The Welfare Effects of Beneficiary Control over the Timing of Cash Transfers

Last registered on May 06, 2025

View Trial History

Pre-Trial

Trial Information

General Information

Title

The Welfare Effects of Beneficiary Control over the Timing of Cash Transfers

RCT ID

AEARCTR-0014679

Initial registration date

October 30, 2024

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

November 15, 2024, 1:01 PM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated

May 06, 2025, 12:01 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Country

Ghana

Region

Primary Investigator

Name

Naureen Karachiwalla

Affiliation

International Food Policy Research Institute

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Francis (Haoyu) Fang

PI Affiliation

University of Washington

Contact Investigator

PI Name

Jason Kerwin

PI Affiliation

University of Washington

Contact Investigator

PI Name

Francis Annan

PI Affiliation

University of California, Berkeley

Contact Investigator

Additional Trial Information

Status

In development

Start date

2024-05-07

End date

2028-12-31

Keywords

Behavior, Finance & Microfinance, Welfare

Additional Keywords

Cash transfers, well-being, credit, insurance

JEL code(s)

D14, D15i, I31

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

To exit poverty, income streams of the poor must align with their liquidity needs (e.g., large investments, consumption-smoothing, saving for shocks, and smaller expenses). Traditional cash transfer programs lack flexibility, offering only fixed payment structures. This project will measure the demand for, and the welfare impacts of, the choice of a fully flexible payment schedule (frequency, timing, and amounts) in partnership with the Government of Ghana's Livelihood Against Poverty (LEAP) Program.

First, we will randomize 800 (of 1,500) households into a control group (C) and 700 to a treatment group will receive cash transfers (T). Next, we will elicit beneficiary preferences over preferred times and amounts of cash transfers and measure respondents' willingness to accept (WTA) switching to the LEAP default timing instead (equal, bi-monthly transfers) using a Becker-DeGroot-Marschak (BDM) mechanism. The BDM mechanism will determine whether a household belongs to one of two treatment groups with the following distribution programs (equal in total value): (T1) smooth, bi-monthly transfers (LEAP default) vs. (T2) fully-flexible payment schedule that provides choice over the timing of transfers. Impacts will be assessed on subjective well-being, consumption expenditures (we will use this to derive marginal utility of expenditures (MUEs) following Ligon (2020) and the treatment’s welfare effects), income, assets, savings and debt, women’s empowerment, and mental health. We will further explore whether liquidity or risk drive these impacts by stratifying treatments on village “financial health” (low, medium, high) (Innovations for Poverty Action IPA 2020).

After evaluating the main effects at endline by comparing the control vs treatment groups and the two treatments to one another (the lump-sum and fully-flexible programs), we plan to later implement a mechanism experiment to causally test whether the main effects are explained by either commitment and self-control motives or consumption self-insurance motives. All households will receive cash transfers. Households will be randomized into a control group receiving no additional programming, a group that will additionally receive an insurance product, a group that will additionally receive a a savings product, and a group that will additionally receive both insurance and savings products. The outcomes will be the same as the main experiment.

This project will thus involve three parts (each with a different pre-analysis plan to be uploaded separately):
1. Measurement and analysis of the respondent's willingness to accept LEAP's default schedule instead of the elicited preferred schedule.
2. Randomized trial with a control group and two experimental groups (fully flexible schedule and bi-monthly equal transfers) with associated analysis.
3. Randomized mechanism experiment providing a free savings or insurance product in addition to the cash transfers.

External Link(s)

Registration Citation

Citation

Annan, Francis et al. 2025. "The Welfare Effects of Beneficiary Control over the Timing of Cash Transfers." AEA RCT Registry. May 06. https://doi.org/10.1257/rct.14679-1.1

Sponsors & Partners

Partner

Name

Government of Ghana, Ministry of Gender, Children, and Social Protection

Type

government

URL

https://www.mogcsp.gov.gh/

Experimental Details

Interventions

Intervention(s)

The LEAP program is the Government of Ghana’s national cash transfer program and began in 2008. It provides transfers to the ultra-poor based on the number of eligible individuals in the household. Vulnerable household members include orphaned and vulnerable children, people with severe disability without any productive capacity, extremely poor or vulnerable households with pregnant women and mothers with infants, and members 65 years and above. Beneficiary selection begins with data from the Ghana Statistical Service for national poverty maps, followed by geographical targeting of districts who have the highest poverty rates, and finally household enumeration using a proxy-means testing questionnaire.

A household with one eligible member receives GH₵512.00, two eligible members receives GH₵608.00, three eligible members receives GH₵704.00, and four or more eligible members receives GH₵848.00. Payments are made electronically (via biometric fingerprints) bi-monthly in equal installments. To remain eligible, households must enroll school-aged children in school, enroll all members in the national health insurance scheme, and children 0-18 months must be registered with the Birth and Deaths Registry, attend required post-natal clinics and complete the Expanded Programme on Immunisation. For the main experiment, the transfers will run for one year and will either be the LEAP default bi-monthly, equal transfers or the beneficiary's preferred transfer scheduled (as determined by the BDM mechanism).

For the mechanism experiment, the same program will be used, but amounts may differ should the Government of Ghana decide. The precise savings and insurance products will be determined at a later date.

Intervention Start Date

2025-07-01

Intervention End Date

2028-07-01

Primary Outcomes

Primary Outcomes (end points)

Marginal Utility of Expenditure (MUE), subjective well-being, willingness to accept (WTA)

Primary Outcomes (explanation)

MUE - see Ligon (2020)
Subjective well-being - see Benjamin, Heffetz, Kimball, and Szembrot (2014)
Willingness to accept (WTA) - Becker-DeGroot-Marschak (BDM) method

Secondary Outcomes

Secondary Outcomes (end points)

Consumption, assets, savings, debt, depression, stress, anxiety, self-esteem, women's empowerment, risk-aversion, ambiguity-aversion, time preferences, cognition

Secondary Outcomes (explanation)

Consumption - mean daily expenditures per adult equivalent unit (food and non-food consumption)
Assets - first principal component of a PCA analysis on a list of 17 assets (productive and durables)
Savings - any savings, amount of savings
Debt - any debt, amount of debt, lender
Depression - PHQ-9
Stress - Cohen's Perceived Stress Scale
Anxiety - GAD-7
Self-esteem - Rosenberg Self-Esteem Scale
Women's empowerment - WEMNS
Risk and ambiguity-aversion - double multiple price list
Time preferences - multiple price list
Cognition - digit span forward and backward

Experimental Design

First, we will randomize 800 (of 1,500) households into a control group (C) and 700 to a treatment group will receive cash transfers (T). Next, we will elicit beneficiary preferences over preferred times and amounts of cash transfers and measure respondents' willingness to accept (WTA) switching to the LEAP default timing instead (equal, bi-monthly transfers) using a Becker-DeGroot-Marschak (BDM) mechanism. The BDM mechanism will determine whether a household belongs to one of two treatment groups with the following distribution programs (equal in total value): (T1) smooth, bi-monthly transfers (LEAP default) vs. (T2) fully-flexible payment schedule that provides choice over the timing of transfers. Respondents will state their maximum WTA to switch away from their preferred schedule, and a spinning wheel will provide a random amount of money between 0-690 GHC. If the respondent's WTA is lower than the amount spun, they will receive their preferred timing schedule, otherwise, they will switch to the LEAP default schedule.

Impacts will be assessed on subjective well-being, consumption expenditures (we will use this to derive marginal utility of expenditures (MUEs) following Ligon (2020) and the treatment’s welfare effects), income, assets, savings and debt, women’s empowerment, and mental health. We will further explore whether liquidity or risk drive these impacts by stratifying treatments on village “financial health” (low, medium, high) (Innovations for Poverty Action IPA 2020).

After evaluating the main effects at endline by comparing the control vs treatment groups and the two treatments to one another (the lump-sum and fully-flexible programs), we plan to later implement a mechanism experiment to causally test whether the main effects are explained by either commitment and self-control motives or consumption self-insurance motives. All households will receive cash transfers. Households will be randomized into a control group receiving no additional programming, a group that will additionally receive an insurance product, a group that will additionally receive a a savings product, and a group that will additionally receive both insurance and savings products. The outcomes will be the same as the main experiment.

Experimental Design Details

Not available

Randomization Method

Stratified, clustered randomization conducted by Stata.

Randomization Unit

Village

Was the treatment clustered?

Yes

Experiment Characteristics

Sample size: planned number of clusters

150 villages

Sample size: planned number of observations

1,500 households

Sample size (or number of clusters) by treatment arms

800 pure control households, 350 households smooth bi-monthly payments, 350 households fully flexible (timing and amount) payments

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Supporting Documents and Materials

IRB

Institutional Review Boards (IRBs)

IRB Name

University of California of Berkeley

IRB Approval Date

2024-01-09

IRB Approval Number

2023-12-16960

Analysis Plan