Benin: IMPACT ASSESSMENT OF THE ADAPTED RURAL FINANCIAL SERVICES DEVELOPMENT PROJECT (PAPSFRA) (IFAD)
Section 1: The PAPSFRA (Adapted Rural Financial Services Development Project) implemented by PROCAR in Benin, was executed from 2013 to 2022. It was set up at the national level to support the microfinance institution of the “Association des Services Financiers in Benin (ASF-BENIN”). The main activities were the restructuring and consolidation of ASF-BENIN, as well as compliance with the new microfinance regulations. ASF-BENIN currently consists of a head office in Parakou and five branches (i.e. Abomey-Calavi, Comé, Djougou, Glazoué and Parakou), under which five corresponding urban branches and 77 rural branches operate.
The ultimate objective of PAPSFRA was to increase the incomes and food security of project beneficiaries. Its specific goal was to contribute to the reduction of rural poverty by improving the supply of financial services tailored to the needs of small-scale producers, including those of women and youth. To accomplish this goal, Project interventions were mostly designed to improve the capacity of the ASF to provide adapted financial services to its clients, operating in accordance with professional standards and with credibility with its partners.
Section 2: The main objective of the survey is to collect quality data at the household level to produce an accurate report of the results obtained by PAPSFRA on IFAD's level II development impact indicators, namely: 2.1. 1) Number of people with increased income, 2.1.2) Number of people with improved production, 2.1.3) Number of people with better market access, 2.1.4) Number of people with greater resilience, 2.1. 5) Number of people with improved nutrition. The conclusion of this impact analysis can generate key lessons for both IFAD and the Government of Benin in supporting accurate rural sector financing policy design.
Section 3: The C4ED, in partnership with the CEED, is collecting the data needed to measure the project's contribution to increasing the income and level of food security of clients benefiting from ASF services. Both quantitative and qualitative methods will be used. Quantitative data are collected at the household and community level. Whereas qualitative inquiry consist of Key Informant Interviews (KIIs) with ASF management staff (e.g. agency directors), and Focus Group Discussions (FGDs) with a sample of ASF clients and control group households.
The sampling frame is made up of ASF clients in the 54 arrondissements selected to be part of the treatment group and eligible households in the control zone made up of 66 arrondissements, i.e. a total of 120 arrondissements selected after a propensity score matching analysis and a statistical power analysis. The sample size was computed using the standard formula of power calculation as follows:
N=[(4σ^2 〖(z_α+z_β)〗^2)/D^2 ]*[1+ρ(1-m)]
where the number of households to be interviewed (N) depends on the minimum detectable effect size (D), measured as the expected average difference in the outcome variable, the standard deviation of the outcome variable (σ), the two parameters α and β, and the clustered sampling composed by the intra-cluster correlation (ρ) and the average cluster size (m). To account for potential selection bias in the intervention assignment, a valid counterfactual will be constructed using data from the comparison group households, which are a robust representation of what ASF clients would have been if ASF had not received the PAPSFRA intervention.
The selection process consists of three stages. The first stage is the design of the sample frame. We sample households from areas that have similar economic characteristics and environmental conditions, ensuring that treated and control households live in similar contexts. To this end, we match treated and untreated arrondissements using Propensity Score Matching (PSM). Then, we randomly select villages among the matched arrondissements using the village list of the 2013 population census.
The second stage is the selection of households to be interviewed. ASF has no specific targeting criteria for its clients. Therefore, we identify comparison households in the control areas using filter questions. These questions control for unobservable characteristics that may affect households’ use of financial services (saving accounts and loan application). Treated households are instead randomly selected from the list of ASF clients.
The third stage is when the data are collected. We match treated and untreated households using PSM to ensure that the treated and control groups are not statistically different from each other in terms of household socio-demographic characteristics, household wealth and economic environment. In other words, the final sample used for estimating impacts is made up of households that are similar in terms of observable factors.
The IFAD department of Research and Impact Analysis (RIA) will use these data to answer the principal research question: To what extent does PAPSFRA intervention contribute to increasing the income and level of food security of clients benefiting from ASF services? The associated outcomes and impacts of interest are listed in section 2 thereof and will be estimated by using the Inverse-Probability-Weighted Regression Adjustment (IPWRA), which is a doubly robust estimation that accounts for both project selection and outcome equations. In addition, a robustness checks will be conducted using the Nearest-Neighbour Matching (NNM).
The two quantitative questionnaires will be deployed into Survey Solutions. STATA software will be used to conduct data analysis.
Section 4: The project is at its early stage. Thus, the activities carried out up to now are those of the beginning of a survey.
Description of actual services provided by your staff within the assignment:
In bullet points, summarize and list all services and activities provided by C4ED, including:
- Contextualization of questionnaires.
- Translation of certain questions into local languages.
- Recruitment of investigators.
- Training of investigators.
- Pre-testing of tools to ensure the functioning of the questionnaires deployed on Survey Solutions.
- Pilot survey.
- Quantitative data collection from 2,640 households and 120 communities using tablets; the CAPI used is Survey Solutions.
- Face-to-face interview.
- Qualitative data collection using an interview guide.
- Focus group type interview and interview with key informants.












