What are work requirements good for?

Stretching back to the establishment of welfare in the United States, politicians have debated both the practical and moral utility of requiring people to work in order to receive government benefits. Since welfare reform in the 1990s gave states wide latitude to create work requirements in the Temporary Assistance for Needy Families cash-assistance program, Republicans have hankered for a chance to extend those requirements to other safety-net programs, as part of their push to “require everyone who can to work.”

The purpose of work requirements in welfare, according to the Congressional Research Service, is “to offset work disincentives in social assistance programs, promote a culture of work over dependency, and prioritize governmental resources,” in addition to helping lift people out of poverty.

Now, Republicans could take one step closer to having that chance, thanks to a last-minute manager’s amendment that was attached to the GOP’s Obamacare repeal-and-replace bill by House Speaker Paul Ryan. In a fashion similar to TANF, the amendment would give states authority to mandate that “non-disabled, non-elderly, [and] non-pregnant” individuals enrolled in Medicaid engage in some amount of hours of “work activities.”

The measure was put forth to appease party concerns about the legislation, before Thursday’s anticipated showdown between the Freedom Caucus and the bill’s backers on the House floor. But beyond the political motives for the Medicaid work requirement, the provision itself requires scrutiny. Do work requirements even do what Paul Ryan wants them to do, and would they work for Medicaid?

The history of TANF doesn’t entirely conform to Ryan’s assertions that work requirements spur more people to find jobs or the larger goal of “promot[ing] a culture of work over dependency.” Most research does indicate that after the passage of the 1996 welfare-reform bill—the Personal Responsibility and Work Opportunity Reconciliation Act—the number of people receiving TANF plummeted, employment among recipients increased, and poverty rates did not increase on account of the requirement. But deeper analysis finds that the employment gains were ephemeral, inconsistent, and have often been questionably attributed to “welfare reform.”

For starters, research from LaDonna Pavetti of the liberal Center on Budget and Policy Priorities indicates that while TANF’s work requirements did translate over the first two years to increased employment or work activity—including volunteer work, community service, and some limited combinations of training, work, and job-seeking—those gains diminished after five years. “You see an initial increase in work, but then people who are not subject to work requirements, their employment rates actually come together,” said Pavetti, who is vice president for family income-support policy at the center. In other words, people who were on TANF and not subject to work requirements saw their employment rates rise in the long-term, even as the people with work requirements lost jobs.

The initial spike from work requirements probably followed proponents’ plan: In the face of pressure—and sometimes aided by state-sponsored training and placement programs—people sought jobs. But those jobs were often low-paying, meaning they could not lift people from poverty as desired, and the newly employed workers were very soon sloughed from employment rolls anyway.

These losses occurred even as employment among people exempt from TANF’s requirements increased over time. That indicates that there were other economic factors that helped sustain long-term employment more than work requirements—including a strong Bill Clinton-era job market and major federal investments in the safety net via Medicaid, the Children’s Health Insurance Program, and child-care grants. The few work requirements that did have some enduring, robust effects, like those in Portland, Oregon, and some communities in California, were coupled with strong training, subsidized work, and placement programs that invested significant time and resources into securing just those results.

Pavetti explained what she described as the central failure of work requirements. “Getting people into the labor system quickly doesn’t necessarily have long-term benefits,” she said. And the reasons are pretty obvious. Low-paying and part-time jobs are usually unstable to begin with, and the vast majority of low-income people who can secure steady employment, including those on Medicaid, already do work.

The remainder of people subject to TANF’s work requirements tend to have major barriers in their lives to employment—including single mothers with young or disabled children, people with mental-health or addiction issues, those who act as primary caregivers to older relatives, or those with functional physical or mental disabilities that don’t meet thresholds to qualify for supplemental insurance. Research from Harvard Medical School researchers in 2006 suggests that at least some of the TANF requirements’ success at shrinking welfare rolls came from simply pushing people already unable to work out of the program and into an even lower rung of poverty.

The only indisputable national success of the TANF work requirement—its reduction of people on welfare rolls—came at a significant cost to millions of people. Today, often because of their lack of access to cash assistance, millions of families and children live in deep poverty, or around $2 a day per person or less, according to researchers Kathryn Edin of Johns Hopkins and H. Luke Shaefer at the University of Michigan. As my colleague Alana Semuels reported last year on Wisconsin’s TANF program, job training and meaningful opportunities for advancement are rare among the program’s recipients, let alone those who become unenrolled from lack of work. In essence, some of the successes attributed to welfare reform in reducing enrollment and costs came from simply making the program less and less accessible for those who fall behind, not from helping people find jobs or lifting them from poverty.

Reports on the application of a TANF-like work requirement for Medicaid forecast some similar effects, but also identified a host of new potential complications given the program’s size, diversity, and complexity. First, research from Hannah Katch, an analyst also at CBPP, indicates that the effect on jobs would be just as inconclusive—or perhaps more so—as the effect with TANF. The economy is much less healthy than it was during the Clinton heyday, and the number of Medicaid enrollees is many times greater than that of TANF. There isn’t a guarantee that the economy could even create long-term stable jobs for the population of Medicaid enrollees, and Medicaid covers people with the kinds of medical issues and disruptions that create job instability in the first place.

Additionally, the AHCA manager’s amendment doesn’t fund any training or placement programs, not to mention the kinds of rigorous initiatives found in the successful TANF work programs in Oregon and California. Without that funding, states may have little incentive to create those programs. Katch believes “there’s actually a disincentive, since the bill itself is reducing overall funding for Medicaid.”

States could hypothetically impose work requirements in a (rather transparent) attempt to cover fewer people, thereby reducing their costs. Also, requiring states to determine who’s eligible for Medicaid based on work status would increase administrative expenses and tax administrative systems that are already overloaded, which could lead to more than a few people losing eligibility despite actually qualifying via administrative error and oversight.

Medicaid enrollees are even more vulnerable to any manipulation by states than TANF recipients are. That’s especially true in states that have expanded Medicaid eligibility to able-bodied low-income people under the Affordable Care Act. Enrollees there are much more likely to be poor men, and thus much more likely to be convicted felons than TANF beneficiaries. That means they are also likely to be barred from many likely employment options; work requirements will only hurt their prospects of ever making it out of poverty.

Individuals who qualify for Medicaid because of a disability, or who receive federal disability payments, are exempted from the AHCA work requirement. But that doesn’t cover every worker with disabilities: The Kaiser Family Foundation reports that many adults covered as “able bodied” under Medicaid nevertheless report significant impairments or disabilities and can’t work. They may be waiting for decisions from the federal government on their disability status or may have a disability that interferes with their line of work but doesn’t qualify for insurance—for example, computer users with severe, chronic migraines triggered by computer use.

Beyond the TANF comparisons, Katch points out that one component in particular makes Medicaid a potentially fraught target for work requirements: Medicaid is itself already a work-support program for low-income people. One of the basic underlying premises of providing public health insurance to is that keeping people healthy, able, and out of the hospital is a pretty good way of ensuring that they show up to work.  The bottom line is that people need to be healthy to work, and people who aren’t healthy can’t.

“Because Medicaid works as a work support, [a work requirement] would be a bit of a vicious cycle,” Katch told me. She outlined a scenario where a hypothetical poor coal miner in West Virginia suffers from two problems rather common to those in his occupation and region—back problems and associated opioid use. He fails to secure work thanks to the back troubles, the addiction, or lack of treatment for either. If he were subject to a work requirement, said coal miner might be kicked off the very program—Medicaid—that pays for treatment that might allow him to re-enter the workforce. That could lead to a downward spiral of sickness and poverty. Medicaid, like the rest of the American safety net, is intended as a failsafe for those not able to find work, and taking it away for those who cannot work subverts its purpose.

Again, most people on Medicaid who can work do, and low-income people working demanding jobs often do so “until their bodies gave out on them,” Pavetti said. There isn’t much evidence that penalizing the Medicaid population for not working will improve the program, its outcomes, or job creation. Indeed, the only real outcome of a Medicaid work requirement is that fewer people will have access to Medicaid, which may be the point.