Work requirements work, but don’t ask MECEP for a fair analysis


Last month, the Maine Center for Economic Policy posted an article on its website chastising Governor LePage and other Republican lawmakers for having the audacity to suggest that able-bodied adults who receive Medicaid benefits should be required to work at least 80 hours every month, or an average of 20 hours per week.

The author of the piece, James Myall, cites research conducted by the Washington, DC-based Center on Budget and Policy Priorities (CBPP) that used large federal datasets to calculate how many workers in the United States would currently fail to meet the minimum of 80 work hours per month for at least one month per year and therefore could lose their Medicaid benefits. The CBPP’s analysis suggests that 9.5 million workers nationwide would fall into this category.

Myall confidently claims that this evidence “demonstrates that even hardworking Mainers would be harmed by the kind of work requirement sought by the LePage administration.”

Since CBPP didn’t compute state-specific figures, it’s unclear how many workers in Maine would potentially be affected, and Myall doesn’t attempt to offer a precise estimate.

More importantly, Myall simply assumes that CBPP’s static analysis accurately reflects what we can expect if Medicaid work requirements are implemented. But that’s silly. It’s like saying that millions of young adults who live with their parents rent-free would be homeless if their parents demanded they pay rent. Why? Because they don’t pay rent today. But most young adults living with their parents would probably find a way to pay their parents rent if the alternative was to find other housing arrangements or become homeless. The fact that few of those young adults pay rent today provides very little information (almost none, really) about how they would respond to a “rent requirement” in the future.

That’s the whole point, of course: to use contingencies to change people’s behavior in a positive way. And rigorous research supports the common-sense notion that work requirements encourage labor force participation.

In 2006, in reference to the historic 1996 welfare reforms that imposed work requirements on welfare recipients similar to those contemplated by the LePage administration, Ron Haskins of the Brookings Institution remarked that “gold-standard studies almost uniformly show reductions in [welfare] caseloads and increases in employment attributable to work requirements.” Haskins also recommended that Congress implement work requirements in other means-tested government programs, saying, “If work requirements are successful in one program, I can see no reason why they cannot be successful in other programs.”

Myall’s argument is also undercut by evidence from Tennessee, which withdrew Medicaid coverage for 170,000 of its residents in 2005. Researchers tracked the job-search and work activities of these individuals to see how they reacted to losing public health insurance. Studies found an “immediate increase in job search behavior and a steady rise in…employment.” The authors suggested that the desire to secure health insurance coverage was a major factor driving these outcomes. It’s clear that people highly value medical coverage and are motivated to enter the workforce in order to maintain that coverage. This suggests that Maine’s work requirements could be highly successful in promoting work among Medicaid enrollees.

The bottom-line is that humans are generally fairly good at maximizing their well-being under outside constraints, and it’s ludicrous to suggest that workers would not adjust their behavior to remain eligible for a government program as valuable as Medicaid (which spends nearly $4,000 per adult recipient, on average).

Unfortunately, this constitutes at least the second recent example of Myall’s tenuous grasp of research methods (the first being his distortion of employment and wage data to argue that Maine’s minimum wage increase is boosting low-income workers’ wages without inflicting any economic harm). MECEP’s lack of methodological rigor should concern those who rely on it to provide fact-based analysis.