Reassessing the Budgetary Treatment of Refundable Tax Credits

Executive Summary:

  • While tax credits can reduce income tax liabilities to zero, low-income workers necessarily pay payroll taxes, which are regressive in structure.
  • Among the rationales for refundable tax credits, which provide income support for low-income workers in excess of income tax liability, is to provide relief against the payroll tax.
  • The budgetary treatment of refundable tax credits has been generally settled since the late 1970s: The portion of a refundable tax credit in excess of individual tax liability is recorded as spending for federal budgetary purposes.
  • More recent approaches to providing payroll tax relief reveal the budgetary tradeoffs that can attend such policies, and can illuminate future approaches to income support programs, such as the child tax credit.


The modern U.S. tax code is a far cry from the modest import duties overseen by Hamilton’s Treasury. Rather, today’s tax code is at once a system for revenue collection; a system for advancing domestic economic and social policy aims including health, housing, and education; and a trade policy. Perhaps no set of policies more clearly evinces the evolution of the tax code from a modest system of levies to a complex delivery system for federal policies than tax expenditures. Among the most popular and costly are refundable tax credits, which can be claimed by taxpayers to reduce income tax liabilities but are refunded to taxpayers to the extent they exceed income tax liabilities. For many workers, the payroll tax is the most significant federal tax liability, but providing payroll tax relief is complicated by the need to finance the Social Security system. Relatedly, the historic budgetary treatment of refundable tax credits as (partially) spending can pose challenges for policy advocates. The evolution of the budgetary treatment of these programs, and recent policy changes, underscore the tradeoffs that must be considered in the design of refundable tax credits.

Tax Expenditures

The modern concept of tax expenditures began to take form in the 1960s, upon the observation of then-Assistant Secretary of the Treasury Stanley Surrey, that exceptions to the taxation under the prevailing tax code resembled spending.[1] Indeed, these exceptions are necessarily considered relative to a reference point. The conceptual reference for the U.S. income tax relies on the Haig-Simons definition of income.[2] Broadly, this measure is defined as consumption plus the net change in wealth over time – a comprehensive income tax would tax this base. To the extent that tax expenditures are considered “tax breaks,” or otherwise an exception from taxation, it is relative to this concept of income and income taxation. So ends the theory; in practice, the modern tax code is replete with deviations from a pure income tax.

The Congressional Budget Act of 1974, the foundational body of law for the modern budget process, sought to clarify this element of tax policy, and defined tax expenditures as “those revenue losses attributable to provisions of the Federal tax laws which allow a special exclusion, exemption, or deduction from gross income or which provide a special credit, a preferential rate of tax, or a deferral of tax liability.”[3]

Refundable Tax Credits

Despite defining tax expenditures in statute, there remains sufficient ambiguity in this definition to allow for alternative interpretations between Congress and the executive branch in the estimation and presentation of the cost of these provisions. The U.S. Treasury Department prepares a table of tax expenditures as part of its contribution to the preparation of the president’s budget. Separately, the Joint Committee on Taxation (JCT) provides its own estimates to Congress. Despite a common definition, differences do occur between the estimates prepared by these institutions despite a common definition.[4]

Conceptually, a tax expenditure is often short-handed to “spending through the tax code.” That is somewhat misleading, as only a particular set of tax expenditures involves federal spending or outlays: refundable credits. Indeed, refundable credits somewhat blur the line between taxation and spending. In general, refundable tax credits are features of the individual and corporation income tax code and provide a credit or a dollar-amount reduction in income tax liability. Unlike typical tax credits, however, to the extent that a refundable credit exceeds a taxpayer’s income tax liability, any excess is paid directly, or refunded, to the taxpayer.

The Tax Reduction Act of 1975 created the first refundable tax credit, the Earned Income Tax Credit (EITC), on a temporary basis. The provision provided a 10 percent tax credit on up to $4,000 in earned income.[5] Uniquely, any credit in excess of the taxpayer’s tax liability would be refunded as a direct payment. While a novelty in the tax code, the EITC emerged from the debates of the 1960s related to negative income taxes, which contemplated delivery income assistance the tax system.[6]

The EITC was eventually made permanent and has been expanded over the following decades. The EITC remained the only refundable tax credit in the tax code until the enactment of the Taxpayer Relief Act of 1997.[7] The law created the Child Tax Credit (CTC), which, similar to the EITC, has been expanded by subsequent legislation in the decades following its enactment and remains a key element in prevailing tax policy debates. Indeed, in budgetary terms, the CTC is significantly larger than the EITC and affects substantially more taxpayers.

The Child Tax Credit

The Child Tax Credit was enacted in 1997 and made available for eligible taxpayers in 1998. The first iteration of the CTC was borne out of a growing, bipartisan interest in establishing a family-based tax benefit. During the George H.W. Bush Administration, a bipartisan commission on children recommended to the president the creation of a $1,000 refundable child tax credit.[8] Over the next several years, Congress and the subsequent Clinton Administration considered alternative approaches to the establishment of a child credit, which ultimately culminated in the enactment of the Taxpayer Relief Act of 1997 and the creation of the CTC.

The Act established a CTC of $500-per-child tax credit ($400 in 1998) available to married couples with incomes of up to $110,000 per year and single parents with income of up to $75,000. The credit phased out at a rate of $50 per $1000 in income above the respective thresholds. While technically the CTC was created as a refundable credit in 1997, refundability was highly limited to a fraction of otherwise eligible recipients. For taxpayers with three or more children, the credit could be refunded, up to the maximum value of the credit, if the taxpayers’ employee or self-employment payroll tax exceeded ETIC payments. For taxpayers with three children, however, the EITC more than compensates for payroll tax liabilities for taxpayers with low incomes. Accordingly, less than 1 million of the more than 26 million CTC claimants in 1999 received a payment for the refundable element of the CTC.[9]

While subsequent legislation in 1999 modestly expanded availability of the credit, it was not until the enactment of the Economic Growth and Tax Relief Reconciliation Act (EGTRRA) in 2001 that both the credit and the refundable element were materially expanded. Under EGTRRA, the credit was scheduled to gradually increase until reaching $1,000 by 2010, while the refundable component – the Additional Child Tax Credit (ACTC) – was substantially reformed. Under EGGTRA, families with earned income over $10,000 could receive the ACTC at a rate of 10 percent of income above the earned income threshold.

The architecture of the ACTC, notwithstanding the credit amounts and relevant income thresholds, has largely prevailed since the enactment of EGTRRA. In 2009, President Obama signed into law the first major legislation of his administration – the American Recovery and Reinvestment Act (ARRA). This “stimulus bill” reduced the earned income threshold to $3,000 for two years. Subsequent legislation, which broadly addressed the expiration of major tax and spending laws that created “fiscal cliffs,” made the $1,000 CTC permanent, as well as the reduced earned income threshold from ARRA.

The last major reforms to the CTC occurred in the 2017 Tax Cuts and Jobs Act (TCJA) and the American Rescue Plan Act (ARPA). The TCJA doubled the CTC to $2,000 and increased the refundable ACTC to $1,400. The law also indexed the ACTC to a measure of inflation, reduced the earned income threshold to $2,500, while significantly increasing the income phaseout thresholds from $75,000 for single filers and $110,000 for married couples to $200,000 and $400,000, respectively. The law also created a new $500 credit for dependents other than children.

Finally, ARPA significantly increased, for tax year 2021, the CTC to $3,000 per child, or $3,600 for any child under age 6, and creates a second phaseout whereby the increased CTC is reduced by $50 for every additional $1,000 in income over $75,000 for single parents, $112,000 for heads of households, and $150,000 for joint filers. The credit continues to be reduced to the current-law $2,000, at which point the current-law phaseout regime applies. ARPA also increases the age limit for qualifying children to include 17-year-olds.

ARPA made the credit fully refundable, such that individuals may receive the full value of the credit in excess of any tax liability and eliminated the $2,500 in earned income threshold to claim the refundable credit. Thus, a single parent with a 4-year-old and a 17-year-old, with $0 earned income, were eligible for a $6,600 benefit under ARPA. The law expired at the end of 2021, reverting the CTC to its pre-ARPA structure.

The Budgetary Treatment of Refundable Tax Credits

Identifying and determining the costs of tax expenditures is sufficiently challenging to produce divergent estimates from the Treasury and the JCT. Refundable credits introduce one additional element of complexity because such credits bridge both sides of the federal ledger. For the major individual refundable tax credits – specifically the EITC and the CTC – the relevant budgetary flows are substantial. According to JCT, 27 million taxpayers would claim the EITC in 2022 at a cost of $68.9 billion – comprised of $59.3 billion in outlays related to the refundable portion of the credit and $9.6 billion in revenue reductions.[10] The CTC has eclipsed the EITC in terms of its availability and budgetary effects, which JCT estimated would be claimed by 47 million taxpayers at a cost of $184.7 billion in 2022 -comprised of $105.3 billion in outlays and $79.4 billion in revenue losses.

Both congressional budget agencies and the executive branch record the refundable portion of refundable credits as spending or outlays. How the agencies present the components of a refundable tax credit differ depending on the forum. For instance, in its cost estimate for ARPA, the Congressional Budget Office (CBO) presented the revenue losses and the outlay increases associated with the CTC in separate sections of the estimate related to revenue and spending.[11] In other outlets, for example the JCT’s tax expenditure report, the combined revenue and outlay effects are listed for the relevant refundable tax credits and the outlay effects are noted in relevant footnotes. Nevertheless, the budgetary treatment of the refundable portion of these tax expenditures is consistently recorded as spending by Congress and the executive branch.

The Evolution of the Budgetary Treatment of Refundable Tax Credits

This budgetary treatment reflects longstanding practice that is nevertheless an evolution of the original approach to recording the costs of refundable tax credits. The modern budget process is only one year older than the EITC, which was the sole refundable tax code for the next 22 years, and somewhat evolved apace with the EITC.

The EITC was created under the Ford Administration and made permanent under the Carter Administration. Over this period, both the executive and legislative branches shifted their respective budgetary treatment of refundable tax credits at least once. Under the Ford Administration, the refundable portion of the EITC was initially treated as a further reduction in tax revenue in its 1976 budget.[12] The Ford Administration later revised the budgetary treatment of the refundable portion of the EITC and recorded it as an outlay or spending.[13] Yet in the Carter Administration’s first substantial budget document after assuming office – the July 1977 Mid-Session Review – the Office of Management and Budget (OMB) reversed course and recorded the refundable portion of the EITC as a further revenue reduction.[14] OMB’s then-Director Bert Lance couched this revised treatment as consistent and comparable with the then-prevailing congressional budgetary treatment for the EITC.[15]

Congress began to reconsider the budgetary treatment of the refundable portion of the EITC over the same time period. For the budget resolutions for fiscal years (FY) 1977 and 1978, the House of Representatives adopted the view that the refundable portion of the EITC should be recorded as spending, rather than a reduction in revenues. For the first 1979 budget resolution, however, the chambers reversed positions, with the House reverting to the view that the refundable credits should be scored entirely as revenue losses, while the Senate adopted the view that the refundable portion of the EITC should be recorded as an outlay.[16] Ultimately, in May 1978, upon agreeing to the conference agreement on the first concurrent budget resolution for FY 1979, the chambers adopted the then-Senate view, and agreed to record the refundable component of refundable tax credits as outlays.[17] The executive branch followed suit and adopted this convention soon after.[18] The practice is now memorialized in the executive branch’s primary budget preparation and execution document, Circular A-11.[19]

 Design Considerations in Refundable Tax Credits

 The first refundable tax credit, the EITC, emerged from a growing interest among federal policymakers to provide additional, reformed federal antipoverty measures. The credit was informed, as noted earlier, from debates related to a negative income tax, as well as concern with overall antipoverty and macroeconomic concerns. Among the considerations in the design of the EITC was relief against the payroll tax.[20] The payroll tax is generally considered a regressive tax, to the extent that the effective tax rate is inversely proportional to income – lower-income workers pay a greater share of payroll tax than do higher-income workers. Historical reforms to the Social Security system have also included payroll tax increases, which have increased the effective tax rates on lower-income workers under the payroll tax.[21]

The payroll tax creates a steep marginal effective tax rate for labor market entrants. In the absence of any other policy, the first dollar of income for new workers is taxed steeply. This disincentivizes the decision to work more so than if the payroll tax were progressively structured. The design of a flat payroll tax and a progressive rate structure was celebrated by President Roosevelt as a feature, rather than a bug, in the overall Social Security system. Such a structure, Roosevelt argued, notwithstanding the fundamental pay-as-you-go financing model, would convey an individual sense of ownership of future benefit payments.[22] This perception is now well-embedded in the policy debate surrounding Social Security generally.

The EITC was not, however, designed to fully mirror the payroll tax system. Indeed, offsetting the regressivity of the payroll tax was but one rationale for its creation. Accordingly, the evolution of the EITC has not been in lockstep with the payroll tax, and the degree to which the EITC offsets payroll tax liability varies by income and family size. In general, the EITC offsets the employee share of the payroll tax for parents (married or unmarried) with at least one child, while single, childless workers typically retain a net positive payroll (employee-side) of the payroll tax.[23] Nevertheless, the enactment of the EITC marked a change in federal policy in its conception of tax-related income support for low-income wage earners facing payroll tax liabilities.

The initial EITC was temporary but was expanded and made permanent in the Tax Revenue Act of 1978. The law was debated by Congress and enacted in the fall of that year. The law created a new mechanism for receiving the credit in advance. In debating the design for the advanceable EITC, the House and Senate settled on a mechanism, with some differences, of using employers as the nexus for providing the advance payments, either through reduced withholding or direct payment. While at the individual level the advance EITC would present somewhat differently from a lump-sum income tax refund, for scoring purposes the CBO, consistent with the revised budgetary treatment of refundable tax credits, considered any EITC payment over an individual income tax liability as an outlay.[24] This prevailed until the advance EITC was repealed in 2010, owing to low uptake and poor compliance, which was estimated to reduce outlays.[25]

Making Work Pay Credit

The EITC has been reformed and expanded since its initial enactment, but policymakers have pursued additional policies that grapple with the similar challenge of the regressivity of the payroll tax. In 2009, President Obama signed ARRA into law. Among the many provisions of the “stimulus act” was the creation of the Making Work Pay (MWP) credit. This provision was a temporary policy that provided a refundable tax credit in the amount of the lesser of either 6.2 percent of earned income up to $75,000 or $400 ($800 for married couples). The credit was paired with an expansion of the EITC, as well. The MWP credit, with a credit amount of 6.2 percent, was conspicuously tied to the employee side of the of Social Security’s Old-Age, Survivors, and Disability Insurance (or OASDI) payroll tax rate. Combined with the EITC, the MWP could fully offset this tax liability up to certain income limits, even for childless adults. The credit was otherwise structured as a traditional refundable credit, and was scored as reducing revenues, and increasing outlays to the extent that it exceeded income tax liability.[26]

 Temporary Payroll Tax Reduction

 At the end of 2010, Congress and the Obama Administration were confronted with the expiration of the prevailing income tax system, largely enacted under the Bush Administration, as well as temporary provisions from ARRA. In December 2010, Congress enacted the Tax Relief, Unemployment Insurance Reauthorization, And Job Creation Act of 2010. The law, among other provisions, largely extended the extant income tax code for two years. The MWP credit was allowed to expire, however. Somewhat in place of this policy, Congress enacted a two-year reduction in the employee-side Social Security tax, from 6.2 percent to 4.2 percent. At the time, the tax covered wages up to $106,800.[27] Under this policy, the Treasury issued additional debt securities to the Social Security trust fund to replenish forgone tax revenue. Unlike a payment to individuals, the mechanism for funding the trust fund does not give rise to outlays or spending but is rather considered an intragovernmental transfer. From a budgetary perspective, the policy scored as a pure payroll tax revenue loss.[28]

 Tax Reform Act of 2014

In 2014, the House Ways and Means Committee’s then-Chairman Dave Camp developed and released a comprehensive income tax reform plan.[29] Among the many reforms in the proposal was one of the EITC to more directly offset the regressivity of the payroll tax, while avoiding certain budgetary concerns that attend to similar policies. Historically, income support programs can take the form of direct payments, such as the Aid to Families with Dependent Children program, which was more commonly referred to as the federal “welfare” program. The budgetary treatment of such programs is straightforward – these are spending programs. Alternatively, income support programs can be delivered through the tax code, such as the EITC, which from a budgetary perspective, produces both revenue losses and spending.

The payroll tax cut of 2010 reflects yet another approach to providing cash assistance to workers, even if they have no positive income tax liability. As noted, this approach is recorded as a pure revenue loss, an apparently straightforward tax cut. The taxes being cut fund Social Security, however. Holding the Social Security trust fund harmless from the revenue loss was essential to the viability of the payroll cut and was accomplished through Treasury issuing additional debt to the trust fund. While this internal transfer ultimately is little different from simply borrowing the difference, it presents an alternative architecture for providing income support through the tax code. The EITC in the Tax Reform Act of 2014 (TRA2014) offers a third approach.

Among other reforms to the tax code, TRA2014 made substantive reforms to the EITC, but also restructured the policy to serve as a credit against the employee share of the payroll tax. All else equal, a payroll tax credit would reduce payroll revenue that funds the Social Security trust fund. Alternatively, the provision could be structured to mirror the payroll tax, but simply be refunded to taxpayers through an income tax refund, which is scored as spending. The TRA2014 reform to the EITC takes a unique, hybrid approach, from a budgetary perspective to providing payroll tax relief, however.

In addition to the substantive reforms to the EITC’s credit amounts and relevant income thresholds, the law created a payroll tax credit that is set as the equivalent of the reformed EITC. The provision is structured such that the payroll credit limits the reformed EITC amount on a dollar-for-dollar basis. Essentially, this structure swaps the EITC for payroll credit, which ordinarily would reduce payroll tax revenue. But the provision also leaves withholding unaffected, and directs that the credit be treated as an overpayment of payroll taxes. The tax code provides for special refunds for such overpayments to be treated as income tax refunds.[30] Income tax refunds score as revenue losses, rather than spending. The TRA2014 provision limited the credit to no more than the employee share of the payroll tax, leaving the employer share unaffected. To the extent that taxpayers were entitled to a further credit under the reformed EITC, they would receive the balance of the credit through the traditional mechanism, which would be recorded as an outlay.[31] From a budgetary perspective, this somewhat swaps the outlays associated with the EITC with revenue losses. The JCT found that the EITC reforms in TRA2014 would reduce revenues by $160.8 billion over 2014–2023 and reduce outlays by $378.0 billion over 2014–2023.[32]

Alternative Scoring Methodology and the Child Tax Credit

When first enacted in 1975, the EITC as a refundable tax credit was a policy novelty. It took several years for the legislative and executive branches to develop a consistent budgetary treatment of refundable tax credits. The current prevailing treatment recognizes the excess credit above income tax liability as spending. The classification of budgetary flows as related to revenue or outlays can meaningfully alter how those flows are perceived. This process establishes the dividing line between characterizing a policy change as a “tax cut” or a “spending increase,” with all the attendant assumptions those terms convey. The budgetary implications of any tax policy are typically viewed to some degree through this dichotomy.

Fundamentally, to provide income support to workers beyond their income tax liability requires a commitment of public resources and a mechanism for delivering them. Whether there is a meaningful distinction between delivering a policy through the tax code or through an explicit spending program is the subject of some debate and study.[33] Indeed, past American Action Forum research has sought to place income support programs entirely on the spending side of the budgetary ledger.[34] These considerations animate the policy debate surrounding the child tax credit.

There remains interest among some policymakers to extend recent expansions in the CTC. JCT estimated that extending the ARPA provisions through 2025 would increase the deficit by $556.008 billion over the budget window.[35] Of this amount, $421.484 billion was an increase in outlays arising from the refundable portion of the CTC given the prevailing structure and historic budgetary treatment of refundable tax credits. By this convention, the CTC expansion would create an expanded tax subsidy that is largely treated as spending.

To the extent that policymakers sought to recharacterize these flows, there exist traditional mechanisms for doing so. Specifically, and this approach reflects the evolution of the scoring of the EITC, Congress could in the same way as was done in the 1979 budget resolution specify a preferred budgetary treatment of the CTC or refundable tax credits generally. It could specify that some or all of the refundable portion should be characterized as a revenue loss. Alternatively, Congress could structure any expanded CTC as a payroll tax credit. To the extent a payroll tax credit reduces Social Security trust fund receipts, Congress would likely, as was the case in 2010, pair such a credit with general fund transfers to the Social Security trust fund. Finally, policymakers could structure a child tax credit as payroll tax credit, but provide a mechanism similar to that found in the TRA2014 EITC proposal to reimburse taxpayers. This approach would likely recharacterize some, but not all, of the outlays associated with a CTC expansion.


Policymakers grappling with providing payroll tax relief to lower-income workers must navigate the budgetary treatment of refundable tax credits. Recent policy developments highlight the alternative design approaches to this shared policy goal. The budgetary implications for these alternative approaches can vary significantly, from the current, conventional approach, which relies on spending to offset payroll liabilities, to a purely revenue-related approach, such as a payroll tax cut and related general fund transfer. Alternatively, policymakers could shift some or all of these budgetary flows from outlays to revenue losses (or the reverse) in the design of the policy.


[2] Robert Haig, The Concept of Income–Economic and Legal Aspectsin The Federal Income Tax 1-28 (Columbia University Press, 1921). See also Henry Simons, Personal Income Taxation: the Definition of Income as a Problem of Fiscal Policy 49 (University of Chicago Press, 1938).





[7] Ibid.


[9] Ibid.




[13] Office of Management and Budget, Mid-session Review of the 1977 Budget (July 1976)

[14] Office of Management and Budget, Mid-session Review of the 1978 Budget (July 1977)

[15] The 1977 Midyear Review of the Economy, Hearings Before the Joint Economic Committee, Congress of the United States, (July 1977)


[17] GPO-CRECB-1978-pt10-9-2.pdf (

[18] U.S. Department of the Treasury, Monthly Treasury Statement (September 1978)






[24] srpt95-1263.pdf (






[30] 26 U.S.C. § 6413(c)(1)