Bryan Olin Dozier/NurPhoto via AP
The U.S. has probably the most complicated tax code and welfare state of any country, and too many obstacles keep some of our poorest citizens from benefiting.
A recent splashy New York Times piece by Jason DeParle, based on a study using data from Columbia University, contained an eye-popping conclusion: American child poverty has fallen by nearly three-fifths over the past quarter-century, from 28 percent in 1993 to 11 percent in 2019. The principal mechanism was expansions of the Earned Income Tax Credit (EITC) and Child Tax Credit (CTC), which provide money to some of the working poor. Sounds pretty good!
Unfortunately, these figures are simply inaccurate. As Matt Bruenig explains at the People’s Policy Project, the study assumes perfect uptake of these credits through a tax simulation, and we know for a fact that this isn’t true. Another study using IRS administrative data found that this assumption overstates the actual anti–child poverty effect of the EITC by 67 percent.
It demonstrates the downside of the American addiction to stuffing every welfare policy into these clunky tax credits.
The EITC in particular exemplifies all the pathologies of America’s default mode of policy. The first is the work requirement: The credit “phases in” starting at your first dollar of labor income, meaning if you don’t work that much, you get very little, and if you don’t work at all, you get nothing. That means for the very poorest people in the country—typically single mothers who only have access to jobs that don’t pay enough for child care—it is no help at all. And contrary to neoliberal notions of poor people being largely able-bodied adults who refuse to look for jobs, about 86 percent of people who don’t work are children, students, the elderly, or disabled—that is, people who either can’t work, or shouldn’t be working.
The second pathology is complexity and concomitant administrative burden. The EITC has a different phase-in schedule depending on whether you are single or married, or whether you have zero, one, two, or three or more children—eight different calculations may be required of applicants. That adds a large bureaucratic headache for both low-income tax filers (who often struggle to fill out complicated forms) and the IRS. That obstacle in turn prevents about 22 percent of eligible people from actually getting the benefit, and fuels a purely parasitic sector of tax prep firms, which about 60 percent of EITC recipients use. As Bruenig points out, the resulting fees eat up something like 13 to 22 percent of the average EITC benefit.
The third pathology is timing. People spend money on immediate needs, but the EITC pays out only during tax filing season in the following year. That means it will both pay money to people who recently got a better job and are no longer in poverty, and fail to pay people who recently lost theirs and have fallen into poverty.
Bruenig estimates in another paper that once you account for all these problems, the overall anti-poverty effect of the EITC is reduced by 47 percent.
The supposed justification for this Rube Goldberg machine is that it will encourage work. There’s just one problem: It doesn’t. Surveys indicate that only a small majority of low-income people are aware of the EITC, and very few of them understand that you have to work to be eligible. Another study carefully examining instances of EITC expansion found they had no effect.
The supposed justification for this Rube Goldberg machine is that it will encourage work.
As an aside, even if the EITC did incentivize work, it’s a profoundly silly way to do it. Far better to simply run the economy hot to create a labor shortage, and thus motivate employers to entice people into jobs with good pay and benefits. The super-low unemployment rate of the current recovery has seen all manner of disadvantaged demographics being gradually drawn into the labor force.
This kind of problem is all over the American state. The EITC is probably in a league of its own when it comes to stupid design, but similar problems of complexity and unfairness plague the Child Tax Credit, retirement and college savings tax credits, the mortgage interest deduction, and so on. Biden’s student loan forgiveness plan is likely going to require burdensome paperwork thanks to its means test. Even when Democrats set up a child allowance–style payment to all families in the American Rescue Plan (alas, since killed by Joe Manchin), they compulsively structured it as a tax credit, which created massive technical problems and drastically reduced uptake among the poorest.
The point of the welfare state is to put money into the hands of people who need it—families with children, the disabled and retired, the unemployed, and so on. It would be far better to follow the social democratic example and simply set up an agency to cut checks to each of those groups automatically, and make the programs fair on the back end through progressive taxation. With a Finland-style welfare state, we could eliminate poverty almost entirely, reduce inequality, and smooth income over the life cycle. All Americans would sleep easier knowing that even if they suffer a turn of bad luck, the state will prevent them from falling into destitution.
Moreover, doing so would play better to our national strengths. The ironic thing about this Kafkaesque tax credit nightmare is that while the U.S. has probably the most complicated tax code and welfare state of any country, our government is also manifestly terrible at implementing complex regulations.
Big and simple, however, we can do reasonably well. Consider the Social Security Administration. Though its old-age benefit program is somewhat complicated and unfair (in particular, it lacks a decent minimum benefit), the agency handles almost all the enrollment and payment procedures itself. Each month, more than 52 million pension payments go out, and the error rate is microscopic—something like 0.17 percent. Not coincidentally, it is also among the most popular government programs. That’s a foundation to build on.
