Targeting welfare programs to construction workers

Last registered on December 29, 2024

Pre-Trial

Trial Information

General Information

Title
Targeting welfare programs to construction workers
RCT ID
AEARCTR-0013166
Initial registration date
December 27, 2024

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
December 29, 2024, 11:12 PM EST

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

Locations

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

Request Information

Primary Investigator

Affiliation
Harvard University

Other Primary Investigator(s)

PI Affiliation
University of Chicago

Additional Trial Information

Status
In development
Start date
2025-01-15
End date
2025-03-31
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
This study examines interventions to improve the delivery and targeting of welfare programs for construction workers in Odisha, India. Construction workers are a vulnerable population exposed to a high degree of physical and expropriation risk, and the state has designed a range of welfare programs intended specifically for them. Despite high awareness, enrollment is constrained by transaction costs and information barriers. Using a cluster-randomized design, we will compare two approaches: (1) reducing transaction costs via government-led onsite registration and (2) addressing information asymmetries through clear, actionable enrollment guidance. We will test treatment effects on a range of outcomes, including labor card enrollment, benefit access, costs of application, labor supply, and earnings. The study contributes to key questions in the public economics and development economics literatures by evaluating cost-effective mechanisms to improve program uptake and reduce exclusion errors in the implementation of welfare programs targeted at vulnerable populations.
External Link(s)

Registration Citation

Citation
Ramachandran, Sabareesh and Kartik Srivastava. 2024. "Targeting welfare programs to construction workers." AEA RCT Registry. December 29. https://doi.org/10.1257/rct.13166-1.0
Experimental Details

Interventions

Intervention(s)
The intervention that we have designed targets construction workers in the state of Odisha, India. The Odisha Buildings & Other Construction Workers Welfare (OB & OCWW) Board issues “labor cards” to construction workers, which enable them to access several public welfare programs targeted specifically to them by occupation. In the status quo, workers face significant barriers to enrolling in these cards, including transaction costs and information constraints with respect to the processes involved. As a result, while awareness of these cards is high, enrollment rates are low. In this proposed project, we want to implement two interventions, designed to tackle both the transaction costs and information constraints.

The first intervention involves top down enrollment, where the focus is on reducing the transaction costs of applying for labor cards. As part of this intervention, government officials will conduct an on site registration drive to enrol workers for the cards. This will remove the travel costs, the opportunity cost of the workers’ time and any side payments that workers would have had to make to formal or informal intermediaries. The second intervention removes the information barriers faced by the applicant. We will share a flyer containing specific information about the application process, with directions for the workers that can be acted upon for them to independently enrol.

We will conduct two rounds of surveys with workers. First, we will conduct an in-person baseline with these workers prior to allocating them to one of the two interventions or a control arm. The intervention will be conducted approximately one week following the baseline. The second round will be an endline phone survey conducted 1 month after the baseline, and 2-3 weeks after the intervention. Lastly, we will also get administrative data on labor card registrations and access to benefits from the Labor Department.
Intervention Start Date
2025-01-20
Intervention End Date
2025-02-28

Primary Outcomes

Primary Outcomes (end points)
Enrollment in labor cards (successful and attempted)
Primary Outcomes (explanation)
Enrollment in labor cards (successful and attempted), which will consist of two binary variables:
- Whether the respondent ever applied for a labor card through any source
- Whether the respondent was ever successful in receiving a labor card (unconditional on having applied for one)

Secondary Outcomes

Secondary Outcomes (end points)
Cost of (attempted) enrollment
Errors of exclusion
Total value of benefits applied for and received
Worker retention
Labor supply
Earnings
Secondary Outcomes (explanation)
- Cost of (attempted) enrollment, which will be constructed as the sum of the following:
-- Formal registration fees
-- Informal and formal side payments made to any intermediaries
-- Costs of collecting or printing documents, and accessing digital portals
-- Cumulative travel costs
-- Opportunity cost of time, valued at their average daily earnings from baseline
- Errors of exclusion, which we will estimate as follows:
-- The rate of rejection for respondents who meet the statutory eligibility requirements
- Total value of benefits applied for and received, which will be estimated using both survey and administrative data; we will report both sets of outcomes (receipts based on surveys and administrative data)
- Worker retention, which will be estimated as following:
-- The probability that the worker remains at the worksite until the spell of employment or task is completed
-- The probability that the worker is hired by the same contractor or employer in the future
- Labor supply, which will be estimated using survey data where we ask respondents how many of the previous 30 days they were working for money, both in construction jobs and otherwise
- Earnings, which will be estimated using survey data where we ask respondents how much money they earned in the previous 30 days

For any outcomes that are defined conditional on enrollment or attempts to enroll, we will specify a two-stage selection model wherein we first predict enrollment rates for the non-enrolled using some assumptions on the functional forms of the costs faced by those respondents.

Experimental Design

Experimental Design
Our intervention will be implemented at the level of worksites, which we have sampled based on a near-census of some neighborhood clusters in Bhubaneshwar and Cuttack, Odisha. Each worksite employs 2-10 workers on an active construction project. Worksites, and all their workers, will be sorted into one of three treatment arms. In the first arm, worksites are assigned to a top-down intervention where last-mile government agents visit the worksite and assist workers with in-person registrations. In the second arm, worksites are assigned to an intervention alleviating workers’ information constraints with respect to the enrollment process. In this arm, we will provide workers with information on where they can enroll in labor cards, without assisting with the actual registration process. The third arm will include control worksites, where we will deliver no intervention.

The randomization of worksites into treatment groups will be stratified at the level of the combination of neighborhood, number of workers, and the expected duration of the worksite. The neighborhood of the worksite will be approximated by drawing fixed gridlines around our study locations. For our surveys, we will attempt to cover the universe of all workers at each sampled worksite.

While we randomize worksites based on our census, it is possible that during our baseline some of these worksites no longer exist because the work was already completed. In these cases the enumerator will contact the corresponding contractor and include a different worksite they are working in then as a replacement. The replaced worksite will have the same treatment status as the original worksite. The enumerator will not be aware of the treatment status of the worksite while picking a replacement.

We anticipate some amount of churn in the workers working in the worksite between baseline and the intervention. In this case the workers benefitting from the intervention would be different from the workers enrolled in the experiment at baseline. Our main analysis will only include the workers enrolled at baseline. In exploratory analysis we will also analyse the impact considering the sample of workers at the intervention worksites to be those present on the day of the intervention (instead of those present during the baseline).
Experimental Design Details
Not available
Randomization Method
Randomization will be conducted using a computer algorithm.
Randomization Unit
Worksites.
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
300
Sample size: planned number of observations
1000 - 1500 (exact number to be established after the baseline)
Sample size (or number of clusters) by treatment arms
100 clusters per treatment arm
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
IFMR Institutional Review Board
IRB Approval Date
2024-11-14
IRB Approval Number
N/A