Gender composition of microentrepreneur teams and business performance in Rwanda

Last registered on August 13, 2018

View Trial History

Pre-Trial

Trial Information

General Information

Title

Gender composition of microentrepreneur teams and business performance in Rwanda

RCT ID

AEARCTR-0003230

Initial registration date

August 12, 2018

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 13, 2018, 2:16 AM EDT

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

Locations

Country

Rwanda

Region

Rulindo and Ruhango districts

Primary Investigator

Name

Manuel Barron

Affiliation

Universidad del Pacifico

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Rowan Clarke

PI Affiliation

University of Cape Town

Contact Investigator

PI Name

Martine Visser

PI Affiliation

University of Cape Town

Contact Investigator

Additional Trial Information

Status

On going

Start date

2016-09-09

End date

2019-03-30

Keywords

Firms & Productivity, Gender, Labor

Additional Keywords

microenterprises

JEL code(s)

J16, J22, O14, D22

Secondary IDs

Abstract

We attempt to answer two key questions. First, How does gender composition of micro-entrepreneur teams affect business performance? Second, how do prosocial expenditures, gender attitudes and time use are affected by female participation in microenterprises. In answering these questions we exploit two margins of comparison: We compare outcomes for female and mixed gender teams with (i) all-male teams, and (ii) prospective micro-entrepreneur teams (all female and mixed gender, respectively) in control villages.

The main outcome variable is business performance, measured by administrative and self-reported data on revenues and recharge frequency. The secondary outcome variables are gender attitudes, aspirations, prosocial expenditures, and time allocation. To answer these questions we implemented a clustered RCT in two districts in Rwanda. The intervention entailed recruiting teams of micro-entrepreneurs experimentally varying their gender composition: all-male, all-female, and mixed-gender (2 males, 2 females).

External Link(s)

Registration Citation

Citation

Barron, Manuel, Rowan Clarke and Martine Visser. 2018. "Gender composition of microentrepreneur teams and business performance in Rwanda." AEA RCT Registry. August 13. https://doi.org/10.1257/rct.3230-1.0

Former Citation

Sponsors & Partners

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

Request Information

Experimental Details

Interventions

Intervention(s)

We set up 136 micro-enterprises in Rwanda, experimentally varying gender composition of the micro-entrepreneur teams in three groups: all males, all females, and mixed gender.

Intervention Start Date

2016-12-10

Intervention End Date

2018-08-05

Primary Outcomes

Primary Outcomes (end points)

Business performance: number of recharges, revenues; both at team and individual level

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

The secondary outcomes are gender attitudes, prosocial expenditure, and child aspirations. These are our secondary outcomes because they are consequences of business performance (increased household income, strengthening of the female figure in the household).

Secondary Outcomes (explanation)

Gender attitudes and child aspirations will be measured by indexes generated through principal component analysis. Prosocial expenditure is expenditure in children's health and education.

Experimental Design

In order to perform an accurate comparison of business performance of male versus female teams, we experimentally varied gender composition of the micro-enterprises, in three types of teams: all-male, all-female, and mixed teams (two males and two females).

We formed teams of prospective micro entrepreneurs randomly varying gender composition. After teams came forward and provided a deposit on the equipment they were to receive, we randomly selected half the teams for the intervention, and kept the other half in the control group. The deposit was returned to the control group teams right after randomization. There are six mutually exclusive treatment arms:

T1: All-males, treatment group (N=45 villages)
C1: All-males, control group (N=45 villages)
T2: All-females, treatment group (N=45 villages)
C2: All-females, control group (N=45 villages)
T3: Mixed-gender, treatment group (N=46 villages)
C3: Mixed gender, control group (N=46 villages)

Experimental Design Details

The intervention was implemented following the same plan of action as our implementing partner, thus allowing for scale up. Their staff (with support from IPA field teams) approached the village leaders to probe for their interest in setting up a recharge station, which implied putting together a four-person micro-entrepreneur team to operate it. Unlike our partner’s standard procedure, we randomly sorted interested villages in three groups (all-male, all-female, mixed) and requested a specific gender composition of the micro-entrepreneur teams in each village. We followed our partner’s standard procedure of requesting 40,000 Rwandan francs (around 50 US dollars) as co-investment (“commitment fee”) from the prospective micro-entrepreneur teams. This fee works as a deposit for the recharging station. In order to find the best counterfactual for the micro-entrepreneur teams, the commitment fee was raised from the prospective teams in all 272 villages prior to treatment assignment. Potential micro-entrepreneur teams were told their village had a 50% chance of being selected for expansion of our implanting partner’s activities. If the village was selected in the control group, the money was returned to the teams. If the village was selected in the treatment group, we randomly allocated 100 lights (one per household) for free and recharge kiosks were set up.

Randomization Method

Randomization was done in office by a computer

Randomization Unit

Village

Was the treatment clustered?

Yes

Experiment Characteristics

Sample size: planned number of clusters

272 villages

Sample size: planned number of observations

1,088 individuals

Sample size (or number of clusters) by treatment arms

45 villages in control 1 (all-male), 45 villages in control 2 (all-female), 46 villages in control 3 (mixed gender), 45 villages in treatment 1 (all male), 45 villages in treatment 2 (all-female), 46 villages in treatment 3 (mixed gender).

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

The expected sample size is 1088 households across 272 villages. There are 136 teams (544 micro-entrepreneurs) in treatment villages and 136 teams (544 prospective micro-entrepreneurs) in control villages. Since there are 45 clusters in each treatment arm, we calculate power with 90 clusters. The other assumptions are 80% power, 5% significance, 4 individuals per cluster, and an intraclass correlation of 0.02 (calculated from baseline data). The resulting minimum detectable effect (MDE) is 0.3 Standard Deviations. At baseline, one standard deviation of household head income is 39,640 Rwandan francs, so we would be able to detect changes of 11,892 Rwandan francs (13.8 US dollars/month).

Supporting Documents and Materials

IRB