Experimental Design
400 villages will be selected from the 8 biggest sending female migrant districts in Indonesia. These villages will be randomly assigned to one of four experimental groups as follows:
T0. Control Group – Migrants in these villages will not receive any information-sharing services under the auspices of this project.
T1. Report Card Treatment Group – Migrants in these villages will receive information-sharing services in the form of agency “report cards” only. The report cards will be introduced during village-level meetings for potential migrants.
T2. Comic Treatment Group – Migrants in these villages will receive information-sharing services in the form of “comic” only. The comic will be introduced during village-level meetings for potential migrants. During these meetings, the facilitators will discuss the content of the comic and relate it to the relationship between agencies quality and employer quality.
T3. Mixed Treatment Group – Migrants in these villages will receive information-sharing services in the form of agencies “report cards” and comic. Both, the report cards and comic book will be introduced during the village-level meetings.
Baseline Survey
In order to construct the PTs rating and gather data prior information sharing services available, we will conduct a baseline survey in all study villages. The baseline survey will include three distinct components as follows:
• Recently-Returned Migrants Survey: The primary objective of this survey will be to collect data from recently (i.e., within the past year) returned migrants about their experience with their PT and employer. Here, we will draw on the insights of the pilot study and collect both subjective and objective measures of PT and employer quality, including those used in our pilot study as well as additional measures suggested to us by local migrant NGO. We will interview up to 20 recently-returned migrants in each study village. Here, we will exclude migrants who used a PT that is no longer officially registered with BNP2TKI. This is necessary since many PTs exited the market following the moratorium on sending informal workers to Saudi Arabia; by limiting the sample to women who migrated with currently operational PTs, we will ensure that we spend baseline resources collecting information that is relevant to future migrants. As part of the village census that precedes the baseline, we will also collect data on the universe of PTs and sponsors (and the particular PTs they work with) that operated in each village over the past 2-3 years (as we did in pilot study).
• Tracking Sample Survey: We will also enroll 12 individuals in each village into a “tracking” sample, which will be targeted for a baseline interview and a subsequent interview during the post-intervention endline survey. Our initial hypothesis is that the impact of the information-sharing service could be quite different for individuals with prior migration experience (who may already be more savvy about picking a PT) as compared to individuals with no prior migration experience (who may be more vulnerable to exploitation). In order to study differential impacts for these two groups, we will therefore include 6 recently returned migrants who intend to migrate again and 6 prospective migrants who have never been abroad before in the tracking sample. At baseline, the tracking sample will be asked about their beliefs regarding the role that PTs and sponsors play in ensuring a good migration outcome and their expectations for experiences and compensation during future migrations. The baseline survey for the tracking sample will also collect data on risk, time preferences, and psychological measures such as locus of control to understand how these parameters correlate with migration behavior and mediate responses to the information intervention.
• Village Leader Survey: This survey will be very short and will collect basic information about the total number of male and female migrants working abroad, and the number of sponsors operating in the village. We will also collect information on whether there are any migration-focused NGOs operating in the area, and if so, who in the village organizes NGO activities.
The endline sample will include three distinct sub-samples. First, we will attempt to re-survey all 12 women per village in the baseline tracking sample. If some of these women remain overseas we will collect information from their families on their location and job description (including the PT they went with). If funding allows, we will attempt to contact them by phone in their destination countries.
It is difficult to predict how many women from the tracking sample will migrate. Although we will target women who state at baseline that they intend to work abroad in the near future, prior research from the Philippines has found low migration rates even among individuals who express interest in migration (Beam et al. 2013). To ensure that we have a large enough sample to detect meaningful changes in migrant welfare, we will therefore canvass each village and interview 5-10 women who returned from a migration that was initiated after the information-sharing intervention took place. The endline migrant’s survey will focus on women’s choice of sponsor and PT and the quality of their experiences abroad. We will also collect data on women’s beliefs about the role of fate in determining migration outcomes, informal information-sharing (e.g. giving and receiving advice within social networks), plans for future migrations, and a range of measures of economic and psychological wellbeing. For non-migrants in the tracking sample, the endline will focus on reasons for non-migration, beliefs about migration, and plans for future migrations.
Finally, at endline we also plan to conduct a short (follow-up) survey of local sponsors. This will help give us an understanding of how the information-sharing service impacted the “demand-side” of the migration market. In particular, the sponsors’ survey will collect data on sponsors’ profits (e.g. payments from PTs, payments to/from TKIs, other operating costs), recruiting areas, and linkages with high- and low-quality PTs. This, combined with administrative placements data, will allow us to paint a broad, detailed picture of the migration market and the varied impacts of the information-sharing service on both supply- and demand-side actors.