Should the child credit pay families who don't work? The 2021 argument

Nobody ever lost an election by mailing checks. That is precisely the problem with arguing about them. A check shows up in the mailbox, it has a child's name on the envelope in spirit if not in print, and the person who objects to it looks like the guy who canceled Christmas.

So let me say at the outset what this post is not. It is not an argument that families with children should get less. It is not a claim that parents who cannot find work do not love their kids. It is an argument about a design question that sounds technical and turns out to be moral: should a federal family credit pay the same amount whether or not a parent has a paycheck? In March 2021, Congress said yes, for one year. The conservative answer is no, and I think the reasons are better than the sticker-shock version your cousin has heard. They are also not the whole story, and I will tell you where our own side is divided.

The strongest case for the 2021 expansion

I will state it the way its best defenders would.

The child tax credit that existed before 2021 had a front door with a lock on it. Under the 2017 rules, the refundable part started only after a family earned $2,500 and then built up at 15 cents on each additional dollar. A parent with no earnings got nothing. A parent with very low earnings got very little. The families locked out are the poorest, and the children in them did nothing to earn their circumstances.

Add the timing. Child costs do not arrive once a year. Diapers, rent, shoes and school supplies come every month. A credit that arrives as a lump sum at tax time asks a family to carry costs on credit cards and payday loans for twelve months. Monthly payments, the argument goes, match the expense to the cash.

And add the human point. A job can disappear for reasons that have nothing to do with a parent's character. A plant closes. A babysitter quits. A parent gets sick. The child at the table is the same child either way.

That is a serious case. It has the virtue of focusing on the child, and I do not wave it off. Here is what the 2021 law did, and then the answer.

What the American Rescue Plan did to the credit

The American Rescue Plan Act was signed on March 11, 2021, and passed on party lines. For the 2021 tax year only, it changed the child credit in five ways.

The monthly arithmetic is worth seeing. For a child under 6, the credit was $3,600. Half, $1,800, was paid in advance over six months: $1,800 / 6 = $300 a month. For a child aged 6 to 17, the credit was $3,000. Half, $1,500, over six months: $1,500 / 6 = $250 a month. The rest came at filing time.

The expansion was written for one year. Congress did not extend it, and the monthly payments stopped after December 2021. Supporters wanted it made permanent, and the fight about that is, in effect, the argument of this post.

The answer, in numbered points

One. The credit is a tax credit, and the tax code runs on income. The Child Tax Credit sits in the Internal Revenue Code because the original idea, in the 1994 Contract with America and the 1997 law, was that the tax system should recognize the cost of raising children for people who pay into it. When a payment goes to someone who earns nothing, it is not a tax credit anymore. It is a child allowance in a tax-credit costume. You can support a child allowance, but you should call it what it is and debate it as one. The history is in the child tax credit from $400 to $2,000, told bill by bill.

Two. The 1996 reform rested on a lesson. Before 1996, the main cash program, AFDC, paid families whether or not a parent worked, and in many cases reduced benefits when a parent started working or married. The 1996 law replaced it with TANF, with work requirements and time limits, because Congress and President Clinton concluded that attaching work to aid works better than detaching it. You can dispute how well it worked, and we lay out the fair criticism from both sides in two vetoes, one signature and the 1996 welfare law that put marriage in statute. But the premise was a bipartisan one. The 2021 design reversed it for the child credit without arguing the premise.

Three. The EITC shows the work-linked model. The Earned Income Tax Credit, enacted in 1975 and expanded in the 1980s and 1990s, pays more as a low-income worker earns more, up to a limit, then phases out. It has drawn support from both parties for decades precisely because the benefit rises with work. Its design says: if you work, the government rewards it. The 2021 child credit, with no earnings requirement and a flat amount, said something different.

Four. A flat payment changes the reward for work, even if it does not stop anyone from working. I want to be careful here. Researchers have disagreed about whether the 2021 payments affected employment, and anyone who quotes a study to you should tell you which one and where to find it. My claim is narrower and does not depend on any study. It is about arithmetic, and I will show it below.

Five. A one-year expansion creates a cliff and a promise. Families that budget around monthly payments are hurt when they stop. Congress should not set up a benefit that families will build their lives around and then end it by omission. If a program is worth having, it should be debated, paid for and made predictable.

The arithmetic of the work link

Here is a worked comparison, using a hypothetical single parent with one child aged 8. The numbers come from the rules in the table above, and to keep the arithmetic clean I assume the parent owes no income tax, so the whole credit is paid as a refund.

Under the 2018 rules, the credit is $2,000. The refundable part is the smallest of three numbers: the unused credit, the $1,400 cap, and 15 percent of earnings above $2,500.

• Earnings $0: 0.15 x ($0 - $2,500) is negative, so the refundable amount is $0.
• Earnings $5,000: 0.15 x ($5,000 - $2,500) = 0.15 x $2,500 = $375.
• Earnings $10,000: 0.15 x $7,500 = $1,125.
• Earnings $20,000: 0.15 x $17,500 = $2,625, which exceeds the $1,400 cap, so $1,400.

The cap is reached when 0.15 x (E - $2,500) = $1,400. Solve: E - $2,500 = $1,400 / 0.15 = $9,333.33, so E = $11,833.33. Between $2,500 and about $11,833, each extra dollar of earnings adds 15 cents of credit. That is a real reward for work.

Under the 2021 rules, the same parent with one 8-year-old gets $3,000 at $0 earnings, $3,000 at $5,000, $3,000 at $10,000, and $3,000 at $20,000. The credit does not change with earnings.

Now be fair. A parent in 2021 who took a job still got her wages, and she could still claim the EITC, which has its own rules and rewards work. The expansion did not repeal the incentive to work. What it did is stop the child credit from pointing in that direction. A federal benefit that is flat across zero earnings and $20,000 of earnings has dropped the one lever that told the family, "this program is on the side of the paycheck."

You can think that is acceptable, and some people do. But you cannot think it is neutral.

Where our own side disagrees: Romney's plan

If the strongest conservative argument were simply "no work, no credit," the story would be tidy. It is not tidy, and honesty requires me to say so.

In February 2021, Senator Mitt Romney introduced the Family Security Act. It would have paid a monthly benefit for every child: about $350 a month for children under 6, and $250 a month for children 6 through 17, which is $4,200 and $3,000 a year. It was designed as a replacement for the existing child credit and some other provisions, and it was paid for in part by ending certain tax breaks and consolidating programs. Romney framed it as pro-family and pro-child.

Here is the point that matters for this argument: Romney's plan did not have a work test either. It was a conservative proposal for a child allowance, paid for with offsets, and it drew criticism from other conservatives on exactly the point this post makes. Other Republicans proposed work-linked alternatives of their own. Within the same party, that is a real, unresolved debate.

How would I choose? I would say: Romney's plan made one thing right, which is that a family benefit ought to be paid for and made permanent, and not a one-year surprise. It made another thing contestable, which is whether payment should hang on work. Those are separable. A reader who agrees with the first and rejects the second is not confused.

Middle paths that keep the work link

If your cousin presses with "so you would leave the poorest kids with nothing," do not give a speech. Offer the middle paths, because they exist and some of them have been law.

1. Lower the earnings floor. The $2,500 threshold in the 2017 law was not sacred. In February 2009 Congress lowered the earnings floor for the refundable part to $3,000 from roughly $10,000, which let more low-earning workers collect. A conservative can argue for a low floor and still insist on a floor.
2. Raise the per-child refundable cap. The $1,400 cap in 2017 was a negotiated number. Raising it, and indexing it to prices, helps the working family that owes little income tax.
3. Tie the refund to payroll taxes. The Rubio and Lee approach of 2017 started from the observation that a modest-wage family pays Social Security and Medicare taxes on every paycheck, and that a credit could be refunded against those. It keeps the work link and meets the family's real tax burden.
4. Spread the payment through the year, for working families. The legitimate point in the 2021 argument is about timing. Nothing in the logic of a work-linked credit prevents adjusting withholding so that a working family sees its credit in each paycheck instead of one lump sum. Whether to build that is an administrative question, and it deserves a real answer.
5. Keep the other tools. The EITC, the child care tax credit and state programs are there for those who cannot work or cannot yet find a job. A family credit is one lever among several, and nobody should pretend it has to carry the whole load.

Notice what these have in common. Each takes the strongest point in the 2021 argument, that costs are monthly and that the poorest are locked out, and answers it without cutting the link to work. That is the kind of reform a divided Congress can pass, which is the lesson of divided government gave us welfare reform and a balanced budget.

A fair test for any family credit

Whoever writes the next child credit, whether it comes from our side or from the other, will be better off if it passes these questions. Use them on any proposal in any year.

1. Does it recognize the cost of raising children? A family with children has less ability to pay than a childless family with the same income. A credit should reflect that.
2. Does it reinforce work? A family credit should reward the paycheck and not replace it. If it does not, the sponsor should say why.
3. Is it paid for? A benefit funded by borrowing is a bill for the child to pay later. Ask who pays.
4. Is it predictable? Families plan around benefits. One-year experiments teach them to distrust the next one.
5. Does it avoid marriage penalties and cliffs? A credit that shrinks when two earners marry, or drops off a cliff at a certain income, punishes the choices we want to encourage.
6. Can it be delivered accurately? Money paid in advance has to be reconciled at filing. If families may have to repay, they should know it in advance.

Notice that the 2017 credit passes most of these and fails one: its refundable cap is lower than many working families would like. And notice that the 2021 credit passes the first and fails the others, in my view. The honest way to argue is to hold every proposal to the same test, including ours.

What to say at the table

If a relative tells you that opposing the 2021 expansion means you want children to go hungry, do not fire back with a statistic. Say this: "I want families with kids to keep more of what they earn, and to be paid for the work they do. I think a credit should point at the paycheck. Here is how the old credit worked and here is what changed. Tell me where you think the right line is." That keeps the conversation on the design, which is where the real disagreement is.

For a related argument about the tax code and the family, see why family policy should not punish the parent who stays home. And for the charge that our side cares about children only until they are born, read 'Pro-life stops at birth' is the line to answer, so let's answer it.

The bottom line

• The 2021 expansion was a real attempt to reach the poorest children, and its supporters deserve a fair hearing. It raised the amounts, paid half monthly, and removed the earnings requirement for one year.
• The conservative answer is about design: a family credit should recognize the cost of children and reinforce work, as the 1997 credit, the 2017 credit and the EITC did, and as the 1996 reform intended.
• Our own side is divided. Romney's 2021 plan offered a permanent, paid-for child allowance without a work test, and not every conservative agrees with it.
• Use the six-question test on any proposal from either party: cost of children, work, payment, predictability, penalties, accuracy.
• Do not take anyone's word for the current rules, ours included. Read the IRS instructions for the tax year in question.