Government Scheme Overload: Take-Up and Targeting in a Multi-Scheme Context

I partner with a Non-Governmental Organization that offers last mile service delivery support to migrants in India. Frontline workers approach migrant communities to disseminate information on scheme eligibility and help register migrants for schemes. After introducing themselves and gaining trust, frontline workers use a mobile application to quickly survey one migrant at a time. This survey includes a variety of demographic questions about the migrant and their family. Upon completing the survey, the mobile application assesses a migrant and their family’s eligibility for different schemes, identified as priority schemes by the NGO.

I classify the schemes migrants are eligible for into two categories:
1. On-the-spot schemes – These are schemes migrants can typically be registered for “on the spot” with minimal documentation. These are primarily composed of schemes for which the migrant can apply to via a government website if they have their Aadhar card (National Identification Card) with them and a phone number linked to it to provide two-factor authentication. Frontline workers explain the scheme to the migrant, scan the migrant’s ID, and register the migrant for the scheme through the government website.
2. Additional schemes – The majority of schemes migrants are eligible for require documentation migrants may not have on their person. These include family ID cards, letters from school, proof of address, ration cards, and birth certificates. In order to apply for these schemes (conditional on eligibility), migrants are instructed to return to the NGO’s office with the relevant documentation, where the frontline worker can either register them on the government website directly or accompany them to a government office where a civil servant will register them.

Migrant workers and their family members are often eligible for both categories of schemes. This study seeks to understand how the quantity of schemes the migrant learns their eligibility about impacts the subsequent take-up of additional schemes via a RCT. The intervention varies both the mode and cadence by which frontline workers share information on schemes to migrant workers. The intervention does not impact the ability for migrants to receive information and apply to “on-the-spot” schemes, but rather, varies how many “additional” schemes they share with each migrant. Additionally, the intervention introduces a “hard-copy” to document what information is shared with the migrant utilizing non-carbon-paper receipt books.

2025-01-25

2025-03-27

1. Demand for information on schemes
2. Take-up of additional schemes
a) Total additional schemes attempted
b) At least one additional scheme attempted

1. Demand for information on schemes - create a binary variable that is 1 if the migrant responded "yes" to wanting more information on schemes and 0 if the migrant responded "no"
2. Take-up of additional schemes
a) Total additional schemes attempted - create a continuous measure for the total number of additional schemes attempted (application submitted, application successful, benefit availed) after the initial registration with the frontline worker
b) At least one additional scheme attempted - create a binary variable that is 1 if the migrant attempted (application submitted, application successful, application received, benefit availed) at least 1 of the additional schemes and 0 if not

1. Interactions with frontline workers
2. Take-up of additional schemes
a) Total applications submitted
b) Total successful applications
c) Total applications
d) Total availed benefits
e) Total attempts
f) Schemes shared
g) Application timeline
3. Take-up of additional schemes (proportions)
4. Cost-benefit analysis

1. Interactions with frontline workers - create a continuous variable for number of interactions the migrant has with frontline workers
2. Take-up of additional schemes
a) Total applications submitted - create a continuous variable for the number of additional schemes for which the migrant has submitted an application
b) Total successful applications - create a continuous variable for the number of additional schemes for which the migrant has successfully submitted an application (application approved, but benefit not yet availed)
c) Total applications - create a continuous variable for the number of additional schemes for which the migrant has applied to (application submitted OR successful)
d) Total availed benefits - create a continuous variable for the number of additional schemes for which the migrant has availed a benefit for
e) Total attempts - create a continuous variable for the number of additional schemes for which the migrant has attempted to avail (application submitted, application successful, OR benefit availed)
f) Schemes shared - create a binary variable that is 1 if the migrant applied to at least 1 of the specific schemes shared by the frontline worker and 0 if none
g) Application timeline - create a variable for the number of days in between the initial registration date and next date the migrant interacts with a frontline worker
3. Take-up of additional schemes (proportions) - create the same variables 2a-e and divide each by the total number of schemes the migrant is eligible for
4. Cost-benefit analysis - average number of migrants registered per day

SAMPLE: The study population consists of internal migrants in the Delhi NC Region (Delhi NCR) who had not previously been registered by a frontline worker. Frontline workers recruit migrant workers via their normal practice of entering migrant-dense communities and introducing the NGO’s work. A migrant community in any given day is assumed to be representative of internal migrants in New Delhi.

RANDOMIZATION: The NGO has 15 offices in Delhi NCR, from which the frontline workers base themselves and visit surrounding communities. I randomly assign each of the 15 offices into one of the three treatment arms.

INTERVENTION: Below are the three treatment arms:
1. Control - no change: Frontline workers register migrants using the mobile application and learn which schemes the migrant is eligible for. The frontline worker registers the migrant for the “on-the-spot” schemes they are eligible for and have documentation for. The frontline worker then fills out the receipt document for the migrant, indicating which schemes they registered for and providing an address of the nearest NGO office. The frontline worker uses their own discretion to verbally share additional information with the migrant on “additional schemes” the migrant is eligible for.

2. Treatment 1 - single scheme cadence: Frontline workers follow the same initial procedure from the control group with respect to registering migrants for “on-the-spot” schemes and recording a receipt of those. In addition, the frontline worker asks migrants, “You are eligible for one additional scheme I strongly recommend. Are you interested in learning about it and how to access it?” If migrants are interested in one scheme, the frontline worker can choose an additional scheme the app has identified the migrant being eligible for that the frontline worker recommends. The frontline worker then records this information on the receipt to describe the scheme, its benefits, and next steps the migrant should take to apply for the scheme. The migrant receives a copy of this document. If frontline workers come to an office, they will learn of another scheme.

3. Treatment 2 - multi-scheme cadence: Multiple Schemes: Frontline workers follow the same procedure registering migrants for “on-the-spot” schemes. Next, the frontline worker notes the total number of schemes the migrant and the migrant’s family is eligible for. With this information in mind, the frontline worker asks migrants, “You and your family are eligible for X total schemes. Are you interested in learning about them and how to access them?” If migrants are interested, they explain the schemes the migrant is eligible for and check the appropriate boxes on the large paper so that the migrant has a record of which (on a list) the migrant is eligible for.

The main difference between the two treatment groups are the quantity of information on “additional schemes” shared with migrants. In Treatment 1, migrants learn of one scheme and receive more detailed information about the scheme on a half-page sheet of paper. In Treatment 2, migrants learn of many schemes and receive a full A-4 sheet of paper containing a large list of all potential schemes with the ones they are eligible for clearly indicated.

Not available

Randomization done in office by a computer

Each office area is randomly assigned to a treatment arm. Therefore, each treatment is clustered at the office area. Within each office area, there are 3 frontline workers who register migrants in the area. Treatment is clustered since frontline workers often travel to the same communities and frequently overlap with each other. Therefore, to minimize spillovers, treatment assignment is done at the office level.

Yes

15 offices

6,750 migrants

2,250 migrants (5 offices) - Control
2,250 migrants (5 offices) - Treatment 1
2,250 migrants (5 offices) - Treatment 2

MDES = 0.1 SD = 1