- Issue: With encouragement from the Trump administration, 14 states have received approval for or are pursuing work requirements for nondisabled Medicaid beneficiaries. The requirements have sparked controversy, including two legal challenges.
- Goal: To predict the effect of work requirements on the insurance coverage of Medicaid enrollees over time.
- Methods: Analysis of the coverage patterns of a national cohort of nondisabled adults in the federal Medical Expenditure Panel Survey. Their experience is applied to a similar cohort of adults in Kentucky (which has received approval for work requirements, subject to a legal challenge) to project the potential effects of work requirements on their insurance coverage.
- Findings and Conclusions: Adding a new administrative hurdle in the form of work requirements in Kentucky would double the number of enrollees who disenroll from the program over a two-year period. We estimate that as many as 118,000 adults enrolled in Medicaid would either become uninsured for an extended period of time or experience a gap in insurance over a two-year period. These findings should be of concern to policymakers: research has found that adults who experience coverage gaps report problems getting health care or paying medical bills at rates nearly as high as those who are uninsured continuously.
With encouragement from the Trump administration, 14 states have gained approval or have submitted applications for requirements that would compel nondisabled adults to work a certain number of hours per week, or engage in another activity such as looking for a job, in exchange for Medicaid coverage.1,2 These work requirements fall under the Section 1115 demonstration waiver authority of the Social Security Act, which allows time-limited experimentation in Medicaid as long as the secretary of Health and Human Services (HHS) determines that it will likely promote Medicaid’s objectives.
The administration’s support of work requirements has sparked controversy. As of August, such requirements were facing legal challenges in Arkansas and Kentucky. In July, a federal district court judge vacated HHS Secretary Alex Azar’s approval of Kentucky’s waiver and sent the matter back to the secretary for further review.3 Azar responded by reposting Kentucky’s original application for a 30-day public comment period, which closed on August 18. More than 11,000 comments were filed in response. The administration is expected to issue a new determination on Kentucky’s application, after which the matter will likely return to court. In the meantime, the National Health Law Program, Legal Aid of Arkansas, and the Southern Poverty Law Center filed a new lawsuit challenging HHS’s approval of Arkansas’s Medicaid work requirement that the state began introducing in June.4
Work requirements present a serious risk — some Medicaid enrollees will not be able to meet the requirements and will lose their coverage. Kentucky estimated in its application that about 95,000 people over four years could disenroll from Medicaid as a result of the requirements. In Arkansas, which is phasing in the new requirements, more than 8,000 Medicaid enrollees have been cut from the program because they have not yet filed the necessary paperwork. Based on this early experience, up to 50,000 people could lose their Medicaid coverage when the requirements are fully phased in next year.5
These estimates do not fully assess the effect of the work requirements because they fail to account for the highly dynamic nature of eligibility and enrollment in Medicaid. For example, the estimates don’t predict the likelihood that people who are disenrolled for failing to meet work requirements will gain other insurance coverage, regain Medicaid, or remain uninsured. Nor do they account for the effect on people who are not currently enrolled in Medicaid but who may become eligible in the future. In this brief, we use data from the federal Medical Expenditure Panel Survey (MEPS) to look at the coverage patterns of Medicaid beneficiaries and then use this information to project the potential impact in Kentucky of work requirements on insurance coverage. (See the Methods and the Appendix for more detail on approach and assumptions.)
Enrollment in Medicaid Is Dynamic
Insurance coverage in the United States is highly dynamic: people move in and out of employer coverage, individual coverage, Medicaid, and Medicare because of life changes as simple as a birthday, moving to a new city, getting a new job, or getting married. This phenomenon is often referred to as “churn.” Income eligibility rules in Medicaid supercharge this dynamic — even a small change in income can make someone eligible or ineligible for benefits. If your income even temporarily ticks above the eligibility threshold for the Affordable Care Act’s Medicaid expansion (about $16,400 for an individual), you are no longer eligible for Medicaid.
Take seasonal farmworkers. They typically lack health insurance while they are working since they are unlikely to receive coverage from their employers. By working hard for several months at a time, they may lose Medicaid eligibility because their incomes rise above the threshold. When their seasonal work ends, their incomes may drop, making them again eligible for Medicaid.
Medicaid enrollees also face significant documentation requirements — both at the time of enrollment and during reenrollment. Prior research by Benjamin Sommers has suggested that administrative barriers, particularly renewal, are likely a key reason adults become disenrolled from Medicaid.6
While some Medicaid enrollees will leave the program for employer coverage or Medicare, research suggests that leaving Medicaid for reasons other than gaining another coverage source, such as termination of eligibility, leaves most enrollees without insurance options. An analysis of the effect of ending Tennessee’s adult Medicaid expansion found that most adults who lost their Medicaid did not gain other coverage.7
One-Third of Adults Enrolled in Medicaid Over Two Years Experienced Churn
To investigate the frequency with which people with Medicaid “churn off” insurance, we looked over a 24-month period at a cohort of nondisabled adults, ages 25 to 63, who were enrolled in Medicaid in January 2014. Of the 9 million people who were enrolled then, two-thirds (66.7%) stayed covered by Medicaid for the entire period ending in December 2015; one-third (33.3%) or 3 million left Medicaid either permanently or for a period of time (Exhibit 1).
Of the 3 million people who left Medicaid over that period (Exhibit 2):
- nearly three-quarters (73.7%) either became uninsured (36.3%) and remained so for the full 24-month period or had a coverage gap (37.4%) before either regaining Medicaid or another source of coverage
- just over one-quarter (26.3%) either went directly to another source of coverage and stayed on that coverage until the end of the period (20.7%) or gained another source of coverage (5.6%) before either regaining Medicaid or Medicare or becoming uninsured
- by the end of the period, more than one-third (36.9%) were uninsured. Another 27.8 percent regained Medicaid and 35.3 percent gained another source of coverage.
Even Temporary Gaps in Coverage Can Reduce Access to Care and Financial Security
Nearly three-quarters of people who churn off Medicaid become permanently or temporarily uninsured; this is a cause for concern. Experiencing even temporary gaps in insurance coverage can be devastating for people, particularly low-income families. The Commonwealth Fund Biennial Health Insurance Survey has found that working-age adults who experience gaps in their health insurance during the year report problems getting heath care because of cost at nearly the same rate as people who are without coverage for the entire year (Exhibit 3). The two groups also report similar rates of problems paying medical bills.
Assessing the Effects of Work Requirements
The way that people enroll, disenroll, and reenroll in Medicaid is pertinent to a conversation about the effects of work requirements for three reasons. First, by introducing work requirements as new eligibility criteria, states will likely add to the number of enrollees who churn off Medicaid each year and increase the number of people who are uninsured or experience coverage gaps.
Second, people who try to reenroll in Medicaid will face a new hurdle: proving they work at least 20 hours per week. This could cause the share of people who ultimately reenroll in Medicaid following a coverage gap to decline further. Consider our example of seasonal workers. If their incomes fall below the eligibility threshold at the end of their working season, they would have to prove they were working at least 20 hours a week to reenroll in Medicaid. This new hurdle could lengthen coverage gaps for people whose sole reason for losing Medicaid in the first place was earning too much during a concentrated period of hard work, only to fail to meet minimum work-hour requirements when that work ended.
Third, our analysis of the MEPS data only looks at people enrolled at a point of time. In each of the 24 months, new people will become eligible for Medicaid because of life changes such as income fluctuation, divorce, or moving from another state. A full assessment of the effects of work requirements also must consider the experiences of people who would become eligible for Medicaid in the future and either would be unable to enroll or would not attempt to enroll because of work requirements.
The Case of Kentucky
Using Kentucky as an example, we looked at MEPS data to project the number of people who might lose Medicaid coverage as a result of work requirements and, of those, how many would become uninsured or experience a gap in coverage.
Work Requirements Would Double the Number of Adults Who Churn Off Medicaid in Kentucky
According to federal data, there were about 325,000 nondisabled adults ages 18 to 64 enrolled in Kentucky’s Medicaid program in 2016–17.8 We estimate that prior to the implementation of work requirements, about 108,000 enrollees would be expected to leave Medicaid over a 24-month period (for reasons discussed earlier in this brief). Of those, about 80,000 would either become uninsured and remain so (39,200) or experience a gap in coverage (40,500).
If Kentucky were to implement work requirements, most of the 325,000 nondisabled adults enrolled in the program would have to demonstrate that they work at least 20 hours per week or engage in other activities, like looking for a job. Based on Arkansas’s experience, MEPS data, and estimates from Randy Haught and colleagues at Dobson, DaVanzo and Associates, about 108,000 people would likely lose Medicaid coverage as result of the work requirements over a 24-month period. Adding work requirements would thus increase the number of nondisabled adults churning off Medicaid in Kentucky from an estimated 108,000 adults to 216,000 — a 100 percent increase.
Work Requirements Would Increase the Number of Adults in Kentucky Who Are Uninsured or Experience a Coverage Gap by 118,000
We used the MEPS data to simulate the coverage experience over a 24-month period of the adults who disenroll from Medicaid as a consequence of work requirements. We estimate that of the 108,000 adults who are expected to disenroll as a result of work requirements, 68,000 to 107,000 will experience a permanent coverage loss and 600 to 40,000 will experience a gap in coverage. The ranges reflect different assumptions about the degree to which people are able to regain Medicaid or gain private coverage (see the Methods for more detail).
We also use the MEPS data to estimate the effect of work requirements on people who disenroll from Medicaid for other reasons, such as income fluctuations. In Kentucky, of the 108,000 adults projected to disenroll for reasons other than work requirements, about 10,000 would be expected to regain Medicaid based on MEPS data. We include a scenario in which all members of this group regain Medicaid under work requirements and one in which none of this group would regain Medicaid.
Thus, as many as 118,000 people would either experience a permanent coverage loss or gap in coverage as a result of work requirements over a 24-month period. Depending on people’s ability to regain Medicaid or gain private coverage with work requirements in place, the number of adults who would remain uninsured at the end of the two-year period ranges from 77,500 to 117,400.
This likely underestimates the number of people who will become uninsured as a result of work requirements because we only examine adults who begin the two-year period with Medicaid and lose coverage as a result of work requirements. We do not estimate the number of adults who may become eligible for Medicaid over a two-year period and might not enroll because of the administrative barrier created by work requirements. These would include people whose income falls sufficiently to make them eligible for Medicaid, adults who move to Kentucky, or young adults who become eligible on their 19th birthdays.
Conclusion and Policy Implications
The Affordable Care Act expanded coverage options in the United States by building on existing insurance sources, including employer-based insurance, the individual insurance market, Medicaid, and Medicare. Where Americans get their insurance continues to be a function of the type of job they or a family member has, how many hours they work, their age, marital status, income, where they live, and other factors. Multiple coverage sources and their complex eligibility rules mean that movement in and out of coverage is a defining feature of our health system. Sometimes movement is seamless but sometimes it includes a temporary or permanent loss of coverage. Commonwealth Fund survey data presented in this brief show that even a short gap in coverage disrupts people’s health care and leaves them exposed to potentially catastrophic medical costs.
The ACA attempted to reduce coverage gaps by making it easier for people to move between coverage options. Several reforms were made to Medicaid aimed at simplifying enrollment, including the elimination of in-person interviews, limiting information requested from Medicaid applicants, and electronic information verification that uses data from other federal and state agencies.9 Adding new enrollment barriers such as work requirements will accelerate churn in the states that pursue them and leave hundreds of thousands without access to health care.
The Medical Expenditure Panel Survey (MEPS) Household Component began in 1996 and collects data from a sample of families and individuals in selected communities across the United States, drawn from a nationally representative subsample of households that participated in the prior year’s National Health Interview Survey (conducted by the National Center for Health Statistics). The survey is a panel design featuring several rounds of interviews with the sample of households covering two full calendar years. We analyzed a cohort of 25-to-63-year-old nondisabled adults who were enrolled in Medicaid in January 2014, examining changes in their coverage over a 24-month period ending in December 2015.
Of the 9 million nondisabled adults ages 25 to 63 years who were enrolled in Medicaid in January 2014, 66.7 percent stayed covered by Medicaid for the entire period ending in December 2015 and one-third (33.3%) or 3 million left Medicaid either permanently or for a period of time (Exhibit 1). Among the 3 million people who churned off Medicaid over that period, 20.7 percent went directly to another source of coverage, either employer coverage (14.3%), individual market (3.3%), or Medicare (3.1%) and stayed on that coverage until the end of the period; 37.4 percent spent a time uninsured before either regaining Medicaid (23%) or gaining employer coverage (11.9%), individual market coverage (2%) or Medicare (0.5%.) (Appendix Exhibit 1). About 5.6 percent gained another source of coverage before either regaining Medicaid (4.8%), enrolling in Medicare (0.2%) or becoming uninsured (0.6%). By the end of the period more than one-third (36.9%) of adults who had churned off Medicaid were uninsured, 27.8 percent had regained Medicaid, 26.2 percent had coverage through an employer, 5.3 percent had individual market coverage, and 3.8 percent had enrolled in Medicare.
We then applied these national estimates of churn to people enrolled in Medicaid in Kentucky. Randy Haught and colleagues at Dobson, DaVanzo and Associates estimate that there were about 325,000 nondisabled adults ages 18–64 enrolled in Kentucky’s Medicaid program in 2016–17. If we assume that churn in Kentucky’s Medicaid program is roughly similar to the MEPS national estimates, of this group, about 108,000 enrollees would be projected to churn off Medicaid over a 24-month period.
Haught and colleagues estimate that most of the 325,000 nondisabled adults enrolled in Medicaid would be eligible for work requirements (Appendix Exhibit 2). The researchers estimate that of those, about 266,000 are working enough hours to satisfy the requirements or would not be working but likely to be granted an exemption. The remaining 59,000 would not be employed.
Based on the MEPS analysis, we assume that of the first group, 24.5 percent would likely churn off Medicaid because of the new administrative barriers. This is the percentage of adults nationally who churn off Medicaid for reasons other than shifting to new coverage. Prior research by Benjamin Sommers has suggested that administrative barriers, particularly renewal, are a key reason adults disenroll from Medicaid.10 Based on the experience of the work requirements in the Supplemental Nutrition Assistance Program in four states, Haught and colleagues estimate that 72.5 percent of those who are not working and subject to work requirements would not submit the necessary documents to satisfy the work requirements and disenroll.
Based on these estimates, about 108,000 people, or 33.2 percent, would not submit the necessary documents for work requirements and lose their Medicaid coverage. Adding work requirements would thus increase the number of nondisabled adults churning off Medicaid in Kentucky over a two-year period from an estimated 108,000 adults to 216,000, an increase of 100 percent (Appendix Exhibit 3).
We then use the MEPS data to simulate the coverage experience of this group that disenrolls from Medicaid as a consequence of work requirements over a 24-month period. First, we examine the 108,000 Medicaid enrollees who are projected to churn off Medicaid because they would not submit the necessary documents to satisfy work requirements. Based on the MEPS data, about 14.4 percent may spend some time uninsured before eventually gaining other coverage through an employer (11.9%), the individual market (2%), or Medicare (0.5%). However, a recent analysis of the suspension of Tennessee’s adult Medicaid expansion, TennCare, in 2005, found no evidence that the adults who lost Medicaid gained private coverage.11 Based on this research, we also estimate a scenario where there is no subsequent gain in private coverage in Kentucky following loss of Medicaid because of work requirements. Finally, because the work requirements will make it more difficult for people who lose Medicaid to regain it, we estimate a scenario where no one who disenrolls regains Medicaid. Thus, we project that 108,000 adults either experience a permanent coverage loss (68,000 to 107,000) or a gap in coverage (600 to 40,000) over a two-year period directly as a result of work requirements (Appendix Exhibit 4).
Second, we examine the experience of the estimated 108,000 enrollees who would be expected to disenroll from Medicaid over a 24-month period for reasons other than work requirements. Under work requirements, people who disenroll from Medicaid for reasons unrelated to work requirements and attempt to regain Medicaid coverage in the future would now have to prove that they were working at least 20 hours per week. We thus estimate a scenario where all members of this group regain Medicaid under work requirements, an estimated 10,000 people, and one in which none of this group would regain Medicaid. Thus, the number of people projected to remain uninsured or experience a coverage gap as a result of work requirements ranges from 108,000 to 118,000 over a two-year period.
The authors thank Sara Rosenbaum of George Washington University and Randy Haught of Dobson, DaVanzo and Associates for helpful comments; and Herman Bhupal and Munira Gunja for research assistance, Deborah Lorber, Paul Frame, and Jen Wilson for editorial support and design, all at the Commonwealth Fund.
1. Commonwealth Fund, “Status of Medicaid Expansion and Work Requirement Waivers,” interactive map, last updated Oct. 4, 2018.
2. Sara Rosenbaum et al., “State 1115 Proposals to Reduce Medicaid Eligibility: Assessing Their Scope and Projected Impact,” To the Point (blog), Commonwealth Fund, Jan. 11, 2018.
3. Sara Rosenbaum, “Stewart v. Azar and the Future of Medicaid Work Requirements: What Are the Takeaways, and What Happens Next?,” To the Point (blog), Commonwealth Fund, July 3, 2018.
4. Jane Perkins, Elizabeth Edwards, and Leo Cuello, Summary of Lawsuit Filed Against HHS Approval of “Arkansas Works” Amendment (National Health Law Program, Aug. 14, 2018).
5. Erin Brantley and Leighton Ku, “A First Glance at Medicaid Work Requirements in Arkansas: More Than One-Quarter Did Not Meet Requirement,” Health Affairs Blog, Aug. 13, 2018.
6. Benjamin D. Sommers, “Loss of Health Insurance Among Non-Elderly Adults in Medicaid,” Journal of General Internal Medicine 24, no. 1 (Jan. 2009): 1–7.
7. Thomas DeLeire, The Effect of Disenrollment from Medicaid on Employment, Insurance Coverage, Health, and Health Care Utilization (National Bureau of Economic Research, Aug. 2018).
8. Randy Haught et al., “The Potential Financial Impact of Medicaid Work Requirement on Kentucky Hospitals,” To the Point (blog), Commonwealth Fund, forthcoming.
9. Sara Rosenbaum et al., Streamlining Medicaid Enrollment: The Role of the Health Insurance Marketplaces and the Impact of State Policies (Commonwealth Fund, March 2016).
10. Sommers, “Loss of Health Insurance,” 2009.
11. DeLeire, Effect of Disenrollment, 2018.