People who receive government assistance ought to do something to show they're trying to lift themselves out of poverty, rather than remaining on the taxpayer dole forever. At least, that's the general idea behind attaching work requirements to welfare programs: If recipients have a job or are engaged in some kind of job training, they might eventually earn enough to not need welfare anymore. In theory, the policy should reduce dependency on government, save taxpayer money and encourage Americans to solve their economic problems through employment.

But what happens when people fail to meet work requirements? What happens when the policy leaves someone in deep poverty? Those questions, explored in a new report this month, have special relevance as 10 states and the Trump administration move to add work requirements to Medicaid, the public health insurance program for low-income individuals and families.

The report, issued by the Center on Budget and Policy Priorities, a left-leaning think tank, looks at four years of data from Kansas, where the enactment of stricter work policies made it easer for the state to remove families from its welfare program, officially known as Temporary Assistance for Needy Families (TANF). From October 2011, shortly before the policies went into effect, to October 2016, the monthly welfare caseload dropped almost 60 percent, from 13,014 families to 5,231 families. 

Then-Kansas Gov. Sam Brownback and allies in the state legislature heralded the changes as a success and a model for the rest of the nation. (Brownback last week was narrowly confirmed by the U.S. Senate to be ambassador at large for international religious freedom.) Last year, Brownback asserted that “Americans freed from welfare don’t just find a way to survive, they thrive.”

The Center came to a different conclusion. Most of the people who lost benefits due to work sanctions were in deep poverty and were no more likely to be working than when they had government assistance. "These folks really aren't doing well," says Tazra Mitchell, a co-author of the report. 

Under Brownback, Kansas made it more difficult to receive TANF benefits without working. For example, when families failed to meet the work requirement, they used to be ineligible only until they met the work requirement again; now they lose eligibility for three months on a first offense, and progressively longer periods for additional offenses. The state also shortened the work exemption period for parents with babies. Under the previous policy, parents didn’t have to work for the first six months after their baby was born; now the exemption ends two months after the baby is born.

Such changes kept many Kansans off the program, but not out of poverty. A year after being removed from the program, three in four Kansas parents who lost TANF benefits had either no earnings or less than half of the federal poverty line. For a family of three, that translates to less than $10,080 per year. Employment rates were roughly the same whether they depended on government assistance or not: 67 percent worked in the year before they exited the program and 68 percent worked in the year after.

The report’s authors caution not to overinterpret their findings. Because they only look at what happened to people before and after a policy took effect, their research design doesn’t establish a causal link between the work requirement and people's economic circumstances. Other factors could have contributed to the low employment rates and incomes of former TANF recipients. Nonetheless, the data do appear to contradict the claim that Kansas residents thrived after the state cut off their benefits due to work sanctions.

The Center’s analysis relies on state data first analyzed by the Foundation for Government Accountability, a free-market think tank started by a former state lawmaker from Maine. In an op-ed last week, a researcher at the Foundation challenged the accuracy of the Center’s results, arguing that "the reality is that incomes went up far more than claimed." The Center, for its part, alleges that the Foundation's original study of the data was flawed, using an extrapolation rather than actual earnings to illustrate income growth over time.  

In the op-ed, Jonathan Ingram, the Foundation's vice president of research, argues that people were better off because one year and four years after being removed from TANF, families’ total earnings were higher than the year before exiting the program. (Both think tanks agree that earnings went up, though they disagree on how much and why.) More importantly, the groups disagree on the relevance of those increases, given that the families started with such low incomes. Four years after losing TANF benefits, families' median annual earnings were $2,175 and about 83 percent of the families were still living in poverty, according to the Center. 

To date, only one state, Kentucky, has received a federal waiver to try a work requirement in the Medicaid program -- and it immediately got caught up in litigation. Nine other states, including Kansas, have asked for a similar waiver. If states do move ahead with a work requirement, it may affect a relatively small portion of the Medicaid population. Nearly eight in 10 non-disabled adults with Medicaid coverage live in households where at least one person works. But among the few who do not work, surveys suggest most can't because they are sick, acting as a caretaker or going to school.