Concept
The Anti-Coercion Principle
Spending power
The Anti-Coercion Principle
The Spending Clause lets Congress tax and spend for the debts, common defense, and general welfare of the United States,1 and the cases read those ends so broadly that Congress holds almost unreviewable discretion to define them.2 Butler went further and held that the spending power is not limited by the direct grants of legislative power found in the Constitution, so objectives outside Article I’s enumerated fields may be reached through conditional grants.3 With the ends so loosely confined, the real limits on conditional spending lie in the conditions themselves.
Four of those limits, taught above, test the offer from the outside: whether its purpose is proper, whether its terms are clear, whether it relates to the program the money funds, and whether it collides with another constitutional provision. A fifth limit tests the offer from the recipient’s side and asks whether the state could realistically have said no.
That fifth inquiry follows from the theory that justifies conditions at all. Spending legislation is “much in the nature of a contract,” and its legitimacy “rests on whether the State voluntarily and knowingly accepts the terms of the ‘contract.’”4 Acceptance carries that weight only if refusal was a live option. When refusing would cost so much that no responsible official could refuse, the offer works as a command, and a command Congress could not issue directly does not become valid because it arrives inside a funding statute. Honoring the limit, the Court has said, is critical to ensuring that spending legislation does not undermine the status of the states as independent sovereigns.5
The principle began as a passing remark and became a holding. Dole acknowledged that an inducement could grow so heavy that “pressure turns into compulsion,” then found nothing of the kind on its facts.6 The challenge to the Medicaid expansion finally applied the warning to strike a funding penalty, and it remains the only modern decision to do so.7 A limit that has decided one case is still a limit, which is why it is now taught as a fifth alongside the four above.
The subsections below follow the order in which the cases built the principle. Dole comes first, where the idea appears and does no work. The Medicaid expansion comes next, where the idea decides a case. Two subsections then isolate the features that drove that result, a threat to an independent grant and the size of the loss, and a last subsection follows the principle into the lower courts, where challengers have rarely won. Keep one working question in view throughout: could the state refuse this condition in practice, or only in form?
Coercion in Dole
Dole introduced the idea in a single observation: a financial inducement from Congress might in some circumstances be so coercive that it passes the point at which “pressure turns into compulsion,” a phrase the Court borrowed from Steward Machine Co. v. Davis.8 The observation marks a boundary, and the boundary follows from voluntariness, because a state’s acceptance carries weight only while refusal remains a live option. The Court then found the highway condition nowhere near it. Congress had directed only that a state wishing to set a drinking age below 21 lose a relatively small percentage of certain federal highway funds, five percent of the funds otherwise obtainable under specified highway grant programs.9
The Court stressed how little South Dakota stood to lose, and later cases measure that smallness against the whole state budget: five percent of highway funds was less than half of one percent of South Dakota’s budget.10 The Court called the argument as to coercion “more rhetoric than fact.”11 A loss that small can be weighed against the value of the state’s own policy, and an official can decide the policy is worth the price. South Dakota could keep its lower drinking age and give up a modest sum, so its choice stayed real.
South Dakota advanced a second route to coercion. It argued that the program’s degree of success showed how coercive it was.12 The Court declined to hold a conditional grant unconstitutional simply because it succeeded in achieving the congressional objective. It rested that answer on Steward Machine’s observation that every rebate from a tax conditioned upon conduct is in some measure a temptation, and that treating motive or temptation as coercion would plunge the law into “endless difficulties.”13 The difficulty becomes plain when the success test is followed to its end. Inducements exist to change behavior, so an effective one always changes it, and if effectiveness proved compulsion, grants would become invalid in proportion to how well they worked. The coercion inquiry avoids that result by asking what refusal would cost the state, which is a different question from how many states accepted.
The Court’s closing description of the statute states the standard it was applying. Congress had offered “relatively mild encouragement” to raise drinking ages, and enacting such laws “remains the prerogative of the States not merely in theory but in fact.”14 The words “in fact” turn attention from the formal availability of refusal to the state’s practical options. An older premise lies beneath that standard. In Oklahoma v. Civil Service Commission, the Court found no violation of state sovereignty because the state could adopt “the ‘simple expedient’ of not yielding” to what it called federal coercion.15 The modern formulation keeps the premise: in the typical case courts look to the states to defend their own prerogatives by adopting “the simple expedient of not yielding” to federal blandishments when they do not want to embrace federal policies as their own.16 Coercion names the situation in which that premise fails.
United States v. Butler reached the opposite conclusion on opposite facts. The government defended the crop-reduction plan as voluntary. The Court replied that a farmer who refused lost the program’s benefits and faced competitors who took the payments, a position that might end in financial ruin, and it called the arrangement “coercion by economic pressure” in which “[t]he asserted power of choice is illusory.”17 Both programs threatened to withhold federal money, so the threat of lost money cannot be what separates them. The difference lies in what the loss did to the recipient’s practical options. For the farmer, refusal meant a competitive disadvantage that could end the business. For South Dakota, it meant a small reduction in one stream of highway money.18 Butler also concerned individual recipients, and the anti-coercion principle has only rarely been applied to void a conditional spending measure aimed at enticing state compliance, because states are expected to refuse what they do not want.19
Because the Court stressed that only a relatively small percentage of funds was at stake, Dole leaves open what it would have done with a larger threat. Its reference to compulsion supplies the basis for treating anti-coercion as a fifth limit: Congress may not structure a program so that states, as a practical matter, cannot resist the inducement to participate.20 In Dole itself the limit did no work, so on an exam the case serves as the baseline for a condition that stays on the permissible side. The next section takes up the case in which the limit decided the outcome.
NFIB and the Medicaid Expansion
Dole’s warning decided nothing until National Federation of Independent Business v. Sebelius. There twenty-six states led by Florida challenged the Affordable Care Act’s Medicaid expansion, and seven Justices agreed that Congress had exceeded its spending power by making noncompliance cost a state all of its Medicaid funding.22
The expansion’s significance depends on what it changed. Medicaid began as a federal-state program covering medical services for four categories of the needy: the disabled, the blind, the elderly, and needy families with dependent children.23 By 1982 every state had chosen to participate, and federal Medicaid funds now make up over 10 percent of most states’ total revenue.24 A program with every state inside it and a tenth of state revenue riding on it hands Congress enormous leverage. That leverage is where the line between encouragement and compulsion has to be drawn.
The Affordable Care Act changed the bargain in two ways. It extended coverage to a far larger group, the nonelderly poor generally, and it offered new federal funds for that coverage.25 It also supplied the stick: a state that did not comply with the new coverage requirements could lose not only the federal funding for those requirements but all of its federal Medicaid funds.26 That threat, attached to money the state already received, is the center of the case.
The Eleventh Circuit unanimously upheld the expansion as a valid exercise of the spending power and rejected the states’ claim that the threatened loss of all Medicaid funding coerced them.27 The Supreme Court disagreed on coercion through two separate opinions. Chief Justice Roberts wrote for himself and Justices Breyer and Kagan.28 Justices Scalia, Kennedy, Thomas, and Alito reached the same conclusion in a joint opinion styled as a dissent, while Justices Ginsburg and Sotomayor dissented on this point.29
Because the seven votes came in two blocs and neither bloc was a majority alone, a student needs a way to tell which opinion states the law. When no opinion commands a majority, the holding is the position of the Justices who concurred in the judgment on the narrowest grounds, and here that is the Chief Justice’s opinion.30 The joint opinion reached the same result by a broader route, so the Chief Justice’s reasoning marks the ground common to all seven. An answer that applies his factors states the rule of the case, while an answer that borrows the joint opinion’s wider theory relies on four votes.
The Chief Justice started from the premise that spending legislation is “much in the nature of a contract,” so a condition is legitimate only if the state voluntarily and knowingly accepts its terms.31 Earlier sections used that premise to explain why conditions must be clear. Here it yields a second consequence: acceptance counts only if refusal is a live option. The premise also bears on political accountability. When a state can choose, voters can fairly hold state officials accountable for accepting or refusing the offer. When the state has no choice, the federal government achieves its objectives without accountability, and the danger grows under the Spending Clause because Congress can use spending to implement policy it could not impose directly under its enumerated powers.32 New York v. United States had drawn the same line. Congress could attach monetary and access incentives to encourage state compliance, but the take-title provision, which made states take title to waste not properly disposed of by a deadline and answer for the damages, went beyond permissible encouragement.33
The Chief Justice found the coercion in two features of the expansion, and the next two sections develop each. The first is what the threat targets. Congress may condition funds on compliance with restrictions on how those funds are used, because that is how it ensures the money is spent according to its view of the general welfare, but conditions that take the form of threats to terminate other significant independent grants are a means of pressuring states to accept policy changes.34 The second is the threat’s size. Dole involved relatively mild encouragement, while a state that opted out of the expansion stood to lose all of its Medicaid funding, which the opinion called “a gun to the head” and “economic dragooning.”35 The opinion declined to fix the outermost line where persuasion gives way to coercion and was content that the Medicaid threat lies beyond it, wherever that line falls.36
The joint opinion arrived at coercion by a different path. It conceded that the Act does not legally compel states to join the expanded program, but it stressed that refusal authorizes termination of all of a state’s Medicaid funding, a sum equal for the average state to more than one-fifth of its expenditures.37 It also read the goal and structure of the Act as showing that Congress understood no state could refuse the expansion, and it concluded that the states had no choice other than to accept the package.38 Its analysis treated the whole package as coercive, while the Chief Justice’s analysis rested on the particular features he identified.
The remedy followed the diagnosis. The offending feature was the threat to withdraw existing funds, so the Court left the expansion available and removed the threat: a state may decline the new coverage requirements without losing the Medicaid money it already receives.39 A common exam error is to say the case struck down the Medicaid expansion. The expansion remains open to every state that wants it.
The two versions of the statute make the line concrete. Under the Act as Congress wrote it, a state that refused lost every Medicaid dollar, federal money that exceeds 10 percent of most states’ total revenue.40 Under the remedy, a state that refuses forfeits only the new money, which is the arrangement the states themselves described when they objected that Congress, instead of simply refusing to grant the new funds to states that would not accept the new conditions, had also threatened their existing Medicaid funds.41 The new coverage obligations and the new money are the same in both versions. The one fact that moved is whether refusal costs the state the grants it already holds, and that fact separates an offer from a command.
NFIB remains the only modern decision to find a spending condition coercive, and it has not been expanded to invalidate less extreme financial pressure on the states.42 The Chief Justice’s refusal to fix an outer line helps explain that restraint. A coercion answer should therefore compare the facts against both anchors, the mildness of Dole’s five percent and the all-or-nothing threat in the Medicaid expansion, because the Court has never announced a dollar threshold.
Threats to Independent Grants
The first feature the controlling opinion identified is the target of the threat, and the opinion sorts conditions by what they do. Congress may require a state to comply with restrictions on how federal funds are used, because that is how Congress ensures the money is spent according to its view of the general welfare. A condition that does not govern use of the funds cannot be justified on that ground, and when it takes the form of a threat to terminate “other significant independent grants,” it is properly viewed as a means of pressuring states to accept policy changes.44
The distinction follows from what justifies a condition in the first place. A use restriction travels with the money it regulates, because the federal interest that supports the grant also supports the rule about spending it. A threat against a different grant the state already receives has no such tie. It shapes no federal dollar’s use, so its only work is leverage: the state must accept the new policy or lose a program that its budget, its agencies, and its residents already depend on. The leverage also reaches back to an arrangement the state joined on earlier terms, which strains the premise that the state knowingly accepted the bargain.45 The states pressed this point in the litigation. Instead of simply refusing new funds to states that would not accept the new conditions, they said, Congress had also threatened their existing Medicaid funds, and the threat served no purpose except to force unwilling states into the expansion.46
Whether the threatened money was truly independent turned on a prior question: how to tell a modification of an existing program from a new program. The government argued that the expansion merely modified Medicaid, since the states agreed when they joined that Congress could change the program’s terms. The Chief Justice answered that the expansion was a shift in kind, not merely degree.47 The original program covered four categories of the needy, and earlier amendments only altered and expanded the boundaries of those categories. The Affordable Care Act instead extended coverage to the entire nonelderly population with income below 133 percent of the poverty level, turning a program for the neediest into part of a comprehensive plan for universal coverage.48 A state could hardly have anticipated that Congress’s reserved right to “alter” or “amend” Medicaid included the power to transform it so dramatically.49 In the Chief Justice’s view Congress was enlisting the states in a new program, so a state that refused could properly be denied only the separate new funding.50
Justice Ginsburg, joined by Justice Sotomayor, saw one program. Medicaid as amended remained “a single program with a constant aim—to enable poor persons to receive basic health care when they need it,” and the Act simply required states to do what they had long been required to do: comply with the conditions Congress prescribes for participation. Congress had alerted the states from the outset of the program that its right to alter or amend was reserved.51 She added that states have no entitlement to Medicaid funds, only the opportunity to accept them on Congress’s terms, so there were no “existing” funds to threaten. She also distinguished Dole. Its condition did not tell the states how to use their highway money, whereas the ACA relates solely to the federally funded Medicaid program, Congress did not threaten funds earmarked for any other program, and Congress did not use Medicaid funding to induce action it could not itself undertake.52
The two opinions thus agree that a threat to a separate program’s money is the problem, and they divide on whether Medicaid before and after 2010 is one program or two. The Chief Justice describes a program by the population it serves and the purpose that population implies, so a change that rewrites both is a new program. Justice Ginsburg describes it by its general aim, so any expansion toward that aim stays inside the original program. The first description decided NFIB, and the second is the answer a defender of a new federal condition would give on an exam fact pattern. Because the line is one of degree, run the analysis in order and let the facts of the particular program supply each answer.
Mayhew v. Burwell shows the steps in use. Maine had voluntarily covered eighteen- to twenty-year-olds, and the ACA required it not to discontinue that coverage on pain of losing all of its Medicaid money.53 The all-or-nothing threat was present, as in NFIB, yet the court found a shift only in degree, an unexceptional alteration of the boundaries of categories of individuals covered under the old Medicaid program.54 The difference from NFIB lay in whether the condition changed an existing program or built a new one around it.
Independence of the threatened grant raises a flag but does not settle the case. In Dole the condition did not direct how highway money was spent, yet the Court upheld it because the pressure was only relatively mild encouragement.55 The Chief Justice relied on several features together in distinguishing the Medicaid expansion: the conditions did not limit how states used the funds, the threat was removal of an independent pre-existing grant, the amount was large, the change was in kind, and the new conditions could not have been predicted when the state accepted.56 One reading treats the holding as limiting federal power only when several of these features appear at once.57 On that reading, an answer that spots a threat to an independent grant has found one feature and must still ask how large the threat is.
Size of the Financial Threat
The second feature is the size of the threat, because how realistic refusal is depends on what refusal costs. A state that rejects a federal offer must either do without the lost money or replace it by cutting other programs or raising taxes,59 and the larger the loss measured against everything the state spends, the less plausible refusal becomes. A very small loss can be weighed against the value of the state’s own policy. A very large one leaves a right to refuse that exists only on paper.
The two anchor cases supply the numbers. In Dole the penalty was five percent of a state’s federal highway funds, less than half of one percent of South Dakota’s budget,60 and the Chief Justice called pressure of that size “relatively mild encouragement.”61 The Medicaid expansion threatened a loss of a different order: a refusing state would lose “not merely ‘a relatively small percentage’ of its existing Medicaid funding, but all of it.”62 Medicaid accounts for over 20 percent of the average state’s total budget, and federal funds cover 50 to 83 percent of those costs,63 so by simple arithmetic the federal money at stake is roughly 10 to 16 percent of a typical state’s budget. One treatise summarizes the figure as about 10 percent of an average state’s total budget.64 The Chief Justice called that threat “a gun to the head” and “economic dragooning.”65 Both threats involve withholding federal money. One is a sliver of a state budget, and the other is a double-digit share of it.
The comparison depends on which percentage is computed, because a loss can be total for one program and trivial for the state. In New York v. Department of Justice the federal government conditioned Byrne program law-enforcement funds on state and local help in identifying undocumented immigrants, so a noncomplying jurisdiction stood to lose every dollar of that program. The Second Circuit nonetheless found the case much closer to Dole than to NFIB, because New York’s loss was under 0.1 percent of its annual $152.3 billion budget, a smaller share even than South Dakota’s, and the state kept a real choice.66 The all-or-nothing structure matches the Medicaid expansion, since refusal forfeits the whole program in both settings. What differs is how much that program weighs in the state’s finances. A student who writes that losing all of a program’s money proves coercion has counted the wrong percentage.
Courts also differ on the denominator. Gruver v. Louisiana Board of Supervisors measured the threatened loss against the funding of the university involved, not the state’s whole budget, and treated a loss of just under 10 percent of that funding as more like Dole’s relatively mild encouragement than the gun to the head in NFIB.67 The comparison is rough for that reason, and the next section returns to the case.
Having set the two anchors side by side, the Chief Justice declined to say where persuasion gives way to coercion. He saw no need to “fix the outermost line where persuasion gives way to coercion,” and the joint opinion drew none either.68 The Court has set out no bright-line test for when an inducement becomes compulsion,69 so the only firm markers are the two cases themselves. A court that decides only that the statute before it lies beyond the line keeps its holding narrow, and the consequence is a wide untested band between half of one percent of a state’s budget and ten percent or more. Most exam hypotheticals fall in that band, so an answer cannot recite a threshold. It must argue by comparison to the anchors.
The joint opinion reached coercion by a route that gives size more of the work. Those four Justices noted that the federal Medicaid subsidy for the average state exceeds one-fifth of the state’s expenditures, and that a state forced out of the program would lose that sum and almost certainly have to increase its own health-care spending, which meant either drastic cuts to other programs or a large tax increase.70 Those new state taxes would sit on top of the federal taxes the state’s citizens already pay to fund Medicaid in other states, so the state could not simply redirect what it had been sending to Washington.71 They added that Medicaid grants make up some 40 percent of all federal money the states take in, so the states are far less reliant on federal funding for any other program.72 The opinions state the loss in different terms, so their percentages do not match, but both place it far above Dole’s fraction of a percent. For these Justices the absence of legal compulsion was not enough when the sanction was that severe. Because the Chief Justice’s narrower opinion pairs size with the threat to an existing independent grant and controls, an exam answer should use the joint opinion’s dependence reasoning as a supporting theme and should not rest on it alone.
Justice Ginsburg attacked the workability of any size-based line. She asked how litigants and judges would tell whether a state has a legitimate choice, and whether courts must count the dollars the federal government might withhold or the portion of the state’s budget at stake. If the budget controls, she asked, which one: that of the lead plaintiff, of all twenty-six challenging states, or of some national median? She noted that fiscal situations differ, since Florida imposes no state income tax and might replace lost federal funds with new revenue, and that the coercion officials actually fear may be punishment at the ballot box for refusing a popular grant. Her conclusion was that the inquiry involves political judgments that defy judicial calculation.73 She wrote as a dissenter who would have upheld the expansion, but the variables she named are the ones any court must handle when it compares a loss to a budget, and no majority has answered her questions.
NFIB in the Lower Courts
The Supreme Court has found a spending condition coercive in only one modern case, and that ruling has not been extended to invalidate less extreme financial pressure on the states.75 The lower courts show why. At least for now, their applications of the anti-coercion principle suggest that the ruling may have more bark than bite.76 In four notable decisions the court asked which of NFIB’s distinguishing features the case before it lacked, and each found at least one missing.
Gruver v. Louisiana Board of Supervisors involved one additional condition attached to Title IX funding. The court observed that Congress changes federal spending programs all the time and that one added condition on a longstanding, unitary set of Title IX funding rules differs from the “brand-new legislative program” the ACA thrust on the states. It also found that the threatened loss of just under 10 percent of LSU’s funding looked more like Dole’s relatively mild encouragement than the gun to the head in NFIB.77 Tennessee’s challenge to a condition requiring Medicaid coverage of resident refugees failed on similar grounds. The court distinguished NFIB because the provision was not part of an “entirely new program”: the state had covered these refugees since the 1970s, and without reimbursement since 1991, and the latest federal change had narrowed the group for which coverage was required.78 The state argued that growth in the number of refugees, and in the cost of covering them, made those features irrelevant. The court answered that the amount of the loss is only one element of the coercion test and that Congress was not using the program’s funding as leverage to force the states to accept a new condition.79 The two cases share a hinge. Each added to a program the state already ran, which is not the same as building a new program around an old one.
Mayhew and New York v. Department of Justice show the same method on different facts, and each turns on the feature that was missing. Maine argued that federal authorities had moved from inducement to compulsion by threatening all Medicaid funding unless it kept covering eighteen- to twenty-year-olds, a practice it had adopted voluntarily.80 The threat of total loss was present, as in NFIB, yet the court held that the condition altered Medicaid only in degree, an unexceptional alteration of the boundaries of categories the old program already covered.81 The Second Circuit in New York reached its result by the other route. A noncomplying jurisdiction stood to lose the whole Byrne program, but the loss was under 0.1 percent of the state’s budget, so the case resembled Dole and not NFIB, and the court closed with the Chief Justice’s reminder that “[t]he States are separate and independent sovereigns. Sometimes they have to act like it.”82 The missing feature differs, a change in kind in one case and a large loss in the other, but the result is the same: a threat to withhold everything in a program does not establish coercion unless other features accompany it.
Read together, the four decisions fit a narrow reading of NFIB. On that reading the case directly supports a finding of coercion “[w]hen Congress takes [1] an entrenched federal program that [2] provides large sums to the states and [3] tells states that they can continue to participate in that program only if they also agree to participate in a separate and independent program.”83 Each element matches something a lower court relied on to refuse an extension. The Gruver and refugee conditions were additions inside a continuing program, the Mayhew condition changed an existing program only in degree, and the New York loss was small against the whole budget. The Chief Justice’s own list of distinguishing features is longer, and one reading treats the case as restricting federal power only when several of those features appear together.84
A broad reading explains why courts have been careful. Read broadly, the Chief Justice’s opinion would seriously threaten the constitutionality of a wide swath of federal spending legislation, including conditional programs in education (the Individuals with Disabilities Education Act and the No Child Left Behind Act) and civil rights (portions of the Civil Rights Act of 1964).85 A court that treats NFIB’s distinguishing features as requirements keeps the principle meaningful for the extreme case without that cost. Whether the Supreme Court will apply the Chief Justice’s test broadly or treat the Medicaid expansion as uniquely coercive has not been decided.
A second extension of NFIB has even less support. The four Justices who wrote the joint opinion suggested that the Spending Clause would not permit Congress to interfere with the states’ “traditional authority in the field of education” by broadly funding local educational programs only so long as state officials accede to federal conditions on curriculum, the hiring and tenure of teachers, school districts, the length of the school day, the calendar, dress code, and student discipline.86 Had five Justices adopted that idea, it would have produced a sixth limit protecting traditional state authority. Only four votes backed it, and it is far from certain that a fifth exists, because the Chief Justice stressed that past rulings had consistently upheld Congress’s authority to condition funds on the states’ complying with restrictions on the use of those funds, and the hypothetical education law would tie local authorities’ federal money to spending that very money in line with Congress’s conditions.87 The Chief Justice’s opinion is the narrowest ground supporting the judgment, so it states the governing law, and the joint opinion’s wider theory is the position of four Justices.88 An exam answer may mention the education hypothetical as a possible direction for the doctrine, labeled as such, and should not list protection of traditional state authority as a recognized limit.
Executive attempts to condition funds raise a different question, which is who may attach a condition at all. In 2017 an executive order announced that jurisdictions identified as sanctuary cities would not receive federal funding and grants, and San Francisco and other jurisdictions challenged it. The district court permanently enjoined the order’s funding provision, and the Ninth Circuit affirmed in San Francisco v. Trump, holding that the order violated the separation of powers because the President’s authority to act must come from an act of Congress or from the Constitution itself, and the administration had not tried to show that Congress authorized withholding grant money on this basis.89 The reasoning follows from the text of the Spending Clause, which vests the power to tax and spend in Congress. A condition drawn from that power must therefore originate in a statute or rest on authority Congress has delegated, and an executive official who adds conditions alone is borrowing a power the Constitution assigns elsewhere. The decision concerns who may impose a condition, so it leaves open how much pressure Congress itself may apply.