Eligibility Restrictions in Recent Medicaid Waivers Would Cause Many Thousands of People to Become Uninsured
August 9, 2018
The Trump Administration has approved Medicaid waivers for several states that create unprecedented barriers to coverage for low-income adults, including work requirements, high premiums for people with very low incomes, and provisions that lock people out of coverage for six months or more if they don’t meet these and other requirements. Independent experts and states themselves have projected that many Medicaid enrollees will lose coverage due to these restrictions. For example, Kaiser Family Foundation researchers have shown how work requirements alone — if imposed nationwide — could cause 1.4 to 4.0 million people to lose coverage, or 6 to 17 percent of those potentially subject to the policy. Kentucky, the first state in which the Administration approved work requirements and other new restrictions, forecasts that its waiver will ultimately reduce adult enrollment by more than 95,000 (15 percent of adult enrollees) in a typical month.Work requirements alone — if imposed nationwide — could cause 1.4 to 4.0 million people to lose coverage.
Partly on the basis of these estimates, the District of Columbia federal district court recently found that the Administration acted in an “arbitrary and capricious” manner in approving Kentucky’s waiver, failing to consider the waiver’s impact on “furnishing medical assistance,” a core objective of the Medicaid program. Some proponents of eligibility restrictions, however, have argued that Medicaid coverage loss estimates greatly overstate the effect of these restrictions on overall uninsured rates. They claim that many of those leaving Medicaid will do so because the new work requirements lead them to obtain jobs with affordable private coverage, not because they don’t meet the new requirements and become uninsured. This argument is not substantiated by any data from the Administration, Kentucky, or other states with approved or pending waivers, and it flies in the face of the available evidence.
- Tens of thousands of enrollees will lose Medicaid because they don’t meet new eligibility or paperwork requirements; there is no reason to expect that many of those enrollees will be able to replace Medicaid with other coverage. Focusing on Kentucky’s work requirement policy, we estimate potential coverage losses based on the Kaiser Family Foundation’s analysis of past experiences with eligibility restrictions in Medicaid and work requirements in other federal programs, an Urban Institute analysis of whom Kentucky’s waiver will affect, and evidence on work patterns among low-wage workers. These suggest that between 45,000 and 103,000 Kentucky Medicaid enrollees could lose coverage because they: (1) work but don’t work the required number of hours each and every month; (2) don’t (or can’t) meet new requirements, despite the threat of losing their health coverage; or (3) don’t submit newly required paperwork, despite remaining eligible for Medicaid. There is no reason to expect that many of those losing Medicaid for these reasons will be able to replace it with other coverage; rather, they will likely become uninsured.
As explained below, these estimates are rough, but conservative in several important respects. They take into account only Kentucky’s work requirement, not the premiums and coverage lockouts its waiver would also impose, which could significantly increase coverage losses; they draw on estimates at the lower end of past experience with eligibility restrictions in Medicaid and other programs; and they assume a higher share of enrollees will successfully comply with new reporting requirements than have done so to date under Arkansas’s similar work requirement.
- In contrast, few enrollees will likely leave Medicaid because the new requirements cause them to find jobs providing affordable health insurance. As discussed below, research on work requirements in other federal programs finds little evidence that they produce sizeable and lasting increases in employment or income. And work requirements and other eligibility restrictions in Medicaid may have even more disappointing results. One reason is that taking away health coverage will likely make it harder for some enrollees to look for work or maintain stable employment. Another is that the recent restrictions come with no new funding for job training, child care, or other work supports.
Moreover, even if some enrollees do find jobs, these will likely be mostly low-wage jobs. Such jobs are unlikely to boost enrollees’ incomes enough for them to shift from Medicaid into subsidized individual market coverage, and the large majority do not offer affordable health insurance. For example, among workers in the bottom quartile of the wage distribution, only 37 percent are offered medical coverage, and less than a quarter actually obtain coverage, presumably in large part because required premium contributions are often higher than they can afford. As an illustrative calculation, if Kentucky’s work requirement had employment effects comparable to the fifth-year effects found in randomized trials of work requirements in cash assistance programs, and if 37 percent of those finding jobs obtained health insurance, fewer than 2,000 enrollees would be expected to obtain employer-based coverage as a result. This compares to tens of thousands leaving Medicaid for other reasons and likely becoming uninsured.
- Coverage gains from the Affordable Care Act’s (ACA) Medicaid expansion translated roughly one-for-one into lower uninsured rates. A different way to evaluate the likely impact of waiver-related Medicaid coverage losses on overall coverage rates is to examine the impact of past changes in Medicaid eligibility. Following the ACA’s expansion of Medicaid to low-income adults, the share of this group covered by Medicaid rose and the share that is uninsured fell by similar amounts, with little or no offsetting decline in employer coverage. This indicates that the large majority of expansion enrollees would have been uninsured but for expansion, contradicting claims by expansion opponents that it would “crowd out” private coverage (that is, cause people to shift from private coverage to Medicaid). That, in turn, suggests that most adult enrollees who lose Medicaid eligibility due to waivers will become uninsured, not gain coverage from other sources. Similarly, past eligibility restrictions and administrative barriers in Medicaid have led many people to become uninsured, though the evidence here is more limited.
- Medicaid coverage improves access to care, health, and financial security, just as private coverage does. Some proponents of eligibility restrictions claim that Medicaid provides low-quality coverage and argue that this is another reason to disregard Medicaid coverage losses. To the contrary, a large and growing body of research on the ACA’s Medicaid expansion finds that coverage gains are generating large gains in access to care and financial security. Notably, a direct comparison between Kentucky’s traditional Medicaid expansion and Arkansas’ expansion through commercial health plans finds comparable gains on these measures, including comparable access to both primary care physicians and specialists. And since Medicaid requires much lower enrollee cost sharing than most employer plans do, it likely offers better access to care and more financial protection to low-income adults than most employer plans would. In addition, evidence on earlier Medicaid expansions to low-income adults, as well as emerging evidence on the more recent ACA expansion, finds that coverage gains have led to improved mental and physical health outcomes.
Few Losing Medicaid Under Waivers Are Likely to Gain Other Health Coverage
In theory, work requirements and other eligibility restrictions could cause people to leave Medicaid for a variety of reasons, with different implications for overall health coverage rates. On the one hand, people might leave Medicaid because they do not or cannot comply with the new requirements — for example because they can’t afford premiums, can’t find a job or a job with sufficient and steady hours, or struggle to provide newly required paperwork. On the other hand, people might leave Medicaid because work requirements cause them to find jobs with higher earnings or employer-sponsored coverage. In practice, there is clear evidence that the first type of Medicaid coverage loss will predominate.
Many Will Lose Coverage Because They Can’t Meet New Requirements
Focusing on Kentucky as an example, we draw on the Kaiser Family Foundation’s analysis of the coverage impacts of Medicaid work requirements and the Urban Institute’s analysis of who would be subject to Kentucky’s policy to estimate that between 45,000 and 103,000 Kentuckians could lose coverage because they work but don’t work enough hours each month, get tripped up by new paperwork requirements, or aren’t working and don’t find work in response to the new rules. Premiums and other policies included in Kentucky’s waiver will result in additional coverage losses. There is no reason to expect that many of those losing Medicaid for these reasons will be able to replace it with private coverage.
Kentucky’s waiver application doesn’t explain the basis for its estimate that the waiver will ultimately (by its fifth year) reduce adult Medicaid enrollment by more than 95,000 in a typical month. However, the analysis above suggests that Kentucky’s work requirement policy could produce Medicaid coverage losses of that magnitude even in the near term, well before the fifth year, and just among people who would likely lose Medicaid without gaining other coverage: that is, those who do not or cannot meet the requirement, apart from any who would leave Medicaid because they gain employment and switch to private coverage.
To arrive at these estimates, we follow Kaiser’s approach in assuming that work requirements could cause 5 to 15 percent of Medicaid enrollees who meet the requirements or should qualify for exemptions, and 25 to 50 percent of other enrollees, to lose coverage. The Kaiser researchers base these estimates on a review of past experiences with eligibility restrictions in Medicaid and work requirements in the Supplemental Nutrition Assistance Program (SNAP, formerly food stamps). They first conclude that “paperwork or administrative requirements in Medicaid or CHIP [Children’s Health Insurance Program] lead to between 5 and 30 percent of people losing coverage” — mostly eligible enrollees tripped up by red tape. Based on that, they conservatively assume that 5 to 15 percent of those who should remain eligible for Medicaid under work requirement policies could lose coverage. (This is also consistent with evidence that people likely eligible for exemptions, for example because they have disabilities or other health limitations, have lost benefits under SNAP and Temporary Assistance for Needy Families [TANF] work requirement policies.)
Similarly, noting that participation in SNAP has dropped by 50 to 85 percent among populations subject to work requirements, the Kaiser researchers conservatively assume that 25 to 50 percent of Medicaid enrollees who appear not to meet work requirements or qualify for exemptions based on survey data could lose coverage. The rest are assumed to find ways to meet the requirements with qualifying work activities other than employment, find ways to satisfy exemption requirements despite not appearing to do so in survey data, or find work.
Urban researchers, meanwhile, estimate that 498,000 non-elderly adults not receiving disability assistance could be subject to Kentucky’s work requirement policy (based on data available from 2016). Of these, about 106,000 appear to be primary caregivers for children, an exemption for which Kentucky should generally be able to screen in its own data, without requiring additional documentation from enrollees. Of the remainder:
- 8,000 to 24,000 enrollees could lose coverage due to failure to successfully claim the exemptions to which they might be entitled. Urban estimates that 160,000 Kentuckians who might otherwise be subject to work requirements are students or else are not working and have a serious health limitation or live with someone who does, suggesting that they might qualify for exemptions based on health status or caregiving responsibilities. Supposing 5 to 15 percent of these enrollees either fail to qualify for or fail to document eligibility for exemptions, 8,000 to 24,000 would lose coverage.
- 5,000 to 16,000 enrollees could lose coverage due to failure to successfully report work hours. Urban estimates that 106,000 enrollees work an average of at least 20 hours per week and worked at least 50 weeks out of the year, likely meeting Kentucky’s 80-hour-per-month work requirement. Supposing 5 to 15 percent of these enrollees fail to properly document work hours at some point during the year, 5,000 to 16,000 would lose coverage.
- 15,000 to 30,0000 enrollees could lose coverage because they work but don’t work the required 80 hours per month every month. Many working Medicaid enrollees are employed in industries such as retail, home health, and construction. Their jobs feature hours that fluctuate to month to month and little flexibility, so illness, family emergencies, or child care or transportation problems can lead to job loss and spells between jobs. About 59,000 Kentucky Medicaid enrollees work but not throughout the year or less than 20 hours per week on average, Urban estimates. Supposing 25 to 50 percent of those in this group do not make up their work hours shortfalls with other qualifying work activities and do not document eligibility for other exemptions, 15,000 to 30,000 would lose coverage.
- 17,000 to 34,000 enrollees could lose coverage because they aren’t working, don’t qualify for exemptions, and don’t find jobs, volunteer work, or job training opportunities. Urban estimates that about 68,000 Kentuckians were not working in 2016 and don’t fall into any of the categories in the bullets above (i.e., they do not appear to qualify for exemptions related to caregiving or student status and do not report having a serious health limitation or living with someone who does), although many face other serious barriers to work. (For example, compared to other enrollees, non-working enrollees are disproportionately older and less educated, and are more likely to lack access to a vehicle.) A 25 to 50 percent coverage loss for this group amounts to 17,000 to 34,000 people.
The bolded figures sum to our estimate that 45,000 to 103,000 Kentuckians will lose coverage because they don’t meet the work requirement. 
These estimates are of course rough. Since work requirements have never before been implemented in Medicaid, assumptions have to be based on evidence from other Medicaid eligibility restrictions and work requirements in other programs. And while Urban’s analysis of the population that could be subject to Kentucky’s waiver is the most granular publicly available, it does not capture all possible exemptions (for example, certain older people who may effectively be exempt by virtue of participating in SNAP, or other categories such as former foster care youth). In addition, our estimates do not take into account that some of those who work but in unstable jobs may also report having a serious illness or living with someone who does. On the other hand, not everyone who appears as if they might be eligible for exemptions in the survey data Urban relies on will actually qualify under Kentucky’s rules; in particular, some of those who report a serious health limitation or living with someone who does will not qualify for Kentucky’s exemptions based on health status or caring for a person with disabilities, and some students will not qualify for Kentucky’s student exemption, which is limited to full-time students.
Notably, however, the analysis is conservative in several important respects. First, as noted, the Kaiser assumptions we use as the basis for the analysis draw on estimates at the lower end of past experience with eligibility restrictions in Medicaid and other programs.
Second, initial experience with Arkansas’ work requirement suggests that Kaiser’s assumption that 85 to 95 percent of eligible enrollees will successfully comply with reporting requirements may be too optimistic. In the first month of the new requirement, less than 6 percent of Arkansas enrollees required to report hours of work or work activities did so, even though a larger share were almost certainly working (or exempt). The share of people reporting may increase over time, but the initial experience points to the risk of even larger coverage losses than those estimated above.
Third, the analysis only accounts for work requirements, not for other eligibility restrictions in Kentucky’s waiver. For example, the waiver would impose premiums of up to 4 percent of income, and enrollees with incomes of 100 to 138 percent of the poverty line would lose coverage if they failed to pay. A large body of evidence finds that even modest premiums result in significant Medicaid coverage losses among low-income people. The waiver also makes it harder for adults to maintain coverage by locking them out of Medicaid for six months for failure to turn in renewal or other paperwork on time. These additional reporting burdens will likely cause even more people to lose Medicaid and becoming uninsured.
Few People Will Lose Coverage Because They Gain Jobs With Health Insurance
Of the various unprecedented eligibility restrictions in newly approved Medicaid waivers, work requirements are the only policy for which supporters have offered a theory by which the restriction could promote economic mobility. But the evidence provides little support for the hypothesis that work requirements will meaningfully increase employment, and even less for the notion that they will cause large numbers of enrollees to gain good jobs with health insurance.
Research on work requirements in federal cash assistance programs —TANF and its precursor, Aid to Families with Dependent Children (AFDC) — finds that employment increases for those subject to work requirements are generally modest, fade over time, and don’t move many families out of poverty. For example, a synthesis of results from randomized trials of 13 programs imposing work requirements in cash assistance programs finds that employment rose by modest amounts in the first two years, but these gains generally faded by year five (to an average effect of about 1 percentage point). Meanwhile, stable employment proved the exception, not the norm, and few enrollees transitioned out of poverty as a result of the work requirements.
For a number of reasons, work requirements in Medicaid will likely have equally or more disappointing results. Sixty percent of adult Medicaid enrollees potentially subject to work requirements already work, and more than 80 percent of the remainder are students or report that they are unable to work due to a disability, serious illness, or caregiving responsibilities. This suggests limited scope for work requirements to increase work participation.
Meanwhile, cash assistance programs generally provide at least some (albeit inadequate) resources for the supportive services that many low-income adults need in order to work, such as child care, job training, and transportation assistance. The more successful experimental programs described above coupled work requirements with robust work supports. In contrast, the Administration’s Medicaid work requirements guidance says that states imposing these requirements need not offer any new work supports and may not use federal funding for such supports. Neither Kentucky nor any other state in which the Administration has so far approved work requirements has committed to providing work supports.
Medicaid work requirements also risk backfiring for some enrollees, because health care is itself a critical work support. As Kaiser Family Foundation researchers concluded from a comprehensive review of the available evidence, “access to affordable health insurance has a positive effect on people’s ability to obtain and maintain employment,” while lack of access to needed care, especially mental health care and substance use treatment, impedes employment. Low-income adult Medicaid enrollees have high rates of chronic conditions and mental illness; for example, 69 percent of adults enrolled in Michigan’s Medicaid expansion have at least one chronic physical or mental health condition. Individuals with conditions like diabetes, heart disease, or depression may be able to hold a steady job if these conditions are treated and controlled, but work may become impossible if conditions go untreated. Consistent with that, majorities of non-working adults gaining coverage through the ACA’s Medicaid expansion in Ohio and Michigan said having health care made it easier to look for work, while majorities of working adults said coverage made it easier to work or made them better at their jobs.
Requiring non-working adults to find work before allowing them to gain coverage and access to treatment could create a catch-22, where people with serious health needs can’t get the medical help they need to find a job unless they first get a job. Meanwhile, work requirements could create a vicious cycle for enrollees who are initially working: if a health setback leads to job loss, that would in turn lead to loss of access to treatment, making it more difficult to regain health and employment. (Premiums and other eligibility barriers included in Kentucky’s and other states’ waivers could also adversely affect employment for some enrollees.)
Moreover, even if some enrollees do find jobs as a result of work requirements, these will probably be mostly low-wage jobs. Such jobs are unlikely to boost enrollees’ incomes enough for them to shift from Medicaid into subsidized individual market coverage, and the large majority do not offer affordable health insurance — meaning most enrollees would still need Medicaid coverage. According to Labor Department data, among workers with earnings in the bottom quartile of the wage distribution, only 37 percent are offered health coverage, and less than a quarter actually obtain coverage, presumably in large part because required employee premium contributions are often higher than low-wage workers can afford. Similarly, only 37 percent of full-time workers with family incomes below the poverty line (and only 13 percent of such part-time workers) are even offered coverage. Consistent with these data, in Medicaid expansion states, 42 percent of workers with family incomes below 138 percent of the poverty line (the income limit for Medicaid in these states) obtain health insurance through Medicaid, more than twice the share that obtain insurance through an employer.
Thus, even if work requirements led meaningful numbers of enrollees to find jobs, such enrollees still could not account for the large Medicaid coverage reductions Kentucky itself projects will result from its policies. As an illustrative calculation, if Kentucky’s work requirement increases employment by 0.9 percentage points (the average fifth-year effect in the cash assistance randomized trials discussed above), if 37 percent of those becoming employed due to the requirement are offered employer coverage (consistent with the overall share of low-wage workers offered coverage), and if all of them sign up for it (despite required premium contributions that may be high as a share of income), fewer than 2,000 people would gain employer coverage. Contrary to supporters’ claims, work requirements are highly unlikely to result in large numbers of people leaving Medicaid because they obtain other health coverage, but will likely lead to tens of thousands of people leaving Medicaid because they do not meet the new requirements.
Increases in Medicaid Coverage from ACA Expansion Led to Comparable Reductions in Uninsured Rates
Another way to project how recent Medicaid waivers will affect overall health insurance rates is to examine the impact of past expansions and reductions in Medicaid coverage. The most directly relevant evidence comes from the ACA’s expansion of Medicaid to low-income adults.
No Evidence of Meaningful “Crowd-Out” From ACA Medicaid Expansion
The ACA offered states the option to expand Medicaid to cover all adults with incomes below 138 percent of the federal poverty line. The primary effect of the restrictive Medicaid waivers that the Trump Administration has approved over the past year would be to take coverage away from a substantial share of this group. If Medicaid expansion significantly “crowded out” other coverage sources for low-income adults, their gains in health coverage from the expansion would be far smaller than their increase in Medicaid coverage. In that case, there would be reason to think that many of those losing Medicaid coverage due to the recent waivers would gain coverage from those other sources. Conversely, if Medicaid overage gains under expansion translated roughly one-for-one into gains in overall health coverage, that would indicate that few of these adults have other viable coverage options, and Medicaid coverage losses from waivers would be expected to translate into higher overall uninsured rates.
Figure 1 shows the relationship between the increase in Medicaid coverage and the decrease in uninsured rates from 2013 to 2016 among states adopting the ACA Medicaid expansion. There is a strongly linear and nearly one-to-one relationship between the increase in the share of a state’s population covered by Medicaid and the decrease in the share uninsured.
That simple relationship is confirmed by more rigorous analyses. As a comprehensive Kaiser Family Foundation literature review concludes, most studies find “no evidence of ‘crowd-out’” while some show “slight declines in private coverage” from Medicaid expansion. In particular, studies have found no evidence that the ACA Medicaid expansion meaningfully decreased employment, and no evidence of decreased employer coverage among those employed.
For example, a study by Urban Institute researchers finds that from 2013 to 2017 (as expansion took effect), the share of non-elderly low-income workers with Medicaid or other public coverage increased by 20.5 percentage points in expansion states, while the share with employer-sponsored coverage remained unchanged. The uninsured rate fell by 22.7 percentage points, reflecting the large increase in Medicaid coverage and a small increase in the share with individual market coverage, presumably through the ACA marketplaces. The study also finds no change in employers offering coverage to low-income workers.
Medicaid Eligibility Restrictions Have Increased Uninsured Rates
While the evidence on how Medicaid eligibility restrictions (as opposed to expansions) affect overall uninsured rates is more limited, research indicates that most people who lose Medicaid coverage due to administrative hurdles or failure to pay premiums become uninsured, rather than transitioning to other coverage.
For example, a study of pre-ACA adult Medicaid enrollees found that 29 percent of those who remained continuously eligible for Medicaid nonetheless lost coverage by the end of a 12-month period, with paperwork burdens associated with coverage renewals and redeterminations likely playing a significant role. Among adults losing Medicaid, about a third re-enrolled within six months, but a substantial majority of the remainder were uninsured six months after exiting the program.
Another study, focused on increased Medicaid premiums for adult enrollees in Oregon, found that the resulting Medicaid coverage losses led to both higher overall uninsured rates and reduced access to medical care. Overall, a Kaiser Family Foundation review of the literature on premiums in Medicaid concludes, “Although some individuals who disenroll from Medicaid or CHIP following premium increases move to other sources of coverage, others become uninsured and face negative effects on their access to care and financial security. . . . Several studies suggest that these negative effects on health care are largest among individuals with greater health care needs.”
Medicaid Improves Access to Care, Health, and Financial Security
Some supporters of Medicaid eligibility restrictions also argue that Medicaid coverage losses are unimportant because Medicaid offers low-quality coverage. In reality, there is strong evidence that gaining Medicaid coverage improves access to care, health, and financial security, including among adult enrollees who would be subject to new requirements. Where valid comparisons to private coverage are available, benefits of Medicaid and private coverage appear similar, and Medicaid offers substantially lower cost sharing, an important advantage for people in or near poverty.
Access to care. A large and growing body of research shows that Medicaid expansion has improved access to care for low-income adults, for example increasing the share with a personal physician, getting check-ups and preventive care, getting regular care for chronic conditions, getting treatment for substance use disorders, and getting mental health treatment; expansion has also lowered the share of low-income adults who delay needed care or do not fill prescriptions due to cost. A study that directly compares Kentucky’s traditional Medicaid expansion with Arkansas’ expansion, which provided coverage through commercial, individual market health plans finds similar gains in access to care and no evidence that Kentucky enrollees had more trouble accessing either primary care physicians or specialists. Likewise, an Urban Institute survey found that low-income adults enrolled in Medicaid were about as likely to have a usual source of care and to have received routine check-ups as moderate-income adults enrolled in employer coverage or subsidized coverage through the ACA marketplaces and were less likely to skip needed care due to cost. Also of note, adult Medicaid enrollees report high satisfaction with their coverage, including satisfaction with their choice of doctors.
Financial security. In addition to providing access to care, another important purpose of health insurance is to improve financial security, preventing illness from leading to financial catastrophe. Expanding Medicaid coverage results in fewer and smaller unpaid medical bills and fewer debts sent to third-party collection agencies, studies have found, and it reduces the share of low-income adults who report struggling to pay medical bills. (In the Urban survey, Medicaid enrollees were less likely to report trouble paying medical bills than people with employer-sponsored or marketplace coverage.) Moreover, by reducing unpaid medical bills, Medicaid expansion improved credit scores among low-income adults, qualifying them for lower-interest mortgage, auto, and credit card loans. The estimated interest savings average $280 per adult gaining coverage per year, and an estimated $520 million across the expansion population.
Quality of care and health outcomes. Four years into the ACA’s expansion of Medicaid, there is growing evidence that improvements in access to care are leading to improvements in treatment and health. For example, recent studies have found increases in early-stage cancer diagnoses and decreases in the share of patients receiving surgical care inconsistent with medical guidelines. Studies have found improvements in overall self-reported health, reductions in the share of low-income adults screening positive for depression, and improvements in diabetes and hypertension control. Meanwhile, studies of pre-ACA expansions of coverage to low-income adults have found substantial gains in physical health, including reductions in mortality.
The bottom line is that there is simply no basis for claims that Medicaid fails to provide the core benefits of health insurance. Rather, all evidence indicates that, if recent waivers cause people to lose or experience interruptions in Medicaid coverage, their access to care, financial security, and health will suffer.
 Rachel Garfield, Robin Rudowitz, and MaryBeth Musumeci, “Implications of a Medicaid Work Requirement: National Estimates of Potential Coverage Losses,” Kaiser Family Foundation, June 27, 2018, https://www.kff.org/medicaid/issue-brief/implications-of-a-medicaid-work-requirement-national-estimates-of-potential-coverage-losses/.
 Kentucky HEALTH Section 1115 Demonstration Modification Request, July 3, 2017, https://www.medicaid.gov/Medicaid-CHIP-Program-Information/By-Topics/Waivers/1115/downloads/ky/ky-health-pa2.pdf.
 Memorandum Opinion, United States District Court for the District of Columbia, Ronnie Maurice Stewart et al. v. Alex M. Azar II et al., https://ecf.dcd.uscourts.gov/cgi-bin/show_public_doc?2018cv0152-74.
 For examples of this argument, see Ricardo Alonso-Zaldivar, “Judge’s Ruling Slows Plans for Medicaid Work Requirement,” Associated Press, July 4, 2018, https://apnews.com/bb481c05694f4b5db7abb7a5ee146eff and David Catron, “How a Judicial Hack Ended Kentucky’s Medicaid Work Plan,” American Spectator, July 2, 2018, https://spectator.org/how-a-judicial-hack-halted-kentuckys-medicaid-work-plan.
 See Garfield, Rudowitz, and Musumeci, op. cit. and Anuj Gangopadhyaya and Genevieve M. Kenney, “Updated: Who Could Be Affected by Kentucky’s Medicaid Work Requirements, and What Do We Know About Them?” Urban Institute, March 2018, https://www.urban.org/sites/default/files/publication/96576/3.26-ky-updates_finalized_1.pdf.
 For a review of this evidence, see Hannah Katch, Jennifer Wagner, and Aviva Aron-Dine, “Medicaid Work Requirements Will Reduce Low-Income Families’ Access to Care and Worsen Health Outcomes,” Center on Budget and Policy Priorities, February 8, 2018, https://www.cbpp.org/research/health/medicaid-work-requirements-will-reduce-low-income-families-access-to-care-and-worsen.
 The analysis also drew on evidence on compliance rates for various provisions of existing Medicaid waivers, such as healthy behaviors incentives.
 The 106,000 figure reflects the 168,000 enrollees Urban classifies as “likely exempt” (from Table 1 of the analysis) multiplied by the 63 percent of this group who appear to be primary caregivers of children (from Figure 2).
 The 160,000 figure is the sum of: (1) the 62,000 enrollees Urban classifies as “likely exempt” who are not parents (168,000 less 106,000) plus (2) the 97,000 enrollees who are not working and report a serious health limitation or living in a household with someone who does (59 percent of the 165,000 potentially non-exempt, non-working enrollees category, as shown in Table 1 of the analysis).
 The 106,000 figure is derived from Urban’s estimates (from Table 1 of the analysis) that 165,000 enrollees were working, of whom 64 percent worked at least 50 weeks and averaged at least 20 hours per week over the course of a year.
 Aviva Aron-Dine, Raheem Chaudhry, and Matt Broaddus, “Many Working People Could Lose Coverage Due to Medicaid Work Requirements,” Center on Budget and Policy Priorities, April 11, 2018, https://www.cbpp.org/research/health/many-working-people-could-lose-health-coverage-due-to-medicaid-work-requirements.
 The 59,000 figure is derived from Urban’s estimates (from Table 1 of the analysis) that 165,000 enrollees were working, less the 64 percent (106,000) who worked at least 50 weeks and averaged at least 20 hours per week.
 The range of 45,000 to 103,000 enrollees represents 9 to 21 percent of the 498,000 non-elderly enrollees Urban estimates would be subject to Kentucky’s new policy based on excluding those receiving Supplemental Security Income or Medicare. As noted, Kentucky’s projections indicate that 15 percent of adult enrollees will ultimately lose coverage as a result of its waiver.
 Jennifer Wagner, “Commentary: As Predicted, Eligible Arkansas Medicaid Beneficiaries Struggling to Meet Rigid Work Requirements,” Center on Budget and Policy Priorities, July 30, 2018, https://www.cbpp.org/health/commentary-as-predicted-eligible-arkansas-medicaid-beneficiaries-struggling-to-meet-rigid/.
 Other states are proposing similar policies, while Wisconsin is seeking to take away coverage from people with incomes as low as 50 percent of the poverty line for failure to pay premiums.
 Samantha Artiga, Petry Ubri, and Julia Zur, “The Effects of Premiums and Cost Sharing on Low-Income Populations: Updated Review of the Research Findings,” Kaiser Family Foundation, June 1, 2017, https://www.kff.org/medicaid/issue-brief/the-effects-of-premiums-and-cost-sharing-on-low-income-populations-updated-review-of-research-findings/.
 Other states are proposing similar restrictions. See Jennifer Wagner and Judith Solomon, “States’ Complex Medicaid Waivers Will Create Costly Bureaucracy and Harm Eligible Beneficiaries,” Center on Budget and Policy Priorities, May 23, 2018, https://www.cbpp.org/research/health/states-complex-medicaid-waivers-will-create-costly-bureaucracy-and-harm-eligible.
 The estimates are also conservative in that Urban may underestimate the share of workers failing to meet Kentucky’s 80-hours-per-month work requirement in one or more months. Due to data limitations, Urban identifies workers who were not working for part of the prior year or who worked less than 20 hours per week on average, not those whose hours fell short in some months while exceeding the threshold in others. Alternative analysis using the Survey of Income and Program Participation suggests that a higher share of workers may fail the requirement in at least one month of the year. See Aron-Dine, Chaudhry, and Broaddus, op. cit.
 See LaDonna Pavetti, “Work Requirements Don’t Work,” Center on Budget and Policy Priorities, January 10, 2018, https://www.cbpp.org/blog/work-requirements-dont-work, LaDonna Pavetti, “Work Requirements Don’t Cut Poverty, Evidence Shows,” Center on Budget and Policy Priorities, updated June 7, 2016, https://www.cbpp.org/research/poverty-and-inequality/work-requirements-dont-cut-poverty-evidence-shows, and LaDonna Pavetti, “Evidence Doesn’t Support Claims of Success of TANF Work Requirements,” Center on Budget and Policy Priorities, April 3, 2018, https://www.cbpp.org/research/family-income-support/evidence-doesnt-support-claims-of-success-of-tanf-work-requirements. See also Ed Dolan, “Do We Really Want Expanded Work Requirements in Non-Cash Welfare Programs?” Niskanen Center, July 23, 2018, https://niskanencenter.org/blog/expanded-work-requirements-in-non-cash-welfare-programs/.
 Jeffrey Grogger and Lynn A. Karoly, Welfare Reform: Effects of a Decade of Change, Harvard University Press, 2005.
 Rachel Garfield, Robin Rudowitz, and Anthony Damico, “Understanding the Intersection of Medicaid and Work,” Kaiser Family Foundation, updated January 5, 2018, https://www.kff.org/medicaid/issue-brief/understanding-the-intersection-of-medicaid-and-work/.
 Centers for Medicare & Medicaid Services letter to state Medicaid directors (18-002), January 11, 2018, https://www.medicaid.gov/federal-policy-guidance/downloads/smd18002.pdf.
 Larisa Antonisse and Rachel Garfield, “The Relationship Between Work and Health: Findings from a Literature Review,” Kaiser Family Foundation, August 2018, https://www.kff.org/medicaid/issue-brief/the-relationship-between-work-and-health-findings-from-a-literature-review/.
 Renuka Tipirneni, Susan D. Goold, and John Z. Ayanian, “Employment Status and Health Characteristics of Adults with Expanded Medicaid Coverage in Michigan,” Journal of the American Medical Association, 2017, https://jamanetwork.com/journals/jamainternalmedicine/article-abstract/2664514.
 Ohio Department of Medicaid, “Ohio Medicaid Group VIII Assessment: A Report to the Ohio General Assembly,” January 2017, http://medicaid.ohio.gov/Portals/0/Resources/Reports/Annual/Group-VIII-Assessment.pdf. See also Kara Gavin, “Medicaid Expansion Helped Enrollees Do Better at Work or in Job Searches,” University of Michigan Health Lab, June 27, 2017, http://labblog.uofmhealth.org/industry-dx/medicaid-expansion-helped-enrollees-do-better-at-work-or-job-searches.
 Bureau of Labor Statistics, Healthcare benefits: Access, participation, and take-up rates, https://www.bls.gov/ncs/ebs/benefits/2017/ownership/civilian/table09a.htm.
 Michelle Long et al., “Trends in Employer-Sponsored Insurance Offer and Coverage Rates, 1999-2014,” Kaiser Family Foundation, March 21, 2016, https://www.kff.org/private-insurance/issue-brief/trends-in-employer-sponsored-insurance-offer-and-coverage-rates-1999-2014/.
 CBPP calculations from Current Population Survey data for 2016.
 This calculation applies the increase to all adult enrollees not receiving disability assistance, even though some would be exempt. Even basing the calculation on the fifth-year effect from the most favorable of the cash assistance trials or on the average initial (first or second year) effect on employment — assumptions likely far too optimistic as applied to Medicaid work requirements, for the reasons discussed above — people leaving Medicaid for other coverage would still account for only a small minority of the total reduction in Medicaid enrollment.
 Unsurprisingly, the declines in the uninsured rate somewhat exceed the gains in Medicaid coverage, since the ACA also made individual market coverage more accessible and affordable.
 Larisa Antonisse et al., “The Effects of Medicaid Expansion Under the ACA: Updated Findings from a Literature Review,” Kaiser Family Foundation, March 28, 2018, https://www.kff.org/medicaid/issue-brief/the-effects-of-medicaid-expansion-under-the-aca-updated-findings-from-a-literature-review-march-2018/.
 See for example Pauline Leung and Alexandre Mas, “Employment Effects of the ACA Medicaid Expansions,” National Bureau of Economic Research Working Paper No. 22540, August 2016, http://www.nber.org/papers/w22540.pdf; Angshuman Gooptu et al., “Medicaid Expansion Did Not Result in Significant Employment Changes or Job Reductions in 2014,” Health Affairs, January 2016, https://www.healthaffairs.org/doi/abs/10.1377/hlthaff.2015.0747; and Robert Kaester et al., “Effects of ACA Medicaid Expansions on Health Insurance Coverage and Labor Supply,” Journal of Policy Analysis and Management, May 2017.
 Adele Schartner, Fredric Blavin, and John Holahan, “Employer-Sponsored Insurance Stable for Low-Income Workers in Medicaid Expansion States,” Health Affairs, April 2018, https://www.healthaffairs.org/doi/10.1377/hlthaff.2017.1205.
 Benjamin D. Sommers, “Loss of Health Insurance Among Non-Elderly Adults in Medicaid,” Journal of General Internal Medicine, January 2009, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2607511/.
 Bill J. Wright et al., “Raising Premiums and Other Costs for Oregon Health Plan Enrollees Drove Many to Drop Out,” Health Affairs, December 2010, https://www.healthaffairs.org/doi/abs/10.1377/hlthaff.2010.0211.
 Artiga, Ubri, and Zur, op. cit.
 See Benjamin D. Sommers et al., “Three-Year Impacts of the Affordable Care Act: Improved Medical Care and Health Among Low-Income Adults,” Health Affairs, June 2017, https://www.healthaffairs.org/doi/10.1377/hlthaff.2017.0293; Antonisse et al., op. cit., Olena Mazurenko et al., “The Effects of Medicaid Expansion Under the ACA: A Systematic Review,” Health Affairs, June 2018, https://www.healthaffairs.org/doi/10.1377/hlthaff.2017.1491, and Hefei Wen, Benjamin Druss, and Janet Cummings, “Effect of Medicaid Expansions on Health Insurance Coverage and Access to Care Among Low-Income Adults with Behavioral Health Conditions,” Health Services Research, December 2015, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4693853/.
 Benjamin D. Sommers et al., “Changes in Utilization and Health Among Low-Income Adults After Medicaid Expansion or Expanded Private Insurance,” JAMA Internal Medicine, August 8, 2016, https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2542420.
 John Holahan, Michael Karpman, and Stephen Zuckerman, “Health Care Access and Affordability Among Low- and Moderate-Income Insured and Uninsured Adults Under the Affordable Care Act,” Urban Institute Health Reform Monitoring Survey, April 21, 2016, http://hrms.urban.org/briefs/health-care-access-affordability-low-moderate-income-insured-uninsured-adults-under-ACA.html.
 Hannah Katch, “Yes, Medicaid Improves Access to Health Coverage and Care,” Center on Budget and Policy Priorities, December 15, 2017, https://www.cbpp.org/blog/yes-medicaid-improves-access-to-health-coverage-and-care.
 Luojia Hu et al., “The Effect of the Patient Protection and Affordable Care Act Medicaid Expansions on Financial Well-Being,” National Bureau of Economic Research Working Paper 22170, April 2016, http://www.nber.org/papers/w22170 and Sommers et al., “Three-Year Impacts of the Affordable Care Act,” op. cit.
 Kenneth Brevoort, Daniel Grodzicki, and Martin B. Hackman, “Medicaid and Financial Health,” National Bureau of Economic Research Working Paper 24002, November 2017, http://www.nber.org/papers/w24002.
 Aparna Soni et al., “Effect of Medicaid Expansions of 2014 on Overall and Early-Stage Cancer Diagnoses,” American Journal of Public Health, February 2018 and Andrew P. Loehrer, et al., “Association of the Affordable Care Act Medicaid Expansion with Access to and Quality of Care for Surgical Conditions,” JAMA Surgery, March 21, 2018, https://jamanetwork.com/journals/jamasurgery/article-abstract/2670459.
 See Sommers et al., “Three-Year Impacts of the Affordable Care Act,” op. cit.; Antonisse et al., op. cit., and Mazurenko et al., op. cit.
 See Benjamin D. Sommers, Sharon K. Long, and Katherine Baicker, “Changes in Mortality After Massachusetts Health Care Reform: A Quasi-experimental Study,” Annals of Internal Medicine, May 2014, http://annals.org/aim/article-abstract/1867050/changes-mortality-after-massachusetts-health-care-reform-quasi-experimental-study and Benjamin D. Sommers, Katherine Baicker, and Arnold Epstein, “Mortality and Access to Care among Adults after State Medicaid Expansions,” New England Journal of Medicine, September 2012, https://www.nejm.org/doi/full/10.1056/nejmsa1202099.