In 2021, the US cut child poverty by as much as 40 percent using one of the most effective anti-poverty tools the country had ever devised: the expanded child tax credit (CTC).
By sending unconditional monthly checks of up to $300 per child to the nation’s poorest families — including those with little to no income who had typically been excluded from such programs — the “child allowance” lifted 2.1 million children out of poverty who would’ve otherwise been left behind.
Arguments against such programs that give unconditional cash usually assert that it’ll drive low-income people to quit their jobs, ultimately harming the economy. But research found little to no drop in employment rates as a result of the expanded CTC. Yet despite a flurry of support from prominent economists and recipients alike, politicians failed to reach an agreement to make the temporary expansion permanent, and Congress let it expire at the end of 2021.
Some concerns that sank the program’s political chances, like Sen. Joe Manchin’s worry that recipients will spend the cash on drugs, don’t hold up to the abundant evidence. But there are still some economists who remain skeptical, unmoved by the steady stream of positive research on short-term programs — which they concede looks very good. Their concern, however, is with the distant future.
“My main problem with a permanent CTC,” said University of Chicago economist Bruce Meyer via email, “is that it would reverse the work-based welfare reforms of the ’90s that dramatically increased employment and were associated with a decline in the share of kids growing up in single-parent families.”
The thinking goes like this: As parents receive strings-free cash, many of their incomes will indeed be pushed above the poverty line. At first. But some will also choose to work less, others will divorce more, and fraying the twin threads of marriage and work in low-income communities will ultimately harm children’s prospects for upward mobility. The initial anti-poverty benefits would, generations down the line, be swallowed up by unintended consequences.
By concentrating their concerns on the far future, they render much of the growing base of short-term research moot, especially in terms of convincing holdout politicians, like Manchin, whose insistence on adding work requirements to any expanded CTC is what killed the child allowance.
Enter a new working paper from Elizabeth Ananat and Irwin Garfinkel, two economists at Columbia University. Expanding on work they first published in 2022, their research surveys long-run cash and quasi-cash transfer programs (like food stamps) in the US in an effort to predict the overall effects of a child allowance over the very long run. Instead of the grim and jobless future forecast by expanded CTC critics, they find that a future shaped by a permanent child allowance is well worth the investment.
Ananat and Garfinkel found that the total long-run benefits to society of making a child allowance permanent outweigh the costs by nearly 10 to 1. While the paper may not sway skeptical economists, the dramatic returns could still help build political momentum to pass the policy. And it at least shows that researchers are now taking the long-run concerns raised by critics seriously.
What a return on investment of 10 to 1 for an expanded CTC actually means
Ananat and Garfinkel’s original cost-benefit calculations did not make for light reading. Although the paper was well-received by her colleagues, the problem, Ananat said, is that no one else read it.
So they built a homepage for their research on Columbia University’s Center on Poverty and Social Policy’s website, and as new relevant studies on the subject come out, they’ve been updating their findings. This most recent 2024 working paper reports on the updated results of their earlier work, while trying to offer a more widely accessible version of their research.
Their promise of a 10 to 1 return is, frankly, massive. For every $100 or so billion the child allowance would cost the government each year, society would reap additional long-term benefits of about $929 billion. Those dollars represent benefits like improved child and parent health and longevity, higher future earnings for children, and reduced crime and health care costs. There would be an effect from the small dip in employment that their calculations predict, and a resulting decrease in tax revenue — but it would amount to just $2.4 billion. That’s a drop in a bucket overflowing with almost a trillion dollars in benefits.
But the nuances of such long-term returns can be difficult to convey. “A little bit shows up in the first few years in the form of reduced [child abuse and neglect], reduced hospitalizations, and those sorts of things,” said Ananat. “But most of it doesn’t show up until the kids grow up. So that requires a very patient type of investor.”
Imagine Vice President Kamala Harris wins the presidential election in November, and immediately upon taking office implements a child allowance (as her recently unveiled economic agenda intends). That would cost roughly $97 billion for 2025 alone. But at the end of the year, if you tallied up all the benefits, you wouldn’t see that $929 billion that Ananat and Garfinkel calculated anywhere. It could be decades before the full value of that nearly one trillion is actually realized.
What is refundability?
Usually, a tax credit gives Americans a break by lowering the amount of taxes they owe. But that means someone who doesn’t owe any taxes — like someone who is unemployed and has no taxable income — won’t benefit from the credit. Currently, the child tax credit (CTC) excludes about 19 million children from the lowest-income families, because they don’t earn enough to qualify for the full benefit.
The American Rescue Plan made the CTC “fully refundable” through the end of 2021, which meant that even parents with no income, so who owe no taxes, still received the full value of the benefit. In other rich countries that style child benefits this way, they’re known as “child allowances.”
Notably, though, it’s not a one-time deal. According to Ananat and Garfinkel, every year that the child allowance is in place will reap another $929 billion in long-term benefits. So if you take the Tax Foundation’s estimate that Harris’s child allowance would cost $1.6 trillion over 10 years, Ananat and Garfinkel’s work suggests over that same time period, it would accrue something like $9.3 trillion in long-term benefits.
“The CTC is worth so much money in the future, that even though some of it only happens 50 years from now, it’s still worth $10 for every dollar you spend today,” said Ananat.
The child allowance is fully refundable, meaning the full value goes to the poorest families, without a work requirement. (The current CTC is only partially refundable, which means that parents must first earn income before qualifying for the benefit.) But Ananat and Garfinkel also crunched a second set of cost-benefit calculations on a partially refundable CTC, matching recent CTC compromises that boost the payment amount while keeping a work requirement of some sort.
They find that doing so would reduce the annual cost to $31 billion, while also reducing the total annual social benefit to $131 billion. That means that adding work requirements shrinks the program from a nearly 10 to 1 return on investment, to just over a 4 to 1 return. And those benefits would exclude children who are in the deepest poverty — the very ones who need such help the most.
Now, converting a variety of outcomes into dollar valuations, and assigning benefits that will come in the future a value for the present, is tricky business. These estimates should be held lightly. But if they’re even in the right ballpark, then every year we don’t implement a child allowance is an absolutely massive missed opportunity. We’d experience as much as a 40 percent drop in child poverty immediately, and begin layering on trillions of dollars in long-run benefits.
Critics and the econometric alternatives
There are basically two ways to try and predict the effects of a child allowance in the deep future. Either way, you’re dealing with serious ambiguity.
Like Ananat and Garfinkel, you can scour the existing evidence from similar long-run programs that have raised family incomes in the US, tabulate a comprehensive list of all their documented benefits and costs, calculate a per-$1,000 effect size for each cost and benefit, and then apply those values to a hypothetical child allowance.
To do that, they looked at past long-run programs like pension programs for mothers from before the New Deal, the 1960s rollout of food stamps, a series of US guaranteed income experiments through the ’70s, and expansions in the 1990s to the earned income tax credit.
Notably, none of these programs are actually a CTC. And although each policy was chosen because it does a similar thing — effectively raises a family’s income — each took place in a different social and economic context than whatever the 2030s and beyond will look like.
The other way to try and predict hypothetical economic futures is to use the economist’s version of a magic eight ball: the econometric model. These are mathematically constructed versions of reality, where fuzzy human behaviors have been translated into precise probabilities and equations. In these computational worlds, you can plug in a variable of interest, like a child allowance, and statistically churn out a prediction of its impact.
Then, economists scrutinize the assumptions of the model, and debate how much insight those results can offer about the actual meatspace world that we all live in.
That’s how Bruce Meyer, Kevin Corinth from the American Enterprise Institute (AEI), and their colleagues predicted that replacing the old CTC, which incentivized work, with the expanded CTC that provides low-income families the benefit no matter whether they’re already working or not, would ultimately “do more harm than good.” Specifically, that it would drive 1.5 million working parents to quit their jobs altogether, and cap the anti-poverty impact at 22 percent, rather than the 40 or so percent seen during the temporary expansion.
Those numbers all hinge, however, on controversial assumptions coded into the model about how likely people — low-income single mothers in particular — are to stop working if they get an extra few hundred bucks per month from the CTC. Other models using different assumptions find much lower reductions in work, including Corinth’s own colleagues from the AEI, who predicted the effect on employment would be only a fifth as strong, with 296,000 parents quitting their jobs rather than 1.5 million.
In each case, these models are reflections of their coded-in assumptions. Like extrapolating from other policies in different historical eras, “Taking a given change in the [child] tax credit’s incentive to work and plugging in a labor supply elasticity is a fraught enterprise,” Jain Family Institute research associate Jack Landry told me last year.
When I asked Meyer about the Ananat and Garfinkel study, he dismissed it as “very selective” in its literature review. And Scott Winship, who directs the Center on Opportunity and Social Mobility at the AEI, said via email that “you’re more likely to find that a policy is worthwhile if you simply assume the largest potential costs don’t exist (in this case, worsened child outcomes in the long run from reduced parental work and increased single parenthood). That’s what they do.”
There is, of course, no perfect way to predict the future. Otherwise I’d have cashed out of the stock market with millions by now. But that’s where critics of a permanently expanded, fully refundable child tax credit are now situating their case. Advocates can either follow them into the future, arguing that long-run benefits will outweigh long-run costs. Or, they could instead focus on building the political momentum necessary to pass a policy that will always have its detractors, while unambiguously helping millions of kids in the present.
Will a child allowance in the US ever exist anywhere other than the future?
Vice President Kamala Harris has already announced that bringing back the expanded CTC is part of her planned presidential economic agenda.
Ananat emphasized that “this isn’t a politically motivated” study, but that even if her research won’t convert skeptical economists, it can still be helpful to those working on political organizing. She said,“People can see this and feel inspired, like, ‘Oh, 10 to 1, that’s worth making more phone calls for.’”
I’m not optimistic that economists will ever strike a unanimous agreement about the future of an expanded CTC (though it’s worth emphasizing that to the degree consensus does exist, it certainly is in support). Even if they did, how much stock should we really place in specific predictions that span decades? The world is unpredictable, and seemingly more so every year.
The decision to implement a child allowance will always have to be made under some degree of uncertainty. Trying to predict the impact of a child allowance decades into the future is always going to be a balancing act of ambiguity and speculation. Peering into the economic models that raise concerns about how desirable a future shaped by a child allowance really is, you find a latticework of tenuous assumptions. Similarly, critics say that when you look under the hood of the 10-to-1 return on investment claim, you find a selective literature review.
But the flood of research on the temporary expansion washed away most uncertainty about the short term. Canada’s had a child allowance for years, and it doesn’t look to be hastening a grim and jobless future. A guaranteed huge drop in child poverty in the short term, plus the potential accrual of trillions of dollars in additional benefits, sounds like a worthwhile bet to me. And even if slouching employment among recipients did start to cause concern in a decade or two, and researchers came to suspect the child allowance was to blame, the wonderfully certain thing about public policy is that it can always be changed.
