Does scarcity reduce cooperation?
Last registered on May 04, 2020

Pre-Trial

Trial Information
General Information
Title
Does scarcity reduce cooperation?
RCT ID
AEARCTR-0005794
Initial registration date
May 03, 2020
Last updated
May 04, 2020 2:00 PM EDT
Location(s)
Primary Investigator
Affiliation
University of Copenhagen
Other Primary Investigator(s)
PI Affiliation
University of Dar Es Salaam
PI Affiliation
University of Dar Es Salaam
PI Affiliation
University of Copenhagen
Additional Trial Information
Status
On going
Start date
2020-05-01
End date
2020-07-15
Secondary IDs
Abstract
Behavioral poverty traps have become a prominent subject of investigation in recent years. Poverty may cause the emergence of certain behavioral traits that in turn prevent people from escaping poverty. In this study, we focus on the causal effect of relative poverty on cooperation, which we measure experimentally in a framed investment game. Cooperation is essential to reap efficiency gains from specialization, not least in poor countries where economic transactions are often informal. Yet, cooperation might be more difficult to sustain under scarcity, since defecting can yield safe, short-run benefits. We leverage variation in relative scarcity induced by the Msimu harvest in rural Tanzania. Before the harvest, farmers face relative scarcity, whereas after the harvest they live in relative abundance. We test whether farmers are less likely to choose the socially efficient but personally risky option in the investment game when they live in relative scarcity. By means of a randomized prime, we further investigate whether this effect is more pronounced when the current economic conditions are made salient. Lastly, we study the effect of scarcity on cooperative behavior with ingroup vs outgroup members.


External Link(s)
Registration Citation
Citation
Agneman, Per Gustav et al. 2020. "Does scarcity reduce cooperation?." AEA RCT Registry. May 04. https://doi.org/10.1257/rct.5794-1.0.
Sponsors & Partners

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

Request Information
Experimental Details
Interventions
Intervention(s)
Naturally induced variation #1

Our primary identification strategy builds on the methodology used by Mani et al. (2013), in that we leverage seasonal variation in scarcity induced by the harvest. More specifically, we collect survey data from credit-constrained farmers in rural areas of Singida, Tanzania, before and after the Msimu harvest. The main crops in the areas are maize and sorghum, harvested from mid-May to June. The first round of our study takes place during the first week of May, just before the harvest. At this point in time, the farmers experience relative scarcity. They have fewer resources at hand, since they have not been able to realize the gains from the harvest yet. The harvest constitutes a substantial shock both in terms of food and cash. The second round of surveying is carried out in late June, when farmers live in relative abundance.

Experimental manipulation #1

Our first source of experimental variation in (perceived) scarcity comes from a randomized prime embedded in the survey. Before participating in the investment game, some farmers are exposed to a survey section asking questions about the current state of their economy and diet, whereas others are asked standard background questions on age, gender, etc. As a consequence of the prime, participants that currently face scarcity have an enhanced focus on this fact when playing the investment game. We test whether the heightened sense of scarcity increases the likelihood that participants choose to abstain from investing, i.e. choose not to cooperate.

Experimental manipulation #2

Our second source of experimental variation consists of randomly varying the counterpart that respondents face in the investment game (i.e., player B) between an ingroup and an outgroup member. While half of the participants are told that Player B is another (anonymous) person from their own village, the rest are told that Player B is from another part of Tanzania. We test whether people invest less when paired with an outsider and whether ingroup favoritism increases with scarcity.

References:

Mani, A., Mullainathan, S., Shafir, E., & Zhao, J. (2013). Poverty impedes cognitive function. science, 341(6149), 976-980.
Intervention Start Date
2020-05-01
Intervention End Date
2020-07-15
Primary Outcomes
Primary Outcomes (end points)
The primary outcome of interest is the decision of "senders" (Player A) in the investment game (i.e., how much they invest).
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
[1] Measures of scarcity. In order to test whether farmers live in greater scarcity during the pre-harvest season, the survey includes a number of questions on current economic conditions.

[2] Senders' expectations of receivers' behavior (i.e., how much senders think receivers will decide to send back for each possible amount they might receive):

The relevant questions are:

(i) If you chose to invest 4000 Ths and that produced a harvest of 12000, how much do you think player B would give back to you?
(ii) If you chose to invest 2000 Ths and that produced a harvest of 6000, how much do you think player B would give back to you?

[3] The decision of "receivers" (Player B) in the investment game (i.e., how much they send back)

[4] Payoff of senders

[5] Participants' decisions in the dice game pre- and post-harvest (which may receive in-depth investigation in a separate paper)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
We conduct our pre- and post-harvest experiment with a randomly drawn sample of approximately 320 farmers from 12 villages in each round, in the region of Singida (Tanzania).

Our main focus is an investment game, but we also conduct a dictator and a dice game.

The investment game portrays a situation where Player A co-owns a farm with Player B. The two players do not know each other's identity (they are only told whether the other player is from the same village or from another part of Tanzania). Player A begins the game with a certain endowment and can choose whether to (a) invest it all to buy seeds for the farm; (b) invest half of it to buy seeds and keep the rest; (c) keep the endowment and not invest.

The investment leads to a harvest that is worth three times as much as the amount invested. The revenue from selling the harvest is given to Player B, who decides how to split it between himself/herself and Player A.

Based on these rules, which are carefully explained by means of examples and visual aids, Player A is asked to indicate how much he/she wants to invest. Player B is asked to indicate how much he/she would give back to Player A for each level of investment Player A could have made (the actual choice made by Player A is not revealed until the end of the game).

Half of our respondents play the game in the role of Player A, the other half play in the role of Player B.

We further manipulate (a) whether respondents receive a prime that makes the current state of relative scarcity/abundance especially salient; (b) whether the other player in the game is from the same village vs another part of Tanzania.

While the main focus is on how each treatment independently impacts behavior in the investment game, we will also consider the interaction between the naturally occurring variation in scarcity and the experimental treatments. In particular, we seek to investigate whether: (i) the combination of an actual state of scarcity and the prime on scarcity reduces the level of investment further, and (ii) the ingroup bias is particularly pronounced in greater scarcity.
Experimental Design Details
Randomization Method
The random selection of villages is conducted by means of computer software. Random selection of village hamlets and participants within hamlets is conducted by means of lotteries conducted on the spot. Sampled participants are randomly assigned to different experimental conditions by means of computer software.
Randomization Unit
Individual
Was the treatment clustered?
No
Experiment Characteristics
Sample size: planned number of clusters
24 villages (12 villages in each of the two survey rounds)
Sample size: planned number of observations
640 individuals (320 in each of the two survey rounds)
Sample size (or number of clusters) by treatment arms
80 in each of the 8 treatments that derive from combining (a) assignment to the role of PLAYER A / PLAYER B; (b) assignment to playing with someone from YOUR VILLAGE / ANOTHER PART OF TANZANIA; (c) assignment to RECEIVING / NOT RECEIVING a prime about scarcity.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Assuming a baseline value of 2600 Ths and a standard deviation of 1275 Ths in the main outcome of interest - the amount sent by Player A in the investment game - the sample for this study was chosen to allow for a minimum detectable effect of 15% (400 Ths) with power of 0.8 and a risk of type II error equal to 0.05. The assumed baseline value of the main outcome is based on piloting the survey.
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
Tanzania Commission for Science and Technology (COSTECH)
IRB Approval Date
2020-03-31
IRB Approval Number
N/A
Post-Trial
Post Trial Information
Study Withdrawal
Intervention
Is the intervention completed?
No
Is data collection complete?
Data Publication
Data Publication
Is public data available?
No
Program Files
Program Files
Reports, Papers & Other Materials
Relevant Paper(s)
REPORTS & OTHER MATERIALS