The economist David Neumark, along with two coauthors, has published a detailed study of how anti-poverty policies affect disadvantaged neighborhoods over long periods of time. The results, released under the aegis of the Employment Policies Institute — that would be the right-leaning EPI, as opposed to the left-leaning Economic Policy Institute — are not good.
Here’s how the researchers themselves put it:
We find evidence that higher minimum wages lead, in the longer run, to increases in poverty and the share of families on public assistance. We find some evidence that the EITC [earned income tax credit, a benefit available to the working poor] has positive longer-run employment effects. We do not generally find significant evidence of longer-run effects of the EITC on poverty or public assistance [except in certain alternative analyses]. . . . Finally, we find evidence that more generous welfare benefits lead to higher poverty and public assistance in the longer-run. Perhaps the most robust important conclusion is that a higher minimum wage and more generous welfare benefits do not reduce poverty and reliance on public assistance in the longer-term.
Other interesting tidbits abound. There’s “some hint that, at least for disadvantaged black neighborhoods, welfare reform [in 1996] may have eliminated the adverse longer-run effects of more generous welfare benefits on poverty.” And oddly enough, minimum-wage hikes correlate with increased employment in the short-term, contradicting previous research that focused on individuals rather than neighborhoods.
As always, the devil is in the details, so here’s a brief dive into the guts of the study. As the authors themselves discuss, there are important limitations and some quirks in the results that justify skepticism. But it’s an important contribution to the debate over the long-term effects of our efforts to fight poverty.
We often think about these policies as federal matters. But since federal policy affects the entire country at once, its effects are hard to distinguish from other trends. So Neumark et al. focus mainly on state policy changes, looking to see what happened in states that went above and beyond what the feds require.
More specifically, they studied disadvantaged areas in states where policies changed, accounting for trends in two different control groups: disadvantaged tracts in other states, and non-disadvantaged tracts in the same states. (These latter tracts shouldn’t be affected by the anti-poverty policies much — as few poor people live there — yet will experience other local economic trends in the states that enacted the policies.) This is a powerful research design, called a “triple difference” approach.
The authors further account for the fact that some national economic trends, such as deindustrialization, affected some census tracts more than others. And they look at two different groups of areas they consider “disadvantaged” as of 1970 — those with low levels of education, and those that were heavily African-American — tracking subsequent developments there through 2010.
But even a carefully designed study like this doesn’t magically turn correlation into causation; in theory, some confounding factor might be driving the results. For one thing, the study does not account for migration — the possibility that people move into and out of these neighborhoods as policies change, rather than becoming better or worse off along with the neighborhoods themselves. Neumark tells me via email, however, that the team has continued its work since submitting the study, and that “controlling for demographic change and for migration doesn’t affect the results.”
And then there are those unusual findings to contend with, including that the EITC reduces employment in the short-term (despite the fact that it subsidizes employment) and increases poverty in the short-term as well. The authors investigate these anomalies in a rather technical fashion and concede that “caution” is warranted when assuming their study has proven causation.
At the very least, though, a well-designed study from a prominent economist has failed to find long-term benefits from minimum-wage laws and welfare, and indeed suggests these policies might cause long-term harm to poor neighborhoods. That’s a disturbing possibility.