Social capital and targeted beneficiaries of a development project: A lab in the field experiment in rural Zimbabwe

How to cite

Belard, Amandine; Farolfi, Stefano; Jourdain, Damien; Willinger, Marc; Manyanga, Mark; Pedzisa, Tarisai. 2022. "Social capital and targeted beneficiaries of a development project: A lab in the field experiment in rural Zimbabwe." AEA RCT Registry. AEARCTR-0010665. https://doi.org/10.1257/rct.10665-1.1

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Amandine Belard* , Stefano Farolfi , Damien Jourdain , Marc Willinger , Mark Manyanga , Tarisai Pedzisa

* Primary investigator

Registry ID
AEARCTR-0010665
DOI
https://doi.org/10.1257/rct.10665-1.1
First registered
Last updated
Status
In development
Trial dates
2022-10-01 – 2023-11-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/10665

Abstract

Biases are inherent to the selection of development project beneficiaries. Three biases have been identified in the literature: 1) self-selection of individuals into projects, 2) the criteria of the implementation agency and 3) the intermediary/delivering agent bias.
This paper looks at the initial level of social capital in targeted beneficiaries of a development project, as it has been showed that social capital is a key ingredient for the success of development programs.
We measure, through survey and incentivized lab-in-the-field experiments, ex-ante social capital levels in beneficiaries of a project aiming at the development of loans and savings associations in rural Zimbabwe.
Our hypothesis is that, due to the combined effects of beneficiaries’ self-selection, targeting criteria and DAs interventions, prior to the start of the project, social capital is already higher in beneficiaries than in non-beneficiaries.
We discuss how, and to which extent, the three mentioned factors (i.e.: self-selection, implementation agency criteria and DA intervention) can be at the origin of these differences, and provide suggestions for future research in order to better disentangle their effects.

Experimental design

Our first hypothesis (H1) is that the ex-ante level of social capital is unequal in the population of potential beneficiaries. The targeted beneficiaries (treatment group) have a higher level of social capital that non-beneficiaries, i.e. there is a selection bias.

According to H1, beneficiaries are more likely than non-beneficiaries to have had social interactions among them before the implementation of the program, as well as different levels/duration of collaborations. For instance, in a project aiming at the development of loans and savings associations, some beneficiaries could have already had previous experiences in similar micro-finance organisations.

Our second hypothesis (H2) refines H1 for the targeted beneficiaries. Targeted beneficiaries that have previously participated in projects or initiatives similar to the project at stake exhibit higher social capital than targeted beneficiaries that have not.

To undertake this survey, we collaborated with an NGO in Zimbabwe that implemented in 2021 a micro-finance project targeting small-holder farmers. The project aims at the creation of associations for loans and savings. The NGO has already been working with rural communities on this matter for a year, but they are extending their number of beneficiaries to more communities. We obtained the list of associations and the number of members that are about to benefit from training and support from the NGO.

Each session involves 5 parts:
1) An experimental part that consists of three incentivized tasks: A Dictator game, a Trust game, and a Public Good game. Participants in a session will play all games in a pencil and paper mode. One of the three games will be randomly selected to be paid out for real. If either the dictator or the trust game is selected, pairs will be randomly formed. If the public good game is selected, groups of four will randomly made. Bots will be used when a group does not have enough participants (this will be common knowledge).
2) A stated preference questionnaire measuring risk-tolerance, patience, and trust, based on a 0-10 integer scale.
3) A questionnaire adapted from Grootaert, Narayan, Nyhan Jones, and Woolcock (2004) aiming at measuring structural social capital.
4) A demographic questionnaire (variables: gender, age, education level, number of cows, crop diversity, food security).
5) A grid adapted from Avdeeenko and Gilligan (2014) that aims at eliciting the stated level of relationships among the participants in a session.

All written document are in English. Oral Shona translation will be available at each session by a facilitator.

We intended to implement our lab-in-the field experiment in February 2023. Due to field constraints we were obliged to anticipate the work in the field in October 2022.

Design details

Randomization method
We randomly selected (using R) 10 associations among the communities that will be working with the NGO in 2022 (treatment group). We compare the levels of social capital of these (future) beneficiaries with the levels in a control group, composed of similar farmers from neighbouring villages. The selection of the control group was done by the NGO in collaboration with the extension officers and the village agents. The instructions were to select farmers with similar characteristics (age, gender) in neighbouring villages.
Randomization unit
The randomization unit is the association.
Clusters
10 associations
Observations
N = 341
Arms
The final sample is composed of 143 treated and 198 control observations over 6 wards. The control group is slightly bigger due to sampling constraints: control sessions were always composed of 20 participants, while treated sessions were composed of the number of members in an association (ranging from 10 to 25).
Primary outcomes
The effect of development projects' selection methods on the level of beneficiaries' ex ante social capital.
Sponsors
Name: Agropolis International Location: Montpellier (France) Url: https://www.agropolis.fr/

Cited by

No citation data available for this registration yet.

Keywords