Here is a small highly curated selection of research on Savings Groups and their impact, and on the people who become their members. These are all worth reading, and we think they are all free from bias, and done by competent researchers.
If you are learning about Savings Groups, this short list is an excellent place to start, and will keep you busy for a while. But seriously - if you have suggestions for things that should be included but are absent from the list, please let us know.
And, if you want more, more, MORE, go the Archive, where you will find hundreds of articles, sorted by category.
A few randomized control tests (RCTs) have been carried out of SG projects. These RCTs have been criticized by some people, who pointed out, among other things, that some of the studies were of extremely short duration, and it was unreasonable to expect to see any results after such a short time. CARE's Uganda study is an example: only two years from start to finish, so that 40% of the groups hadn't even distributed a single time. That point is well taken, except that even during those short periods, the studies found substantial positive impacts. This is truly remarkable - compare similarly rigorous studies of microcredit, and you'll see the difference. Credit programs help a few people, and hurt a few people, and overall they don't have much net impact. SGs seem to be a better deal. The disappointment in the RCTs is perhaps due to the fact that some have oversold the concept of Savings Groups, making extravagant claims. Here's the deal: SGs don't necessarily "alleviate poverty", at least, not over the periods of time that has been studied. They don't turn people into entrepreneurs and lead to miraculous business start-ups. However, they do seem to make the lives of poor people easier, and they allow them to build up assets so that they can do whatever they want - maybe investment, maybe school fees, and maybe even ceremonies. SEEP published an excellent summary of the various RCTs, by Megan Gash and Kathleen Odell.
I had the privilege of working with FSD Kenya on The Quality of Delivery Study (QDS) which looked at how different delivery channels perform. By delivery channels, we mean the differences between the approaches of different NGOs, their corporate culture, incentives, agent training, group training, group methodology and so on.
About five percent of the QDS respondents reported they had lost money in their groups; some people think that statistic represents good news ("Wow! Only five percent!"). However, it's good to remember that the people who report losing money tend to be poor women, who ended up in SGs because they followed our advice. Ouch. And five percent, or one in twenty, turns out to be many tens of thousands of people.
Now, the important fact from the QDS is that the rate of loss very strongly correlated with the origin of the group. CRS had a tiny loss rate, less than one percent, while the spontaneous groups had by far the greatest percentage of members-with-problems, with over 11% of members reporting having lost money. Another large NGO was in the middle between those two extremes. These findings are statistically significant, and they show what should be obvious, that program design and care in implementation make a huge difference in outcomes. Anyone forming SGs should read this report carefully, and heed its lessons. There is a big big difference between the best and the worst programs.
An interesting 2013 study on the Post Project Replication of Savings Groups in Uganda was conducted for VSL Associates by Datu Research. This paper makes a convincing case that there is a lot of post-project replication, with about two new groups formed post project for every group trained by the project. The study goes on to break down what percentage of groups were eventually trained by a trainer, and which were simply trained by members of other groups.
The study found a great deal of multiple membership, and it's a good reminder that we would do well to distinguish between those who save in multiple groups (no one ever got in trouble because they saved too much) and those who borrow from multiple groups, which creates risks for the poor borrower who typically pays ten percent interest a month on multiple and increasing loans. In fact, as the study points out, multiple borrowing creates risks not just for the member, but for the groups, and even for an entire network of groups.
Reading the Datu study left me wishing the interviewers had been able to take the time to track down groups that no longer exist to find out why they are no longer meeting: we often bias our studies by concentrating on the survivors and ignoring the failures.
Portfolios of the Poor is important for anyone working in this field and will reward the reader who can plow through a scholarly book that feels a little long at times. The authors - Daryl Collins, Jonathan Morduch, Stuart Rutherford and Orlanda Ruthven - draw on financial diary research in three countries - India, Bangladesh and South Africa - and paint a detailed picture of people's financial lives. The book shows that poor people have both an illiquidity preference (they like to lock money up so they or their family can't spend it) and a liquidity preference (they want to be able to lay their hands on money when they need it). The result of these contradictory desires is that informal groups have a variety of instruments and rules that both encourage poor people to save, and allow them to borrow. Note that the book is about people, not about financial institutions, and it never mentions savings groups. No matter. This is an important book.
Another document that I have appreciated a lot: Saving for Change in Mali: A study of some Atypical groups from Sikasso to Kayes, by Roanne Edwards. The 24 groups studied were atypical either because they succeeded beyond expectations, or because they failed, or because they took the methodology in unexpected directions. (Actually, there weren’t really 24 groups: the study includes a couple of villages that refused the SfC methodology before the agents could even train one group).
There are many good lessons and queries in the paper concerning everything from the importance of gaining the support of the men and elders in the village, to the key difference that dynamic leaders can play.
Many of these outlier groups define their success by their ability to collaborate with other agencies. They succeeded because the women, organized in a group, had the confidence and the resources necessary to take advantage of other opportunities that were already present in the village. For example, Benkadi group in Kayes is systematically harvesting various government subsidies to run a mill, and buy a generator, plows and a pump. Whatever you think of subsidies, you can’t criticize the groups that take advantage of them. The striking element of this group is that it has drifted away from the SfC methodology: they have not distributed after three years, because they have so many other things they want to do with their funds. For that reason, some would consider the group suspect, while the author considers it a success.
In other cases, outside aid came in and created multiple problems, including unanticipated expenses, and disagreements among members. This is one of the rare efforts to look at the groups that go off in their own direction and that evolve in ways that the planners and trainers didn’t expect. And as groups get older, it becomes more and more likely that they will innovate in an effort to meet their evolving needs. As much as it is encouraging about the possibilities for group evolution, the study is also full of warnings of pitfalls.
Finally, two studies by Julie Zollman: Waiting for Rain, Reaching for Mangoes is a portrait of some groups in Swaziland. As in the study of atypical groups, above, the author neither hides the groups' flaws nor exaggerates their virtues, and her style is always compelling and easy to read. Second, the FSD Kenya-sponsored Financial Diaries Study: this opens a window into the financial lives - in fact, all aspects of the lives - of poor people in one region of Kenya.
Start with these studies. Suggest others. And go to the archive. Thanks!