Leadership and Organizational Structure in Trade Unions in Myanmar

We conduct experiments with two groups: (1) union leaders; and (2) workers in unionized factories (both union members and non-members). With group (1), we conduct a mobilization and a real effort experiment where union presidents invite union leaders to mobilize workers to come on Sundays to participate in the CTUM’s research activities. We also conduct a “leading by example” experiment where union leaders receive an extra sum of money and then are asked to donate to CTUM-related public good under different conditions of observability by workers. With group (2), we conduct a discussion experiment where we experimentally manipulate the presence of union leaders to facilitate small group discussions on the preferred and expected minimum wage revision among diverse workers; a public good game where workers receive an extra sum of money and they are asked to donate to CTUM-related public good under different signals of union leaders’ contributions; and a “mobilization and coordination” experiment in which union line leaders motivate and/or coordinate members of small groups of workers to participate in a living cost survey. Note: At time of registration, due to possible time constraints, there is uncertainty whether the “mobilization and coordination” experiment will be implemented. This protocol will be updated as soon as possible to reflect the whether it remains part of the study.

2019-12-15

2020-05-31

We conduct experiments with two groups: (1) union leaders; and (2) workers in unionized factories (both union members and non-members). We study three key outcome variables for group (1): Number of workers recruited to participate in Session 2; number of posters prepared in Session 1; and amount of additional payment donated to CTUM public good. We study four key outcome variables for group (2): Expectations regarding the 2020 minimum wage adjustment; preferences regarding the 2020 minimum wage adjustment; participation in the living cost survey; amount of additional payment donated to CTUM public good. Note: At time of registration, due to possible time constraints, there is uncertainty whether the “mobilization and coordination” experiment will be implemented. This protocol will be updated as soon as possible to reflect the whether it remains part of the study.

Leveraging on CTUM’s network, we liaise with 28 unions at the factory-level that are affiliated with the CTUM. This sample consists of most of the garment sector factories affiliated with CTUM in the Yangon and Bago Regions of Myanmar. Each union participates in data collection sessions over the course of two sessions held across two weekends. In Session 1, the union’s president and its leadership team, “line leaders” (LLs) are invited to participate. In order to select LLs to participate, we randomly select sewing lines where LLs work to participate in the study. To randomly select sewing lines, we stratify lines by quartiles of proportion of workers unionized. In factories with low (<20%) or high (>80%) unionization rates, we oversample from the underrepresented group of workers (union members or nonmembers, respectively).

In Session 1, union leaders take a survey, and we conduct two experiments. We conduct a mobilization and a real effort experiment where union presidents invite LLs to mobilize workers to come on Sundays to participate in the CTUM’s research activities. We also conduct a “leading by example” experiment where union leaders receive an extra sum of money and then are asked to donate to CTUM-related public good under different conditions of observability by workers. We stratify LLs’ assignment to these experiments by their union.

In addition, to ensure sample representativeness, we randomly sample workers inside the factories, stratifying on factory, sewing line, union membership, and skill level. LLs are responsible to invite those workers on their line who are randomly selected to participate.

In Session 2, we implement two surveys: one on union membership, leadership and skill assessment and one on living costs. We leverage this data collection to embed experiments among the set of CTUM’s activities dedicated to raise workers’ awareness and elicit workers’ views on the minimum wage level: (1) a discussion experiment where we experimentally manipulate the presence of LLs to facilitate small group discussions on the preferred and expected minimum wage revision among diverse workers; (2) a public good game where workers receive an extra sum of money and they are asked to donate to CTUM-related public good under different signals of union leaders’ contributions; and (3) a “mobilization and coordination” experiment in which union line leaders motivate and/or coordinate members of small groups of workers to participate in the living cost survey. Note: At time of registration, due to possible time constraints, there is uncertainty whether the “mobilization and coordination” experiment will be implemented. This protocol will be updated as soon as possible to reflect the whether it remains part of the study.

For Session 1, we randomly select sewing lines where LLs work to participate in the study. To randomly select sewing lines, we stratify lines by quartiles of proportion of workers unionized. In factories with low (<20%) or high (>80%) unionization rates, we oversample from the underrepresented group of workers (union members or nonmembers, respectively). The randomization is done in office by a computer.

At Session 1, we stratify by factory and randomly assign LLs to different experimental conditions. This randomization is done in the field during Session 1 by computer.

For Session 2, we randomly select workers after stratifying on factory, sewing line, union membership, and skill level. This randomization is done in the field during Session 1 by computer. LLs are responsible to invite those workers on their line who are randomly selected to participate.

At Session 2, we stratify by factory and union membership status and randomly assign workers to different experimental conditions. This randomization is done in the field during Session 2 by computer.

Individual

No

Not applicable.

Expected numbers (actual will vary): 1792 workers; 308 line leaders (or “EC Members” depending on union); 28 presidents.

Line leaders:
Mobilization and leading by example experiment: 77 per treatment arm (2x2 design, can pool observations)
Public good experiment: 103 per treatment arm

Workers:
Group discussion experiment: 597 per treatment arm
Public good contribution experiment: 448 per treatment arm
Mobilization and coordination experiment: 299 per treatment arm (2x3 design, can pool observations)