Cash Transfers and Community Participation in Public Affairs: A Village-Level Randomized Controlled Trial in Kenya

How to cite

Orkin, Kate; Walker, Michael. 2018. "Cash Transfers and Community Participation in Public Affairs: A Village-Level Randomized Controlled Trial in Kenya." AEA RCT Registry. AEARCTR-0002918. https://doi.org/10.1257/rct.2918-7.1

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

Kate Orkin* , Michael Walker

* Primary investigator

Registry ID
AEARCTR-0002918
DOI
https://doi.org/10.1257/rct.2918-7.1
First registered
Last updated
Status
On going
Trial dates
2014-09-15 – 2024-12-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/2918

Abstract

We provide well-identified evidence on how a large cash windfall from direct cash transfers from an NGO affects political processes and their outcomes for the poorest in rural communities in Kenya. We exploit village-level variation in the rollout of an unconditional cash transfer programme from two village-level randomized controlled trials, covering 1,097 villages. In each trial, we randomly assign half of the villages in an area targeted by the NGO GiveDirectly to an unconditional cash transfer programme, after stratifying by village economic status and size. We examine effects on poorer households which would have been eligible to be recipients of the cash transfer programme, richer households who would not have received the programme, non-recipients, and village-level outcomes. We focus on the extent of clientelist exchanges and on citizens' civic engagement and involvement in group and community initiatives.

Intervention

The NGO GiveDirectly provides large, unconditional cash transfers to poor households in rural Kenya. GiveDirectly identifies villages in which they are willing to work, and in order to facilitate research on cash transfers, these villages are randomly assigned to treatment or control status. Within treatment villages, GiveDirectly then identifies all households that meet their eligibility criteria, enrolls and verifies the eligibility of households, and sends cash transfers to all eligible households via the mobile money system M-Pesa. Eligible households receive a one-time of around USD 1,000 made in a series of three payments. GiveDirectly is responsible for the intervention.

The cash transfers used in this program were originally distributed as part of two previously-registered AEA trials. The earlier trial, ”General Equilibrium Effects of Cash Transfers in Kenya” (GE), examines effects on prices, wages and economic growth in 653 villages, and began giving out cash transfers in September 2014 (Haushofer, Miguel, Niehaus and Walker, 2014, available at https://www.socialscienceregistry.org/trials/505). The later trial, ”Promoting Future Orientation Among Cash Transfer Recipients”, started disbursing transfers in November 2016 in 416 villages (Orkin, Garlick, Haushofer, Mahmud, Sedlmayr, and Dercon, 2016, available at https://www.socialscienceregistry.org/trials/996). It examines effects of cross-cutting a goal-setting exercise with the cash transfer.

We are collecting new data on political and civic engagement outcomes across both trials. The combined study increases the number of villages and administrative units to study to increase power for village- and administrative unit-level analysis.

Experimental design

Randomization of treatment on the village level was conducted separately by the two study teams and is outlined in the PAPs linked to above.

In both studies, in cash treatment villages, GiveDirectly enrolls and distributes cash transfers to households that meet its eligibility criteria.

Design details

Randomization method
Stratified randomisation in an office by a computer. Randomization of treatment on the village level was conducted separately by the two study teams and is outlined in the PAPs linked to above.
Randomization unit
Villages and sublocations
Clusters
1,097
Observations
Estimated average of 18 households per village in 416 villages, plus 1,111 village elders and roughly 100 assistant chiefs across all villages, and approximately 65 candidates for the Members of the County Assembly position or past incumbents of the position for wards in which the 1,111 villages are located.
Arms
We located 1,097 of the 1,111 village elders in our sampling frame. There are 552 villages in the control group, 545 in the cash treatment group.
Primary outcomes
See pre-analysis plan
Secondary outcomes
See pre-analysis plan

Cited by

No citation data available for this registration yet.

Keywords