NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Understanding Lean Season Behavior in Ghana
Last registered on July 08, 2020


Trial Information
General Information
Understanding Lean Season Behavior in Ghana
Initial registration date
July 06, 2020
Last updated
July 08, 2020 5:23 PM EDT
Primary Investigator
Yale University
Other Primary Investigator(s)
PI Affiliation
Northwestern University
PI Affiliation
University of Ghana, Legon
PI Affiliation
University of Ghana, Legon
PI Affiliation
Northwestern University
PI Affiliation
Massachusetts Institute of Technology
Additional Trial Information
On going
Start date
End date
Secondary IDs
The Lean Season Grant (LSG) Project studies the ways in which seasonality in income leads to costly income- and consumption-smoothing behavior among rural low-income households in Ghana, an whether a “graduation” anti-poverty program shapes the subsequent investment decisions of these individuals when they receive an additional transfer of capital. In particular, our Trial Registry involves two sets of analyses we intend to conduct through an experiment conducted in Ghana.

In the first paper, “Understanding Lean Season Behavior”, we will study whether seasonal variation in income causes households to prioritize labor and production decisions oriented towards earning immediate cash in the lean season, and away from activities more likely to maximize longer-term profits. Similarly, we will test whether households make consumption decisions in the lean season that imply either a high discount rate or some set of constraints that prevents them from smoothing consumption between the period immediately after the harvest and the lean season.

In a second paper, we are measuring the long-run impacts to a graduation program, in which households received a transfer of productive assets and training. As part of this paper, we will include a section on “Graduation and the Returns to Capital,” in which we will test whether or not having previously received the graduation program subsequently affected how individuals used a subsequent transfer of capital. For example, to the extent there are non-convexities in investments, graduation treatment households might be above some threshold at which they are better able to use subsequent funds, or the graduation program might have strengthened individuals’ skills and/or their beliefs in their own abilities.
External Link(s)
Registration Citation
Banerjee, Abhijit et al. 2020. "Understanding Lean Season Behavior in Ghana." AEA RCT Registry. July 08. https://doi.org/10.1257/rct.6124-2.0.
Experimental Details
In February 2019, we attempted to inform 950 households (previously part of a randomized evaluation of a "graduation program" that they would be receiving a one-time transfer of 700 Ghana Cedis (roughly 150 USD at the time) via mobile money transfer in June of 2019.
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
We have four key families of outcomes: (1) investment, (2) labor supply, (3) asset stocks and decumulation, (4) expenditures and welfare
Primary Outcomes (explanation)
(see pre-analysis plan for full description)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
This experiment was conducted as an overlay experiment on a previous RCT (with initial randomization conducted in 2011), in which households were randomized to be (a) to receive a "graduation" program, (b) to be a control household in a treated community, or (c) as a control household in a community in which no households were treated. Using our existing data, we randomly assigned 950 households to receive the grant, of which 305 were households who received the graduation program, 172 from control households in treated communities, and 473 households from pure control communities.
Experimental Design Details
Not available
Randomization Method
Randomization was conducted via Stata
Randomization Unit
Individual level (previous randomization in 2011 was two-stage, at first the village-level, then individual level within treated communities).
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
Sample size: planned number of observations
Sample size (or number of clusters) by treatment arms
950 treated, 1656 control
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information
IRB Name
Innovations for Poverty Action
IRB Approval Date
IRB Approval Number
Analysis Plan

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information