Work Requirements and the Data

One of the few major pieces of legislation moving this summer is the farm bill. Versions of the farm bill have passed both the House and the Senate, and a conference committee will begin the process of reconciling their differences shortly. Among the most striking and contentious differences are the House reforms to the Supplemental Nutrition Assistance Program (SNAP, or “food stamps”) that include work requirements. The Senate bill contains no work requirements.

There are long-standing work requirements for cash welfare (named Temporary Assistance for Needy Families, or TANF). One of the features of cash assistance is that it is cut back as incomes rise. As a result, more work and a bigger paycheck is offset in part by reduced cash assistance from the government. These “implicit taxes” can be quite high — much higher than the top income tax rate — and deter work. Work requirements are one way to offset these damaging incentives and are complementary to economic self-sufficiency, which is the only permanent route out of poverty.

Nevertheless, discussions of work requirements are fraught with emotion and controversy, and often uninformed by any sense of the magnitudes involved. From that perspective, I was interested in the report released yesterday by the Council of Economic Advisers (CEA) on the topic of “Expanding Work Requirements in Non-Cash Welfare Programs.” Three key insights jump right out.

First, the majority of recipients of non-cash welfare appear capable of working. In Medicaid, 61 percent of recipients are working age (i.e. between 18 and 65) and not disabled. The corresponding fractions for SNAP and housing assistance are 67 and 59 percent, respectively.

Second, of those in these three programs who are capable of working, between 50 and 60 percent do not work at all and only a bit over 20 percent work full time.

Third, work participation is markedly different among those not on the non-cash aid programs. In the non-disabled working-age population (those between ages 18 and 65) that is not receiving aid, 77 percent worked and 57 percent worked full time. Indeed, even among those with children, the rate of full-time work was nearly 60 percent.

(There is one major caveat to all of these figures. Due to data limitations, the CEA used datafrom December 2013. Doubtless the specific rates would change as the economy recovered, but the patterns are likely similar.)

As a matter of arithmetic, work is less prevalent among those on aid programs — they would not qualify if they were working. But the data suggest that there is considerable opportunity to raise the work activity of aid recipients. That leaves one with at least two interesting value judgments to consider. First, is it appropriate to require work effort (or schooling or training) as part of the social contract that supports the safety net? Second, is it appropriate to adopt a paternalistic policy of “making” people work because it will be good for them to have work skills and experience? Or, should society respect the individuals’ decisions not to work?

The debate continues.