Would The Expanded Child Tax Credit Be Bad For The Economy?

A series of reports in The Wall Street Journal informs readers that U.S. employment will decrease by 1.5 million if, as Democratic leaders wish, the expanded child tax credit (CTC) that was temporarily in effect in 2021 is reenacted and replaces the current-law child credit.

In April 2021 President Biden proposed a five-year extension of the expanded credit. On a party-line vote, Democrats in the House included a one-year extension (through 2022) of the expanded CTC in the Build Back Better Act (H.R. 5376) that passed November 19, 2021.

But in the Senate, Democrats couldn’t get the 50 votes required to include the credit in final legislation, the Inflation Reduction Act (P.L. 117-169), largely because of the objection of Sen. Joe Manchin III, D-W.Va., about the absence of a work requirement.

The Joint Committee on Taxation estimates that a permanent extension would cost $1.4 trillion over 10 years. Democrats continue to seek extension of the credit in the lame-duck session, perhaps in a deal that would include repealing amortization of research expenses and extension of 100 percent bonus depreciation.

The 1.5 million reduction in employment estimate comes from a recent study by University of Chicago economists. If that estimate is correct, it would mean slower overall economic growth. And the poverty-reducing effect of the credit, estimated by the Census Bureau as cutting child poverty almost in half, would be much less.

So we have the usual trade-off between economic efficiency and vertical equity. (Kevin Corinth et al., “The Anti-Poverty, Targeting, and Labor Supply Effects of the Proposed Child Tax Credit Expansion,” Becker Friedman Institute Working Paper (Oct. 2021).)

But in today’s environment, there is another dimension to the economic implications of the credit: It could have the inflation-inducing effect of reducing labor supply.

Former Treasury Secretary Lawrence Summers often says the best indicator of future price inflation is wage inflation. In recent months, wage inflation has been stoked by a tight labor market in which the number of job openings far exceeds the number of unemployed.

In his November 30 speech, Federal Reserve Chair Jerome Powell picked up on this theme. He informed his many listeners that the shortage of workers was a major contributing factor to current inflation and that it would be something the Fed will be closely monitoring in its decisions on where to set interest rates.

Powell cited studies that estimate the labor shortage at about 4 million workers. So, according to the University of Chicago estimates, the expanded child credit would make the shortfall 40 percent larger.

Powell said, “Policies to support labor force participation could, over time, bring benefits to the workers who join the labor force and support overall economic growth.”

With the labor force 1.5 million workers smaller, the Democrats’ proposal would do the opposite of supporting growth in the labor force and would provide a significant boost to inflation.

The Wall Street Journal is a must-read newspaper, but on the economics of the child credit, it leaves a misleading impression. It refers only to unfavorable research on the employment effects of the credit.

There is in fact considerable uncertainty and controversy on this topic. Among available estimates, the 1.5 million is at the upper bound, and more than a few estimates are only a small fraction of that figure.

Incentives Up-Down/On-Off

To understand what underlies these estimates requires a bit of explaining. Let’s begin with the three different categories of incentive structures in programs providing assistance to low-income families:

(1) benefits that phase out with income,

(2) benefits that phase in and then phase out with income, and

(3) benefits that are (for practical purposes) lump sum.

Assistance that is phased out as income rises is the most common in both nontax (for example, Medicaid and food stamps) and tax (health insurance premium tax credits) low-income assistance.

As income rises, benefits are decreased perhaps by 10 or 20 cents on the dollar, effectively creating a tax (negative benefit) for marginal increases in income. Given the large number of programs for which a low-income family could qualify, it is possible in some circumstances that the combined effect of programs can result in marginal tax rates so large that the reduction in benefits from rising income is more than the increase in income, so it pays more to work less or work not at all.

The current-law CTC (as enacted into law in the Tax Cuts and Jobs Act) and the earned income tax credit have a phase-in/phase-out structure that, at least in part, addresses the problem of negative employment effects. Without working, you don’t qualify for these credits. And as earned income rises, so do the credits.

So, instead of a penalty, there is benefit from increased labor supplied. But this marginal benefit exists over an income range (the phase-in range) only until a maximum credit amount is reached. Then there is an income range in which the benefit doesn’t change.

And following that, the marginal benefit declines (the phase-out range). Over the phase-out range, there is the same type of employment disincentive as in the first category of assistance program.

That decline in marginal benefit reduces employment incentives. If we graph the benefit against income as in the figure, we get a trapezoidal shape. The marginal benefit, as measured by the slope of the line, is first positive, then zero, and then negative.

So in theory, although certainly less detrimental to work incentives than straight income-phased-out assistance, the overall effect on employment of these trapezoidal credits is ambiguous. (Please see the note at the end of this article about the relatively small negative impact on employment from the “income effect” present in all these credits.)

For a family with two children, the rate of the EITC is 40% of income until the maximum credit amount, which is $6,614 in 2022. The income phase-out range is from $20,070 to $49,339 at a rate of 21%.

The refundable portion of the TCJA CTC ($1,400 per child) — which is the most relevant portion of the total credit ($2,000 per child) to low-income families without tax liability — increases by 15 cents for each dollar of income above $2,500. For 2022, the CTC starts phasing out at an income of $200,000 for single filers and $400,000 for joint filers.

The third category of low-income assistance, lump sum, isn’t conditional on income (or at least not until high levels of income are reached). The American Rescue Plan Act of 2021 temporarily (for 2021 only) increased the amount of the child credit to $3,600 for each child under 6 and $3,000 for each child ages 6 to 17.

And, critically for this discussion of incentives, ARPA removed the phase-in for the refundable portion of the CTC. Because families with no or very low income could qualify for the full credit, the ARPA CTC could do a great deal more to reduce poverty than the TCJA CTC. But also, on the downside, eliminating conditionality on income at low levels eliminates the positive employment incentive of earning extra income.

In the figure, the upward-sloping phase-in line at lower income levels becomes a horizontal line when the TCJA CTC is replaced with the ARPA CTC (and the horizontal line shifts up because per-child benefits are larger).

The Controversy

The prior paragraphs tell us that the TCJA CTC (like the EITC) has positive price effects over low income and negative price effects over high income, and zero price effects in between, while the ARPA CTC has only negative or zero price effects. So we expect a proposed switch from the TCJA CTC to an ARPA CTC would reduce employment — that is, to the extent recipients respond at all to the incentive effects.

In simplified terms, the University of Chicago economists calculate the difference in the TCJA CTC and ARPA CTC after-tax, after-benefit wages and then multiply that difference by a response to changes in participation from similar wage changes (estimated in prior studies) to estimate change in participation from the proposed change.

The estimated responsiveness (“elasticity,” economists call it) used by Corinth et al. is large, which means the negative response of employment to the elimination of net marginal benefits will be large, accounting for most of the estimated 1.5 million reduction in employment discussed above.

The authors cite several other studies that use or estimate elasticities of similar magnitude, including a 2019 paper by former Treasury economist Kye Lippold (“The Effects of the Child Tax Credit on Labor Supply” (Nov. 2019)).

In his November 28 Wall Street Journal opinion piece, Scott Hodge cites work by the JCT that estimates that switching from the current-law TCJA CTC to an ARPA CTC would reduce labor supply by 0.2%. That may seem like it confirms the University of Chicago results.

But according to the Bureau of Labor Statistics, the U.S. workforce was 164.5 million in November. Multiplying that amount by 0.2% gives us 329,000 — only a fraction of the estimate from that study.

Much of the relevant empirical work on this topic comes from studies of the EITC. It makes sense to use this work to evaluate the current-law child credit because the structure and incentive effects of the two credits are qualitatively similar and adjustments can be made for the quantitative differences.

The conventional wisdom based on this research is that the EITC overall has been effective in increasing employment. More specifically, the positive employment effect on single mothers (in the phase-in range, with low income, and having an incentive to enter the workforce) more than offsets the negative employment effects on married women (considered second earners and therefore at moderate income levels, with benefits reduced for working).

Many Republicans and most Democrats support the EITC and the TCJA CTC, and they accept this positive evaluation of the credits. But there are some doubters in the economic community. Simultaneous with the expansion of the EITC in the 1990s were both a reduction in traditional welfare benefits and a booming economy.

Some commentators believe that some or all of the estimates that attribute increases in employment to the EITC may really be attributed to the major changes in welfare rules and economic conditions. (Henrik Kleven, “The EITC and the Extensive Margin: A Reappraisal” (Sept. 2022).)

In economics, it is rare not to find controversy around the issues of measuring elasticities — especially when evaluation of the effectiveness of policies is the subject of political debate. It would be nice if, for example, there were a single, agreed-upon elasticity of labor supply (that is, an estimate of the responsiveness of hours worked to a change in the after-tax, after-benefit wage rate).

But elasticities are often estimated with different data sets, in different periods, using different measures — so different results should be expected (even if there are no errors in method).

And then when those estimated elasticities (from some past period) are used to estimate the impact of a new policy (to take effect in the future), there is no guarantee that elasticities will not change over time or be different for different groups, or that the amount of labor is measured in a consistent way. Another factor in the lack of consensus is the possibility of political bias, both subconscious and otherwise.

At least four recent studies using statistical analysis of data on recipients of the extended credit in 2021 find little or no negative effect on employment. Estimated elasticities from the past aren’t used (so the question about the correct elasticity is sidestepped).

These studies leave open the question about the effects of a long-term or permanent extension of the ARPA CTC on employment, which we would expect to be larger. But the consistency of the results is striking. By neglecting to even mention any of these studies, the Wall Street Journal articles give a one-sided presentation of the economic research.

Value of Home Work

Here we have focused on the economic impact of the CTC on labor supply. Of course, poverty reduction also matters. The effect of the credits on poverty is closely tied to their effect on labor participation because if benefits are combined with reductions in wage income, the net impact (dollars of benefits less reduced wages) of the credits on poverty may be significantly less than the easy-to-measure direct impact (dollars of benefits).

So, in effect, determining the impacts on labor supply and poverty are two sides of the same coin.

The EITC and child credits affect factors other than those taken into account in traditional economic models, such as childhood education, food insecurity, and infant health. And we shouldn’t forget that many activities of unemployed individuals — such as caring for disabled people, older adults, and small children — are important nonmarket activities not included in GDP even though they are critical to the short- and long-term well-being of Americans.

Note on Income Effect

What we haven’t discussed here yet is that in addition to the marginal work incentive effects for each additional dollar of earned income (which can be positive or negative), all these credits have what is known as an income effect.

This is a negative incentive on workforce participation that results from X dollars of benefits. (That is the only relevant incentive effect in a lump sum credit like the ARPA CTC.)

To illustrate the difference between an income effect and the marginal benefit from extra work, consider this example. Assume you receive $5,000 from the government. Because it is a flat amount of extra income, you may work, but the amount you work doesn’t change the credit. You might decide to take more time off as a result of this extra income and work, say, 1,950 hours per year.

Now assume the government gives you $5,000 for the 2,000 hours you now work each year and pays you an extra dollar for every extra hour you work. You decide to work 2,100 hours. That 2,100 hours would be the net change of -50 hours from the income effect and an offset of an extra 150 hours from the price effect.

Although there is significant disagreement about the size of the marginal effect, the income is generally agreed to be small. (In the University of Chicago study, the income effect accounted for only 10% of the total 1.5 million.) Thus, we have left it out of this discussion to simplify the explanation for readers not expert in these matters.

Additional References

Elizabeth Ananat et al., “Effects of the Expanded Child Tax Credit on Employment Outcomes: Evidence From Real-World Data From April to December 2021,” National Bureau of Economic Research Working Paper No. 29823 (Mar. 2022) (“Our analyses of real-world data suggest that the expanded CTC did not have negative short-term employment effects that offset its documented reductions in poverty and hardship.”).

Michael Karpman et al., “Child Tax Credit Recipients Experienced a Larger Decline in Food Insecurity and a Similar Change in Employment as Nonrecipients Between 2020 and 2021,” Urban-Brookings Tax Policy Center (May 2022) (“If the ARP’s temporary changes to the CTC were made permanent, it is unclear whether the credit’s design would discourage adults from being employed. . . . We found no significant differences in the changes in employment between December 2020 and December 2021 for adults who received the payments and adults who did not receive the payments.”).

Ben Lourie et al., “Effects of the 2021 Expanded Child Tax Credit” (Apr. 4, 2022) (“We find evidence of either no change, or even an increase, in employment and wages among lower-income recipients relative to moderate-income recipients, with neither group decreasing employment.”).

Natasha Pilkauskas et al., “The Effects of Income on the Economic Wellbeing of Families With Low Incomes: Evidence From the 2021 Expanded Child Tax Credit,” NBER Working Paper No. 30533 (Oct. 2022) (“We find no evidence that the monthly CTC benefits led to a reduction in employment or labor force participation in the six months during which the benefits were distributed.”).