The Washington PostDemocracy Dies in Darkness

The battle over Biden’s child tax credit and its impact on poverty and workers

(Demetrius Freeman/The Washington Post)

“The money is already a life-changer for so many working families. This will help cut child poverty in half this year, according to the experts.”

— President Biden, remarks on his economic agenda, Oct. 28

One of the most consequential questions faced by lawmakers and presidents is: What will be the impact of this new policy?

It’s especially tricky in the realm of economics and taxes. A relatively simple change ideally would reverberate across the economy, raising incomes of Americans. But if a proposal is poorly designed, it could backfire and have unanticipated negative effects.

Policymakers rely on economists to model the potential effects. But right now, there’s a fierce battle within the economics world over Biden’s expanded child tax credit. At stake: whether the president’s proposal would reduce poverty as much as advocates say it would.

The economists involved are well respected and care deeply about poverty issues. But tensions have risen, even sparking a formal letter requesting a correction in an influential report on options for reducing child poverty issued by the National Academy of Sciences.

The Fact Checker cannot presume to settle an economic dispute of this magnitude. But we’ve dug deep into the literature and discussed at length the dispute with the economists involved, in an effort to help explain it to readers without getting too much into the weeds.

The Facts

As part of the coronavirus relief package, Biden took the existing child tax credit, expanded it and in effect made it a sort of monthly child allowance. This is a novel concept in the United States, but not in other parts of the world, where 108 countries have a periodic child or family allowance anchored in national legislation, according to UNICEF.

Biden also gave the benefit to people who owe no income taxes — “fully refundable” is the technical term. Previously, in most cases, you had to pay some income taxes to get the child tax credit. Now parents under a certain income limit receive the full credit — $3,600 per child ages 5 and younger and $3,000 per child ages 6 to 17 — regardless of the amount of income taxes they owe or how little work they do. (Most low-income families, however, also incur payroll taxes.)

This change makes a big difference in how much money a parent could receive. Under the traditional child tax credit, a single mother with two children above the age of 6 working full time at the minimum wage, earning $14,500 a year, would receive $1,800, or $900 per child, when they file their taxes. The size of the child tax credit increased as earnings rose. Now, under Biden’s expanded credit, that parent would receive $6,000 — $500 a month.

This child credit expansion is only in effect through the end of this year, but Biden’s “Build Back Better” plan would extend the more generous amount for a year and make full refundability permanent.

Now let’s introduce an economic concept: elasticity. Generally this reflects how the demand for a product can change if its price rises or falls. But the same concept can be applied to labor economics, reflecting how supply of labor (or how much more or less people want to work) changes when wages rise or fall.

For instance, the earned income tax credit is widely considered the best anti-poverty provision in the tax code. It reduces the taxes owed by low- and moderate-income workers. The EITC increases the after-tax wage on the first dollar of work, translating to a refund for each hour worked for even those with the lowest income.

The EITC is believed to be an incentive for work. People won’t earn a tax break — giving them more cash — unless they work. Up until this year the child tax credit was similar to the EITC in that you only got it if you worked. The Biden proposal would get rid of that feature and give it to everyone regardless of work. The question is whether this expanded child tax credit eliminates an incentive to work and thus would encourage people to drop out of the workforce.

In economics, two factors are considered to answer this question. One is known as the “substitution effect.” If wages go up, people tend to work longer — but if someone can get enough money without working, fewer people might work. There is also the “income effect,” in which more money makes things more affordable, including a decision to not work.

A panel of experts writing the Academy report, published in 2019, concluded that a child allowance policy of $250 a month would have a “negligible” impact on employment (about 150,000 fewer workers) and that it would reduce poverty 36 percent and deep poverty even more. The peer-reviewed Academy study was so influential that 462 economists cited its findings in an endorsement of the Biden plan.

“The panel of experts who reviewed this issue for the National Academy of Sciences concluded that a universal child allowance would have a negligible effect on employment,” the Sept. 15 open letter to Congress said.

But a new working paper, written by Bruce D. Meyer, Kevin Corinth and two colleagues at the University of Chicago, came to a different conclusion. Their calculations, based on a restricted-use data set that combines information from the Current Population Survey with administrative data on earnings, retirement income, federal benefit programs and taxes, initially confirmed that child poverty would drop 34 percent with Biden’s expanded child tax credit if no one changed their behavior as a result of the new policy.

But then they applied the substitution effect. They found that 1.5 million people would drop out of the workforce, 10 times the level estimated by the Academy, because of the money provided via the child tax credit. (To be precise, about 1.3 million workers would drop out because of the substitution effect. About 140,000 workers would respond to the income effect, which is similar to the Academy calculation.)

With so many people not working, Meyer and Corinth say, child poverty would only fall 22 percent — and deep poverty not at all. Moreover, they said, the policy would erase most if not all of the gain in employment by single mothers since the 1990s.

Obviously, as with any economic models. these calculations depend on the inputs — especially the number (essentially a ratio) that reflects the elasticity of employment participation. There is a rich paper trail of estimates of the elasticity of workers who receive the EITC, which could be applied to the child tax credit. But since this is a novel policy, economists have to make choices about the right number to use.

That’s a big part of the dispute, which we will discuss in detail below.

Meyer and Corinth also have charged that the Academy panel, in “a consequential error,” never applied the substitution effect to its calculations on the child tax credit, only the income effect, even though it did so elsewhere in the report with regard to a proposed expansion of the EITC. In effect, the Academy panel estimated the positive effects of the EITC for work but not the negative effects of the child tax credit.

For instance, the Academy report assumed that $1,000 of extra EITC will generate a 7 percent (5.6 percentage point) increase in employment for women with some college or less. In theory, then, converting the child tax credit to an allowance should be like taking away $2,000 of work incentive (for one child) or even more (for multiple children).

Meyer and Corinth have sent a letter to the National Academies of Sciences, Engineering and Medicine demanding a correction for the public record.

Panel members acknowledged that they did not model the substitution effect for the $3,000 child tax credit, explaining the 2017 tax law was passed in the midst of doing the report. Before 2017, the maximum child tax credit was $1,000 and the Academy team argues that this amount was small enough to not have a large effect on the analysis.

“Almost all of our thoughts were about families who had very low or zero earnings who would not work or reduce work effort,” said Timothy Smeeding of the University of Wisconsin at Madison. “We did not have any credible estimate of the substitution effect for families earning $25,000 to 40,000, and so we ignored it.”

Hilary Hoynes of the University of California at Berkeley and Robert Moffitt of Johns Hopkins University, the panel members most responsible for modeling behavior, sent a joint statement to the Fact Checker.

“Given the existing law, and the policy we were expanding, it was the assessment of our committee that the ‘income effects’ of the policy change would be more important and the substitution effect would be muted. Why? Because going from $1,000 to $3,000 is a large increase in income,” they wrote. “We judged that most people would notice a large $2,000 increase in income more than they would notice a very small reduction in the working gain, and that they would respond more to the former.”

The Academy panel members are critical of the Meyer-Corinth numbers because they say the report concluded that people making more than $50,000 would drop out of the workforce in response to child tax credits, which they say is not credible.

Meyer said that if people making more than $50,000 were excluded from the calculations, his research would still yield a drop of 1.1 million. That’s still much higher than the Academy panel estimates.

“The bottom line is that we don’t think we were wrong in our assessment, although we are open to the possibility that the employment effects may be a bit bigger than we estimated,” Hoynes and Moffitt said. “Like all good scientists, we are willing to alter our opinion if shown hard evidence that we were incorrect. But we cannot imagine ever being persuaded of employment effects in the Corinth-Meyer range, whose analysis we think is fundamentally off base.” (Their full statement can be found at this link.)

Ultimately, as we noted above, the result depends on how much an economist believes workers will be affected by the incentives and disincentives in tax policy to decide whether to work. That would be their elasticity. Plug in a different number and you get a different result.

Among the questions to consider: Have economists previously overestimated the impact of the EITC on work? (Possibly, but that has been questioned.) Are women becoming more like men and thus are less influenced by changes in wages? (Or, as an economist would put it, have labor supply elasticities for women declined substantially in recent decades?)

Meyer and Corinth choose one elasticity that applied only to single mothers who receive the EITC (about 10 million people) and then a much lower one for everyone else.

Another working paper, recently updated, found that, including the substitution effect, 500,920 to 578,197 parents would choose to leave the labor force in response to Biden’s expanded child tax credit. That’s much higher than the Academy panel but about one-third of the Meyer-Corinth result.

An earlier version of this paper appeared to support the elasticity used by Meyer-Corinth for single women receiving the EITC — a critical element in Meyer-Corinth’s conclusions. But in the new version that sentence has been removed and a footnote criticizes the Meyer-Corinth methodology. The authors said they did a more rigorous review of the literature as part of their update, resulting in a smaller elasticity.

“We’ve since done a lot of research into where those elasticities are coming from and we have concluded that the numbers [found in the literature they had previously quoted] are too high for this exercise for a few reasons,” said Katherine Michelmore of the University of Michigan, one of the co-authors. She cited both the perceived change in labor supply elasticities for women as well as adjustment to using post-tax income. (One paper cited in the updated version as important studied only married couples, but the authors assumed that unmarried women and men have the same elasticities as married men.)

Meyer and Corinth have posted a detailed discussion of the elasticity estimates and why they believe their choices are supported in the economic literature. In particular, they charge that Hoynes — who helped organize the letter to Congress that cited the Academy study — miscalculated elasticity in an earlier paper and so her numbers cannot be trusted in this context.

“I’m completely flabbergasted that Hilary, who I have known for 30 years, had miscalculated the elasticity,” Meyer told the Fact Checker. “This is not rocket science. Hilary is not getting the most simple thing right. I’m dumbfounded and sad.” He suggested that a more accurate figure “would give her the number she did not want.”

“Putting aside his flowery and unprofessional language, the answer is yes: I reject his assertion that I miscalculated the elasticity,” Hoynes said. (Readers interested in the technical aspects of this dispute can view this document and this document.)

Beyond this dispute over which numbers are correct, there is a meta-question of how to evaluate the reduced work that might result from the expanded child tax credit. If secondary earners or even primary earners reduce their hours, it may be a benefit, not a cost, depending on how people use their time. (For instance, parents could devote more hours to educating their children, resulting in more productive workers in the future.) In economics, giving people more choices generally is considered a good thing.

The Bottom Line

In our college macroeconomics course, the professor once started a lecture with an old joke. A scientist, an engineer and an economist are stuck on an island with one tin of food and need to open it. The scientist proposes to harness the sun’s rays and the engineer wants to finagle a solution with a coconut shell. The economist brightly says: “Let’s assume we have a can opener.”

The assumptions make all the difference in whether one concludes that the expanded child tax credit would be a disincentive to work. All agree there would be some disincentive, but whether it is a meaningful figure is in dispute. Depending on the elasticity selected and the types of workers it is applied to (single mothers, parents, etc.), one can get a range from 200,000 to 2.5 million.

That range raises the political stakes. Unfortunately, the actual number will never be known unless the policy is implemented. Assumptions can only take you so far.

(About our rating scale)

Send us facts to check by filling out this form

Sign up for the Fact Checker weekly newsletter

The Fact Checker is a verified signatory to the International Fact-Checking Network code of principles