Final Impact Evaluation of the Saving for Change Program in Mali, 2009-2012

Background:
The Saving for Change program, which began in Mali in 2005, enables women to organize simple self‐managed savings and credit groups. The program is meant to provide a locally appropriate tool for villagers to manage their finances and to reach populations rarely served by formalized institutional lenders. At weekly meetings, each woman contributes a mutually agreed upon savings amount to a communal fund, which grows in aggregate size each time the group meets. When a woman needs a loan, she proposes the desired amount and intended purpose to the group. The group collectively discusses whether there are enough funds and how to divide funds and prioritize requests. Loans must be repaid with interest, at a rate set by the members. The commitment of regularly saving in a group mobilizes greater savings than each woman could feasibly save individually. Each group manages its own funds that are mobilized entirely from the internally generated savings with no matching or external loans provided. At a predetermined annual date, the group divides the entire fund among members in a process termed the ‘share‐out.’ Each member receives all her savings plus a share of the income from interest on loans made to members over the year, fines for missing payments and meetings, and the income generating activities carried out by the group proportionate to the amount she saved over the year. The annual return on savings is 30% to 40% or more.

Purpose/objective:
The overall research objectives were: to understand the role of SfC within the context of larger household livelihood strategies and existing forms of savings and credit in Mali, better understand who participates in the program and why, provide rigorous evidence on the socioeconomic impacts of the SfC program on the lives of the people in the study region, and discern the relative merits of two different program replication strategies (structured and organic) to see if the cost of providing additional training and support to the volunteer replicating agents through structured replication yielded higher levels of outreach and program success as compared to providing replicating agents with informal one‐on‐one training through organic replication. 

Methodology:
The methodology was designed to provide complementary quantitative and qualitative data sets to allow for statistical and case study analyses of SfC program participants and non‐participants in treatment and control villages. By adopting a rigorous mixed methods approach, the overall study developed a robust data set through which project impacts could be triangulated from a variety of perspectives.

The quantitative study (IPA) was based on a randomized controlled trial (RCT) that measured the socioeconomic impacts of the program in 500 villages (6000 households) over a three‐year period with detailed socioeconomic surveys administered at the household level in 2009 (baseline) and again in 2012 (endline). This dataset provides a dynamic picture of the households studied over time as a complement to the baseline and endline evaluation data. 

The BARA study included a purposive sample of 19 total villages, 15 of which were selected from among the IPA sites in Ségou and four of which were located outside the RCT zone and had participated in SfC since 2005. The study sites were chosen to maximize diversity in terms of geographic location and accessibility, ethnic composition of villages, major livelihood strategies, and representation of Malian NGO partners involved in the program. Qualitative data were collected with community interviews, focus group discussions, key informant interviews during a period of 2‐3 days spent in each village in 2009 and 2012.

Key findings:
IPA found that the women who became members of SfC in treatment villages were on average slightly older, more socially connected and wealthier than non‐members, a finding consistent with the impressions of technical agents in the BARA interviews. This does not indicate that SfC is not reaching the poorest of the poor given that all villages in the sample are very poor and many are very remote. Moreover, participation rates are fairly similar across the entire wealth distribution. IPA data also shows that as the program matures, SfC members are more representative of all women in the village.

Recommendations:
Both BARA and IPA would emphasize that SfC is an effective program providing real socioeconomic benefits to its intended populations. Research from 2009‐2012 consistently indicates that some of the major strengths of the SfC program include its inherent accessibility and appeal to poor and illiterate women, especially those who live in remote areas and do not have access to formal financial services, ease of replication, and potential for long‐term sustainability.

Comments from the organisation, if any: