The Roles of Information and Search Frictions in Determining Working Conditions in Bangladesh’s Apparel Sector

How to cite

Boudreau, Laura; Heath, Rachel; Ahmed, Md. Shakil. 2023. "The Roles of Information and Search Frictions in Determining Working Conditions in Bangladesh’s Apparel Sector." AEA RCT Registry. AEARCTR-0012216. https://doi.org/10.1257/rct.12216-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Laura Boudreau* , Rachel Heath , Md. Shakil Ahmed

* Primary investigator

Registry ID
AEARCTR-0012216
DOI
https://doi.org/10.1257/rct.12216-2.0
First registered
Last updated
Status
On going
Trial dates
2023-07-11 – 2024-06-30
Countries
Bangladesh ()
JEL codes
J22, J28, J31, J32, J63, J81
Registry record
http://www.socialscienceregistry.org/trials/12216

Abstract

The emergence of low-skill manufacturing sectors in developing countries can provide economic benefits, especially for women, but working conditions are often difficult. We propose a cluster RCT to experimentally investigate whether information and search frictions in Bangladesh’s labor market contribute to inefficient matching between garment workers and firms. We will use data from a baseline household survey to measure working conditions and wages across garment factories in local labor markets. We will then provide this information on job characteristics, job openings, or both, and assess the impacts of these treatments on workers’ beliefs about working conditions (and wages) in garment factories nearby to their home, mobility, wages, and working conditions.

Intervention

There are 4 experimental conditions:
1. Scorecard: Workers are provided a scorecard with information about working conditions, wages, and travel distances for factories within 5 km of their neighborhood.
2. Vacancies: Worker a provided a pamphlet with information about job vacancies in factories within 5 km of their neighborhood.
3. Scorecard and Vacancies: Worker receives both (1) and (2).
4. Control: Worker does not receive information.

For the scorecard, because there is no publicly data available on working conditions in a representative sample of garment factories, we began with a survey, representative at the geographic level (“representative survey” henceforth), that asked workers about working conditions in their factories. We took information about working conditions and wages from this survey and turned this into an overall measure of working conditions and wages in the factories for the scorecard treatment.

Our vacancy information is collected via interviews with HR managers in partnership with our research partner BRAC Centre for Entrepreneurship Development (CED) which has extensive experience collecting data from managers at factories. We then collate this information into a pamphlet of all open vacancies at surveyed factories within 5 kilometers of the neighborhood for the vacancy treatment.

Both the scorecard and the pamphlet referenced a hotline number that participants can call for more information.

Experimental design

We randomly assign participants to 1 of the 4 different treatment conditions described above.

Prior to conducting the experimental interventions, we conducted a baseline survey with experimental participants. We will conduct follow-up surveys after 1, 2, 3, and 6 months to measure participants’ employment outcomes.

Design details

Randomization method
The randomization is done by a computer.
Randomization unit
The unit of randomization is a “neighborhood.” In this study, "neighborhood" does not connote a community; we define a "neighborhood" as a circle of Bangladesh's surface area with a radius of 0.5 kilometers (km). We implement a stratified randomization by geographic area (Gazipur, Savar, or Narayanganj), above/below median average beliefs about working conditions at the worker's factory compared to other factories nearby, above/below median count of factories within 5 km of the neighborhood, and above/below median count of eligible workers in the neighborhood cluster.
Clusters
270 neighborhoods.
Observations
Wave 1: 2,821. Wave 2: 1,620. Total: 4,441.
Arms
Control: 83 neighborhoods. Vacancies: 62 neighborhoods. Scorecard: 63 neighborhoods. Vacancies and scorecard: 62 neighborhoods.
Primary outcomes
1. Belief about quality of non-pecuniary job aspects of baseline factory relative to other factories within a 60-minute walk of participant's home. 2. Total monthly earnings. 3. Job quality index [for participants employed in garments sector]. 4. Mobility (a binary variable measuring whether the participant has changed jobs in the previous month).
Secondary outcomes
1. Belief about wages of baseline factory relative to other factories within a 60-minute walk of participant's home. 2. Job search: A binary variable measuring whether the participant has searched for jobs in the prior month. 3. Mental health: An index of mental health survey questions (Generalized Anxiety Disorder 7-item (GAD-7) scale)). 4. Travel time to respondent’s workplace (among the sample of participants who work in a factory in the MiB list, we can triangulate these reports with distance from their house using the geolocation data from MiB).
Sponsors
Name: Gender and Labor Markets in Low-Income Countries Initiative Url: https://g2lm-lic.iza.org/ Name: International Growth Centre Url: https://www.theigc.org Name: Private Enterprise Development in Low-Income Countries (PEDL) Url: https://pedl.cepr.org
Partners
Name: BRAC Institute of Governance and Development (BIGD) Partner_type: ngo Url: https://bigd.bracu.ac.bd Name: BRAC Centre for Entrepreneurship Development (CED) Partner_type: ngo Url: https://ced.bracu.ac.bd Name: Development Research Initiative (dRi) Partner_type: private_company Url: https://dri-int.org

Cited by

No citation data available for this registration yet.

Keywords