Virginia Work Requirements Experiments

How to cite

Engel, Jordan; Layton, Timothy; Leive, Adam; McIntyre, Adrianna. 2024. "Virginia Work Requirements Experiments." AEA RCT Registry. AEARCTR-0012966. https://doi.org/10.1257/rct.12966-1.0

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

Jordan Engel* , Timothy Layton , Adam Leive , Adrianna McIntyre

* Primary investigator

Registry ID
AEARCTR-0012966
DOI
https://doi.org/10.1257/rct.12966-1.0
First registered
Last updated
Status
In development
Trial dates
2024-03-31 – 2025-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/12966

Abstract

Work requirements are increasingly common in major public assistance programs, initially implemented in cash support (TANF) and food assistance (SNAP) programs, and recently proposed in Medicaid. Proponents of work requirements contend that these policies increase beneficiary engagement in work and community activities and may put them on a trajectory to self-sustainability. Opponents of work requirements suggest that many low-resource households will lose much-needed benefits, without commensurate improvements in employment. There is limited evidence on the causal impact of work requirements in modern public assistance programs, even as more than a dozen states have proposed new Medicaid work requirements in recent years, while other states and the federal government have expressed interest in removing work requirements from other programs. We propose a randomized-controlled trial of work requirements in Virginia, leveraging state support for exempting a portion SNAP recipients from the state’s scheduled work requirements. We will be able to separate the causal impact of work requirements from another barrier to enrollment – frequent recertification requirements – by cross-randomizing exemptions to each type of requirement. We will study impacts of both policies on program enrollment, employment, earnings, food insecurity, and health care utilization using a mix of administrative and survey-collected data.

Intervention

Target Population: SNAP beneficiaries subject to work requirements: Able-bodied adults without dependents (ABAWDs). ABAWDs consist of individuals aged 18 to 49 who are not pregnant, medically certified as unable to work, living in a household that includes a child under the age of 18, or caring for an older dependent.
Status quo policy: ABAWDs are subject to a 6 month recertification period, meaning they must recertify their SNAP eligibility every 6 months. If they do not recertify their eligibility, they are removed from the program. Part of eligibility recertification is reporting participation in work or other activities that meet the work requirement – such as education, job-training, volunteer work, etc. – for each of the previous 6 months. The work requirement involves working or participating in an approved work program for 20 hours or more per week. ABAWDs can also satisfy the requirement by volunteering for at least the number of hours equal to their SNAP benefits divided by the federal minimum wage.
If the ABAWD does not complete recertification, he is removed from the program. Similarly, if the ABAWD is determined to not have been in compliance with the work requirements throughout the previous 6 month recertification period, he is removed from the program. Each ABAWD is allowed 3 months of exemption from the work requirement, called “Y1” months in the case management system. ABAWDs can use these Y1 months at any time during their enrollment spell. At the time of recertification, the ABAWD reports his work activities for each month. It is then determined whether the ABAWD was in compliance with the work requirement for each month or has a Y1 exemption month available. If the ABAWD is deemed in compliance or exempt for each of the 6 months, his program eligibility is recertified for an additional 6 months. After those 6 months, the ABAWD again goes through the recertification process.
Some localities (counties and independent cities) have historically received exemptions from the ABAWD work requirements due to weak labor market conditions. In particular, qualifications include having a 3- month average unemployment rate above 10 percent, a 24-month average unemployment rate that is above 20 percent of the national average during the same period, or being designated as a labor surplus area by the Department of Labor, among others.
Treatment: Each year, Virginia receives an allotment of work requirement exemption months from the federal government to use as it wishes. These exemption months are allocated based on a fraction of the state’s prior caseload of ABAWDs. In the case management system, these are referred to as “E9” months.
These exemption months roll over from year-to-year, and Virginia has accumulated a large number of months. The Virginia Department of Social Services (DSS) has agreed to deploy many of these exemption months in a randomized controlled trial (as described below) to assess who benefits most from these exemptions. They will use the results of the RCT to determine how best to allocate these exemption months in the future.
Table 1: Study Arms
Arm 1 – Control group (status quo): 6-month recertification period, no additional work requirement exemptions
Arm 2 – Intervention Group 1: 6-month recertification period, 6- month work requirement exemption
Arm 3 – Intervention Group 2: 6-month recertification period, 12- month work requirement exemption
Arm 4 – Intervention Group 3: 12-month recertification period, 6- month work requirement exemption
Arm 5 – Intervention Group 4: 12-month recertification period, 12- month work requirement exemption
We will cross-randomize ABAWDs to receive work requirement exemptions of 6 or 12 months and to receive recertification extensions from 6 to 12 months. Specifically, ABAWDs will be randomly assigned to one of 5 study arms, described in Table 1.
We will implement the work requirement exemptions via a new rule in Virginia’s case management system. For anyone assigned to Arm 2, just before the 6 month recertification, the system will erase any “Y1” months that have been used and replace them with “E9” months. This has the effect of exempting the ABAWD from work requirements during the previous recertification period, by allowing them to have additional Y1 months to allocate. Case workers and the case management system know that if an ABAWD has Y1 months available, they are allowed to remain on the program for an additional recertification period. So, by erasing the Y1 months and replacing them with other exemption months, the individual will be allowed to remain on the program even if they were not working during the previous recertification period. For anyone assigned to Arm 3, Y1 months will also be replaced just before both their first 6 month recertification and just before their second 6 month recertification, effectively exempting them from work requirements for a full 12 months, and allowing them to stay on the program for 18 months without working.
We include treatment arms with both 6 month and 12 month work requirement exemptions in order to determine which groups benefit most from shorter versus longer exemptions. Because DSS has a limited number of “E9” months to allocate in the future, they wish to determine whether it is better to allocate more exemption months to a smaller set of ABAWDs or fewer exemption months to a broader set of ABAWDs. Currently, they do the latter, and this study design will allow us to determine whether that is the optimal approach.
ABAWDs assigned to Arms 4 and 5 will also receive the same work requirement exemptions as those assigned to Arms 2 and 3, but they will also receive a recertification extension from 6 months to 12 months. This will also be implemented via a new rule in the case management system that will change the recertification date for anyone in these arms to be 12 months later instead of 6 months later. ABAWDs assigned to these arms will still be required to complete a brief “interim report” around month 7, as required by federal law, but they will not be required to report compliance with work requirements or a variety of other detailed eligibility criteria.
Randomization and Implementation: All SNAP work requirements are currently on hold due to the public health emergency (PHE) related to the COVID-19 pandemic. The study will begin the month that the PHE ends and SNAP work requirements are reinstated. Implementation will proceed as follows.
In study month 1 (the first month after the end of the PHE) we will identify all ABAWDs in their first month of a six-month recertification period. These individuals will be called “cohort 1,” and they will consist both of individuals who just enrolled in SNAP and individuals who have been enrolled in SNAP for some period of time but who just completed an eligibility recertification. We will randomize these individuals to either the control group or one of the four intervention groups. To randomize, we will randomly assign all combinations of 4 digits to one of the 5 arms. We will then provide a crosswalk between 4 digit numbers and arms to DSS. DSS will then use that crosswalk to assign ABAWDs to arms based on the last 4 digits of their case number in the DSS enrollment system. Study month 1 will be referred to as “cohort month 1” for individuals in cohort 1. We will then repeat this study enrollment process in each month until we enroll 40,000 ABAWDs in the study, in each month enrolling and randomizing all ABAWDs in the first month of a 6 month recertification period. Based on prior SNAP data, we predict that there will be 6-7 cohorts.
After a given cohort’s randomization, the intervention will proceed as follows.
• Communication – In cohort month 1, just after randomization, we will send communications to recipients assigned to the intervention groups informing them that they have been selected to have their work requirements waived. Recipients assigned to intervention Arm 2 will be told that their work requirements have been waived during this 6-month recertification period. Recipients assigned to intervention Arm 3 will be told that their work requirements have been waived for this 6-month recertification period and the following 6-month recertification period. Recipients assigned to intervention Arm 4 will be told that the length of their recertification period has been changed, and they will not be required to recertify until 12 months later instead of 6 months (they will also be told that they will need to complete a brief interim report at 6 months). Recipients assigned to intervention Arm 5 will be told that the length of their recertification period has been changed to 12 months AND their work requirements are waived during that entire 12 month period.
• Intervention – In cohort month 1 we will change the recertification dates for individuals assigned to Arms 4 and 5 in the SNAP system. In cohort month 5, we will replace Y1 months in the SNAP system. Specifically, we will replace Y1 months with E9 months for individuals assigned to intervention Arms 2 and 3. In cohort month 6, individuals assigned to intervention Arms 2 and 3 (and Arm 1, the control group) will have their 6 month recertification. In cohort month 7, individuals assigned to Arms 4 and 5 will be sent interim reports to complete. (We may also cross-randomize an intervention to help people remember to complete their reports.) In cohort month 11, we will again replace Y1 months in the SNAP system. Specifically, we will replace Y1 months with E9 months for individuals assigned Arms 3 and 5. In cohort month 12, all groups will have a recertification. After this point, all individuals in the cohort will return to the status quo, meaning that all study participants will be subject to the 18 month recertification and will be required to be fully compliant with work requirements at that time in order to remain enrolled in SNAP beyond that time.

Experimental design

Newly enrolled SNAP Able-Bodied Adults without Dependents (ABAWDs) will be randomized to one of five treatment arms that alter participants’ recertification period and work requirement exemption. We will then study the effects of work requirements and recertification periods on in SNAP on program participation, labor supply, and well-being.
The study’s target enrollment will be 30,000 single-adult households. Starting on March 31, 2024 all new SNAP cases that are certified and eligible for work requirements (ABAWDs) will be randomly assigned to one of five study groups with equal probability on a rolling basis until we achieve the target enrollment. We thus expect 6,000 participants per study arm.

The study arms are as follows:
1. Control group (Status Quo): 6-month recertification period, no additional work requirement exemption months
2. Intervention group 1: 6-month recertification period, 6-month work requirements exemption
3. Intervention group 2: 6-month recertification period, 12-month work requirement exemption
4. Intervention group 3: 12-month recertification period, 6-month work requirement exemption
5. Intervention group 4: 12-month recertification period, 12-month work requirement exemption
We will also conduct a survey among a smaller sample of participants. The survey sample will start from a pool of 15,000 participants, yielding an expected 5,000 respondents. The survey sample will include the first 15,000 participants enrolled in the study.

Design details

Randomization method
We used simple random sampling without stratification to assign all possible 5 digit numbers to treatment arms. At the time of SNAP enrollment certification, all clients are assigned a client ID. The eligibility and enrollment system will automatically crosswalk the last 5 digits of the client ID to treatment arms to assign clients to treatment arms. The crosswalk assigns 5 digit numbers to the study groups with equal probability. We will construct balance tables to verify that randomization successfully resulted in balance of observable characteristics. Observable characteristics will include controls listed below plus predicted earnings and SNAP participation constructed by regressing these outcomes on ex ante participant characteristics in the status quo arm.
Randomization unit
Single-adult SNAP households
Clusters
30,000 single-adult SNAP households
Observations
30,000 single-adult SNAP households
Arms
6,000 per treatment arm: 6,000 Control Group 6,000 Intervention Group 1 6,000 Intervention Group 2 6,000 Intervention Group 3 6,000 Intervention Group 4
Primary outcomes
1. SNAP enrollment at 7 months post-certification, 13 months, and 19 post-certification 2. Employment (in UI-covered jobs) during first 6 months post-certification, first 12 months post-certification, and first 18 months post-certification
Secondary outcomes
• Earnings in UI-covered employment • Predicted earnings based on firm age, firm size, and NAICS industry codes (“job quality”). • Labor supply measured in IRS tax records • Earnings and employment in all jobs measured in survey (survey) • Number of employers (survey) • Food insecurity (survey) • Use of food pantry (survey) • Self-reported health (survey) • Self-reported mental health (survey) • Volunteer activities (survey) • Educational activities (survey) • Credit reports • Self-reported financial difficulties (survey) • Evictions (court records) • Forced relocation (survey) • Geographic mobility • Index of all outcomes based on standardized treatment effects (Kling, Liebman, Katz 2007; Finkelstein et al. 2012) • Administrative burden of complying with the work requirements • Flexibility of job hours/scheduling (Schneider & Harknett, 2019) • Selling plasma (survey) • Insurance coverage (survey) • Indicator for inpatient use (APCD) • Indicator for emergency department use (APCD) • Total health care spending (APCD) • Food insecurity (survey) • Self-reported health (survey) • Self-reported mental health (survey) • Self-reported financial difficulties (survey)

Cited by

No citation data available for this registration yet.

Keywords