March 18, 2021

Four Questions Treasury Must Answer About the State Tax Cut Prohibition in the American Rescue Plan Act

Key Findings

  • The American Rescue Plan Act’s restriction on states’ Fiscal Recovery Funds being used to directly or indirectly offset a net tax cut is vague and raises difficult questions of interpretation and application. A broad interpretation of this prohibition may be unconstitutional.
  • This restriction potentially implicates a wide range of tax decisions, not just rate reductions but also federal tax conformity, excluding unemployment benefits from income taxation during the pandemic, adjusting standard deductions, or awarding discretionary tax incentives.
  • States will require answers to many questions, but particularly: (1) what constitutes a net tax reduction? (2) how is a net tax reduction determined to have resulted from a policy change? (3) which potential expenditures could be deemed to create fiscal capacity for a net tax cut? and (4) how would offsetting a tax reduction be defined, especially across multiple years?
  • U.S. Department of Treasury guidance will be crucial as states seek to navigate this new environment.


States are set to receive $195.3 billion in fiscal relief under the American Rescue Plan Act (ARPA), equivalent to 20 percent of the annual tax collections of state governments.[1] With state revenues essentially flat in 2020 (a net decline of less than 0.2 percent), the greatest challenge for states may be figuring out what to do with it. The State and Local Fiscal Recovery Funds can only be used for certain enumerated types of expenditures, and they specifically cannot be used to cut state taxes (there is no similar prohibition for localities), either directly or indirectly, or for deposits into pension funds.[2]

That prohibition on indirectly offsetting a state tax cut is extremely vague and potentially quite expansive, and it has created significant consternation in state capitols. A Treasury spokesperson stipulated on Wednesday that the provision does not prohibit states from enacting tax cuts so long as those reduction do not rely on federal aid.[3] This may signal that Treasury guidance will construe the provision narrowly. If this assurance is to mean anything, however, Treasury will need to answer several important questions, outlined in this paper.

The restriction on the state tax cuts is brief and vague, but potentially quite broad:

“A State or territory shall not use the funds provided under this section or transferred pursuant to section 603(c)(4) to either directly or indirectly offset a reduction in the net tax revenue of such State or territory resulting from a change in law, regulation, or administrative interpretation during the covered period that reduces any tax (by providing for a reduction in a rate, a rebate, a deduction, a credit, or otherwise) or delays the imposition of any tax or tax increase.”[4]

At the same time, however, the list of expenditures to which states can put federal aid is relatively short, particularly given that most states have either no revenue shortfall or only a modest one, and that many have struggled to spend the $150 billion in Coronavirus Relief Fund assistance appropriated under the CARES Act for economic and public health responses to the pandemic, a purpose to which state Fiscal Recovery Funds can also be dedicated. States may use this latest aid to:

  1. Respond to the public health emergency or its negative economic impacts, which includes aid to households, businesses, and impacted industries, and is likely to include covering the compensation of state health and public safety officials and current unemployment benefit claims, consistent with Treasury’s guidance on the Coronavirus Relief Fund, which was dedicated to similar purposes;
  2. Supplement the pay of essential workers;
  3. Pay for general government services to the extent of any pandemic-induced revenue losses in the most recent full fiscal year; and
  4. Make necessary investments in water, sewer, or broadband infrastructure.

Notably, assuming that this list is all-encompassing, states would be prohibited from using the funding to cover general operating expenses beyond what is necessary to backfill revenue losses. They cannot use the money to increase their general fund expenditures, nor, presumably, could they deposit the aid in a rainy day fund or use it to replenish their depleted unemployment compensation trust funds. And they certainly cannot use it on tax cuts (or pension plans, where deposits are also prohibited).

The restriction on direct use to facilitate a tax cut is largely uncontroversial. While some policymakers might have contemplated giving away the state aid in the form of temporary tax cuts absent the restriction, the federal government’s prohibition is reasonable and clearly within its authority to impose.

Because money is fungible, however, it is very difficult to be sure what sort of uses of state aid might be interpreted as indirectly offsetting a net tax cut, even if the state had the resources to cut taxes in the absence of the federal assistance. Violation of the provision would result in the U.S. Department of the Treasury recouping the funds, previously transferred to a state, which were deemed to have been instrumental in facilitating the tax cut.

A great deal is at stake. At a theoretical level, the principle of fiscal federalism is implicated here, as a broad interpretation of the federal prohibition would represent an unprecedented degree of state entanglement in state fiscal policy, using federal dollars to dictate state policy in a way that vastly exceeds what has been attempted in the past. And at a practical level, the potential field of preemption is vast, not just what most people might think of as tax cuts.

What might run afoul of the prohibition? A non-exhaustive list would include states acting to:

  • Inflation-adjust their standard deduction or personal exemption;
  • Expand the earned income tax credit;
  • Tinker with sales tax exemptions;
  • Adjust tax incentives;
  • Follow the federal government’s lead in excluding $10,200 of unemployment compensation from taxation;
  • Offset local property tax burdens;
  • Conform to the federal government’s revised treatment of forgiven loans under the Paycheck Protection Program; or
  • Take administrative action to stave off a large unemployment insurance tax increase that might otherwise be triggered by the insolvency of their trust funds.

It is, therefore, vital that the federal government provides greater clarity on what does and doesn’t constitute indirectly offsetting a net tax cut.

Four Questions

Treasury guidance should provide greater clarity for states. Particularly, states would benefit from answers to the following questions:

1. What constitutes a net tax reduction?

The bill language prohibits offsetting a “reduction in the net tax revenue of such State or territory.” No baseline is specified. If a current policy baseline is intended, then any provision that would result in a reduction in revenue compared to the continuation of the status quo would be subject to the restriction, even if state revenues increased.

Imagine, for instance, that state revenues were projected to rise $500 million and the state increased a deduction for low-income filers, resulting in an increase in only $450 million. Because this is lower than the amount that would have been raised in the absence of a policy change, is this a net tax cut (the current policy baseline position), or because net tax revenue is higher than in the previous year, is it outside the prohibition?

Since the intention of the restriction is to prevent the recovery funds from facilitating a tax cut, using a collections baseline is eminently more sensible. Clearly, in the above scenario, the tax reduction was facilitated through state revenue growth, not any assistance provided by the federal government. But while this seems clearly right as a matter of policy and intent, it is not as clear as it might be in the language, even though the inclusion of the word “net”—in addition, presumably, to allowing changes in multiple taxes to offset each other—does suggest it. Treasury should make the logic explicit and provide important clarity for states on this point.

2. How is a net tax reduction determined to have resulted from a policy change?

This may seem like a foolish question, and in some instances it would be. The law speaks to reductions in net tax revenue “resulting from a change in law, regulation, or administrative interpretation during the covered period [March 3, 2021 through December 31, 2024] that reduces any tax (by providing for a reduction in a rate, a rebate, a deduction, a credit, or otherwise) or delays the imposition of any tax or tax increase.” This raises two questions, one of measurement and one of timing.

On measurement, how does Treasury propose to determine exactly how much revenue was forgone by the implementation of a tax reduction? States make estimates, of course, both before and after, but how much of any change in overall state revenues came from a tax reduction cannot be known with precision. How does Treasury plan to calculate it for purposes of deciding whether to claw back state aid, and how much?

On timing, the question is, what constitutes a change in law, regulation, or administrative interpretation during the covered period. If a state adopted a multiyear phasedown of a tax rate prior to the covered period, set to continue in the coming years, would the implementation of these reduced rates—provided for in law, prospectively—constitute a change in law? If a revenue agency had to sign off on the reduction by confirming that certain preestablished conditions had been met, but lacked discretion beyond that administrative determination, would this count as a regulation or administrative interpretation? What governs in determining what took place prior to the covered period and what takes place within it, current law or current policy?

Or if a state has rolling conformity with the Internal Revenue Code and the federal government amends its provisions in a way that might generate a net tax cut—say, by excluding a portion of unemployment benefits from tax (a provision to which many states conform) or by expanding the child tax credit (which has some limited mirroring in state laws)—does the automatic update to these new provisions implicate the provision? This would be particularly ironic, given that ARPA both threatens to claw back state aid if states implement net tax cuts (at least under some circumstances) and provides for tax reductions, within the covered period, to which some states automatically conform.

Again, relying on the presumption that the prohibition is only intended to apply to tax reductions states could not undertake with their own resources, future changes already enacted or contingently provided for under existing law should not come under the restrictions spelled out in ARPA. This too, however, is ambiguous, and Treasury should clarify this point.

3. Which potential expenditures could be deemed to create fiscal capacity for a net tax cut?

It is easy to imagine indirectly offsetting a tax cut in a way that Congress might legitimately wish to proscribe. If these new federal relief dollars cannot be directly funneled into tax relief, it is rational that Congress would not want states to be able to play a shell game by which they use federal aid to cover expenses in the general operating budget, using the state revenue freed up by this infusion to cut taxes. Gimmicks like this were undoubtedly at the heart of the restriction not only on direct but also on indirect offsetting of state tax cuts, and guardrails against this activity may be appropriate—and are certainly obligatory under the legislation.

But with such broad and vague language, it is equally easy to imagine scenarios in which the mere concept of the fungibility of money is sufficient to trigger the restriction, even where federal aid did not create the fiscal capacity for a tax change.

It matters, therefore, what states can spend the money on, and which of these expenses might reasonably be interpreted as offsetting any policy a state may adopt resulting in a reduction in net tax revenue. Is the list of four categories of expenditures exhaustive? It would seem to be. Will those categories be interpreted expansively? That remains to be seen. Nothing in the language of the provision would seem to allow depositing money into any sort of fund for future use, for instance, yet Congress expressly prohibits deposits into pension funds. Does that mean that other deposits, perhaps into a rainy day fund or an unemployment compensation trust fund, might be permissible as responses to the negative effects of the pandemic, particularly if those funds were depleted in the past year?

Some officials, moreover, seem to regard this relief as their ticket for funding long-languishing transportation projects or growing the state’s annual budget. It is hard to understand how these expenditures could meet with approval under the text of the bill, but some, at least, are hoping for a remarkably generous Treasury ruling on this question.

Assume, for now, that the enumerated expenses are the only ones on which this federal aid can be spent, and that the most straightforward interpretations (if such a thing is possible) are adopted. The only federal aid dollars that can be clawed back are those which are unused by 2024 or are put to an impermissible use, like offsetting a tax cut or spending in a way not authorized under the bill.

It seems implausible that a new round of grants to businesses or individuals, permitted in the first bucket (responding to the public health emergency and its negative economic effects), could be understood as offsetting a tax cut, since such grants are not ordinary governmental expenses or part of a general operating budget. They are supplemental to the state’s ordinary activities and specific to the coronavirus crisis, so the expenditure creates no fiscal capacity for a tax cut, provided the state doesn’t use this aid to allow a reduction in the state’s general outlays. (How would this be determined?) Presumably aid money spent this way could not be recouped if a state implemented a net tax cut because, by definition, it did not contribute to it (even indirectly). States would appreciate having this assurance from Treasury.

The same can likely be said of supplemental wages to private sector essential workers. This is not an expenditure currently found in any state budget, so offering this assistance with federal aid doesn’t free up a single dollar for tax cuts. It would be extremely surprising if aid spent this way could be recaptured.

The limited set of infrastructure projects authorized under the bill—water, sewer, and broadband—pose a more interesting question. Infrastructure projects usually come out of the capital budget, which is distinct from the general fund budget. Some taxes flow directly to capital projects, while most general taxes go to the general fund, which is for operating expenses. (Terminology sometimes varies across states, but the concepts are the same.) A state might already have, say, a rural broadband project in the works, and could fund it with these aid dollars. That frees up money in the capital budget, but not the general fund budget. Were a state to cut a tax that flows into the general fund (like, usually, income or sales taxes), could that lead to the recoupment of aid which created additional capacity in capital accounts? (Presumably it would for taxes that flow to capital projects, like, for instance, the motor fuel tax.)

States can theoretically shift money from the general fund to the capital budget, or divert tax revenues from general taxes there, and sometimes do. Does the mere possibility of a future transfer govern, or would a state actually have to do so during the covered period to create a situation in which an authorized infrastructure investment could be deemed to offset a net tax cut?

That leaves two remaining allowable expenses. The first is a subset of a prior category. Assuming states can cover payroll for public health and public safety officials, this reduces state outlays in a way that creates fiscal capacity—money they would have budgeted otherwise, but which is now available. This might well be interpreted as providing an indirect offset for a net tax cut. The second is the final broad category, but one only available to some states: the relief can be used to cover general government expenditures to the extent of revenue losses arising from the pandemic, and if a state turned around and provided a net tax cut at the same time, that would almost certainly violate the prohibition.

However, that the above seems logically consistent with the provision, and maybe even necessary for making it logically coherent, does not mean that such an interpretation is assured. Treasury should provide clear guidance on which expenses can and cannot be deemed to contribute to offsetting a tax cut.

4. How would offsetting a tax reduction be defined, especially across multiple years?

Imagine that a state experienced a revenue loss of $200 million due to the economic effects of the COVID-19 pandemic, a loss it is permitted to offset with the Recovery Funds. (This raises an additional question: how does Treasury propose to determine what portion of any revenue losses in the relevant fiscal year are attributable to the coronavirus crisis?) In subsequent years, however, this state experiences substantial growth, with revenues rising $400 million the next year and a further $100 million the year after that. The revenue baseline is now $300 million higher than it was pre-pandemic. If the state were to cut taxes by $100 million per year after that, this would still leave it with revenues above not only its pandemic-level receipts but also its pre-pandemic high. The $500 million in revenue growth is more than enough to facilitate the tax cut—if the state prioritizes it over government growth—whether or not the state had ever received that $200 million to offset pandemic losses.

Money, however, is fungible. If the state never received that aid, it would still have enough money to cut taxes without reducing revenues, but in that case, it would have had less to set aside for additional government expenditures or to set aside for a rainy day. Does the original $200 million get clawed back?

And if it has the potential to—contingent on whether this is even a net tax cut (see the first question)—does it matter if the expenditure and the cut were in different years? In other words, can federal aid used in one year be deemed to offset a tax cut made in another, or is each year (or possibly, in states with biennial budgets, each budget cycle) discrete?

Each of these questions and many more deserve to be answered. The constitutionality of the provision, moreover, may well hinge on how these ambiguities are resolved. The federal government undeniably has the power to impose certain conditions on the direct expenditure of the funds it provides states, but it does not possess the blanket authority to commandeer state policy as a condition of receiving federal funds. States unambiguously have the power to establish their own revenue levels. It would certainly raise red flags, for instance, if the federal government withheld Medicaid funding from states which legalized marijuana or tried to dictate state immigration policy via Community Development Block Grants. Such provisions would be presumptively unconstitutional.

Clawing back federal aid under a broad interpretation of indirect offsets to tax cuts would run into the same constitutional problems. If funding is contingent on states accepting a constraint on their policymaking ability, that constraint must be unambiguously related to the federal interest served by the funding, and not be so coercive as to reach the point where pressure turns into compulsion, under the precedent set in South Dakota v. Dole (1987).[5] In that case, the Supreme Court approved a federal provision that made 5 percent of federal highway funding contingent on adopting 21 as the drinking age, but cautioned that if significantly more of the funding had been predicated on it, that would have been unduly coercive.

In National Federation of Independent Business v. Sebelius (2012),[6] meanwhile, the Court held that the federal government could not make Medicaid funding contingent on a state’s willingness to implement Medicaid expansion. Although that case is a maze of plurality opinions, this particular holding was considerably less controversial, with three justices in the 5-4 majority upholding the Affordable Care Act agreeing that this provision was unconstitutional. A broad policy tying states’ hands for three years, therefore, is constitutionally dubious at best, while a narrow interpretation—a prohibition on directly funding a tax cut, which would be inconsistent with Congress’s purpose, and some narrow guardrails to avoid shell games—would likely stand.


Given the narrow range of possible uses of the federal aid that could reasonably be interpreted as offsetting a tax cut, many states might be at minimal risk regardless of the tax policy choices they make. However, the vague language of this prohibition has every state nervous and unsure of what it can do.

Can California, which suspended certain structural deductions within its tax code when it anticipated a massive revenue shortfall, reverse those changes now that it has a $26 billion surplus without incurring a reduction in its federal aid? Can governors dip into their “opportunity funds” to provide discretionary tax incentive packages to businesses, or does that result in a dollar-for-dollar recapture? Does the decision to conform to the provisions of ARPA itself potentially run afoul of its tax provision by reducing taxes on unemployment benefits?

States need answers to these questions, and they need them soon.

[1] For state allocations and revenue gains or losses, see Jared Walczak, “State Aid in American Rescue Plan Act Is 116 Times States’ Revenue Losses,” Tax Foundation, Mar. 3, 2021,

[2] H.R. 1319 (2021), Section 9901.

[3] David A. Lieb, “Treasury says state tax cuts OK if separated from virus aid,” Associated Press, Mar. 17, 2021,

[4] H.R. 1319 (2021), Section 9901.

[5] South Dakota v. Dole, 483 U.S. 203 (1987).

[6] National Federation of Independent Business v. Sebelius, 567 U.S. 519 (2012).

Banner image attribution: nickjene, Adobe Stock

Was this page helpful to you?


Thank You!

The Tax Foundation works hard to provide insightful tax policy analysis. Our work depends on support from members of the public like you. Would you consider contributing to our work?

Contribute to the Tax Foundation

A sales tax is levied on retail sales of goods and services and, ideally, should apply to all final consumption with few exemptions. Many governments exempt goods like groceries; base broadening, such as including groceries, could keep rates lower. A sales tax should exempt business-to-business transactions which, when taxed, cause tax pyramiding.

A tax credit is a provision that reduces a taxpayer’s final tax bill, dollar-for-dollar. A tax credit differs from deductions and exemptions, which reduce taxable income, rather than the taxpayer’s tax bill directly.