With the unemployment rate around 4 percent and people flooding back into the workforce, the question facing economic policymakers has changed: Once the cyclical recovery has fully run its course, what can be done to help into employment Americans still left behind?
The answer offered by many Republicans is to increase work requirements on low-income benefits. The House farm bill would tighten eligibility for the Supplemental Nutrition Assistance Program (SNAP), also known as food stamps. States are doing the same for Medicaid under federal waivers. But such steps would jeopardize the health and nutrition of millions of Americans, while doing little to increase work.
Of the people who would be targeted by work requirements for means-tested benefits, a large majority already are working. SNAP is like Milton Friedman's negative income tax: a near-cash benefit that is provided broadly to people with low incomes and then phased out as their finances improve. Unlike Friedman's proposed unconditional benefit, however, SNAP requires most working-age, nondisabled beneficiaries to register for work with a state agency and accept any reasonable job offer.
In fact, 74 percent of working-age SNAP recipients are employed within the year before or after they received benefits, and most of the rest have health issues or are caring for family members. Some of the SNAP statistics tossed around miss this because they focus on whether people were working in the month they received benefits.
Grading SNAP recipients based on whether they worked in a single month is particularly problematic because, as economic research by Kristin Butcher and Diane Schanzenbach has shown, the jobs available to this population are often temporary and unstable. "Working" often means involuntarily cycling into and out of employment.
Policies that fail to appreciate this can be more punitive than pro-work. Under the House plan, an able-bodied adult who spends a single month without working 20 hours a week would lose SNAP benefits for a year. For a second offense, the penalty is three years.
Medicaid work requirements have similar problems. Kentucky's plan, currently being stayed by the courts, would require beneficiaries to work 80 hours each month or lose their health coverage. Again, the narrow measurement is a problem. One-quarter of beneficiaries who work at least 1,000 hours annually—averaging to 80 hours a month—could be denied coverage for part of the year, according to the Center on Budget and Policy Priorities. What these people lack is not motivation, but stable, consistent, decently paid jobs.
Partly as a result, Kentucky itself estimates that more than 95,000 people would lose health coverage under its policy. Taking medical care away from low-wage workers in unstable jobs is not only ineffective in promoting employment, it may be counterproductive, since losing access to regular treatment can make it harder for people with serious health needs to look for and keep a job.
The best direct evidence on work requirements comes from the National Evaluation of Welfare-to-Work Strategies, a set of 11 randomized trials done during the 1990s in a half dozen cities across the country. These trials found modest increases in work over the short run, but the effect faded out by the end of five years, generally leaving the treatment group no more likely to be employed than the control group. A notable exception was Portland, Ore., which, along with the work requirement, provided intensive support to train and place people into jobs. Still, on average in all the cities, households ended up with low-paid work that did not make up for the money they lost on their benefits, leaving more people in deep poverty.
Employment rates did increase a lot after President Clinton and Congress reformed welfare in 1996. But research by Jeffrey Grogger, an economist at the University of Chicago, found that it was thanks less to work requirements than to the roaring economy and the expanded earned-income tax credit. This insight should be the basis of an alternative approach: Government can do more to ensure that work pays, that people who want employment have the support they need, and that jobs will be available for them to find.
The most promising ideas put the private sector first. Numerous proposals have been made to expand the earned-income tax credit, including for workers without dependent children who get essentially nothing today. Rep. Ro Khanna, a Democrat who represents Silicon Valley, has suggested temporary subsidies for employers who hire unemployed and disadvantaged workers. Others are proposing additional support for job training and placement, child care, paid leave—all of which can be effective in enabling people who want to work to do so.
As a last resort, providing federal jobs may have to be part of the policy mix. This is an idea advanced not only by progressives but also by Kevin Hassett, chairman of President Trump's Council of Economic Advisers. In 2013 Mr. Hassett advocated "a short-run jobs program that recruits the long-term unemployed to assist with the normal functions of government."
The largest progressive idea, a federal jobs guarantee, is not ready to become law today and will have to reckon with the increasingly problematic budget outlook. But enabling more people to work—and thereby boosting economic growth—isn't a partisan objective. Anyone who shares this goal should welcome the debate over big ideas to achieve it with work incentives and supports. In the meantime, it would be a mistake to put poor families into an even deeper hole with ineffective and damaging work requirements.