The debate is no longer academic over whether work should be required for “able-bodied” adults seeking Medicaid coverage. The Trump administration approved three state Section 1115 demonstrations (also called waivers) to test Medicaid work requirements; eight additional states have submitted similar applications. On April 10, President Trump issued an executive order to agencies to develop work requirements for other benefits, such as low-income housing and food stamps. The administration’s immigration policy favors people who are “skilled, who want to work, who will contribute to our society.” As such, Medicaid work requirements fit within a broader vision of government benefits that are merit-based rather than universally available to eligible people.

Denying public benefits due to failure to work is hardly a new idea. It was evident in the mid-twentieth century precursors to Medicaid, which restricted eligibility to people who were aged, blind, or mothers (usually unwed) and their dependent children; non-disabled, low-income men were presumed to be able to work and thus “not worthy” of health coverage. Explicitly linking benefits to work requirements surfaced in the 1980s and 1990s, and was adopted in the 1996 welfare reform legislation. While, some states, starting in that period, sought Medicaid waivers to test work requirements, previous administrations rejected them because, as one official explained, they “could undermine access, efficiency, and quality of care provided to Medicaid beneficiaries” (which has become part of the basis for a legal challenge to the Trump administration’s policy).

The Affordable Care Act (ACA) intensified interest in Medicaid work requirements. The law aims to make health care a right, not a privilege; to do so, it expanded access to and financial assistance for coverage, along with a requirement that people who can afford coverage buy it (the “individual mandate,” whose penalty was eliminated in the tax cut legislation). The ACA has expanded Medicaid in thirty-two states and the District of Columbia to low-income adults. Since many in this newly eligible group could work, some argue, they must work in return for this generous coverage. While the ACA’s premium tax credits reduce the costs for people with income below 400 percent of poverty (about $83,000 for a family of three), middle-income people without this help have faced high and rising premiums—partly due to actions by the Trump administration. Concerns about affordability have led to questions of fairness: why should a person at a low-paying job be required to pay a large amount for health coverage while a person who is not working gets it for free?

In addition to values, debate has swirled around the consequences of Medicaid work requirements: will they help or hurt eligible people? Administration guidance on these demonstrations argues that work requirements will increase employment and thus health status. Experts agree on the correlation between work and health (people who work are healthier than those who do not), but not necessarily on causation—that work improves health. Moreover, evidence is slim on the efficacy of conditioning the receipt of benefits on work. Experience in other programs suggest that it has modest, and merely temporary, effects on employment.

Some argue that the administration has it backwards: rather than work improving health, they say, health and health coverage improve work. The majority of Medicaid enrollees in two states reported that having coverage made it easier to continue working and helped them do their job better. About one-third of working-age people on Medicaid have a mental illness or an addiction. For them, treatment covered by Medicaid may be an essential support for their seeking and keeping a job. Making health coverage a reward for work instead of a prerequisite may trap people in illness and poverty. The administration’s guidance excludes people with certain types of disabilities and other vulnerable groups from work requirements and offers alternatives to work like community service; the efficacy of such policies is as yet unknown.

Irrespective of its impact on work, the effect of Medicaid work requirements on enrollment is undisputed. State waiver applications assume savings from reduced coverage. Consistent with this, the public believes that the policy is motivated by a desire to reduce government spending (41 percent) rather than helping to lift people out of poverty (33 percent). People losing eligibility for Medicaid are likely to become uninsured, since few qualify for job-based insurance or financial assistance in the ACA’s Health Insurance Marketplace.

Which leads to the question: is coverage important to health, irrespective of whether a person works? Researchers have answered this question. Starting in 2001, the Institute of Medicine commissioned a six-volume series reviewing the evidence on the impact of coverage. It concluded that uninsured people get less—and less timely—health care, have worse health outcomes, and are at greater risk of death than insured people. Health reform in Massachusetts contributed to fewer avoidable deaths. And, among others, a study found that the ACA’s Medicaid expansion significantly increased access to care (e.g., primary care and regular care for chronic conditions). By far, more and better evidence supports the claim that coverage improves health than the claim that work improves health.

The concerns motivating these waivers could be addressed more effectively in other ways. Work incentive programs, including paying cash to people to work, may yield a higher return on investment than a Medicaid work requirement, which has a high cost of implementation and operation. On the subject of fairness, Medicaid enrollees could be asked to pay—rather than work—for their health benefits: a number of states charge near-poor Medicaid enrollees premiums comparable (as a percent of income) to those paid by Marketplace enrollees. Better yet, policy makers could close the perceived fairness gap by reducing premiums for middle-income people. Congress could eliminate the arbitrary upper-income limit on premium tax credits for Marketplace coverage. States could block the Trump administration’s expansion of unregulated “short-term” plans, which would significantly lower individual-market premiums.

But if the real goal of Medicaid work requirements is to generate savings by reducing enrollment, evidence shows these demonstrations are putting people’s health at risk.