Europe’s Future: A Federal Alternative to Differentiation

The article aims to show the negative side-effects of differentiated integration (DI) for the democratic quality of the European Union (EU), to then investigate an alternative model that, theoretically, promises to promote integration and to uphold the said democratic quality. The article’s argument is that DI is the response to the resistance that the integration of policies that partake to the sovereignty of Member States (i.e., core State power – or CSP – policies) generates in some Member States. The accommodation of that resistance has required the adoption of an intergovernmental logic to manage it. The paper’s research questions are thus the following: what are the institutional consequences of DI? Is there an alternative theoretical model for integrating CSP policies? 

Here is the article’s road map. First, I will discuss why sovereignty-induced DI has triggered the intergovernmental development of the EU, while this has not been the case with capacity-induced DI (which has mainly concerned the regulatory policies of the single market or SM). Second, I will highlight the inability of the intergovernmental regime to guarantee the accountability of its decision-makers (as well as their effectiveness). Third, I will delineate an alternative model to DI to integrate policies that are traditionally close to areas of national sovereignty, namely the federal union model (as a distinct model from the federal state) and its problematic features. Here I will discuss the theoretical model, not the strategy for implementing it (the “how to go from here to there”). The Conclusion summarises the argument developed in the article. 

Sovereignty-induced differentiation

Differentiated integration (DI) is both a theory and a practice. As a practice, it has become the official public philosophy of the EU ¹ . National leaders and supranational actors have come to share the view that DI constitutes the only political strategy to adapt the process of integration to an increasingly plural landscape of national conditions and preferences. Particularly with the enlargements of the 1990s and 2000s, the argument goes, the lack of homogeneity among the EU Member States has increased dramatically, to the point of requiring an adaptation of the process of integration to that plurality of national interests and preferences. Through DI, different combinations of EU Member States participate in different policy regimes, although these regimes are open to changing compositions and all Member States are expected to join later. For instance, the economic and monetary union (EMU) is constituted by 19 of the EU 27 Member States; 22 out of the 27 Member States (plus four States not members of the EU) participate to the Schengen agreement, which abolished national borders; the Permanent structured cooperation in security and defence policy (PESCO) registers the participation of 25 out of 27 EU Member States; the Prüm Convention on cross-border cooperation for contrasting terrorism concerns seven out of 27 EU Member States. These are all examples of legalised differentiated policy regimes. However, there are also forms of DI that have not acquired any legal distinction, such as the 2019 Malta agreement consisting in a “Joint declaration of intent on a controlled emergency procedure – voluntary commitments by Members States for a predictable temporary solidarity mechanism”, entered into between four out of the 27 EU Member States. Undoubtedly, DI has worked quite effectively in aggregating a growing number of Member States within the EU’s legal framework. It has been the political formula for combining the deepening of institutional integration and the widening of EU membership. The Brexit experience confirmed the convenience, for the same Eurosceptic Member States, of using DI to protect their policy preferences, given that exiting from the EU would imply much higher costs (in social and economic terms). 

The practitioners’ view has been shared by scholars as well. Distinguishing between differentiation (as a generic phenomenon of all modern political systems) and DI (as a specific feature of the European integration process) ² , scholars have conceptualized the latter as a permanent feature of the integration process, particularly after the 1992 Maastricht Treaty ³ . The two mainstream theories of European integration have stressed its physiological character. For neo-functional scholars ⁴ , differentiation consists of different timing in Member State adaptation to EU laws, although all Member States are considered to share the direction of the process of integration (that is, the formation of a supranational organisation). For liberal intergovernmental scholars ⁵ , differentiation reflects the irreconcilable preferences of Member States, acting as rational actors, in dealing with the dilemmas of cooperation. If neo-functionalism scholars are motivated by a teleological view of European integration whose outcome is inevitably the formation of an encompassing and cohesive supranational organisation, liberal intergovernmentalism scholars are motivated by a realist view of European cooperation, where differentiation is the tool for solving inter-State negotiations. While for the former group of scholars, the EU is a State-in-the-making (although of a federal type, to use Lijphart’s terms ⁶ ), the latter view it as an international organisation based on States-as-they-are. For the former, DI is the strategy for accommodating different Member State preferences and perspectives in participating, although with different timing, in the building of a supranational State-like entity (multi-speed Europe) ⁷ . For the latter, DI is the ad hoc solution to keep different national interests around the table, with the EU being interpreted as an international organisation consisting of autonomous European states (Europe à la carte) ⁸ . 

The more recent literature on DI has helped identify its various forms ⁹ . Schimmelfennig and Winzen have distinguished between instrumental differentiation and constitutional differentiation or between internal (to the EU treaties or legal order) and external (through international treaties signed by EU Member States) differentiation ¹⁰ . Winzen had previously conceptualised the reasons and forms taken by DI, distinguishing between a DI motivated by a Member State’s lack of capacity and resources in adapting to the rules and practices that constitute the acquis communautaire and a DI which emerged out of the Member State’s resistance to participation in a specific policy regime which might jeopardize its control over national resources ¹¹ . DI induced by a lack of capacity may be detected in several policy areas and in different periods of time, but it is generally a problem that it is destined to be solved. It emerged as an effect of the enlargement process with new Member States being unable to meet some of the policy requirements of EU membership, because of domestic administrative and institutional hurdles in adapting to EU regulations. In general, having recourse to directives rather than regulations has historically helped guarantee flexibility in the implementation of single market rules by Member States ¹² . Quite different is the DI triggered by a Member State’s defence of its claims to sovereignty. This type of DI has emerged only in the post-Maastricht period when policy fields traditionally making up CSP were brought to Brussels to be managed. Sovereignty-induced DI was further claimed during multiple crises in the 2010s and the 2020 pandemic, considering that those crises touched on areas falling under the scope of national sovereignty prerogatives.  

The contribution of the more recent literature on the conceptualization of DI has been remarkable ¹³ . Particularly, the contribution by Winzen has opened a promising theoretical road for conceptualizing the peculiarity of sovereignty-induced DI ¹⁴ . However, the governance implications of the latter have not been sufficiently discussed. In fact, the recourse to sovereignty-induced DI has required the use of a specific intergovernmental governance regime, based on national governments’ voluntary coordination rather than on legally regulated inter-institutional relations. Thus, differentiated policy regimes are correlated to differentiation in governance regimes. In the next section, I will investigate such differentiation in governance, connected to sovereignty-induced DI, to delineate its institutional features and democratic consequences. 

The governance of sovereignty-induced DI

A large part of the literature on DI assumes that differentiated policy regimes are managed by a single decision-making structure expressing a supranational logic of integration ¹⁵ . Certainly, it is acknowledged that inter-institutional relations might change in dealing with the different policy responsibilities of the EU, although it is assumed that those inter-institutional relations are constrained by the same logic of integration ¹⁶ . 

However, what has come to characterise the post-Maastricht EU has been the institutionalization of distinct decision-making regimes. Heuristically, one may distinguish between two decision-making or governance regimes, one dealing with issues of low domestic political salience regarding single market regulatory policies and the other dealing with policies of high domestic political salience which entered the EU agenda after the end of the Cold War. 

The 1992 Maastricht Treaty was the turning-point for this differentiation in governance structures. Through that Treaty it was formally recognised that the EU could proceed in the integration process of domestically crucial policies (the CSP policies close to national sovereignty) ¹⁷ provided that the Member State governments were guaranteed an exclusive or predominant decision-making role (through the Council and European Council) over those policies. 

The 2009 Lisbon Treaty constitutionalised the distinction between different decision-making regimes in relation to distinct policies. It strengthened the supranational decision-making regime for single market policies, which became the ordinary legislative procedure (based on the triangulation among Commission-Council-European Parliament (EP)), and it institutionalized an intergovernmental decision-making regime for policies traditionally partaking to areas of national sovereignty ¹⁸ , with the European Council as a collegial executive. According to Fossum, the (supranational) Community system and the (intergovernmental) Union system have thus come to coexist within the same political and legal order ¹⁹ .

The distinction between the two governance regimes is necessary in order to conceptualise the correlation between the institutionalisation of intergovernmental governance and the Europeanisation (integration but not necessarily supra-nationalisation) ²⁰ of CSP policies, a correlation that is missing in DI literature. Those policies, in fact, substantiated the reality and the symbols of national sovereignty, the historical raison d’être of the State’s existence. Theoretically, those policies too might have been managed by the supranational governance introduced in the 1957 Rome Treaties. However, the latter has a logic of integration that does not fit with the need to control, by the single Member State government, their Europeanisation. Supranational governance, in fact, has made possible the formation of the most integrated single market in the world (much more integrated than the US) ²¹ thanks to a regulatory approach that has standardised and homogenised its policies. Overloading the single market’s formation of quasi-constitutional aims (the “ever closer union”), the supranational approach has administratively centralised its functioning through the transfer to the Brussels’ triangle of institutions embodying the community method of most market-related policies. As the United Kingdom (UK) had to acknowledge, the four freedoms constituting the single market (free movement of goods, capital, persons, and the freedom to establish and provide services) are not negotiable, as if they represented constitutional principles and not macro-economic criteria. 

This could hardly be the case for CSP policies. Their standardisation and homogenisation would have depleted national sovereignty without compensation. Member State claims on preserving their control over national sovereignty’s resources and symbols were and are structural, not contingent, because they express consolidated national preferences, views or identities. For this reason, national governmental leaders adopted a different model for governing the Europeanisation of CSP policies, different from the supranational one ²² . Without the formation of the intergovernmental regime, it would have been implausible to have DI of CSP policies. The most significant cases of DI in these policy regimes display an intergovernmental logic that is alien to the logic presiding over the decision-making process in single market regulatory policies. The latter are integrated through the ordinary legislative procedure, with its compulsory character, while CSP policies are integrated through voluntary coordination among national governments. 

The EMU economic (and not monetary) policy, the Schengen asylum policy, the Permanent structured cooperation in security and defence policy (PESCO), the Prüm Convention on cross-border cooperation for contrasting terrorism, have all been made possible by the voluntary nature of the policy-making process and all have come to be organised by one form or other of intergovernmental governance, i.e., by the logic of voluntary policy coordination. Few policies started as differentiated policy regimes thanks to their intergovernmental logic, although they then evolved in a supranational direction (as in the cases of reinforced cooperation on divorce and European patents). 

The voluntary basis of CSP policymaking is guaranteed by the consensual logic of the deliberation that takes place within the European Council (or those Council formations dealing with CSP policies, such as the Eurogroup of Economy and Finance ministers of the EMU or the Foreign Affairs Council). In regulatory single market policies, a Member State can have extra time to adapt the EU laws to its domestic structures, but it cannot call into question the constitutive rules (the four freedoms) that keep the single market together. On the other hand, sovereignty-induced DI consists in the institutionalisation of a distinction between a majority of Member States agreeing to manage specific CSP policies in an intergovernmental fashion in Brussels and a minority of Member States unwilling to follow suit but allowed to opt out of the process without halting it. Only a decision-making method based on voluntary coordination could make this possible, without disrupting the integration canvas. If policy differentiation on CSP and intergovernmental governance are intertwined, then the understanding of the former cannot be separated from the latter. 

With the multiple crises in the 2010s, as well as with the 2020 pandemic, that affected areas following in the scope of CSP policies (e.g., fiscal, migratory, security and health policies), the European Council has emerged as the unquestioned central decision-making institution of the EU. Indeed, it was the European Council that decided how to deal with the financial crisis and how to respond to the possible financial bankruptcy of Greece in the early 2010s, to strike a deal with the Turkish government for containing the flux of Syrian refugees in 2015, to sterilize the Commission’s proposal to redistribute refugees in Member States according to objective criteria in 2016. The European Council has become not only the executive centre of the EU, but it has come define the direction to be followed by national legislatures in implementing its political decisions, even to dictate how to interpret judicially the legislative measures taken through the triangulation of the Commission, the Council and the EP. In the attempt to neutralize the veto threatened by the Hungarian and Polish prime ministers to the rule of law conditionality attached to Regulation no. 2020/2092 relating to the use of EU funds, in the meeting’s conclusions of 10-11 December 2020 “the European Council underlines that the Regulation is to be applied in full respect of Article 4(2) TEU, notably the national identities of Member States inherent in their fundamental political and constitutional structures (thus stressing that, ed.) the guidelines will be finalised after the judgment of the Court of Justice so as to incorporate any relevant elements stemming from such judgment”. In instructing the Commission and the EP, but also the ECJ, on how to interpret the Regulation, the European Council overstretched again its role, according to a pattern already identified by Fossum ²³ . 

However, the European Council has also displayed structural limits from a democratic perspective. It has shown the ineffectiveness of a decision-making process which depends on unanimity but also the unaccountability of the real decision-makers. Indeed, the EP could not play a checking role on the European Council, nor was that role exercised by a collection of national Parliaments. Above all, the European Council, in claiming to be the only EU institution legitimised to manage sovereignty-based issues, ended up politicising them, thus rendering their management even more difficult ²⁴ . 

This raises the following question: is there an alternative model to be considered for integrating sovereignty-induced claims?

Sovereignty in democratic federations

After the 1950s (particularly after the 1954 rejection of the European Defence Community by the French legislature), federalism (as a theory) has gradually disappeared from the debate on European integration (notwithstanding the fact that federal principles have continued to influence “the building of Europe”) ²⁵ , with neo-functionalism and liberal intergovernmentalism becoming the mainstream theories for interpreting (and influencing) that process. The affirmation of the latter theories and the decline of federalism has had an impact on the conceptualisation of European integration. Contrary to federalism’s predisposition to privilege the analysis of the institutions of integration, neo-functionalism and liberal intergovernmentalism have rather a predisposition to focus on the process of integration. 

The absolutisation of processes over institutions has led the two mainstream theories to under-conceptualise the consequences of the various solutions or compromises agreed upon (for instance, the DI), as well as their implications for the democratic quality of the EU ²⁶ . Certainly, federalism has continued to inspire politicians ²⁷ , representing a sort of underlying culture of the EP’s mainstream parties. The reference, for this federal political culture, has continued to be the experience of the most influential European federal state, the post-WWII Germany, also because of the highly influential role played by representatives of the German Christian Democratic Union within the main EP party, the European People’s Party. Indeed, it might be argued that the German federal State’s model ²⁸ has inspired the logic of administrative centralisation and fusion between levels of governments pursued in the making of the single market ²⁹ , if not the latter’s ‘over-constitutionalization’ ³⁰ . 

However, the federal State model cannot answer the question epitomised by DI, namely the integration of sovereignty claims within the Union. The identification of a federal alternative to DI requires a comparative investigation of the experience of democratic federations ³¹ . Since the pioneering works of Sbragia ³² and Stepan ³³ , the experiences of federations have been classified according to their historical path of formation and consolidation. Sbragia distinguished between federations by aggregation and federations by disaggregation, while Stepan conceptualised the basic distinction between coming-together and holding-together federations. For both scholars, the former expresses the aggregation of previously independent territorial States and the latter the territorial disaggregation of a previously unitary State. The institutional structures of the two federal models, and their underlying rationale, reflected that founding logic. Coming-together federations are much less centralised than holding-together federations, because the units that activated the aggregation aimed to maintain as much power (and competence) as possible in their own hands. They even constrained the centre from within, separating its decision-making institutions. On the contrary, holding-together federations kept the territorial disaggregation under the pre-eminence of a federal centre and organised the latter according to the logic of fusion of powers. In ideal type terms, we might call the former federal unions and the latter federal States. Along the entire (democratic) federal spectrum, the United States (US) and Switzerland are closer to the federal union pole, while post-WWII Germany, Austria and Belgium are closer to the federal State pole, with Australia and Canada in between the two ³⁴ . Although all federations are based on the twin principle of self-rule and shared rule ³⁵ , in federal unions, self-rule and shared rule are institutionally separated (each level has its own competences and institutions for carrying them out), while in federal States the distinction between self-rule and shared rule is less compelling because competences are mainly shared between levels of government and their management implies the latter’s close cooperation. In other words, multiple separation of powers keeps federal unions together, cooperation and fusion of powers keeps federal States together.

This institutional distinction between the two models epitomises different interpretations of federal sovereignty ³⁶ . In holding-together federations, federal sovereignty belongs to the central institutions of the federation. In post-WW II Germany, sovereignty is represented by the Bund, i.e., by the institutions of the parliamentary government (Bundestag and Bundeskanzler), integrated by the institution representing the Länder’s executives (Bundesrat). The Bundestag is the political chamber of the federation, in the sense that it is elected periodically by the German voters with the aim of forming a parliamentary majority supporting the Chancellors and their governments. The Bundesrat (constituted by representatives of the Länder’s governments) has no say in the formation of the federal/parliamentary majority. None of the Länder, as a single unit, can constitutionally claim to be sovereign, although it has “the constitutional right to conclude treaties” based on its “quality of State (…) in fields of exclusive Land competence” ³⁷ . The quality of State has institutional, not sovereignty’s, implications. Considering both exclusive and shared competences, the federal centre handles a large part of federal policies, although the Länder contribute to the policy-making process in matters relevant for them through the Bundesrat, their representatives in the committees connecting the two levels of government, and their power to implement the policies decided centrally ³⁸ . In Germany, the growing policy responsibilities of the centre have been achieved through a cooperative logic between shared rule and self-rule, although federal government has retained the final say ³⁹ . This model of cooperative federalism has an internal dynamic to standardise and homogenise policies across the Länder, although it allows flexibility in the implementation of federal policies at the level of single Länder  (as happened with the five eastern Länder, plus eastern Berlin, entering the federation in 1990) ⁴⁰ . Differentiation here was capacity-motivated, the solution being the transfer of resources to a needy Land to speed up its homogenisation with the other Länder. No Land has ever claimed a specific differentiated regime on behalf of its own sovereignty. This fused and centralizing system has guaranteed the accountability of the federal decision-makers. The federal governments, supported by their majority in the Bundestag, are responsible for governmental decisions and respond for their effects to voters in periodic parliamentary elections. 

The federal union approach to sovereignty claims

In coming-together federations, such as the US and Switzerland, instead, the place of sovereignty has continued to be a disputed issue. Since it was the States/cantons that decided to aggregate, they tried to keep as many competences as possible ⁴¹ , although they acknowledged the need to share the policies that could guarantee their security ⁴² . For this reason, sovereignty was divided among the federal centre and the federated States/cantons, although that distribution passed through a constitutional act that formally kept sovereignty undivided. Through the constitution, the federation’s founders transferred sovereignty to the new centre only insofar as necessary to protect the federation, or to enable its economic development, keeping for the States/cantons all other policy responsibilities. In the US, federalism consisted in the shared rule (at the federal centre) of the CSP policies for guaranteeing the federation’s security from external and internal threats and in the self-rule (in the federated States) of market-making policies in accordance with each States’ cultural traditions and economic structures. Certainly, the federal centre could claim (as made possible, in the US, by the Commerce Clause of the 1787 Constitution) the control of market-regulating policies, when that was necessary to guarantee the federation’s economic expansion and security. In the US and Swiss federations, there is not a single (State/canton or federal centre) level of government that might claim to represent federal/national sovereignty in its entirety, since the latter is represented by the States/cantons and the centre together. 

At the same time, the power of the federal centre has been constrained from within, through different strategies of separation of powers aiming to guarantee the reciprocal independence of the executive and legislative institutions. Separation of powers is a recipe for preventing centralisation. In the US and Switzerland, there is no equivalent of a Bundeskanzler representing the federal government. Indeed, there is no government as such, because the government is constituted of “separated institutions sharing power” ⁴³ . The government is a process rather than an institution. In the US, the President is indirectly elected every four years (renewable once) through the Electoral college which dissolves after the election, with the two chambers of the Congress separately elected by State voters (for a six-year mandate, in the case of the Senate) and districts-within-State voters (for a two-year mandate, in the case of the House of Representatives). Horizontal separation of powers means that Congress cannot vote the President down (it can impeach the latter for constitutional, not political, reasons), nor does the president have the power to dissolve Congress, although a complex system of checks and balances incentivise (or should incentivize) those institutions to cooperate in taking decisions. In Switzerland, the executive is a collegial institution (a directoire, or federal council, of seven members) elected by the two chambers of the federal legislature (the Federal Assembly being made up of the National Council, representing Swiss voters, and the Council of States, representing the cantons), but (once elected) no longer dependent on the latter’s confidence ⁴⁴ . 

As is apparent from these examples, federal unions could not centralise authority, since sovereignty is a divisive issue.

Owing to the asymmetry in demographic sizes of the States/cantons and their different national/cultural identities, federal unions have set up an internally divided centre to reduce the possibility that one State/canton (or group of them) could gain control and thus impose its will on the other States/cantons. The horizontal separation of powers, combined with the limited and enumerated policies assigned to the federal centre, has reduced the incentive to seek opportunities to opt out from the centre’s shared rule. Moreover, States/cantons have kept (or have tried to keep) identity issues within the realm of their sovereignty (in the US, issues connected with the ‘racial identity’ of the States, in Switzerland with the religious identity of the cantons). Differentiation has thus developed between the various cantons/states’ self-rules, since each state/canton can pursue specific policies in the fields of its competences, according to the preferences of domestic voters or governors, in compliance with the State’s/canton’s constitution. If a policy is not part of the (limited) shared rule package, then a State/canton can claim to manage it, although that should be justified constitutionally (indeed, as has happened in the US with the civil rights policy, that justification might become highly disputed). 

In the US, the distinction between shared rule and self-rule policies has been challenged by the growth of the country’s international exposure and its internal complexity. The dramatic development of public responsibilities (since the 1930s and particularly after the Second World War) has led to the affirmation of what has been called the ‘policy state’ ⁴⁵ , characterised by a growing role for the centre over the States (to the point of configuring a sort of cooperative federalism) ⁴⁶ , although State sovereignty has continued to be protected by the Constitution and related practices. Indeed, States can even behave as independent actors at the international level, contradicting decisions taken by the federal authorities (as was the case with environmental policy, when California decided to participate to the 2015 Paris Agreement after the Trump administration decided to withdraw from it in 2017). Thus, the cooperation required by the implementation of federal policies could not change the internalised logic of competition and conflict between the separated levels of government regarding ‘who should do what’ ⁴⁷ .

Federal unions have further constitutionally constrained the policy capabilities of the centre through a horizontal separation of powers that institutionalises the powers of the States/cantons (in the US, through a confederal Senate consisting of two senators per State, regardless of its demographic size; in Switzerland, through the referendum power that canton citizens may use to vote on contested federal laws). Federal States, instead, have constitutionally empowered the centre by enlarging the shared power of its parliamentary government, although mitigated by the federal representation of the territorial units in the upper chamber. In both types of federalism, there is only one constitutionalised governance system for dealing with federal shared rule policies, although in federal unions the voters elect (directly or indirectly) the representatives of separate institutions, whereas in federal States they can elect a parliamentary majority. 

Thus, the two types of federalism meet in different ways the accountability requirements of a democratic federation. Their single decision-making framework, for dealing with policies of federal competence, is demos-constraining in federal unions and demos-enabling in federal States. 

The EU as a federal union?

A union of highly demographically asymmetrical States, with deep-rooted distinct national identities, such as the EU, cannot be federalised according to the federal State model, so only a federal union seems possible. But experience shows that federal unions do not emerge from organic evolution, nor from the sum of the solutions to the periodic crises the union has had to face: federal unions require a preliminary constitutional choice by political elites, which separates shared rule from self-rule, supported by the consent of the citizens they represent ⁴⁸ . 

In a federal union, although Member States would have their own Constitution, the latter’s values and procedures would need to be coherent with those upheld by the constitutional pact. Different national interpretations of the basic principles of individual freedoms and the rule of law could only poison a federal union (as shown by the dramatic experience of the US Civil War, 1861-65). The constitutional pact could celebrate the plurality of sovereignties of the federal union, rather than its single sovereignty character, only if there is a coherence of values among those sovereignties. Because sovereignty is the core issue, federal unions are very fragile experiments, systemically exposed to conflicts that could disrupt them, and only the existence of political elites willing to compromise have saved existing federal unions them from this outcome. Here, the net distinction between shared rule and self-rule is fundamental: a federal union is a compound polity where national units can differentiate themselves in their self-rule policies, but they should participate equally in the shared rule policies of the federal centre ⁴⁹ . 

To deal with sovereignty issues, federal unions should enlarge the number of policies under self-rule control (vertical separation of powers) and internally separate the institutions governing the enumerated policies under shared rule control (horizontal separation of powers). In the case of the EU, this model would imply the bringing of enumerated CSP policies to the federal level, but also the devolution of unnecessary supranational/regulatory policies to the control of national authorities and voters. 

At the same time, at the centre, federal unions cannot centralize governmental power in the same way federal States do. The latter can set up a parliamentary government because they do not have sovereignty issues to settle. As soon as those issues emerge (as in Belgium or Canada, or in the quasi-federal Spain), parliamentary government and its drive towards fusion in levels of government are called into question. In federal unions, the centre should be organised according to a form of separation of powers. In the EU, it is unwarranted to centralize power in the European Council as well as in the EP.  A constitutional truce would be necessary between the European Council and the EP to set up a unitary executive power independent from both. Indeed, the single governance regime cannot be organised according to the logic of the fusion of powers (or parliamentary government), since it would increase the fear of disavowing sovereignty-induced claims from Member States, nor can it maintain the current intergovernmental character, which has transformed those sovereignty-induced claims into incentives for increasing decision-maker unaccountability. A separation of powers would make it possible to set up an executive power independent from both national governments (as represented by the European Council) and the popular legislative chamber (the EP) and connected to both through some forms of checks and balances. 

In the past, the model of the multiple separation of powers has been instrumental for establishing a political union constituted of demographically asymmetrical states with distinct national identities. However, in the case of the EU, that model should be devised in such a way as to enable the formation of popular majorities across the separate European institutions, to promote the accountability of decision-makers in areas of shared-rule policies. In any case, the viability of the federal union would require State political elites to share basic values and to be willing to compromise on interest divisions. 

The latter condition is not currently guaranteed in the EU, where some Eastern European national governments (such as the Polish and Hungarian ones) explicitly contest the basic values enshrined in the Treaties, thus advancing an illiberal constitutional identity at odds with the liberal constitutional culture of the EU. A federal union could not survive a clash between alternative constitutional identities. For this reason, if the division on constitutional values persists, the federal union should emerge from the decoupling of the EU in order to separate the Member States aiming to build ‘an ever-closer union’ from those interested only in participating in a customs union ⁵⁰ . New forms of economic cooperation between the two groups of States might be then devised, although the respect of the rule of law should also be guaranteed in a common market.

Conclusion

This article has argued that, in the EU, DI has emerged due to the intergovernmentalisation of CSP policies. DI has been made possible by governance differentiation, specifically by the institutionalisation of an intergovernmental governance regime based on voluntary coordination. Whereas single market policies imply the acceptance of its legal requirements to be part of it, this is not the case for participating in CSP policies. The single market is regulated by law (directives or regulations), whereas CSP policies are rather managed through voluntary coordination. To let one or other Member State opt out from one or other CSP policy has been the strategic device to Europeanise crucial sovereignty-sensitive policies. However, that device has also had unappealing consequences. It has strengthened the intergovernmental logic to the detriment of the supra-national logic, to the point of institutionalising, within the EU, an intergovernmental union, epitomized by the decision-making centrality acquired by the European Council, whose political accountability is all but inexistent.

The article has thus discussed a federal alternative to DI, able to promote integration and to uphold its democratic quality. Since DI is due to the necessity to accommodate sovereignty-induced claims, the federal alternative cannot have the features of a federal State model (where there is no sovereignty dilemma). Instead, through the adoption of a federal union model, it is possible to integrate the sovereignty claims of the Member States, while trying to ensure an accountability at the level of the federal authorities. In federal unions, in fact, shared rule policies (generally CSP policies) are managed through a single governance regime that prevents domination, while self-rule policies are left to the control of each Member State and can be highly differentiated. Moreover, even regarding CSP policies, Member States can retain important resources under their own control. The federal union is designed to discourage centrifugal pressures from Member States fearing threats to their sovereignty prerogatives, through a federal governance organised according to a demos-constraining logic. A logic that, however, has displayed crucial limits, namely the difficulty not only in identifying a cross-institutional majority that is accountable to voters but also in preventing unilateral decisions by a Member State or a group of Member States to call into question the constitutional coherence of the federal union.Federal unions institutionalise a tension between the centre and the States, as all of them contribute to the sovereignty of the union. Indeed, a federal union can be defined as a sovereign union of sovereign states ⁵¹ , inasmuch as the member States are sovereign on specific policies (self-rule) and the centre is sovereign on other policies (shared rule). The boundary between self-rule and shared rule will continuously shift, thus requiring a constant renegotiation between the two levels of government and a general attitude of compromise among the elites operating at the various levels and their sharing of basic constitutional values. In conclusion, the article has developed a theoretical argument on the plausibility of an alternative model to DI to integrate the sovereignty claims of the EU Member States, showing at the same time its problematic features and its systemic fragility.

1.  See, European Council (2017) The Rome Declaration, 25 March, https://www.consilium.europa.eu/en/press/press-releases/2017/03/25/rome-declaration/; see also, European Commission (2017) White Paper on the Future of Europe, 1 March, https://ec.europa.eu/info/news/commission-presents-white-paper-future-europe-2017-mar-01_en.
2.  See, John Erik Fossum, ‘Democracy and Differentiation in Europe’ (2015) 22 Journal of European Public Policy 799.
3.  See, Alexander CG Stubb, ‘A Categorization of Differentiated Integration’ (1996) 34 Journal of Common Market Studies 283.
4.  See, Alec Stone Sweet, Neil Fligstein and Wayne Sandholtz, ‘The Institutionalization of European Space’ in Alec Stone Sweet, Neil Fligstein and Wayne Sandholtz (eds), The Institutionalization of Europe (Oxford University Press 2001).
5.   See, Andrew Moravcsik, ‘What Can We Learn from the Collapse of the European Constitutional Project?’ (2006) 47 Politische Vierteljahresschrift 219.
6.  See, Arend Lijphart, Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries (Yale University Press 1999).
7.  See, Jean-Claude Piris, The Future of Europe: Towards a Two-Speed EU? (Cambridge University Press 2012).
8.  See, Giandomenico Majone, Rethinking the Union of Europe Post-Crisis: Has Integration Gone Too Far? (Cambridge University Press 2014).
9.  See, e.g., Dirk Leuffen, Berthold Rittberger and Frank Schimmelfennig, Differentiated Integration: Explaining Variation in the European Union (Palgrave Macmillan 2013).
10.  See, Frank Schimmelfennig and Thomas Winzen, Ever Looser Union?: Differentiated European Integration (First edition, Oxford University Press 2020).
11.  See, Thomas Winzen, ‘From Capacity to Sovereignty: Legislative Politics and Differentiated Integration in the European Union: From Capacity to Sovereignty’ (2016) 55 European journal of political research 100.
12.  See, Michelle Egan, ‘The Internal Market: Increasingly Differentiated?’, in Ramona Coman, Amandine Crespy and Vivien A Schmidt (eds), Governance and Politics in the Post-Crisis European Union (Cambridge University Press 2020).
13.  See, Vivien A Schmidt, ‘The Future of Differentiated Integration: A “Soft-Core,” Multi-Clustered Europe of Overlapping Policy Communities’ (2019) 17 Comparative European politics (Houndmills, Basingstoke, England) 294.
14.  See, Winzen (n 12).
15.  See, Bruno de Witte, Andrea Ott and Ellen Vos (eds), Between Flexibility and Disintegration: The Trajectory of Differentiation in EU Law (Edward Elgar Publishing 2017).
16.  See, Helen Wallace, ‘The Institutional Setting: Five Variations on a Theme’, in Helen Wallace and William Wallace (eds), Policy-Making in the European Union (4th ed., Oxford University Press 2000) 3–27.
17.  See, Philipp Genschel and Markus Jachtenfuchs (eds), Beyond the Regulatory Polity?: The European Integration of Core State Powers (Oxford University Press 2014).
18.  See, Uwe Puetter, The European Council and the Council: New Intergovernmentalism and Institutional Change (Oxford University Press 2014).
19.  See, John Erik Fossum, ‘EU Constitutional Models in 3D: Differentiation, Dominance and Democracy’ (2021) EU3D Research Paper 15.
20.  See, Sergio Fabbrini and Uwe Puetter, ‘Integration without Supranationalisation: Studying the Lead Roles of the European Council and the Council in Post-Lisbon EU Politics’ (2016) 38 Journal of European Integration 481.
21.  See, Matthias Matthijs, Craig Parsons and Christina Toenshoff, ‘Ever Tighter Union? Brexit, Grexit, and Frustrated Differentiation in the Single Market and Eurozone’ (2019) 17  209.
22.  See, Sergio Fabbrini, Which European Union?: Europe after the Euro Crisis (Cambridge University Press 2015).
23.  See, John Erik Fossum, ‘Politics versus Law: The European Council and the “Balancing European Union Style”’ (2020) Arena Working Paper.
24.  See, Andrew Glencross, ‘The European Council and the Legitimacy Paradox of New Intergovernmentalism: Constitutional Agency Meets Politicisation’ (2016) 38 Journal of European Integration 497.
25.  See, Michael Burgess, Federalism and European Union: The Building of Europe, 1950-2000 (Routledge 2000); David H McKay, Designing Europe: Comparative Lessons from the Federal Experience (Oxford University Press 2001).
26.  See, R Daniel Kelemen, ‘Federalism and European Integration’ in Antje Wiener, Tanja A Börzel and Thomas Risse (eds), European Integration Theory (3rd edn, Oxford University Press 2018).
27.  See, Guy Verhofstadt, Europe’s Last Chance: Why the European States Must Form a More Perfect Union (Basic Books 2016); Romano Prodi, Europe as I See It (Polity Press 2000). See also, Joschka Fischer, From Confederacy to Federation: Thoughts on the finality of European integration, speech given at the Humboldt University in Berlin, 12 May 2000, available at https://ec.europa.eu/dorie/fileDownload.do?docId=192161&cardId=192161.
28.  The post-WWII German federation is the outcome of the Allied Authorities’ decision to decentralize the powers of the Third Reich’s highly centralized state rather than of the autonomous decision of independent Länder to aggregate to form a federation. Indeed, the territorial borders of previous Länder were recognized in just a few cases, but many new Länder were formed by either aggregating or disaggregating previous Länder to guarantee a reasonable balance between them (in any case to prevent a Land having more than 30 per cent of the federation’s total population).
29.  See, Jan‐Werner Müller, ‘In the Shadows of Statism: Peculiarities of the German Debates on European Integration’ in Kalypso Nicolaïdis and Justine Lacroix (eds), European Stories: Intellectual Debates on Europe in National Contexts (Oxford University Press 2010).
30.  See, Dieter Grimm, ‘The Democratic Costs of Constitutionalization— The European Case’, The Constitution of European Democracy (Oxford University Press 2017).
31.  See, John Kincaid (ed), A Research Agenda for Federalism Studies (Edward Elgar Publishing 2019).
32.  See, Alberta M Sbragia, ‘Thinking about the European Future: The Uses of Comparison’, in Alberta M Sbragia (ed), Euro-Politics: Institutions and Policymaking in the ‘new’ European Community (Brookings Institution 1992).
33.  See, Alfred C Stepan, ‘Federalism and Democracy: Beyond the U.S. Model’ (1999) 10 Journal of democracy 19.
34.  See, Arthur Benz and Jörg Broschek (eds), Federal Dynamics: Continuity, Change, and the Varieties of Federalism (Oxford University Press 2013); Johanne Poirier, Cheryl Saunders and John Kincaid (eds), Intergovernmental Relations in Federal Systems: Comparative Structures and Dynamics (Oxford University Press 2015); John Erik Fossum and Markus Jachtenfuchs, ‘Federal Challenges and Challenges to Federalism. Insights from the EU and Federal States’ (2017) 24 Journal Of European Public Policy pp467; Kincaid (n 32).
35.  See, Daniel Judah Elazar, Exploring Federalism (University of Alabama Press 1987).
36.  See, Dieter Grimm, Sovereignty: The Origin and Future of a Political and Legal Concept (Columbia University Press 2015).
37.  See, Roland Lhotta and Julia von Blumenthal (2015) ‘Intergovernmental Relations in the Federal Republic of Germany: Complex Co-operation and Party Politics’, in Poirier, Saunders and Kincaid (n 35) 226.
38.  See, Arthur Benz and Jared Sonnicksen, ‘Advancing Backwards: Why Institutional Reform of German Federalism Reinforced Joint Decision-Making’ (2017) 48 Publius 134.
39.  See, Fritz W Scharpf, ‘The Joint-Decision Trap: Lessons from German Federalism and European Integration’ (1988) 66 Public Administration 239.
40.  See, Fritz W Scharpf, ‘Community, Diversity and Autonomy: The Challenges of Reforming German Federalism’ (2008) 17 German Politics 509.
41.  See, Joseph M Parent, Uniting States: Voluntary Union in World Politics (Oxford University Press 2011).
42.  See, William H Riker, Federalism: Origin, Operation, Significance (Little, Brown 1964).
43.  See, Richard E Neustadt, Presidential Power and the Modern Presidents: The Politics of Leadership from Roosevelt to Reagan (New York : Free Press ; Oxford : Maxwell Macmillan International 1991) 29.
44.  See, Adrian Vatter, Swiss Federalism: The Transformation of a Federal Model (1st edn, Routledge 2018).
45.  See, Karen Orren, The Policy State: An American Predicament (Harvard University Press 2017).
46.  See, Robert Schuetze, From Dual to Cooperative Federalism: The Changing Structure of European Law (Oxford University Press 2009).
47.  See, Donald F Kettl, The Divided States of America: Why Federalism Doesn’t Work (Princeton University Press 2020).
48.  See, Ronald L Watts, Comparing Federal Systems (3rd edn, McGill-Queen’s University Press 2008).
49.  See, Sergio Fabbrini, Compound Democracies: Why the United States and Europe Are Becoming Similar (Revised, Oxford University Press 2010).
50.  See, S. Fabbrini, Europe’s Future: Decoupling and Reforming, Cambridge University Press, 2019, op. cit. 
51.  See, Peter S. Onuf, The Origins of the Federal Republic: Jurisdictional Controversies in the United States, 1775-1787 (University of Pennsylvania Press 1983).