The rise and rise of coercive federalism

In 1984, the US Congress passed the National Minimum Drinking Age Act, requiring states to raise their minimum drinking age to 21. States that did not comply would lose part of their federal highway funds. South Dakota, which allowed 19-year-olds to buy certain alcoholic beverages, refused to change its law and sued the federal government, arguing the law violated the Tenth and Twenty-First Amendments. Elizabeth Dole, US Secretary of Transportation, was taken to court to defend the federal policy. In 1987, the Supreme Court ruled on whether Congress could withhold federal highway funds from states that did not raise their minimum drinking age to 21. A 7-2 majority upheld the law, provided the conditions pursued general welfare, were unambiguous, reasonably related to federal interests, did not require actions violating constitutional provisions, and the financial inducement did not amount to coercion.Footnote ⁷

Dole fleshed out a test to handle the rise of ‘coercive federalism’ in the US. In the American context, this label was first used in John Kincaid’s ‘From Cooperative to Coercive Federalism’, published in the Annals of the American Academy of Political and Social Science in 1990 to describe the emergence of a new phase in US federalism.Footnote ⁸ After decades of relentless national power growth, by the late 20th century, states found themselves at the mercy of a powerful central government, equipped with an unprecedented array of tools to persuade, entice, or pressure them into adopting and enforcing federal policies. Reluctant states often faced the federal government’s strength, more than capable of ‘coercing’ them into compliance.

The increased centralisation of the US has been beyond dispute for many years, and it has been documented by political scientists.Footnote ⁹ A variety of tools have facilitated this drift to the centre. To begin with, since the New Deal, a broad interpretation of Congress’s power to regulate interstate commerce (Article I, Section 8, Clause 3) has been used as a basis for enacting legislation in virtually any field remotely related to the economy, effectively granting near-plenary legislative power.Footnote ¹⁰

Second, federal statutes have frequently been used to totally or partially pre-empt state legislation. By one account, while only 29 pre-emption bills had been enacted before the year 1900, between 2000 and 2007 more than 100 pre-emption acts were passed.Footnote ¹¹ The resulting displacement of state law is hard to overstate. Through the so-called ‘federal mandates’, namely federal statutes or regulations that force state or local governments to take specific actions, the central government has also increasingly imposed on states or local governments obligations to pursue specific objectives, to undertake precise executive actions or to adopt certain legislation, at times even forcing them to bear the costs (so-called ‘unfunded mandates’).Footnote ¹²

Even the incorporation of the federal Bill of Rights has contributed to this centralising drift by imposing nationwide standards on sensitive fields like freedom of speech or criminal procedure.Footnote ¹³

Perhaps more than anything else, it was the power of the purse that made US federalism coercive. Since the ratification of the Sixteenth Amendment in 1913, which allowed the federal government to levy taxes without apportioning them among the states, the federal budget has steadily grown. With these vast resources, national financial assistance has become a major revenue source for the states.Footnote ¹⁴ In the 2020s, federal grants make up more than 30% of the state budgets on average, although with considerable variation from state to state.Footnote ¹⁵ With these levels of federal funding, states are often vulnerable to ‘coercion’, namely the choice between implementing federal policies they do not agree with and facing the loss of substantial shares of their budgets. As federal grants fund an extremely wide array of sensitive state policies, from education to labour training, from health care to infrastructure, conditions attached to them can thus be used by the centre to purchase the submission of the constituent units.Footnote ¹⁶ At times, the inflow of federal funds is conditioned on the execution of specific federal mandates or on the implementation of specific projects (categorical grants); in other cases, the conditions for disbursement are more generic (block grants).

Given the magnitude of federal grants, state governors and assemblies often have little choice but to do Congress’s bidding. This is even more problematic as federal agencies interact directly with state agencies in charge of administering the funds, circumventing local politicians.Footnote ¹⁷ In the US, the purse has become an instrument of federal coercion as powerful as the sword.

The limits of conditionality: South Dakota v Dole

As the traditional guardians of federalism – the Senate and political parties – have become national actors with modest interest in protecting the states from federal coercion, the definition of the latter’s limits moved to the courts.Footnote ¹⁸ Although a few earlier cases had already touched upon the problem, it was only in the late 1980s, after the explosion of federal grants, that a workable test was provided.Footnote ¹⁹ Enter the main character of the play, the Dole test.

As mentioned, South Dakota faced the loss of federal highway funds for allowing individuals under 21 to purchase alcoholic beverages. The state argued that setting minimum drinking ages was a subnational power and that the National Minimum Drinking Age Act usurped this authority. A majority, led by Chief Justice Rehnquist, sided with the federal government and, in the process, established the judicial test for determining the coercive nature of federal funding withdrawals. The Dole test, designed to ensure the constitutionality of conditions attached to federal funds, consists of five steps.Footnote ²⁰

General welfare. First, the congressional spending power is limited by the so-called general welfare clause (Article I, Section 8, Clause 1), stating that ‘Congress shall have Power to lay and collect Taxes, Duties, Imposts and Excises, to pay the Debts and provide for the common Defence and general Welfare of the United States’. The construction of this clause has been controversial since the country’s founding, with prominent politicians like Madison arguing that it allowed the federal government to spend only to carry out powers elsewhere conferred in the Constitution and others, like Hamilton, claiming that the federal government could spend for any purpose, as long as it was general rather than local.Footnote ²¹ In 1936, the Supreme Court endorsed Hamilton’s view in Butler. The general welfare clause was read as meaning that, while the federal power to regulate was limited to areas explicitly conferred in the Constitution, the power to spend knew no such limits, provided that it was aimed at the general welfare. Regulatory, not spending, powers of Congress were enumerated.Footnote ²² Dole confirmed this second interpretation. Moreover, the majority also clarified that ‘whether a particular expenditure is intended to serve general public purposes’ had to be left to the discretion of Congress and was not subject to judicial determination.Footnote ²³ If Congress deemed that interstate divergences in drinking ages required a national solution, that was final.

Unambiguous expression. Second, Dole established that conditions had to be unambiguous. The receiving states had to be able to understand ex ante the consequences of accepting the federal funds. This, the majority argued, was indisputable in the case at stake: the minimum drinking age at 21 sanctioned by state law was a crystal-clear requirement.

Germaneness. Third, the conditions had to be related to the federal interest in particular national programmes or projects, namely a reasonable connection between the congressional aims and the conditions. Funds directed at primary education could not be withdrawn to induce the state to increase the minimum drinking age, as the two fields are entirely disconnected. However, in the mind of Dole’s majority, a standardised minimum drinking age across the nation was reasonably germane to the distribution of highway construction funds, for reasons of public safety. Congress deemed younger voters more prone to drinking and wanted to ensure safer road conditions by raising the minimum age to 21. In practice, this step of the test merely means that Congress ‘must name and allocate funds in such a manner that a challenging party cannot demonstrate that the conditions are unrelated to congressional goals’.Footnote ²⁴ Congress is free to identify federal aims and deem the conditions related to their attainment, knowing that strong judicial deference prevents rigorous scrutiny of the aims-means connection.

No unconstitutional conditions. Fourth, states must not be lured into actions that would be unconstitutional under an independent bar. For instance, federal funds could not be used to induce invidiously discriminatory state action or cruel and unusual punishment. In Dole, however, a minimum drinking age at 21 would violate no independent constitutional provision.

Coercion. At the end of the judgment, not formally listed as one of the conditions and added almost as an afterthought, came the most consequential requirement of all: the financial inducement must not be so intense as to turn pressure into coercion. The last stage is the coercion test proper, and it depends on the hardly quantifiable yet clearly perceptible sense that the financial pressure may be so high as to leave no choice at all to the state. In Dole, the limited consequences of refusal, namely losing 5% of the funds allocated to highways, were deemed too modest to be coercive by the Supreme Court, but a coercion test had been established.

Sebelius: applying the test

Dole provided a test to assess the constitutionality of financial conditions. For years scholars debated the merits of the test, some finding it hopelessly vague,Footnote ²⁵ others appreciating its potential for protecting the states’ prerogatives.Footnote ²⁶ In the meantime, though, the Supreme Court showed little interest in using it to control conditions on federal funds. Dole barked but did not bite for a long time, until the matter came to a head in 2012.

Barack Obama’s signature legislation was the Patient Protection and Affordable Care Act, also known as the Affordable Care Act or Obamacare (2010).Footnote ²⁷ The reform aimed to expand the coverage of the 1965 Medicaid programme to all Americans under 65 with incomes up to 133% of the federal poverty line. As Medicaid is a federal spending programme administered by the states, the Affordable Care Act stipulated that the federal government would fully cover the cost of this expansion until 2016, after which it would cover about 90% of the costs indefinitely. States would thus be left to cover only 10% of the costs of the newly eligible population after six years. States that opted out of the Affordable Care Act, however, would lose all Medicaid funds. Medicaid spending accounted for over 20% of the average state’s total budget, with federal funds covering between 50% and 83% of those costs, so that the threatened withdrawal amounted to over 10% of a state’s overall budget.

Litigation brought by 26 states, largely along party lines, eventually led the case to the Supreme Court. The Sebelius judgment ultimately made Dole’s test operational.Footnote ²⁸ While the Affordable Care Act was declared constitutional thanks to Chief Justice Roberts’ defection to the liberal camp, the specific clause which required the states to either accept the expansion or lose the entirety of Medicaid’s funds was severed and voided.Footnote ²⁹ Incidentally, this part of the holding was sustained by the entirety of the conservative bloc in the Court, but also joined by two liberals, Elena Kagan and Stephen Breyer.Footnote ³⁰

Roberts connected the prohibition of financial coercion to the dual structure of US federalism. As in the case of the anti-commandeering doctrine, according to which the federal government cannot compel state governments to enforce or administer federal laws, the states must also have a legitimate choice to agree on any spending programme. Otherwise, political decisions would be taken at the federal level but attributed in the eyes of the citizens to the states, a fatal blow to political accountability.Footnote ³¹ Even more, the ‘danger [of unaccountability] is heightened when Congress acts under the Spending Clause, because Congress can use that power to implement federal policy it could not impose directly under its enumerated powers’.

In the Chief Justice’s view, in Sebelius the pressure exerted by the threat to withdraw the federal funds had crossed the line between enticement and coercion. While in Dole South Dakota risked losing less than 1% of its total budget, in Sebelius ‘[t]he threatened loss of over 10 percent of a State’s overall budget, in contrast, is economic dragooning that leaves the States with no real option but to acquiesce in the Medicaid expansion’. As a result, ‘the financial “inducement” … is much more than “relatively mild encouragement” – it is a gun to the head’.Footnote ³²

This ‘gun to the head’ argument has naturally been fiercely criticised, even on the right.Footnote ³³ The most insightful critique came from Justice Ruth Bader Ginsburg (joined by Sonia Sotomayor), who recast the inducement-coercion threshold as a sorites paradox: where, exactly, does pressure tip into compulsion? In Ginsburg’s words:

In addition to this argument, Ginsburg made another point worth recalling: states have the resources to counter the alleged federal coercion. If Medicaid’s expansion is unacceptable, they can refuse the funds, raise the necessary resources by increasing state taxes or cutting other expenditures, and compensate for the loss of income. Florida, for instance, imposed no income tax: nothing prevented it from using this fiscal space to do so. Florida’s taxpayers would surely be paying both the federal tax funding the expanded Medicaid and the new state tax needed to set up whatever substitute programme might be required, but this would be a political rather than legal question:

In Ginsburg’s opinion, any coercion test is not only volatile but also, and perhaps most importantly, unworkable in court. Any judicial determination inevitably requires assessing whether it is fair to put a government in the position of facing its electors and justifying why the money was refused. This is an inherently political judgement that should not be left to the courts.

Be that as it may, federal coercion is nowadays a structural feature of US law. Lower courts revived Dole in the 2017 ‘sanctuary cities’ litigation, finding Donald Trump’s executive-order defunding threats insufficiently related to the affected programmes and again a ‘gun to the head’.Footnote ³⁴ The use of the purse to again threaten and coerce sanctuary cities and even universities in the early stages of the second Trump administration is another example of federal coercion.Footnote ³⁵ These later developments, however, lie beyond the scope of this paper. What matters is that, in the end, Dole has been established as the standard judicial test to assess the legality of financial inducement. It is an overall functioning assessment whose transferability is worth evaluating.

Boosting the EU’s spending power

Coercive federalism is not limited to the US. As mentioned in the introduction, similar phenomena have also been envisaged elsewhere.Footnote ³⁶ Here, however, our emphasis is on the EU.

Spending conditionality has characterised the EU for decades.Footnote ³⁷ During the financial period 2014-2020, it became a standard feature of most EU funds.Footnote ³⁸ Still, to the extent that a discussion on coercive federalism in the EU is taking place, it is directed at the reforms of recent years. Since the 2021-2027 budgetary cycle, and in part to respond to the pandemic-induced economic crisis, substantial reforms have been introduced. Specifically, three new regulations were adopted.

In December 2020, the new Rule of Law Conditionality Regulation (2020/2092, hereafter the Conditionality Regulation) was enacted. It empowered the Commission, with the Council’s approval, to suspend the disbursement of any EU funds when rule of law violations were found to affect or seriously risk affecting in a sufficiently direct way the sound financial management of the EU budget or the Union’s financial interests.Footnote ³⁹ In February 2022, the Court of Justice rejected Hungary’s and Poland’s actions for annulment of the Regulation.Footnote ⁴⁰ In December of the same year came the first application of the suspension mechanism (against Hungary).Footnote ⁴¹

Additionally, in 2021 the Council and the Parliament also adopted the so-called Common Provisions Regulation (2021/1060), positing common rules for the governance of a wide array of EU funds.Footnote ⁴² Article 9 of the Common Provisions Regulation established an ‘enabling condition’ which, when read in the light of Article 15 and Annex III to the Regulation, also required compliance with the EU Charter of Fundamental Rights when implementing projects financed by the named funds.

Finally, the EU adopted the Recovery and Resilience Facility Regulation (2021/241), intended to distribute approximately €720 billion of EU resources borrowed by the Commission from the financial market to the member states during a period of six years (2021-2026).Footnote ⁴³ The inflow of funds was conditional on each member state submitting a national recovery plan outlining the reforms and investments to be financed, subject to assessment by the Commission and approval by the Council. Suspension or freezing of funds, upon a Commission proposal and Council decision, is also possible while the programme is ongoing. Negotiations with EU institutions are thus crucial to steer the member state in the desired direction.Footnote ⁴⁴

As a result, although the growth of EU funds has been substantial, it has also been accompanied by a reformed governance of their distribution. The Recovery and Resilience Facility Regulation, in particular, appears to have introduced a new method of economic policy governance, where broad priorities are defined at the EU level, detailed implementation is outlined in national plans, ongoing and performance-based monitoring is carried out by the Commission, and the final decision on any suspension of funds rests with the Council.Footnote ⁴⁵ In the case of the Conditionality Regulation too, the final word belongs to the Council acting by qualified majority upon a Commission proposal, while the latter decides alone on the possible suspension of funds under the Common Provisions Regulation.Footnote ⁴⁶

Based on this architecture, several times in recent years EU funds have been suspended.Footnote ⁴⁷ In 2022, the Conditionality Regulation was applied for the first time, targeting Hungary. The Commission cited a lack of anti-corruption enforcement, arbitrary use of executive power, and insufficient judicial protection. In December 2022, the Council approved the Commission’s request, though it reduced the amount of frozen funds.Footnote ⁴⁸ Around the same time, the Commission also withheld €21.9 billion from Hungary based on the Common Provisions Regulation for violations related to LGBTQ+ rights, academic freedom, the right to asylum, and judicial independence.Footnote ⁴⁹ €10.2 billion of that sum were unblocked in December 2023, after the Commission positively assessed a series of Hungarian reforms aimed at restoring judicial independence.Footnote ⁵⁰ The European Parliament later challenged the Commission decision: while the action is pending at the time of writing, on 12 February 2026 Advocate General Ćapeta endorsed the Parliament’s position.Footnote ⁵¹ Under the same legal framework, approximately €76.5 billion were not awarded to Poland, although they were disbursed in February 2024.Footnote ⁵² As for the Recovery and Resilience Facility, it was only in February 2024 that the Commission positively assessed Poland’s recovery plan as including sufficiently effective reforms to strengthen judicial independence. More than €35.4 billion, namely the value of the original plan approved in June 2022, had remained unclaimable until then. The total Recovery and Resilience Facility allocation for Poland, including subsequent revisions and REPowerEU, amounted to approximately €59.8 billion.Footnote ⁵³ Similarly, €10.4 billion in Recovery and Resilience Facility grants and loans were made conditional upon Hungary’s achievement of the 27 super milestones included in the component ‘Governance and public administration’.Footnote ⁵⁴ Such milestones are ‘identical … to the requirements adopted under the Common Provisions Regulation as part of the Charter [horizontal enabling condition] for Hungary’.Footnote ⁵⁵

All these suspensions share a common flavour: they are essentially directed at taming the rule of law crisis. This is no place for a detailed explanation of the crisis.Footnote ⁵⁶ It is enough to notice that the typical institutional guarantees of liberal democracy have been eroded in some member states. In the particularly pronounced cases of Hungary and Poland, this has meant a majority-backed constituent process, the capture of the media by the government, the sidelining of the parliamentary opposition, and the dismantling of a professional civil service. Most notably, several courts have been transformed into enablers of the government.Footnote ⁵⁷ The increased power of the EU purse has been aimed at addressing these deviations.

Towards coercive federalism?

Is the boosted conditionality of EU funds so pronounced as to be coercive? For the first time, the hypothesis that conditions on EU funds may be coercive has been advanced by scholars.Footnote ⁵⁸ Member states must coordinate their fiscal policies, with the Commission playing a key role in evaluating national budgets within the European Semester. Additionally, the approval of recovery plans is a prerequisite for the disbursement of EU funds under the Recovery and Resilience Facility Regulation, with assessments carried out by both the Commission and the Council. Through negotiations, EU institutions influence the reforms that member states adopt. For example, digitalisation must account for at least 20% of national recovery plans; such reforms are essential for accessing the funds. It has been argued that, similar to the US, the EU’s increased spending power enables Brussels to influence and even dictate policies in diverse and often divisive areas, including labour, health, and education.Footnote ⁵⁹

While EU conditionality can also seek to enforce existing EU legislation or even soft law rules (enforcement function), the crux of the matter here is that it can lure the member states into adopting legislation (regulatory function).Footnote ⁶⁰ The distinction between enforcement and regulation is clear in theory but blurred in practice and, in my opinion, it is the latter that dominates the worries about the authority of EU law through funding. Some of the most sensitive suspensions of EU funds, directly connected to the rule of law crisis, require reforms of national law regulating the judiciary or at least the repeal of existing national law. As Garben recently argued, the EU’s response to the rule of law crisis amounts to indirect law-making under the cloak of enforcement.Footnote ⁶¹ Although the Conditionality Regulation has an enforcement function, the norm it enforces is the highly abstract and value-laden Article 2 TEU. It is the substantive reforms this requires at the national level that spark the most controversy.Footnote ⁶²

Enforcement conditions raise what we might call the issue of ‘differentiation’: if member states can be sanctioned by withdrawing money when they do not implement EU law, what use is there for the specific procedures like the infringement action? The Court’s answer, examined below, is that such enforcement is allowed only where it stays distinct in aim and subject matter from the Treaty procedures it duplicates. Other than that, the key issue raised by the growth of EU spending conditionality is still competence creep.

Given this articulated framework, has the EU gone too far in its efforts? Is this illegal federal coercion as in Sebelius?

For the time being, EU law lacks a test comparable to Dole. The purpose of this article is to assess whether such a transplant may be feasible. Before addressing this hypothetical, however, it is worth noting that the first seeds of an EU-Dole test have already been planted in a particularly famous pair of judgments by the Court of Justice. As mentioned earlier, Hungary and Poland, the predictable targets of the Conditionality Regulation, filed separate actions for annulment immediately after its approval. In the twin cases Hungary v Parliament and Council and Poland v Parliament and Council, the Court of Justice rejected their claims and confirmed the legality of the Regulation.Footnote ⁶³ However, in doing so, it also offered a couple of useful remarks on the general regime of EU spending under the current framework of the Treaties.

Hungary and Poland raised various challenges, two of which are crucial. First, the lack of an appropriate legal basis to impose rule-of-law sanctions; second, the circumvention of the procedure under Article 7 TEU.

The first argument is the most relevant for our purposes. The Conditionality Regulation was enacted based on Article 322(1)(a) and (b) TFEU, a provision designed to empower the European Parliament and the Council to adopt the rules necessary for the implementation of the EU budget. According to the Court, this also includes financial rules that ensure the sound financial management of EU resources by member states when implementing the budget in cooperation with the Commission. Since violations of the rule of law, such as the lack of an independent judiciary and the absence of effective judicial review, may pose a threat to the sound management of EU resources, Article 322(1)(a) and (b) was deemed an appropriate legal basis. To the extent that breaches of the rule of law in member states affect, or seriously risk affecting, the sound financial management of the Union budget in a sufficiently direct way, a condition linking the inflow of funds to respect for the rule of law is legitimate.Footnote ⁶⁴

This way, the Court set up a germaneness test. A condition, even a horizontal one, binding in all areas of Union action and on all types of funds, is allowed if functional to the pursuit of a power conferred on the EU. Conditioning access to EU funds on respect for the rule of law, pivotal as this value may be, would not be allowed if it was not likely that by protecting the rule of law the EU would also preserve its finances, a power surely conferred by Article 322 TFEU. Competence creep is thus the one concern the Court is keen on taming, and the long-term result may well be a general obligation of germaneness between the objective pursued through conditionality and the powers conferred on the EU, even beyond the Conditionality Regulation.Footnote ⁶⁵ As a result, in the context of the regulatory function of EU conditionality, the Court set up a test requiring a ‘sufficiently direct link’ between the conditions and the objective of the legal basis grounding the relevant expenditure.Footnote ⁶⁶

The second argument is less interesting yet not irrelevant. The point made by the plaintiffs was that the one procedure to preserve the values of the Union enshrined in Article 2 TEU is the political mechanism of Article 7 TEU, which requires unanimity in the European Council. The Court of Justice replied that the Treaties do not prevent the establishment of disparate procedures to preserve the values of Article 2 TEU, provided that they are sufficiently different in aim and subject.Footnote ⁶⁷ Conditional spending cannot be used to circumvent procedures established in the Treaties and powers specifically conferred on given institutions.Footnote ⁶⁸ This closely resembles an independent constitutional bar test: conditionality cannot incentivise the violation of the Treaties and, specifically, of the principle of institutional balance enshrined therein. Moreover, and as previously mentioned, this also sets up a differentiation test useful in the context of enforcement conditions: enforcement via withdrawal of funds is permissible if the conditionality clauses are different and independent in their aim and subject matter from the competing procedures established in the Treaties.Footnote ⁶⁹

As a result, the proto-Dole test of the EU is particularly focused on the connection between the conditions and the goals which the EU is allowed to achieve (germaneness) and on the prohibition on circumventing the institutional balance enshrined in the Treaties (independent bar). It remains to be seen whether a more general Dole test can move to the EU.Footnote ⁷⁰ Can Mrs Dole go to Brussels?

General welfare

The need to pursue the general welfare is little more than ‘a rhetorical assertion that spending must promote the welfare of the nation as a whole’ in the US.Footnote ⁷⁷ De facto, it only requires that Congress is able to identify some form of nationwide rather than local welfare which might benefit from federal funding. For instance, a higher drinking age may translate into greater sobriety on the highways of neighbouring states.

The EU’s legal capacity to establish funds for the general rather than local welfare is subject to more stringent yet not impassable limitations. The principle of conferral as in Article 5(2) TEU requires that ‘the Union shall act only within the limits of the competences conferred upon it by the Member States in the Treaties to attain the objectives set out therein’. The provision thus commits the EU to act rather than just regulate within the scope of the conferred powers. This clearly rules out any US-style quasi-universal spending power.

The principle of conferral requires the identification of an appropriate legal basis for each fund, which in turn must be based on objective factors amenable to judicial review, in particular the aim and content of the measure.Footnote ⁷⁸ Several provisions have anchored EU expenditure and pursued the Union’s policies through conditionality. For instance, in the 2014-2020 cycle alone, Article 43 TFEU on common agricultural policy has justified measures to prevent pollution of groundwater; Article 77 TFEU on border control grounded the request to declare compliance with the Schengen acquis; Articles 82, 84, and 87 TFEU as part of the Area of Freedom, Security and Justice have conditioned the inflow of funds on compliance with the Charter, the European Convention on Human Rights, and the 1951 Geneva Convention on the Status of Refugees.Footnote ⁷⁹

Nowhere did the EU come closer to a US-style general spending power than in the field of cohesion. To begin with, Article 177 TFEU allows the establishment of cohesion funds which, in turn, have justified a variety of requests for national measures, ranging from indicators to measure progress on digital literacy to requirements for the energy performance of buildings, from the promotion of youth employment to measures to stimulate efficiency in the healthcare sector.Footnote ⁸⁰

Most importantly, the closest to an EU ‘general welfare clause’ is Article 175(3) TFEU, according to which ‘[i]f specific actions prove necessary outside the Funds and without prejudice to the measures decided upon within the framework of the other Union policies’, the Parliament and Council can adopt them with the ordinary legislative procedure and after consulting the Economic and Social Committee and the Committee of the Regions. The aims for which Article 175(3) TFEU can be used are set out in Article 174 TFEU and in particular cover economic, social and territorial cohesion. As it has been argued, ‘[c]ohesion policy is the Union expenditure item that most closely resembles the policy objectives of (regional or federal) States’.Footnote ⁸¹

The debate on how to interpret Article 175(3) TFEU is open. According to a restrictive view, Article 175(3) TFEU should be associated with territorial cohesion between rich and poor regions, as the letter and original meaning of the provision would suggest. Any wider interpretation would require treaty reform for reasons of democratic accountability.Footnote ⁸²

Others argue that, although this is certainly problematic given the letter of the Treaty, cohesion has not been solely about territorial cohesion for a long time. Since the Maastricht Treaty, cohesion funds have been used to finance projects aimed at economic growth and higher employment rather than territorial cohesion. By the 2000-2006 budgetary cycle, only 63% of cohesion funds were allocated to territorial cohesion, and this figure decreased to 51% in the following cycle. Even extremely wealthy areas such as Bavaria or Alto Adige/Südtirol have benefited from cohesion funds. Most of the EU’s key political actors, namely the member states, the European Council, and the Commission, have contributed to this broad interpretation of ‘cohesion’.Footnote ⁸³ It would be hard to justify a much stricter definition of Article 175(3) TFEU, given how liberal cohesion policy has been so far.

Moreover, in recent years, spending on various areas has been authorised based on Article 175(3) TFEU. The entire Recovery and Resilience Facility, which covers funds allocated to the green and digital transition, policies for the next generations such as education and training, and crisis preparedness, is based on this clause. The REPowerEU Regulation (2023/435), which amended the Recovery and Resilience Facility Regulation, has financed a variety of programmes aimed at reducing dependence on fossil fuels and accelerating the green transition.Footnote ⁸⁴ Even the planned boosting of defence spending, the so-called Readiness 2030 plan, may be financed via cohesion funds.Footnote ⁸⁵

A certain degree of realism suggests that the second interpretation is the most likely reading of the first step of an EU-Dole test. Too much past practice, directly or indirectly sanctioned by national authorities and EU institutions, would otherwise be retroactively considered illicit. This is an argument from stability: there is normative strength in a consolidated institutional practice.Footnote ⁸⁶ ‘Cohesion’ is broad enough to accommodate multiple meanings, and there is no a priori reason to prefer an ‘original meaning’ over consistent practice. The former may be deemed more in line with the democratic principle, for it was in the mind of the treaty-drafters, but the practice has been shifting for decades. Throughout the years, on multiple occasions treaty reform was on the table, and never was the reversal of this shifting practice considered. Nor have the member states, whose mastery of the budget is absolute, ever rolled the practice back by reducing the amount of cohesion funds. If consensus is persistent, we must assume a certain agreement among the states. Absence of evidence is not the same as evidence of absence. If it was, this should also be true of principles such as direct effect and primacy, whose normativity rests entirely on long-lasting practice and silent acceptance. Of course, the normative pull of prolonged consistency is not so pronounced that it could not be overturned if the practice was severely incompatible with the values of the system, but I do not believe that this is the case of ‘cohesion’ at the moment. Cohesion is simply not (only) about territorial cohesion anymore, and Article 175(3) TFEU is, if not a proper EU spending clause, at least a ‘cohesion flexibility clause’.Footnote ⁸⁷

As a result, before assessing whether a certain decision to lend or grant funds is legitimate, as a first step we need to consider whether it serves an EU-wide interest in territorial, social and economic cohesion. Unlike in the US, this is not a rhetorical question, since in the EU some non-economically related areas, such as family law or public security, would likely fall outside the test. However, it is not a demanding test either.

Unambiguous expression

Second, conditions must be unambiguous. This is probably the most straightforward of all conditions, little more than an iteration of Lon Fuller’s statement that ‘obscure and incoherent legislation can make legality unattainable by anyone’.Footnote ⁸⁸

In the US, it is sufficient for Congress to specify the conditions in the relevant bills.Footnote ⁸⁹ That funds would be withdrawn if the minimum drinking age was not set at 21 in state law was perfectly transparent and made so by adequate legislative specification.

In the EU, the picture is again more complex, but does not necessarily provide an impediment. This is due to the political guarantees of European federalism. While the Senate has shifted from an assembly of states’ delegates to a national institution in the US, in the EU the representation of the states is robust, and perhaps nowhere more solid than when deciding on the budget.

The Own Resources Decision establishes the EU budget by determining the share of customs duties, agricultural levies, value added tax, and national contributions to the EU budget. According to Article 311 TFEU, this is decided unanimously in the Council, the EU’s ‘second chamber’, and must then be approved by each member state according to their specific constitutional requirements. As budgetary procedures go, this resembles treaty amendment more than ordinary legislation. The Multiannual Financial Framework, which sets the upper limit of yearly commitment appropriations for each sector of activity, is again determined unanimously by the Council, though with the consent of the European Parliament under Article 312 TFEU. The annual budget is also determined with the Council’s approval. Moreover, like all EU legislation, the legislative instruments that have so far imposed special conditions on EU funds, such as the Conditionality Regulation or the Common Provisions Regulation, have also been approved by the Council.Footnote ⁹⁰ Finally, whenever the new method of government inaugurated in Next Generation EU is applied, the member states negotiate their national plans individually with the Commission, discussing the milestones and targets to be achieved in exchange for the EU loans or grants.Footnote ⁹¹

As a result, not a single cent can be spent without the involvement of the member states’ representatives in the Council: the EU purse remains under tight intergovernmental control.Footnote ⁹² While the American states do not participate directly in the establishment of federal conditions, in the EU the member states have ample opportunity to demand clarification.

This does not mean that the conditions are always clear. For instance, the aforementioned Article 9 of the Common Provisions Regulation requires compliance with the Charter when implementing projects financed by the named funds. It also demands that ‘Member States and the Commission shall ensure that equality between men and women, gender mainstreaming and the integration of a gender perspective are taken into account and promoted throughout the preparation, implementation, monitoring, reporting and evaluation of programmes’. These conditions are sufficiently vague to represent a delegation of power to the Commission, which in turn will have considerable powers of specification. The exact meaning of clauses like Article 9 may not be unambiguous.

A member state which insisted on this ambiguity and thus refused to approve the Regulation, one that had concerns about clarity and still voted for it for unrelated reasons, and even a member state originally convinced of the viability of the conditions and confronted with the reality of ambiguity later on, should be allowed to prove that the conditions were unclear. In a hypothetical application of the EU-Dole test, the member state should thus be permitted to demonstrate that the conditions were ambiguous when challenging the measure. The burden of proof, however, should lie with the member states in all such cases.

To summarise, the ample involvement of national representatives in the determination of the conditions on EU funds suggests that they know their meaning or had opportunities to ask for clarification. However, the member states should be left free to demonstrate that in the case at hand this was not true.

Germaneness

Third comes germaneness. In essence, this is the problem of connection between means and aims, one as old as federalism itself. Ever since McCulloch v Maryland, the notion that federal aims entail the (implied) powers functional to their realisation has been a typical characteristic of compound systems.Footnote ⁹³ The vexed question is how direct this relationship must be before federal action qualifies as overreach. Most frequently, this is more a political than a legal question: depending on the historical context and the consequent need for centralisation, sufficiently flexible clauses will be read in a broader or narrower manner to fit the needs of the time.Footnote ⁹⁴

In the US, this has been most apparent in the already mentioned broad interpretation of ‘commerce’, which has favoured the rise of a quasi-plenary national legislative power.Footnote ⁹⁵ When it comes to funds, conditions must be functional to federal programmes or projects. In Dole, the lack of homogeneous rules on drinking age was deemed disruptive of the federal interest in a well-functioning interstate highway system. While this was the majority opinion, in her dissent Justice Sandra Day O’Connor noted that this ‘attenuated and tangential’ connection would allow Congress to regulate anything related to federal highways, but did so to no avail and, perhaps most importantly, with no workable alternative.Footnote ⁹⁶

In the EU, the growth of central powers functional to the aims of the Union is no novelty either: ERTA established implied powers in external relations parallel to explicit internal competences,Footnote ⁹⁷ the EU’s ‘Necessary and Proper clause’ in what is today Article 352 TFEU has been used to establish an EU environmental policy before any mention in the Treaties,Footnote ⁹⁸ and most notably the EU’s ‘Commerce clause’ in today’s Article 114 TFEU has been liberally used to harmonise market-related areas.Footnote ⁹⁹ Tobacco Advertising I famously erected guarantees against Article 114 TFEU’s overreach: harmonisation must tackle real obstacles to trade or appreciable distortions of competition, and only likely future disparities qualify.Footnote ¹⁰⁰ In reality, however, Article 114 TFEU has been used for all sorts of legislation, a federalising tool barely patrolled in court.Footnote ¹⁰¹

As previously mentioned, however, the judgments on the Conditionality Regulation have started tackling the problem of germaneness of conditions: conditions must be directed at a competence assigned to the EU. Subjecting the inflow of EU funds to the guarantees of judicial independence is feasible only to the extent that this is also functional to a financially sound EU budget, a power explicitly recognised in Article 322 TFEU. That this may also be beneficial to the value of the rule of law, arguably an objective of the Union, is only tangential.

EU spending conditionality must follow the legal basis of the fund to which it is attached.Footnote ¹⁰² Although the judgments about the Conditionality Regulation do not amount to a general test for spending conditionality, they seem to imply that it must also pursue the aim of the legal basis in a sufficiently direct manner.

Compared to Dole’s test of relevance, the EU equivalent is more rigid.Footnote ¹⁰³ In the US, where Hamilton’s view eventually prevailed over Madison’s, the spending power is disconnected from the enumerated federal legislative powers. Federal spending can be directed at any aim, provided that it pursues the nation’s general welfare. The only limit is that Congress must identify some connection, however remote, between the aims of the funds and the conditions. Dole’s streamlining of the drinking age indirectly served one of the purposes of highway spending, the promotion of safety.

Due to the principle of conferral, in the EU this seems to transform into an analysis of the relationship between the objectives of the legal basis and the conditionality: the latter must be functional to the former.Footnote ¹⁰⁴ However, it remains to be seen how the notion of ‘sufficiently direct connection’ will be interpreted and applied.

For starters, purposes ancillary to the primary goal of a measure’s legal basis can be pursued, so long as the measure can be autonomously justified: if an EU measure harmonises the advertising of tobacco products and can be grounded in Article 114 TFEU, it can also have the purpose of secondarily safeguarding public health.Footnote ¹⁰⁵ Likewise, where conditions are merely ‘accessory’ components of the act governing the fund or programme, their inclusion does not shift its centre of gravity and may rest on the same legal basis as the predominant objective.Footnote ¹⁰⁶

Second, the legal bases are often vague. The case of ‘cohesion’ is a paradigmatic example, but other provisions are also open to debate, and so is the question of what serves their purposes adequately. If there is a sufficiently direct link between the objective of the legal basis and those of the conditions, the latter are lawful. For instance, it does not seem arbitrary to condition the inflow of funds based on Article 43 TFEU concerning common agricultural policy on the adoption of measures preventing the discharge of dangerous substances: a clean environment may be conducive to a rational development of agricultural production.

The connection between the legal basis and the conditions becomes increasingly feeble when the latter are specific to a separate thematic area, linked to distinct EU spending programmes or policies, or even to cross-cutting parameters like the Charter or the European Convention on Human Rights.Footnote ¹⁰⁷ However, a connection between the aims of the legal basis and these ‘horizontal’ conditions may at times still be envisaged. For instance, the primary aims of Articles 82, 84, and 87 TFEU are judicial cooperation in criminal matters, crime prevention, and police cooperation, but it does not seem unreasonable to imagine that respect for the Charter and the European Convention on Human Rights, which do not exclusively affect these areas, is a condition sufficiently connected to these primary aims.Footnote ¹⁰⁸ The delicate rights at stake in criminal law and procedure may be seen by some as more directly connected to a bill of rights compared to more mundane areas like transportation development. The connection between the legal basis and the conditions is subject to considerations of degree and to subjectively different assessments.

Finally, where that link is assessed as too remote, which is certainly more likely in the case of horizontal conditions, entirely separate legal bases become the only option.Footnote ¹⁰⁹

One lesson we might extrapolate from the other side of the Atlantic is that the decision about what connection is ‘sufficiently direct’ belongs to the legislative power, with a relatively modest degree of judicial review. In the US, the connection between the federal aims and the function of the conditions is almost entirely in the hands of Congress. In the EU, the ‘federal aims’ are predetermined in the legal bases, but the identification of the connection, however ‘sufficiently direct’, between the conditionalities and the aims of the legal bases should still be left to the Parliament and Council. For all their limitations, and despite having weaker credentials than the US Congress, these are democratically legitimate institutions, whether directly or indirectly. Whatever deference is accorded by the Court of Justice to these institutions when choosing the legal basis, drafting legislation, and pursuing ancillary purposes, it should also be recognised when allotting funds and establishing conditions. A more demanding scrutiny may be required when the conditions are set not by the legislature but by the Commission, which does not enjoy the same legitimacy, and when the degree of connection to the aims of the legal basis is prima facie weak, as in the case of horizontal conditions. However, given the inherently political nature of the identification of a condition’s link to the legal basis, in principle the judicial deference should be substantial.Footnote ¹¹⁰

No unconstitutional conditions

No condition can lead to the violation of an independent constitutional rule. The fourth condition borders on the trivial and, despite attempts to interpret it more broadly, in the American context it seems to merely prohibit luring the states into violating explicitly enumerated constitutional rights by threatening to withdraw federal funds.Footnote ¹¹¹ Naturally, the same would be true in the EU: offering or withdrawing EU funds to adopt measures which violate the Treaties or EU law at large would be illicit. This seems to have already been suggested in the judgments on the Conditionality Regulation when arguing that conditionality as a tool to preserve the values of Article 2 TEU is a viable option only to the extent that it is sufficiently diverse in aims and subject matter from other mechanisms already established in the Treaties like Article 7 TEU. Providing otherwise would mean circumventing an independent constitutional rule, namely the principle of institutional balance, in EU jargon.

The problem here is one of enforcement. In the US, anyone whose rights may be affected can go to court and try to have the illicit condition quashed. In the EU, the guardian of the Treaties under Article 258 TFEU is also the subject which would most likely induce the illegal behaviour, namely the Commission. The incentive to bring the case to court is minimal.

Nevertheless, alternative routes are available. The other member states can denounce the resulting national action via Article 259 TFEU, just like the Commission. Moreover, the EU measure imposing the illicit condition can be challenged via an action for annulment by the threatened member state as a privileged applicant and by the affected natural or legal persons too, if the EU measure is of direct and individual concern. Finally, preliminary references raised by national judges may lead to the disapplication of the illicit national action or to the declaration of invalidity of the EU measure demanding it.

Coercion

Lastly, the financial inducement must not be so intense as to turn pressure into coercion or, to quote from Sebelius, must not be a ‘gun to the head’. I would argue that this last condition should be part of the EU Dole test, but would only apply in extremely unlikely circumstances, for very little financial coercion is currently credible in the EU. This is true for four reasons: financial capacity; power of apportionment; the political safeguards of the member states; and the EU’s type of executive federalism.

To begin with, the credibility of the ‘gun to the head’ in Sebelius derived from the fact that the states would face a loss of about 10% of their total annual budget, compared to the less than 1% of Dole. In a context in which federal funds make up a third of the annual state budget, the power of the federal purse is substantial. At this stage, this is not true of the EU. In 2021 the EU’s share of public spending in the member states was on average just below 2%, reaching 2.7% when adding Next Generation EU resources. Truly, some member states are far more reliant on EU funds than others, with countries like Lithuania, Croatia and Latvia having between 10 and 12% of their public spending financed by EU funds, but this remains well below US levels.Footnote ¹¹² Moreover, to match the level of financial pressure in Sebelius, all EU funds would have to be suspended at once. Not even the most blatant violations of the horizontal conditions on the rule of law have justified anything similar so far. The closest we get in the EU is the enabling condition to the Common Provisions Regulation, which refers to several funds simultaneously, but it does not encompass the entirety of the EU financial contribution. Moreover, even if it did, such a gargantuan suspension would only coerce the most reliant member states. For the time being, the suspension of EU funds is more likely to impact like Dole rather than Sebelius: the EU budget simply is not large enough to coerce the member states. Nor is it going to be anytime soon: by estimates of the European Parliament, even the budgetary expansion proposed by the Commission for the 2028-2034 cycle, which is far from certain to be approved, does not cross 1.26% of the bloc’s gross national income.Footnote ¹¹³

Second, the US Congress has vast autonomous power over apportionment through taxation or debt, which the EU does not possess. The member states remain the masters of the EU budget. This is true not only because more than 60% of the EU budget currently comes from national contributions rather than autonomous levies, but also because, as previously mentioned, the member states decide how the EU budget is shaped, both in size and distribution.Footnote ¹¹⁴ The EU cannot simply decide to levy taxes or issue debt bonds and raise the resources necessary to coerce its constituent units: quite the opposite, the member states retain unrivalled powers to tax, borrow, and spend. This makes the argument advanced by Justice Ginsburg in Sebelius particularly compelling in the case of the EU. If Florida can be asked to face the consequences of its refusal to adhere to the expanded Medicaid by using its autonomous taxing power to raise the missing resources, a fortiori so can the EU member states. Let them refuse the money and answer for it before their voters.

To some extent, this is happening already. Hungary, one of the most financially dependent member states and the most frequent addressee of withdrawals for violations of rule of law-related and fundamental rights-related conditions, has been willing to endure part of the loss of EU funds due to the violation of the Conditionality Regulation.Footnote ¹¹⁵ During Orbán’s tenure as Prime Minister, the government deemed the political cost acceptable, alternative routes of financing viable, and the aim of controlling the media or pushing back asylum-seekers worth the price.Footnote ¹¹⁶ The electors eventually ousted Orbán’s government in April 2026: relations with the EU were central to the campaign, and tellingly one of new Prime Minister Péter Magyar’s first acts was a visit to Brussels, where on 29 May 2026 he secured the Commission’s agreement to begin unblocking some €16.4 billion in frozen recovery and cohesion funds, subject to the completion of rule-of-law and anti-corruption reforms.Footnote ¹¹⁷ Ultimately, the decision on what financial loss is too much belongs to the electors.

Third, the institutional setting of the EU gives the member states robust political guarantees of their prerogatives and consequently a bargaining power which simply does not exist in the US. Hungary was able to get itself off the hook by merely threatening to veto one round of financial aid to Ukraine in 2023. The silent deal which likely followed freed half of the suspended cohesion funds in exchange for the lifting of this veto.Footnote ¹¹⁸ As long as the decision-making process in the EU is still partially intergovernmental, even economically weak member states can easily counter any form of EU financial coercion.

The three arguments put forward so far suggest that federal coercion in the EU is unlikely. An additional argument against an EU ‘gun to the head’ step hinges on its normative foundations. As was mentioned when discussing Sebelius, the limit of coerciveness in the US has been developed to complement the prohibition of commandeering, a doctrine rooted in the Tenth Amendment and in the dual structure of US federalism. As the Supreme Court articulated it in New York v United States (1992), Congress may not commandeer state regulatory processes by ordering states to enact or administer a federal programme.Footnote ¹¹⁹ Printz v United States (1997) clarified that Congress may not circumvent that prohibition by conscripting state officers directly, such commands being, in the Court’s view, fundamentally incompatible with the US constitutional system of dual sovereignty.Footnote ¹²⁰ The doctrine thus presupposes that the federal and the state levels each retain their own legislative and implementing apparatuses, so that the federal power cannot, in principle, conscript state institutions to administer its programmes. Just as the federal government cannot use state officials to implement its policies, it also cannot circumvent this prohibition by financial incentives. As Roberts said, such an arrangement would prevent sound accountability for decision-making.

This is not the case on the other side of the Atlantic. In the EU, national administrations are the natural enforcers of EU law, with centralised enforcement being the exception.Footnote ¹²¹ Even national legislation serves the purpose of implementation, most notably when transposing directives. Despite criticism of the resulting blurring of political accountability in the EU,Footnote ¹²² such pronounced executive federalism is anchored in the constitutional principle of subsidiarity.Footnote ¹²³ The cultural and linguistic proximity to the citizens guaranteed by subsidiarity has a weight in Europe which is unknown to the US.

Thus, the cooperative nature of EU federalism, as opposed to its dual counterpart in the US, is what makes it fundamentally distinct. In light of this, the American rationale that financial coercion is prohibited because it would remove the clear boundaries between federal and state governments and prevent the public from knowing who the actual decision-makers are, is less relevant in the EU.

For all these reasons, a credible ‘gun to the head’ is extremely unlikely in the EU (the first three arguments), and in any case it would be based on weaker normative justifications (the fourth argument).

Still, there is some limited space for a quasi-gun to the head step in the EU too. Imagine a fictional scenario in which a highly dependent member state has all its EU funds frozen by the Commission and Council under the Conditionality Regulation due to minor issues with its anti-corruption authorities. If the country were to lose a substantial portion of its budget overnight, say around 10%, then it might be possible to imagine a form of coercion. In the jargon of EU law, this would likely be considered a severe violation of the principle of proportionality, and more precisely a failure at its necessity prong.Footnote ¹²⁴ Suppose the suspension were triggered by a narrow, readily curable shortcoming, say a delay in submitting the asset declarations of a handful of officials, or an integrity authority left temporarily understaffed. Freezing the entirety of the envelope in response would not be necessary: a targeted or partial suspension was plainly available as a less onerous measure that would have equally protected the Union’s financial interests. It is only in such cases of manifest excess that something resembling the Sebelius ‘gun to the head’ could re-enter EU law through the door of gross disproportionality. Conceptualising such withholding of funds in these terms is also advantageous for the Court of Justice, since the latter is accustomed to intervening only where the institutions manifestly exceed the limits of their discretion, in any case placing the burden of proof on the claimant.Footnote ¹²⁵ Such additional precautions are justified by the many ways in which the member states may, as we have seen, escape any pressure from Brussels. At any rate, this scenario is very unlikely to materialise in the foreseeable future, unless there are reforms that massively increase the size of the EU budget, grant the EU appropriation powers, and reduce the instances of unanimous decision-making in the Council.