Treasury's Noble Effort to Save the Net Tax Revenue Provision

It's an eminently reasonable interpretation of the American Rescue Plan's ambiguous text. That doesn't necessarily mean it will survive in the courts, but it ought to

The Treasury Department today released its much-anticipated regulations fleshing out the state tax cut limitation contained in March’s $1.9 trillion stimulus package.¹ On a scale from 1 to 10, with 1 being “we’re throwing up a white flag and giving up on enforcing the tax-cut limitation,” and 10 being “we’re not going to let states cut taxes if they take stimulus money,” this is approximately a 3. I think Treasury is making the right call in that regard: an overly aggressive interpretation of the state tax cut limitation almost certainly would have lost in court. Treasury faced a tradeoff between (a) actually trying to ensure that stimulus dollars don’t go to finance state tax cuts and (b) running a constitutional-law suicide mission. It does about as much as possible to accomplish the former goal while still preserving some chance of surviving judicial review.

This post provides an overview of the statutory framework and the new Treasury regulations, outlines the key legal challenges facing Treasury’s rule, and explains why I think Treasury ought to win this one (though I don’t think anyone can confidently predict that Treasury will prevail).

The statutory framework

The American Rescue Plan Act, signed into law by President Biden on March 11, adds a new section 602 to the Social Security Act that appropriates $219.8 billion for payments to states, territories, and tribal governments. (Section 602 is a relatively small piece of the $1.9 trillion total—though really big relative to the $700 billion or so of federal aid that state governments could expect to see in a pre-pandemic year.)

The new section 602 specifies four permissible uses of these funds (with the money remaining available through the end of 2024):

1. To respond to the COVID-19 “public health emergency” or “its negative economic impacts, including assistance to households, small businesses, and nonprofits, or aid to impacted industries such as tourism, travel, and hospitality”;

2. To provide premium pay to essential workers or grants to employers of essential workers;

3. To make up for reductions in revenue resulting from the COVID-19 crisis; and

4. To invest in water, sewer, or broadband infrastructure.

In addition, and most importantly for present purposes, the statute includes a subparagraph (section 602(c)(1)(A)) that reads:

A State or territory shall not use the funds provided under this section … 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.

That language has come to be known as the “net tax revenue” provision (or, in conservative legal circles, the “tax mandate”).

The statute requires states to certify compliance with the net tax revenue provision periodically. Moreover, it provides for recoupment of any funds that a state uses to offset a net tax revenue reduction. And it delegates authority to the Treasury Secretary to “issue such regulations as may be necessary or appropriate to carry out” the new section 602.

Treasury’s interim final rule

The Treasury Department issued those regulations today in the form of an interim final rule. An interim final rule takes effect immediately, but Treasury has opened a 60-day public comment period and will consider whether to revise the rule on the basis of those comments. (Presumably any revision will go only in the direction of making the rule more lenient: after states accept funds, Treasury would run into legal trouble if it tried to tighten strings ex post. My understanding is that folks at Treasury “get” this point, and that any changes we might see in the future will be in a state-favorable direction.)

Treasury’s interim final rule establishes a five-step framework for assessing compliance with the net tax revenue provision:

Step 1: Has there been a “covered change” in state tax law (legislative or regulatory) that reduces state tax revenue? Note that income tax changes that simply conform to corresponding federal income tax changes (e.g., excluding the first $10,200 of unemployment insurance from adjusted gross income) won’t count as covered changes. If a state has made a revenue-reducing covered change, then it must self-assess the amount of revenue that it will lose as a result. That amount is the maximum that the state will need to “pay for” through recoupment. If the state hasn’t made a revenue-reducing covered change, then it will be in full compliance with the net tax revenue provision, though it still has to run through the other steps.

Step 2: Calculate the state’s baseline for the relevant reporting year. The baseline is the state’s fiscal year 2019 tax revenue indexed for inflation using the Bureau of Economic Analysis’s implicit price deflator. If revenue loss from revenue-reducing covered changes is less than 1 percent of the baseline, then a state will be found in full compliance.

Step 3: Compare actual revenue to the baseline for the relevant reporting year. If actual revenue is greater than the baseline, then a state is in full compliance—even if revenue-reducing covered changes exceed 1 percent of baseline. (This could be the case if, e.g., organic revenue growth offsets revenue-reducing covered changes.)

Step 4: Identify tax increases and cuts in own-source spending that offset revenue-reducing covered changes. This is where things get trickier. If revenue-reducing covered changes exceed 1 percent of the baseline and a state’s revenue comes in under baseline, the state needs to show that it’s making up for the revenue-reducing covered changes without relying on American Rescue Plan funds. It can do this by totaling up the following:

• increases in tax revenue resulting from other tax law changes;

• spending cuts in departments or agencies that aren’t relying on section 602 funds;

• spending cuts in department or agencies that are relying on section 602 funds, minus the amount of section 602 funds used in that department or agency.

Step 5: If (a) a state has made revenue-reducing covered changes that exceed 1 percent of the baseline AND (b) actual revenue is less than the baseline AND (c) the state can’t identify tax increases and spending cuts that offset revenue-reducing covered changes, then the state is subject to recoupment. The recoupment amount is the lesser of (a) the actual revenue shortfall relative to the baseline or (b) the amount of revenue-reducing covered changes not offset by tax increases or spending cuts.

Immediate effect on pending lawsuits

The first thing to say about Treasury’s interim final rule is that it’s much more lenient than Republican state officials feared.

For example, Ohio—which was first out of the gate in March with a lawsuit challenging the net tax revenue provision as unconstitutional—argued that “every change in tax policy that leads to a decrease in tax revenue violates the Tax Mandate.” That’s certainly not how Treasury sees it. Under the interim final rule, all manner of state tax cuts can be compliant with the net tax revenue requirement as long as they (a) don’t reduce revenue by more than 1 percent relative to the inflation-adjusted FY 2019 baseline or (b) aren’t offset by own-source spending cuts within the department or agency in which section 602 funds are spent.

Treasury’s rule also addresses—and largely moots—the concerns of some states who argued that the net tax revenue was fatally ambiguous. For example, West Virginia and a dozen other states argued in lawsuit last month that they “have no ex ante assurance of what they must do to comply” with the net tax revenue provision. But now with Treasury’s interim final rule, it’s pretty clear what specific steps states can take to avoid recoupment.

Indeed, today’s rule casts doubt on whether some of the states suing Treasury over the net tax revenue provision even have standing to be in court. To establish standing, a plaintiff in federal court must show that it has suffered an “actual or imminent” injury. But if a state won’t run afoul of the Treasury rule, and so won’t be subject to recoupment, then it’s hard to see how it has been injured by the net tax revenue provision. (The Ohio House, for example, has passed a tax cut, but the House speaker says that “[w]e’ve funded our tax cut with state dollars,” in which case Ohio wouldn’t have reason to worry about—or sue over—Treasury’s rule.)

Maybe states will argue that the theoretical possibility of recoupment is enough to establish standing, though—at least based on the court’s past standing decisions—that would be quite a stretch. At the very least, today’s rule means that the states suing Treasury will have to revise their arguments substantially.

The long run

Unfortunately, although Treasury’s interim final rule strikes me as eminently reasonable, I don’t think that today’s rule means states will all drop their lawsuits. (In addition to the Ohio and West Virginia-led lawsuits, Arizona filed a suit of its own in April, and Texas filed a lawsuit on behalf of itself, Louisiana, and Mississippi last week.) Will any of those states succeed in getting a court to strike down the net tax revenue provision?

That’s the multibillion-dollar question. I see four major buckets of issues:

1. Pennhurst/Chevron. In a 1981 case, Pennhurst State School & Hospital v. Halderman, the Supreme Court held that “if Congress intends to impose a condition on the grant of federal moneys” to states, “it must do so unambiguously.” On its own, and prior to Treasury’s interim final rule, the net tax revenue provision very likely violated Pennhurst: states really had no idea what would qualify as a reduction in net tax revenue (relative to what baseline?), and they would have had no clue what counted as an “indirect[] offset.” The question now is whether Treasury’s rule cures the Pennhurst problem.

The Supreme Court’s 1984 Chevron decision holds that when Congress explicitly or implicitly delegates authority to an agency to interpret an ambiguous law, federal judges should defer to the agency’s view as long as the agency has adopted a reasonable interpretation of the statutory text. (Courts have carved out exceptions to Chevron’s general rule that are too numerous to list here.) One not-quite-resolved question in constitutional law is whether an agency can invoke this interpretive authority in the context of an ambiguous federal spending condition in order to supply the clarity that Pennhurst demands. (I think there’s a reasonable argument that the Court’s 1985 decision in Bennett v. Kentucky Department of Education implicitly answers this question in the affirmative, but it’s hard to find language in Bennett that totally resolves the issue.)

In an excellent (and now highly relevant) essay in the Yale Law Journal in 2001, George Washington University law professor Peter Smith set out the terms of this debate:

Resolution of [the Pennhurst/Chevron] question depends principally on how one characterizes Pennhurst’s clear-statement rule. The first approach, which I call the “accountability model,” … treats Pennhurst’s rule as a structural mechanism to ensure congressional accountability when Congress imposes burdens on the states. Under this framework, Congress (which, at least theoretically, represents the interests of the states), in contradistinction to administrative agencies (which do not), must unambiguously decide whether to impose a particular burden on the states. The accountability model thus expresses a concern about delegation: Congress, and not agencies, should make important decisions of policy, including whether to alter the federal-state balance, because only Congress is electorally accountable for such decisions. Accordingly, under the accountability model, reasonable agency interpretations of statutory grant conditions are not entitled to deference ….

The second approach, which I call the “state choice model,” views Pennhurst’s rule as a means to ensure notice—and thus fairness—to the states when a federal grant program imposes a burden on the state recipients. Under this account, the question of which federal actor (that is, Congress or the agency) has imposed a condition on the state is not determinative; the inquiry focuses instead on whether, in light of the information available when the state accepted federal funds, the state can fairly be said to have understood the nature of the bargain, and thus had the opportunity “freely” to "choose" whether to accept the funds. Because a state can just as readily “ascertain what is expected of it” from an agency regulation as from the statutory text itself, the state choice model accords Chevron deference to reasonable agency interpretations of statutory grant conditions.

In a nutshell: If the accountability model is the right interpretation of Pennhurst, then Treasury’s interim final rule doesn’t save the net tax revenue provision. If the state choice model is the right interpretation, then it potentially does.

Smith makes a compelling argument that the state choice model is the better interpretation of Pennhurst (see especially pp. 1214-1221 of his article). As a doctrinal matter, the state choice model better aligns with the Supreme Court’s explanation for the Pennhurst rule in the first place. According to Justice Rehnquist’s majority opinion in that case: “The legitimacy of Congress’ power to legislate under the spending power … rests on whether the State voluntarily and knowingly accepts the terms of the ‘contract.’” Since Treasury’s interim final rule is coming before the disbursement of American Rescue Plan funds, states have the opportunity to make the voluntary and knowing choice that Pennhurst requires.

Smith also makes a compelling case that the state choice model provides a normatively more attractive interpretation of Pennhurst. As Smith emphasizes, requiring Congress to resolve all potential ambiguities in statutory text would “impose[] an unrealistic burden of legislative specificity on Congress.” Moreover, allowing agencies to fill in gaps via regulation allows them “to bring their expertise to bear on problems that they are charged with addressing.”

I’d add to that one more point in favor of the state choice model: Allowing agencies to supply via regulation the ex-ante clarity that Pennhurst requires will likely serve the ultimate interests of states. If Congress must resolve all ambiguities itself in the statutory text—an impossibly high bar in many contexts—Congress will effectively face a choice between giving states a blank check and giving them nothing at all. States will probably receive more federal funding in the long run if Congress can task agencies with fleshing out the details of conditional spending programs.

To be sure, the most relevant question at this point is not whether I think Treasury’s interim final rule satisfies the Pennhurst test but whether the federal courts will think it does. If the issue makes its way to the Supreme Court (as is possible though not certain), then Treasury will encounter a conservative majority that is quite skeptical of the Chevron doctrine. And although the court might not be ready to overrule Chevron outright, it probably will be reluctant to extend Chevron’s domain any further. So Treasury faces an uphill battle on the question of whether Chevron can, in effect, trump Pennhurst (though it has done itself quite a favor by adopting a reasonable and quite permissive interpretation of the net tax revenue requirement).

2. Inflation. Even if a court concludes that an agency regulation potentially can supply the clarity that Pennhurst requires, there remains the question of whether Treasury’s rule survives Chevron scrutiny. One of the possible vulnerabilities, I think, is Treasury’s use of an inflation-adjusted baseline to judge compliance.

Recall again that under section 602(c)(1)(A), a state cannot use American Rescue Plan funds to “offset a reduction in the net tax revenue of such State … resulting from a change” in tax law. If there has been a change in tax law but no reduction in net tax revenue, then the limitation doesn’t bind. But does “reduction in net tax revenue” mean a reduction in nominal revenue or a reduction in real (inflation-adjusted) revenue?

My own view is that (a) the statute is ambiguous, (b) either the nominal-revenue or real-revenue reading is reasonable, and so (c) courts ought to defer under Chevron to Treasury’s choice to use inflation-adjusted revenue. [Aside: I’ll admit that there’s some tension between this view and an argument that now-National Economic Council deputy director David Kamin and I made when Trump administration officials were considering a Treasury regulation that would inflation-index the basis of property for capital gains tax purposes. We said “cost” in section 1012 of the Internal Revenue Code couldn’t reasonably be read by Treasury to mean inflation-adjusted cost. I think the tension is resolvable based on differences in legislative history and statutory structure—there’s really strong evidence that Congress didn’t understand itself to be giving Treasury carte blanche to index basis for inflation, whereas there’s really strong evidence that Congress did understand itself to be delegating authority to Treasury to fill in the details of the net tax revenue provision. But it’s a good reminder that with changes in administration, what goes around inevitably comes around.]

The question of whether the FY 2019 baseline is or isn’t inflation-adjusted might seem highly technical, but it’s a really big deal. At inflation of 2 percent per year, the difference between a real and nominal baseline would be more than 10 percent by FY 2024. A state would have to go full Grover Norquist for its FY 2024 revenue to be below the FY 2019 baseline in nominal terms. Without adjusting the baseline for inflation, the net tax revenue provision will be nonbinding as a practical matter by the later part of the period covered by the American Rescue Plan.

3. Granularity. Another possible vulnerability is Treasury’s aggregation of spending cuts and section 602 funds at the level of the state department, agency, or authority. Imagine, for example, that a state pays for after-school sports programs through its education department. If the state (a) cuts taxes by $50 million, (b) cuts after-school sports spending by $50 million, and (c) uses $50 million of section 602 funds to provide premium pay for K-12 teachers, then it would appear to be on the hook for $50 million of recoupment (provided that it’s not eligible for an off-ramp under steps 2 or 3) even though it has offset the $50 million tax cut with $50 million of spending cuts to after-school sports.²

Treasury presumably decided to aggregate at the level of the department, agency, or authority—rather than at a more granular level—in order to have a bright-line rule that could satisfy Pennhurst’s clear-statement requirement. From a litigation-risk perspective, I think that was the right choice, but I can see states arguing that aggregation at the department/agency/authority level is too ungenerous. Again, if you think (as I do) that Chevron applies to conditional spending requirements, then Treasury’s level-of-granularity choice seems like a reasonable reading of ambiguous text. From an academic perspective, I’ll note that this is an interesting example of an instance in which Pennhurst’s clear-statement demand might have led to the adoption of a less state-favorable rule (because the more state-favorable alternative—aggregation at the program level—would have been somewhat unclear on what constitutes a distinct program).

4. Tenth Amendment. A final argument that states might make is that even if the net tax revenue provision—as clarified by Treasury—satisfies Pennhurst’s clear-statement requirement, the provision still intrudes too far into a traditional domain of state authority. (The argument might be framed in terms of the anti-coercion doctrine, the anti-commandeering doctrine, or the unconstitutional conditions doctrine—all of these framings land in more or less the same place.)

The West Virginia lawsuit, for example, makes an argument along these lines:

The States have always retained their sovereign right to determine their own taxation and fiscal policies, which represents an important structural check on the federal government. … [F]orbidding States to exercise their sovereign power to set tax and fiscal policy is not a permissible condition that the federal government can set on its distribution of federal funds.

The problem with this argument is that it’s totally ahistorical. Congress regulates state taxation in all sorts of ways—for example, forbidding state sales taxes on Internet access and air travel, banning state income taxes on Treasury bond interest, and limiting state tax authority in literally dozens of other ways. What’s unusual about the American Rescue Plan is that Congress typically regulates applications of state tax, whereas the American Rescue Plan regulates reductions of state tax. But what’s good for the revenue-raising goose should be good for the tax-cutting gander.

It’s possible that the Supreme Court will use a challenge to the net tax revenue provision as an opportunity to curtail Congress’s authority over state taxation. (Some language in Justice Alito’s opinion in Murphy v. NCAA arguably gestures in this direction.) If applied symmetrically (to congressional limitations on revenue-raising measures and congressional constraints on state tax cuts), that might not be such a bad thing. But section 602 would be a funny place to start. In other instances, Congress regulates state tax authority outright—and permanently. Here, Congress is attaching a tax-related string to spending that states aren’t obligated to accept—and the string applies only temporarily. If the net tax revenue provision violates the Tenth Amendment, a whole host of other provisions that limit states’ revenue-raising authority would seem to be constitutionally vulnerable too.

Bottom line

All in all, I think Treasury did a really nice job threading the needle on the net tax revenue provision—crafting a regulatory scheme that leaves states with lots of freedom to maneuver while maintaining meaningful limits on the use of section 602 funds to pay for tax cuts. In a just world, the rule would survive judicial review. (But of course, that doesn’t necessarily mean it will.)

One final note: While I support the policy underlying the net tax revenue provision, Congress did a truly terrible job of drafting it. There are so many ambiguities baked into the text that it almost reads like an issue-spotter on a law school exam. Progressive legislation will likely suffer some defeats over the next several years given the ideological composition of the federal courts. But we really ought to avoid own goals. Treasury deserves praise for its regulatory handiwork here. The drafters of section 602(c)(1)(A) do not.

1. “Much-anticipated,” at least, in the small community of policy wonks and academics who pay close attention to issues of cooperative federalism and state budgets.

2. The state in this example might seek to invoke § 35.10(d) of the interim final rule, which allows a state to request relief from recoupment under a facts-and-circumstances standard.