Some Safety Net Programs Depress Wages

by Matt Unrath

Are taxpayers propping up the profits of large corporations by supporting low-wage workers? Advocates and labor leaders think so, and have pointed to research produced by the UC Berkeley Center for Labor and Education to help make their case. Last year, for example, the Labor Center released a report to great attention that tallied up the cost of in-work social programs like Medicaid, the Supplemental Nutrition Assistance Program (SNAP) and the Earned Income Tax Credit (EITC).

While the authors are careful not to make the argument explicitly, the report implies that, by propping up paychecks, the taxpaying public is subsidizing the profits of low-wage workers’ employers, like fast-food chains and big box retail stores. That was how The New York Times covered the report (also eliciting an editorial in The Times that made the argument even more directly). The report has provided an effective talking point for ongoing campaigns to raise wages and labor standards for workers at these companies.

The research has raised a valuable question: is the welfare system a subsidy for low-wage companies like McDonald’s and Wal-Mart? Some economists are skeptical. For a program to qualify as a subsidy, they point out, it must have the effect of encouraging people to work more, thus reducing the wages that firms need to offer to attract workers. Indeed, economic theory suggests that the opposite is more probable: public assistance could incentivize a worker to reduce their labor. Employers would need to raise wages to incentivize folks not to stay home.

UMass Amherst economist Arin Dube summed up the debate by simply noting that different programs have different labor supply effects.  The EITC,  a standard example of a “safety net” program, increases labor supply and reduces wages.  However, Dube argues that SNAP and housing vouchers do not necessarily encourage workers to work more — meaning it would be hard to subsidize businesses this way. Some concluded that he had said that the safety net does not have any positive effect on labor supply; and therefore no negative effect on wages.

Unfortunately, the debate’s conclusion misconstrues Dube’s analysis and ignores some important empirical analyses. Large parts of the current safety net certainly subsidize low-wage employers, but much of the rest of the effects of these subsidies remain, at the very least, unclear.

Let’s first consider the Earned Income Tax Credit. Georgetown economist Nada Eissa and UC Berkeley economist Hilary Hoynes find that the EITC does incentivize eligible workers to enter the workforce and work more hours – though the increased income might encourage some to work slightly less. This increase in labor supply results in firms’ ability to lower the wages needed to recruit workers. UC Berkeley economist Jesse Rothstein estimates that for every additional dollar spent on the EITC, employers would receive $0.72. Since employers are able to lower their wages for everyone, workers ineligible for the EITC actually see their total income fall by $0.43 for every dollar we spend on the EITC. According to Rothstein’s calculations, the $60 billion the federal government spends on the EITC annually lowers collective wages by about $24 billion.

Now, critics of the Labor Center report do acknowledge the EITC is a program that plausibly functions as a subsidy to employers. But the EITC accounts for about a full third of the spending tallied in the Labor Center report. An estimated 70 percent of EITC spending – nearly $38 billion – may be captured by low-wage employers. At the outset, acknowledging this effect of the EITC must mean acknowledging that low wages can come at some public cost.

The most-costly programs are Medicaid and the Children’s Health Insurance Program (CHIP). The Labor Center report estimates that these two programs provide about $69.2 billion in insurance coverage to 34 million individuals in working families. These programs are traditional out-of-work benefits: work is not required. The programs should therefore decrease labor supply and should not act as subsidies to employers.

Some evidence, however, suggests that Medicaid can crowd out private health insurance coverage. Significant expansions of Medicaid or CHIP, for example, might induce employers to rescind private health insurance coverage to workers, believing they will now be served by public insurance. If firms do not replace the equivalent value of that health insurance via wages one could make the argument that Medicaid serves as a subsidy to these employers. Though not definitive, the well-studied expansion of Medicaid in Oregon had no effect on employment or wages for recipients, meaning no evidence of upward wage pressure.  Furthermore, initial analyses of the impact of other expansions of Medicaid also reveal little effect on labor force participation, employment and earnings. A study of a similar expansion in Tennessee, previous to the ACA, however, did find substantial reductions in labor supply.

The two other programs evaluated in the Labor Center report are SNAP and TANF. Generally, recipients of both of these forms of support are eligible for support without working, suggesting that recipients would not necessarily increase their labor supply and so have the effect of lowering wages. Dube directly addresses food stamps in his post, pointing to a study by Hoynes and Schanzenbach that concluded a decrease in labor supply due to the initial roll out of food stamps in the 1970s. What this argument overlooks, however, is the strict work requirements for TANF and the popular advocacy for adding similar requirements to SNAP. At minimum, 17 states already require that single adults receiving SNAP demonstrate they are enrolled in training, have a job or are demonstrably seeking one within three months.

The extent to which these benefits induce some individuals to seek employment and disincentivize others might be an empirical question – and likely dependent on the eligibility requirements imposed on the different classes of recipients. Still, Dube’s and others’ argument that these programs “are not tied to work” is an oversimplification. Relative to the base-case scenario in which these programs are true out-of-work benefits, the imposition of work and training requirements on these programs will encourage labor supply.

The Labor Center report also does not account for, due to data limitations, the public costs of child care subsidies, job training, mass transit and parking subsidies, government-sponsored matched savings programs for retirement savings, and other work-dependent income support programs. The cost of these programs may be dwarfed by the size of the EITC and Medicaid, but they are still publicly supported and work-dependent. The fact that these are necessarily tied to work would qualify them as programs that would theoretically increase labor supply and function as subsidies to employers.

In summary, the EITC and several other work-related safety net programs likely lower wages and so subsidize firms. It is possible Medicaid and CHIP have similar effects. SNAP and TANF should not increase labor supply for most recipients. But the safety net programs that are directly tied to work would assumedly reduce wages. The claim that the safety net does not reduce wages is at best unclear, but likely wrong. An honest policy discussion must weigh the benefits and costs of proposals to increase labor standards and support work through public assistance. A complete rejection of the wage-depressing effects of some parts of the safety net does a disservice to that debate.

Matt Unrath is an MPP Candidate at the Goldman School of Public Policy and is currently a researcher at the UC Berkeley Center for Labor Research and Education.