Development Impacts of Asset Transfers to Vulnerable Populations

How to cite

Rozelle, Michael; Voors, Maarten; Meriggi, Niccolò; Bulte, Erwin. 2025. "Development Impacts of Asset Transfers to Vulnerable Populations." AEA RCT Registry. AEARCTR-0015422. https://doi.org/10.1257/rct.15422-1.0

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

Michael Rozelle* , Maarten Voors , Niccolò Meriggi , Erwin Bulte

* Primary investigator

Registry ID
AEARCTR-0015422
DOI
https://doi.org/10.1257/rct.15422-1.0
First registered
Last updated
Status
On going
Trial dates
2023-12-18 – 2025-02-28
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/15422

Abstract

We use a field experiment to study the impact of an asset transfer program targeting food insecure women across 71 communities in Sierra Leone. The primary intervention consists of a livestock bundle, including productive dual purpose poultry, training, and a starter package of high quality feed. We measure the effects of this intervention on four primary outcomes: dietary quality, livestock income, farm investment, and mental wellbeing. A second treatment group receives an additional "demand" component of the intervention, which we test for marginal effects on these outcomes. We furthermore test for the conditionality of these effects on certain baseline characteristics (recent employment experience, household asset level, and proximity to the nearest major road) and economic preferences elicited at baseline (both risk and time). We also assess within-village spillovers, relying on individual level random assignment to the intervention bundle in treatment villages.

Intervention

In this experiment, our main intervention consists of the transfer of high-productivity chickens, combined with training in poultry management and a starter package of high quality feed, to selected food insecure women as part of a transfer program intended to support their livelihoods. These high-productivity chickens belong to the SASSO T-451 breed, which is a dual-purpose bird capable of growing to a much larger size, and laying many more eggs, than typical local fowl. The breed is known as being "ideal for village conditions" partly due to being "robust" and "easy to manage". As well as the breed's genetic advantages, these chickens received a full course of vaccinations and were raised to the age of approximately 7 weeks at a professional poultry farm prior to distribution.

The study sample comprises 2,596 women across 110, predominantly rural, communities in western Sierra Leone. Within each community, we conducted baseline surveys with approximately 25 of the most food insecure women, selected on the basis of a short ”screening survey” given to women who registered their interest in participating in the project. Communities were assigned by a stratified random process to one of three groups: Control, Transfer Program, and Transfer Program + Demand. Within treatment group communities (ie. communities not assigned to Control status), we conducted a further stratified individual-level randomisation of baselined subjects to a treatment or control group. Of the 2,596 study participants, 311 were assigned to receive the asset transfer program, comprising 155 subjects in the Transfer Program arm, and 156 in the Transfer Program + Demand arm.

The training sessions in poultry management were curated by a professional veterinarian with extensive poultry experience. These one-day sessions focused on simple and practical actions that the attendees could take in order to improve the health and performance of the chickens provided through the intervention. Professional-grade poultry feed was distributed to the treatment group for the first three months after receipt of the chickens in regular monthly deliveries, during which the distribution team provided updated guidance on the recommended feeding quantities and schedule for treatment group subjects' birds. This was designed to gradually taper the proportion of all feed that the provided chickens derived from professionally mixed feed, so that the birds could slowly learn to substitute this for scavenging locally if necessary.

Treatment group subjects in the Transfer Program + Demand arm received an additional component of the intervention wherein they were given the opportunity to sell the eggs laid by their provided chickens at regular weekly intervals to the implementation team. These individuals were told, at the moment that their chickens were distributed to them, that when their hens began their laying period, the implementation team would make weekly visits to their community to offer to purchase their eggs, and that they would be able to sell their SASSO eggs to the implementation team at a constant price per egg. They were also told that the implementation team would visit them directly within the community when offering to purchase the eggs, thereby avoiding transaction costs to the individual. When delivering the final instalment of professional-grade feed, it was communicated to these individuals that the price offered per egg would be 4 leones (approximately 0.17USD at the time of writing). This price was set at rough equivalence to the price of chicken eggs in regional markets, minus a small nominal cost of transportation. Once the treatment group's provided hens began their laying period, the project implementation team began to make weekly visits to each individual in the Transfer Program + Demand arm with remaining SASSO hens for this purpose.

Experimental design

110 communities were sampled for this study across four districts of Sierra Leone. These 110 communities were then randomly assigned to one of three treatment arms by stratifying on geographical location (by dividing the 110 communities into 3 clusters by a k-means algorithm), road proximity (by a dummy variable equal to 1 if the community centre is within 1km Euclidean distance of an A or B road, and 0 otherwise), and population (by a dummy equal to 1 if the community has a population equal to or above the median in our sample). These arms were Control (where no individuals would be given the asset transfer program), Transfer Program (where up to 5 randomly selected individuals would be given the asset transfer program, consisting of birds, training, and feed), and Transfer Program + Demand (where up to 5 randomly selected individuals would receive the intervention as in the Transfer Program arm, but also given the opportunity to sell the eggs their provided chickens laid at regular weekly intervals to the project implementation team).

In communities randomly assigned to either the Transfer Program or Transfer Program + Demand treatment arms, one further, individual-level, randomisation was conducted. This randomisation was stratified on the subject's household asset index at baseline (below and above the median asset index within the subject's community). This selected up to 5 individuals from the community to receive the relevant intervention bundle according to the randomly assigned community treatment arm. Baselined individuals within the treated community who were not selected in the individual-level randomisation did not receive any intervention.

Design details

Randomization method
Randomisation was done on a computer in Stata.
Randomization unit
The cluster-level randomisation unit is the community, and the second-level randomisation unit is the individual.
Clusters
110
Observations
2596
Arms
Control: 39 communities Transfer Program: 35 communities Transfer Program + Demand: 36 communities
Primary outcomes
Dietary quality: consumption of animal proteins, as well as the diversity of food groups in the participant's diet. Livestock income: income generated from the sale of livestock and livestock products, as well as the total value of owned livestock. Mental wellbeing: life satisfaction, likelihood of extended worry or anxiety, self-esteem, and aspirations for the future. Investment in farm production: use of recommended farming equipment and practices, and willingness-to-pay for the same bird in the future.
Secondary outcomes
Individual income: The sum value of the individual’s income from livestock and employment activities. Financial Inclusion: Whether the individual has any personal savings; the value of these personal savings; whether the individual possesses a formal bank account; and whether the individual has personally taken a recent loan. Food Security: The incidence of adverse food security-related events

Cited by

No citation data available for this registration yet.

Keywords