Cognitive constraints in consumption smoothing

Last registered on November 23, 2022


Trial Information

General Information

Cognitive constraints in consumption smoothing
Initial registration date
August 18, 2019

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
August 20, 2019, 9:38 AM EDT

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

Last updated
November 23, 2022, 4:49 PM EST

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



Primary Investigator

UC Santa Barbara

Other Primary Investigator(s)

Additional Trial Information

Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Consumption seasonality is prevalent across poor and rich countries, from African farmers to US food stamp recipients. Such consumption fluctuations—in which households seem to “over-consume” in early periods and have “too little” saved in later periods—present a puzzle, violating neoclassical consumption smoothing. This project posits that cognitive constraints lead to mis-optimization in dynamic intertemporal problems such as consumption smoothing, in ways that can generate declining consumption profiles over time.

We implement a field experiment in rural Zambia to test for the role of cognitive constraints. In this setting, farmers harvest maize once per year, and then consume it over the next 12 months, mimicking a standard “eat-the-pie” problem. Existing data shows that households substantially reduce food consumption in the months leading up to the next harvest, a period that is locally known as the annual “hungry season”.

Households are randomly assigned to a control condition or a treatment condition, which manipulates the cognitive burden of solving the consumption planning problem after harvest. We will measure impacts on (i) demand for planning aids, (ii) accuracy of savings forecasts and beliefs about future consumption, savings and labor supply (iii) demand for luxury goods (iv) consumption profiles, (v) savings and (vi) labor supply
External Link(s)

Registration Citation

Jack, Kelsey. 2022. "Cognitive constraints in consumption smoothing." AEA RCT Registry. November 23.
Experimental Details


Broadly, our null hypothesis is that cognitive constraints have no effect on consumption decisions. Building on a model of cognitive constraints and decision making, we are further interested in pinning down one form of cognitive costs: individuals forget some of the expenditures they will face in the future.

To test for the relevance of these costs, our experimental intervention manipulates the cognitive burden of making consumption decisions. We purposefully construct our intervention so that individuals only need to pay the cognitive cost once to formulate a consumption plan for the year, and then can record this plan visually so that they do not need to think through the problem again in later periods. To accomplish this, we take advantage of the fact that our study households store maize in large canvas bags—providing a naturalistic vehicle for recording a visual representation of one’s planned consumption and expenditures.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
1) Demand for labels: Both the treatment group and the control group will be offered labels that can be used to track how they use their maize. They will be offered the opportunity to take-up these labels, or a small gift. Demand for the labels will serve as an outcome
2) Willingness to pay for luxury goods: Following the intervention (post-harvest), we will conduct a real-stakes willingness to pay exercise for a luxury good. This exercise will be repeated during the hungry season.
3) Consumption and savings: Using surveys, we will measure consumption and savings during the post-harvest and hungry seasons.
4) Labor supply: Using surveys, we will measure household labor sold to the market during the post-harvest and hungry seasons.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
• Control. The surveyor offers households a choice between a packet of colored labels, where each color corresponds to one of six expenditure categories, and sugar. Households can use these at any time to label their maize bags according to what they would like to spend on each category. Households are told that if they choose the labels, the surveyor will help them affix the labels to their maize bags.
• Treatment. Costly effort. The surveyor asks households to think through allocating their maize endowment among each major category for the upcoming year. Specifically, the surveyor undertakes a structured exercise that prompts respondents to remember the universe of possible expenses (e.g. school fees for each term, each type of farm input cost in each month, etc.) This forces respondents to remember each possible upcoming expense that will contribute to each category. Households are then given the choice to take up the same packet of labels as in Control, and can attach these to their maize bags to visually represent their plan (or ask the surveyor to do it for them).

n follow up data collection with a different sample of households, we examine two types of budget boards: a 2-category board on which households can allocate maize between non-food and food expenses for the coming year, versus a budget board that is the same as in the main study (multiple categories of non-food expenses and month-by-month food expenses). We assign one group to allocate expenses for the coming year across the 2 categories in the 2-category board, after which they undertake this allocation across the full budget board. We assign the remaining households to allocate expenses for the coming year to the full budget board only. Our primary interest is in comparing the allocation of savings across categories and WTP (see primary outcome #2) between the two boards. As a supplementary suggestive exercise, we also ask households to list which specific expenses they had in mind when undertaking the exercise on each type of board in order to better understand what kinds of expenses are less versus more likely to be neglected.
Experimental Design Details
Randomization Method
Randomization will be done by a computer.
Randomization Unit
Randomization will be at the household level.
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
900 households
Sample size: planned number of observations
900 households
Sample size (or number of clusters) by treatment arms
450 households per arm
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Institutional Review Boards (IRBs)

IRB Name
UC Berkeley Committee for Protection of Human Subjects
IRB Approval Date
IRB Approval Number
IRB Name
University of Zambia Biomedical Research Ethics Committee
IRB Approval Date
IRB Approval Number


Post Trial Information

Study Withdrawal

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

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials