Sara Rosenbaum, Rebecca Morris, Erin Brantley, and Leighton Ku
February 8, 2021
On March 29th the United States Supreme Court brief will hear oral arguments in Azar v Gresham and Philbrick v Gresham, which focus on the legality of using § 1115 of the Social Security Act to conduct Medicaid work experiments, under which work or other “community engagement” activities are compelled as a condition of eligibility. Section 1115 is a research statute. Therefore, for the HHS Secretary to act within the permissible bounds of the law, the agency’s action be reasonable in light of the law’s experimental purpose; indeed, 1115 is not simply license to run a Medicaid program that alters requirements that may clash with political preferences. Like other experiments, those conducted under 1115 must test hypotheses that rest on reasonable underlying assumptions and reliable evidence.
A central question is whether stated hypotheses were reasonable in the case of 1115 experiments to test the impact of compelled work as a condition of eligibility, given the evidence on which they rest. Other key elements of a reasonable experiment would be an assessment of whether the possible benefits outweigh the risks, a question that goes to the heart of whether an experiment “promotes the objectives” of the Medicaid program. As with experiments generally, under the express terms of the statute, an 1115 experiment also must include an evaluation to reliably measure its effects. Sound experimentation itself thus provides the basis on which actions become reasonable and lawful under 1115.
Viewed through a research lens, compelled work experiments are unreasonable.
Research aims and hypotheses. The Trump administration offered various research justifications for its experiment. The original 2018 State Medicaid Directors Letter inviting work experiments rested on a central working assumption: work improves health https://www.medicaid.gov/federal-policy-guidance/downloads/smd18002.pdf. This assumption in turn served as the basis for the administration’s hypothesis that compelled work would improve health. Later in its brief to the Court, however (filed one day before President Trump left office) the Justice Department argued instead https://www.justice.gov/sites/default/files/briefs/2021/01/25/20-37tsunitedstates.pdf that the experiment rested on the hypothesis that compelled work promotes access to private insurance, thereby saving state resources. Here, of course, underlying assumptions would be that work requirements would increase employment and that the jobs held by people subject to compelled work requirements would provide sufficient income and workplace benefits to provide access to commercial insurance. In granting certiorari, the Court has asked for the briefs to address whether the “court of appeals erred in concluding that the Secretary may not authorize demonstration projects to test requirements that are designed to promote the provision of health-care coverage by means of facilitating the transition of Medicaid beneficiaries to commercial coverage and improving their health.”
Thus, a key question is whether evidence supports an assumption that compelled work experiments will produce access to commercial insurance to any significant degree. The best available evidence showed that work requirements fail to increase employment in a sustained or meaningful way. Even if Medicaid work requirements were successful in causing an increase in employment for some low-income people, a basic understanding of the private insurance market would indicate that these few individuals would still be unlikely to gain private health insurance.
Moreover, it is important to understand that Medicaid and commercial insurance are not mutually exclusive. Just as people enrolled in Medicare also can be enrolled in Medicaid as dual enrollees, low-income workers may have both Medicaid and employer-based insurance, if their employers offer insurance. In that case, Medicaid serves as secondary insurance to the private coverage. In addition, recent research has demonstrated that the Medicaid expansions did not discourage unemployed workers from finding or keeping jobs (https://www.nber.org/papers/w26553). A larger problem, which was a fundamental rationale for Medicaid expansion, was that too many low-income working adults were unable to get private insurance. The Administration’s work requirements projects did nothing to try to improve that problem or to stimulate employers to offer insurance to their low-wage workers.
Pathways to commercial insurance
There are essentially two pathways to commercial insurance in a post-ACA world: employer-sponsored coverage; and individual policies purchased through the health insurance marketplace.
Marketplace coverage. In order to qualify for a subsidized marketplace plan, an individual would need to project annual income above the marketplace subsidy threshold (138 percent of the federal poverty level in ACA expansion states and above 100 percent of poverty in non-expansion states). These thresholds translate into $30,305 and $21,960, respectively, for a family of three. Either threshold is far higher than part time work (per the CMS requirement) at low wages; furthermore, evidence from past work requirement experiments https://www.urban.org/sites/default/files/publication/98086/work_requirements_in_safety_net_programs.pdf suggests that people transitioning to work from means tested public benefits (e.g., cash welfare, SNAP) usually work in low-wage jobs with only limited change in income.
Workplace benefits. The other source of commercial insurance of course is workplace benefits. As a general matter, low wage employers are significantly less likely to offer job coverage. http://files.kff.org/attachment/Report-Employer-Health-Benefits-Annual-Survey-2017 In addition, few firms offer health insurance to part-time workers: only 12% of small firms and 31% of large firms did so in 2017. Therefore, as illustrated by the accompanying table on employer coverage in states with approved or pending Medicaid compulsory work experiments, low wage work, especially part-time work, typically produces at best limited access to employer coverage, either because coverage may not be offered at all to low wage workers or because, if it is, they cannot afford to purchase it. (Lower-wage workers enrolled in employer coverage pay more in premiums than better-paid workers, despite having fewer resources.) For example, in Arkansas, the first state to implement compulsory work, a part time low wage worker in 2017 had a slightly-better-than 1 in 6 chance (16%) of having employer coverage. In New Hampshire, despite a strong economy and virtually full employment at the time the work experiment was approved, a part time low wage worker had a 1 in 4 chance of having employer insurance (26%). For full-time, low wage workers, the figures rise somewhat in New Hampshire to 1 in 3 (33%) but actually fall to 1 in 8 in Arkansas (13%).
Furthermore, even these low figures may be misleadingly high. These data do not indicate whether workers’ coverage was derived through their status as dependents on another family member’s employer plan as opposed to primary participant coverage under a plan sponsored by their own employer.
The Trump administration offered various justifications for pursuing experiments that compel community engagement (defined as part time work, either paid or unpaid, serving as an unpaid caretaker, or attending school) as a Medicaid condition of eligibility. The administration justified the experiment to the Court on the ground that it would yield access to commercial insurance. Yet several types of CMS’ permissible community engagement activities – volunteer work, going to school, being a caretaker – offer virtually no pathway either to sufficient income to purchase marketplace coverage or an employer-sponsored plan. Furthermore, part-time paid work is unlikely to produce either workplace coverage or sufficient income to purchase a subsidized employer plan. Even a cursory examination of the evidence surrounding who has access to commercial insurance and how commercial insurance markets work demonstrates the irrationality of the hypothesis the administration purported to test, raising the fundamental question of whether the experiments were grounded in evidence and reasonable theory to begin with, or simply permission to states to run a Medicaid program more to the administration’s political liking.
Table on the % Workers Aged 19-64 Using Employer-Based Health Insurance by Full-Time and Poverty Status is available here.