The Year Child Poverty Fell by Nearly Half, and Why It Came Back

On July 15, 2021, a deposit landed in tens of millions of American bank accounts that most of the recipients had never asked for and many did not know was coming. It was labelled, in the dry language of bank statements, as a payment from the Treasury. For a parent with a toddler it was up to $300; for a parent with a ten-year-old, up to $250. The same deposit arrived on the fifteenth of the next month, and the month after that, through December. Then, in January 2022, it did not arrive at all. No vote was taken to stop it. Nobody gave a speech announcing the end. The law that created it simply ran out, and the Senate could not agree to extend it. When the Census Bureau later counted what had happened to American children over those two years, the numbers looked like a controlled experiment that nobody had meant to run.

In short: The expanded child tax credit of 2021, paid monthly and available to families with little or no income, pushed child poverty to the lowest rate the Census Bureau's fuller measure has ever recorded. When it lapsed, child poverty more than doubled within a year. This post argues that a permanent, fully refundable credit is one of the most proven anti-poverty tools the country has, and answers the strongest objections to it.

What the 2021 credit actually changed

The child tax credit was not new in 2021. It was created in 1997, and the 2017 tax law raised it to $2,000 per child. But the old credit had a design flaw that most middle-class families never noticed: it was built around earnings. A family had to have income to get the full amount, and the families with the least income got a partial credit or nothing. By the estimates of several policy researchers, roughly a third of American children lived in families that did not earn enough to receive the whole thing. The children who most needed the help were the ones the credit was least likely to reach.

The American Rescue Plan, signed in March 2021, changed three things at once. It raised the amount to $3,600 for children under six and $3,000 for children six through seventeen. It made the credit fully refundable, so a family with no earnings received the same amount as a family with modest earnings. And it sent half of the money in advance, in monthly installments, rather than as a lump sum at tax time.

The third change is the one people tend to forget, and it may have mattered most for how families felt. Rent is monthly. Groceries are weekly. A refund in February does not help much with a car repair in September. For six months, parents had a predictable sum arriving on a known date, and they could plan around it.

The number the Census Bureau reported

Every September the Census Bureau releases two measures of poverty. The official measure dates to the 1960s and counts only cash income before taxes, which means it cannot see a tax credit at all. The Supplemental Poverty Measure, published since 2011, counts what families actually have to spend after taxes, tax credits and benefits such as food assistance, and it adjusts for costs like housing and medical care.

Under the supplemental measure, child poverty in 2021 fell to 5.2 percent, the lowest rate since the measure's history begins. The Census Bureau's own analysis credited the expanded child tax credit, along with stimulus payments, with much of that drop. A year later, with the expanded credit gone, the same measure put child poverty at 12.4 percent. That is more than double, in a single year, in a labour market where jobs were plentiful and wages at the bottom were rising.

It is worth being careful about what this does and does not show. 2021 was a strange year, full of pandemic programmes that also ended. Stimulus checks, expanded food benefits and other help were all part of the picture. Researchers who have tried to separate the pieces still give the child tax credit a large share of the credit, and the Census Bureau's own estimates of how many children each programme lifted out of poverty put the expanded credit near the top. But anyone who tells you the credit alone explains every point of the drop is overselling it, and the case does not need that.

What families did with the money

Between the monthly payments, the Census Bureau was running a fast, large survey called the Household Pulse Survey, built to track how households were coping during the pandemic. Researchers used it to see what changed when the payments started and again when they stopped. The most consistent finding was about food. Reports of households with children not having enough to eat fell after the payments began and rose after they ended.

Surveys of how families spent the money showed it going overwhelmingly to the ordinary things: food, rent, utilities, school clothes, paying down debt, child care. That matters because the most common worry about cash aid is that it will be wasted. The evidence from 2021 suggests the opposite. When you give a parent with a tight budget some breathing room, the breathing room tends to go to the children.

There is a quieter finding, too. Families with the lowest incomes, the ones who had been left out of the old credit, received the largest increase relative to what they had before. A credit that had been designed around rewarding work became, for six months, a credit designed around children. That shift is the whole argument in miniature.

The strongest case against it

The objection to the expanded credit is serious and deserves a fair hearing, because it is the reason the programme died.

Critics, including a number of economists and Senator Joe Manchin of West Virginia, whose vote Democrats needed in an evenly split Senate, argued that making the credit fully refundable cut the link between benefits and work. The 1996 welfare law had been built on the idea that aid should encourage employment, and the earlier credit followed that logic. A group of economists at the University of Chicago published an estimate that a permanent, fully refundable credit could lead well over a million parents to leave the workforce. Manchin also raised concerns about cost, about inflation, and about whether the money would go to families who did not need it. When the Build Back Better bill, which would have extended the expanded credit, stalled in the Senate in December 2021, the credit's extension went with it.

That argument deserves an answer, not a sneer, because the value behind it, that work matters to people's dignity and to their prospects, is one most progressives share.

Here is the answer. The prediction was testable, and in 2021 it was tested. Researchers at Columbia University and elsewhere looked at employment among parents before and after the monthly payments began and found no meaningful drop in work. Other studies using different data reached similar conclusions. Six months is a short window, and a permanent programme could in theory behave differently, but the large exits from work that the critics forecast did not show up in the data we have. Meanwhile, the evidence that poverty in childhood harms health, schooling and adult earnings is substantial. A credit that reliably keeps children out of deep poverty is not only charity. Many economists who study it argue it is an investment that pays back over a lifetime.

The cost objection is real too. A permanent expanded credit costs a great deal, on the order of a hundred billion dollars a year or more depending on the design. I think that is a choice worth making, and the comparison is instructive. The tax cuts of 2017 and 2025 cost far more over a decade, and most of their benefits went to households at the top.

Where things stand in 2026

The idea did not disappear with the monthly payments. In January 2024 the House passed a bipartisan tax bill, the Tax Relief for American Families and Workers Act, by a lopsided margin, 357 to 70. Among other things it would have made more of the credit available to lower-income families. The bill stalled in the Senate.

Then in July 2025 President Trump signed the large tax and spending law that Republicans passed on party-line votes. It raised the maximum credit to $2,200 per child and indexed it to inflation. But it kept the basic structure that leaves out the poorest children: the refundable portion still depends on a family's earnings, so the families with the least income still get less. Supporters described the change as help for working families, which it is for many. It does not restore what the 2021 credit did for children at the very bottom.

There has also been a quiet experiment happening below the federal level. Several states, including some led by Republicans, have created or expanded their own child tax credits since 2021. Some of them, like Colorado's and New Mexico's, are aimed at families with low incomes. They are smaller than the 2021 federal credit, but they show that the idea has an appeal that crosses party lines, and they keep producing evidence.

Why the design questions decide everything

If there is a lesson progressives should take from 2021, it is that details are destiny. Three choices separated the credit that cut child poverty from the credits before and after it.

The first is refundability. A credit that requires earnings will always miss the children whose parents are sick, caring for a relative, between jobs, or working off the books. Full refundability is what made the 2021 credit reach the families in deepest need.

The second is frequency. Monthly payments matched the rhythm of family budgets. They also had administrative costs and caused some confusion at tax time when families had to reconcile what they had already received. Those problems are fixable, and a permanent programme would have time to fix them.

The third is durability. A credit that lasts six months cannot change a family's long-term plans. Parents could not count on it when deciding whether to take a lower-paying job with better hours or move closer to family. A permanent credit, like Social Security, changes what people can plan for.

There is also a political lesson, and it is not flattering to the side that passed the law. The expanded credit was written as a one-year measure inside an emergency package, partly to hold down its official price. That made it easy to let lapse. Many families never knew where the money came from or why it stopped. A programme that people do not understand is a programme that is hard to defend. If Democrats return to this in the new Congress, which convenes in January 2027, the case will need to be made out loud, to parents, in plain language, long before any vote.

The families in the numbers

It is easy to talk about this in percentages and forget who sits behind them. Children in poverty are disproportionately in rural counties in the South and in Appalachia, in tribal nations, and in the poorer neighbourhoods of large cities. Their parents are often working, frequently at jobs with unpredictable hours. Many are single mothers. A significant number live in states that have voted for Republicans for decades, which is one reason the credit has defenders on both sides of the aisle.

When the 2021 payments arrived, they arrived in all of those places at once. A family in rural Kentucky and a family in the Bronx received the same amount for the same-aged child. That is the quiet radicalism of a universal child benefit: it treats every child as worth the same. Most wealthy democracies already pay one. For six months in 2021, the United States did too, and the results were about as clear as social policy ever gets.

The argument for doing it again is not that it is generous. It is that we already ran the test, the test worked, and we stopped anyway.

What you can do

• Tell your members of Congress what you saw in 2021. The new majority will set its priorities early in 2027. Use the representatives lookup with your address to get the phone numbers and contact forms for your House member and both senators, and check their latest votes before you call. A short message about a family you know, and a clear ask for a fully refundable credit, carries more weight than a form letter.
• Find out whether your state has its own credit. State child tax credits are growing, and state legislators are reachable in a way federal ones often are not. The same lookup shows your state legislators. If your state has no credit, ask why not; if it has one, ask whether families with the lowest incomes can get all of it.
• Help a family claim what already exists. Many eligible families miss tax credits because they do not file a return. Free tax-help sites and anti-poverty groups run outreach every filing season, and you can find them through the organizations directory, which lists nonprofits by cause in every state.
• Follow the argument as it develops. The fight over the credit runs through the larger debate about the safety net and what families can afford. Read across both before the next tax bill is written, so you can answer the work-incentive objection with the evidence instead of a slogan.