varies with the gender of the account holder.

4. Tim Ogden Says:
   July 15th, 2009 at 10:55 am

   Pascaline and Jon,

   thanks for the response. A few more questions/thoughts:

   * It’s easy (at least for me) to forget the relatively short time scales–certainly not enough for those who did not open an account or made only one deposit to see the gains of participants and copy them. It will be interesting to see whether use expands over time and whether returns decline if more businesses use the accounts to accumulate capital.

   * My question about low take-up was really about low usage. If you combine those who did not take up with those who only made one deposit, that amounts to about 75% of the sample right?

   * One of the things I’m wondering about these days is the evidence in Portfolios of the Poor that those living below the $2/day threshold are actually pretty sophisticated users of financial services, and the generally held assumption that lack of familiarity with banks inhibits take-up of financial services.

   Tim

5. Michiko Says:
   August 6th, 2009 at 2:06 am

   David,

   I wonder how “small” the study budget was: certainly less than the 7,000 field interviews, I suppose. I’ve just skimmed through a RCT proposal to study 41 farmer groups (a total of about 2,400 farmer interviews) over three years in Pakistan with a price tag of about $390t (plus evaluators’ fee): definitely not cheap.

   Michiko

We value frank and constructive exchanges and encourage you to use your real name in your comments.