Differentiation and EU Governance: Key Elements and Impact

ABSTRACT Existing literature on differentiated integration has made an important contribution in theorising and operationalising its regulatory dimension. However, in order to fully evaluate the impact of differentiation on EU governance, this approach needs to be enriched with additional elements. The organisational element allows us to grasp the different forms, venues and actors of differentiation. The constitutional element connects the different forms of differentiation to the foundations of EU constitutionalism and identity. The socio-political element goes beyond the analysis of differentiation as a policy practice and qualifies it as a policy choice. On this basis, a generalisation is proposed resulting from a comparative analysis of a wide range of contributions in different policy sectors in order to assess differentiation in the EU’s governance against three main criteria: effectiveness, sustainability and accountability/legitimacy.

In this article, we aim to offer an innovative and comprehensive assessment of the impact of differentiation on European Union (EU) governance. This assessment encompasses the constitutional, social, local and external aspects of differentiation and addresses degrees and types of differentiation according to both their legal and organisational dimensions, including accountability mechanisms in differentiated integration. The final aim is to provide an overall appraisal of how much and what form of differentiation propels European integration forward -as a whole and for specific policy areas -and what kinds of differentiation should be avoided to prevent disintegration in future.
For our analysis, we rely on the impressive amount of empirical research conducted in the framework of the EU IDEA -Integration and Differentiation for Effectiveness and Accountability project funded by the European Commission under the Horizon 2020 programme and on the contributions collected in this Special Issue. In particular, we look at differentiation both as policy practice in different sectors -Economic and Monetary Union (EMU) and the Single Market; foreign, security and defence policy; and Justice and Home Affairs (JHA), including migration -and as a policy choice based on the preferences of policy-makers and citizens in EU member states that participate in a differentiated policy or institution, in EU member states that have opted out and in differentiated non-member states (that is, non-EU members that have opted in).
In the first part, we deconstruct differentiation and discuss three factors to be considered beyond the traditional regulatory approach: the organisational factor, the constitutional factor and the socio-political factor. In the second part, we conduct a comparative analysis of current and prospective forms of differentiation implemented in the different policy areas based on the case studies produced as part of EU IDEA and this Special Issue. The scope of this analysis is to provide a generalisation based on a wide range of contributions in order to assess differentiation in the EU against three main criteria: effectiveness, sustainability and accountability/legitimacy.
Finally, in the conclusion, we identify some key elements to assess how much and what form of differentiation is desirable, sustainable and legitimate. Compared to previous research, this approach is designed to offer a multi-dimensional analysis that encompasses various aspects of differentiation through a constant engagement between theoretical and normative work and everyday policy practice.

Deconstructing differentiation: factors that affect its impact on EU governance
Previous scholarship has made significant progress in theorising and operationalising the regulatory dimension of differentiated integration (Holzinger and Schimmelfennig 2012;Leuffen et al. 2013). In contrast, the organisational dimension has received much less attention Lavenex and Schimmelfennig 2009;Lavenex 2011). To provide a comprehensive assessment of the impact of differentiation on EU governance, we should go beyond the regulatory approach and enquire into the main elements affecting differentiation, including organisational, constitutional and socio-political factors.

The organisational factor
In this Special Issue, we use a broad definition of differentiation that refers to any modality of integration or cooperation that allows states (members of the EU and nonmembers) and sub-state entities to work together in non-homogeneous and flexible ways (Lavenex and Križić 2022, this Special Issue). This definition combines the attention of traditional approaches to the regulatory dimension of differentiation (Leuffen et al. 2013) with a focus on organisational aspects and the interplay between the two. Thus, such a definition allows for a comprehensive multi-dimensional approach that is key for assessing how the dynamics of differentiation impact on EU governance.
With reference to the regulatory dimension of differentiation, we mean the traditional focus on its legal aspects, which literally define 'who is bound by which policies and how' (Lavenex and Križić 2022, this Special Issue). From a regulatory point of view, different models of differentiation can be distinguished on the basis of the scope and quality of the legal act. The scope includes full participation, partial opt-out, punctual opt-in (or selective participation) and no participation. The quality ranges from the community method, which uses the ordinary legislative procedure (Art. 294 of the Treaty on the Functioning of the European Union [TFEU]), to intergovernmental cooperation, where the European Council generally acts unanimously and the European Parliament has a purely consultative role, and transnational cooperation.
From an organisational point of view, we look at the procedural dimension of differentiation, namely 'who participates in the production and implementation of differentiated policies and how' (Lavenex and Križić 2022, this Special Issue). Regulatory approaches do not grasp the different actors involved and the entirety of the process. Through the organisational dimension, by contrast, we can capture the non-homogenous participation of different actors (EU members, non-EU members, sub-national actors) in multiple institutional venues where EU-related policies are designed and implemented. The analysis thus includes different stages of the policy cycle, from policy-shaping (agenda-setting) and policymaking (output) to policy implementation (outcome) and problem-solving (impact), which involves formal decisionmaking structures as well as secondary organisations involved in the preparatory or implementation phases, such as committees, policy networks, regulatory agencies and non-EU bodies.
The focus on and the combination between the regulatory and the organisational dimensions allow us to capture the various forms that differentiation takes in the EU's context and their impact on governance. First, it allows us to account for the commonly held understanding of (horizontal) differentiation (Leuffen et al. 2013) that includes internal and external dimensions. Differentiation occurs when some EU members cooperate beyond the existing EU acquis, or partly opt-out of it, or when non-EU members selectively join existing EU arrangements, which creates arrangements whose membership differs from formal EU membership.
Second, through this approach, we can capture different actors involved in differentiation in the EU that are not necessarily states but include various sub-state actors, such as regulators or local authorities. This could also include transnational networks of subnational authorities (primarily cities and regions) that are part and parcel of Europe's system of multi-level governance, insofar as they participate in the EU's policy cycle at various stages and in various ways (Hooghe 1995;Perkmann 2007;Tortola 2013;2017;Tortola and Couperus 2022, this Special Issue). Still, unlike what happens at the state level, where we usually find one main cooperation or integration arrangement per policy area, at the sub-national level, several networks usually coexist within each policy area, creating multi-level differentiation.
Third, such an approach is open to a crucial yet widely neglected feature of differentiation, namely differentiated cooperation. Differentiated cooperation does not necessarily manifest in different levels of legal integration, but also takes place at an organisational level through non-homogeneous participation in the institutional venues where EU-related policies are designed and implemented, such as in the case of EU cooperation with third countries or informal cooperation of member states in the foreign policy field.
Finally, the selected approach allows us to detect incongruence caused by differentiation between those who take decisions and those affected by them (Fossum 2015;Heermann and Leuffen 2020) that leads to a mismatch between decision-makers and decision recipients. Decision-makers in differentiated policy sectors can represent either more (in the case of internal differentiation) or fewer (in the case of external differentiation) citizens than those affected by the policy in question. Thus, we should assess governance arrangements and types of accountability in order to detect how they vary across differentiated policies and institutions.

The constitutional factor
Given that 'united in diversity' is the overarching narrative of European constitutionalism, differentiation can be perceived either as a challenge to unity or as a tool for managing heterogeneity among member states, thus reconciling diversity, cohesion and unity. Assessing whether differentiation can be a tool for integration -and for avoiding disintegration -requires an analysis of whether and how different forms of differentiation safeguard the core constitutional elements from which the EU and its member states cannot deviate without compromising the essence and functioning of the Union (De Witte at al. 2017).
Hence, we need to take into consideration two foundational elements of EU constitutionalism, namely: (i) political unity, referred to "the unity of its policies, laws and institutions; and to any prospect of it developing into a political community based on shared rights and obligations of membership" (Lord 2015, 784); and (ii) legal uniformity, anchored in the acquis communautaire, which covers not only primary and secondary law, but also the "adherence to the aims of political, economic and monetary union and the administrative capacity to effectively apply and implement the acquis" (Council of the European Union 2005).
These essential elements regulate the relationship between the EU and its member states and among member states, which are engaged in the process of "creating an ever closer Union among the peoples of Europe" (Art. 1 Treaty on European Union [TEU]). As such, they constitute the building blocks of the EU's common identity and citizenship.
The political option of differentiation has been implemented in the past to manage heterogeneity and accommodate diversity among member states, for example, to make possible the Big Bang enlargement in 2004-2007 through (mainly temporary) exceptions or to react to the impact of the 2008-2011 economic and financial crisis in the Eurozone, most notably through additional differentiated integration arrangements in the EMU such as the Stability and Growth Pact and the Fiscal Compact. As analysed by Funda Tekin and Vittoria Meissner (2022, this Special Issue), if the narrative of political unity among elites has turned from a positive ('yes, we can') to a negative legitimation discourse ('yes, we must'), political unity has remained the ultimate goal through the commitment of member states to an ever closer Union. In fact, reality has shown that shared rights and obligations for all members can only be partially achieved, and often only through patterns of differentiated integration.
As a result, differentiation has gradually become the 'new normal' in EU integration. We cannot exclude that heterogeneity may increase in the EU as a consequence of current crises, thus revamping old and new proposals for further differentiation. Based on previous experience and related literature, different forms of differentiation could help once more safeguard the constitutional diversity in the EU's sui generis system (Tekin 2012), and thus preserve its political unity, or at least set boundaries to the amount of political disunity that the EU can tolerate.
Differentiation has also become a normal feature of EU law, while legal uniformity remains an ideal goal. Not only does primary and secondary law allow for derogations from the acquis, for example, in the form of opt-outs or transitional arrangements for new member states; its uniform application is impeded both by the fact that national authorities and courts do not always apply it faithfully and correctly, and by the fact that differential application is allowed by the EU, such as in the case of directives (De Witte 2017, 9-10). As a consequence, uniform application of EU law as defined by the Court of Justice in 1972 -rules of Community law being "fully applicable at the same time and with identical effects over the whole territory of the Community" -1 has clashed with a mounting practice of differentiation, but the principle of the EU's unitary legal order has not been challenged per se.
Nevertheless, a number of dilemmas persist over the nature and depth of differentiation compatible with the EU's constitutional integrity. The EU's integration system is based on the fundamental premise that member states share a set of core values on which the EU is founded, as defined in Art. 2 TEU, and envisages mechanisms for the EU's institutions to address potential breaches. A notable example is rule of law, which is safeguarded through the provisions of Art. 7 TEU 2 and the Rule of Law Mechanism. 3 Although we consider the EU's fundamental values as an element that cannot be compromised if we want to ensure the survival of the EU's project, we cannot exclude differentiation as a political option in the field of rule of law, since it is an area that relates to core state powers and policies (Genschel and Jachtenfuchs 2014). In fact, we have witnessed a high level of heterogeneity in judicial governance, both internally among the member states and externally in relations to the candidate countries (Damjanovski et al. 2020, 20). In the cases of Hungary and Poland, alternative conceptions and application of EU rules have put rule of law under severe strain. Poland has gone so far as to question the principle of primacy of EU law, when Poland's Constitutional Tribunal ruled that the country's constitution takes precedence over some EU laws in October 2021. 4 As argued by Ivan Damjanovski et al., this form of differentiation does not relate to the substantive (policy) aspects of member states' participation in EU integration, such as in the cases of Schengen or EMU, but to its structural (systemic) aspects. If the divergence in the member states' conception of EU membership is not a new phenomenon, its nature and depth might have a potentially disruptive impact on the EU legal order and the functioning of the EU more generally (22).

The socio-political factor
Another paramount factor shaping EU differentiation is the level of support differentiation and integration enjoy among national and local political elites as well as the wider public. For a long time, differentiation was perceived rather negatively by European elites as an (unwanted) by-product of integration (Koenig 2015). Today, after the end of the years of permissive consensus (Hooghe and Marks 2009), differentiation and integration need to gain broad and explicit support from large socio-political segments of the European populace in order to be a viable policy choice and, thus, a legitimate, effective and sustainable option for EU governance. In other words, we should assess whether political actors and the wider public regard differentiated integration as a suitable and effective alternative to homogeneous integration that could foster integration within the EU (and wider Europe) or at least be a realistic solution to move integration forward over time.
Against this background, two recent datasets presented by Vittoria Meissner and Funda Tekin (2021) and Anna Stahl (2021) suggest that today differentiated integration is already considered as an inevitable and pragmatic, if not desirable, form of integration among EU national politicians and citizens. The majority of interviewed political actors pointed out that differentiated integration should not be exclusive and maintain an open character, allowing other countries to join later on (Meissner and Tekin 2021, 6-7). In addition, the degree of support for differentiation varies across policy areas, as political actors perceive differentiated integration as effective only in certain sectors, for example, in foreign, security and defence policy. Among EU 27 public opinions, differentiated integration enjoys the support of a majority of citizens, although the understanding of the concept of differentiated integration may vary considerably across countries and demographics (Stahl 2021, 7). Moreover, while one can argue that national attitudes seem to have a more positive perception of differentiation than in the past (Stahl 2021;Meissner and Tekin 2021), what remains crucial is to assess the nature (pro-European or Eurosceptic) and extent (across policies and countries) of such a positive attitude toward differentiation.
Indeed, the nature of this support for differentiated integration seems to vary a lot, as this (general) notion suggests different things to different people. If we look at the perception of differentiation among the public, supporters include both pro-European and Eurosceptic citizens, as differentiation seems to be perceived as a tool both for deepening integration and for opposing it (Stahl 2021, 7-8). This is often the case also among political elites, where a majority of both Europhile and Eurosceptic political actors support differentiation as a conscious choice in specific policy fields, albeit for very different reasons (Meissner and Tekin 2021, 10). All this suggests that we should check, case by case, how distinctive types of differentiation are understood by different people and generate different degrees of support or aversion to integration and differentiation.
Differentiation also produces an impact on European identity. Based on relevant literature, one can argue that European identity is "heavily contingent on the ways in which national identities are discursively articulated" by political actors in EU member states and candidate countries (Aydın-Düzgit et al. 2020, 2-3). As claimed by Senem Aydın-Düzgit, Jan Kovář and Petr Kratochvíl, the impact of differentiated integration on European identity is perceived by member and candidate countries in terms of "preserving European identity (deepening), the dilution of European identity (disintegration) and the territorial/geographic limits of European identity (widening)" (4), and this perception varies considerably depending on the policy area in question (28). Therefore, the assessment of the impact of differentiation on European identity should be based on both the national and the policy factor -by taking into consideration the articulation of political discourse at the national level and the perceived implications of differentiation in different policy areas.

Measuring the impact of differentiation on EU governance
The factors presented above are essential elements to be considered alongside the traditional regulatory dimension of differentiation in order to conduct a comparative analysis of the various forms of differentiation implemented in the different policy areas, with a view to measuring their overall impact on EU governance. We take into consideration three policy areas that are particularly relevant for the functioning of the Union as a whole: economic governance; foreign, security and defence policy; and JHA, including migration. Our comparative analysis aims to assess to what extent each form of differentiation is effective -inasmuch as it introduces a useful degree of flexibility that facilitates policymaking, policy implementation and problem-solving -sustainableinasmuch as it preserves the homogeneity required to avoid disintegration -and legitimate -inasmuch as it is recognised by EU citizens, member states and affected third partners as appropriate.

Effectiveness
As defined by Sandra Lavenex and Ivo Križić (2022, this Special Issue), effectiveness entails generating policy outputs whose outcome is to solve policy problems. In order to define differentiation as effective, we assess it against two main benchmarks: (i) a relative policy change scenario -did differentiation produce an improvement compared to a hypothetical state of affairs without differentiation? -and (ii) an ideal scenario -did differentiation constitute an appropriate or even ideal solution to the underlying policy issue? (Underdal 2002;Lavenex and Križić 2022, this Special Issue). Also, we evaluate the effectiveness of differentiation both within the various differentiated arrangements and in the Union as a whole.
If we apply a combined regulatory-organisational approach to the study of differentiation in the EU as a whole, we find an incredibly varied array of forms of differentiation across policy sectors. In the economic policy field, it is possible to detect a number of differentiated arrangements with a varying degree of institutionalisation, including internal and external dimensions and involving the participation of national and subnational actors (Eisl and Rubio 2022, this Special Issue). In the foreign policy field, differentiation often occurs at the margins or outside EU institutional structures or legal frameworks, in the form of regional groupings, contact and lead groups, as well as various defence initiatives, while Treaty-based formal differentiated frameworks such as constructive abstention or enhanced cooperation have rarely been used (Grevi et al. 2020, 4). In the realm of defence policy, some forms of differentiation involving legally binding commitments have taken place in accordance with Treaty provisions -notably the Permanent Structured Cooperation established in 2017 by 25 member states. In the field of border, asylum and police cooperation, differentiation started with strictly intergovernmental arrangements. The Lisbon Treaty finally brought this area within the standard EU legal framework, so that the ordinary rules on decision-making and jurisdiction of the Court of Justice are now largely applied to this area. However, special arrangements still remain: Denmark has an opt-out for the entire area, Ireland and pre-Brexit UK were granted opt-outs with an opt-back-in clause on a case-by-case basis, while non-members continue to participate in the Schengen area, namely Norway, Iceland, Switzerland and Liechtenstein. In addition, the EU developed an external dimension of border and migration policy through flexible cooperation with neighbouring transit countries (Comte and Lavenex 2022, this Special Issue).
The hypothesis formulated by Lavenex and Križić (2022, this Special Issue) is that more stringent institutional frameworks should lead to better policy performance, based on the assumption that they are more likely to ensure members' compliance with agreed rules. If we look at the effectiveness of various differentiation arrangements in terms of policy outputs and policy outcome, this holds true for Treaty-based cooperation schemes among states in supranational policy areas, both internal, such as in the cases of the Banking Union, and internal/external, such as the European Economic Area (EEA) and the Energy Community (Mack 2020;Eisl 2020;Franza et al. 2021). However, the effectiveness of formal differentiated arrangements in intergovernmental EU policy areas such as diplomacy, defence and migration is negatively affected by the lack of sanctioning mechanisms and by a culture of non-compliance, for example, in the implementation of the obligations deriving from the Permanent Structured Cooperation and the EU asylum legislation (Biscop 2019, 7;Comte 2020). In the case of JHA, differentiation has reached effectiveness through strictly intergovernmental arrangements, since most EU members gave up their systematic internal border control, but also because it was complemented by a successful process of convergence thanks to the transfer of responsibility to EU institutions, which delivered new policy outputs in terms of harmonisation and the creation of new agencies and instruments (Comte and Lavenex 2022, this Special Issue).
In the foreign policy field, informal arrangements of differentiation that take place outside EU Treaties, such as lead groups, have allowed restricted numbers of member states to circumvent the shortcomings of unanimity-based decision-making in EU foreign policy by joining forces and enabling a European response on specific issues (Siddi et al. 2022, this Special Issue). The degree of institutionalisation is not always important as other factors such as policy alignment also when we look at external differentiation in foreign policy: this is exemplified by the case of Turkey, which is formally a candidate country but far from EU positions in foreign policy matters, as compared to the UK, which has no formally established relationship with the EU in foreign and security policy matters but remains very important and contributes to the effectiveness of EU foreign policy in key international contexts (Aydın-Düzgit et al. 2020, 14;Alcaro and Siddi 2020). Also, forms of sub-national differentiated cooperation such as the Covenant of Mayors and similar networks, which are primarily focused on promoting best practices and policy learning, seem to be better served by more informal arrangements that favour an inclusive and open-ended approach to membership (Tortola and Couperus 2022, this Special Issue).
While we can conclude that various forms of differentiation are likely to produce a relative improvement in terms of policymaking and implementation, their effectiveness is more contested if we evaluate their overall impact in providing a solution to specific policy issues. In particular, we find that formal differentiation arrangements, such as those introduced in the early 2010s, during or in the aftermath of the economic and financial crisis, ultimately strengthened the functioning of the EMU and/or reduced its vulnerabilities, thus making the Union more effective in its policymaking and less vulnerable to future crises (Pilati and De Angelis 2020, 18). However, we are still far from an ideal situation, since the introduction of policy innovations -for example, the Fiscal Compact -led to a political backlash in some countries which is still threatening EU political cohesion. As for differentiated cooperation in the field of foreign policy, such as the lead groups on Iran (E3 between Germany, France and the UK) and Ukraine (the Normandy format between France and Germany), the engagement of larger EU member states in major international crises has led to positive policy change, but has not provided an adequate or ideal solution (Siddi et al. 2022, this Special Issue). As far as border, asylum and police cooperation is concerned, the differentiation arrangements have been able to achieve the abolition of internal border controls to create a borderless European Single Market, but they have not been not able to proceed beyond a certain point. This is confirmed by the unilateral reintroduction of border controls in the Schengen area in reaction to crises in Libya and Syria and, more recently, to the Covid-19 emergency, by the failure to find political agreement to reform the Dublin system and ultimately by the controversial impact of external differentiation on the management of migration, notably through the migration deal with Turkey (Comte and Lavenex 2022, this Special Issue).
Empirical research suggests that there is no one-size-fits-all solution to make differentiation effective, and shows the importance of having an institutional set-up that is adequate for the identified objectives (Eisl and Rubio 2022, this Special Issue). If more stringent institutional frameworks are suited for Treaty-based differentiation schemes among member states in supranational policy areas, looser forms of differentiated cooperation seem to better serve the objectives of intergovernmental EU policy areas, including in the external dimension, as well as for differentiation schemes involving subnational actors.
In the case of more informal differentiated cooperation, it was found that the positive impact has been stronger when it has adhered to common European values and identity (as expressed in the EU policy discourse), thus revealing the importance of the constitutional factor, and when built on long-established EU positions. This can be favoured by the creation of an institutional link with relevant EU decision-making bodies, as witnessed with the Iran nuclear deal, where this link was established through the association to the E3 of the office of the High Representative of the Union for Foreign Affairs and Security Policy, which ensured that informal differentiated cooperation fed into the agenda of the Foreign Affairs Council and the European Council. Cooperation with (and opposition of) major powers such as the United States has also been an important driver for (or challenge to) effectiveness (Siddi et al. 2022, this Special Issue).
Another finding is that the effectiveness of differentiation arrangements depends on their adaptability over time (Eisl and Rubio 2022, this Special Issue). This is the case for third-country access to the Single Market, due to the constantly developing EU regulatory environment and the necessity to adapt objectives and instruments to evolving circumstances. Those agreements that do not facilitate dynamic adjustments, such as in the case of the EU-Swiss agreement, tend to be ineffective (Eisl 2020). The Trade and Cooperation Agreement (TCA) between the EU and the UK has "an in-built dynamism" (Wachowiak and Zuleeg 2022, this Special Issue, 148) in the form of several grace periods, transitional periods and reviews of (part of) the Agreement. However, whether these technical mechanisms are able to ensure effectiveness, or even avoid collapse of part of the current arrangement, in a highly politicised context such as the one characterising EU-UK relations remains to be tested, highlighting the importance of socio-political factors.

Sustainability
Sustainability of differentiation is used as a benchmark to identify the scope and form of homogeneity that is required to make heterogeneity a tool for integration (internal and external) rather than disintegration. Sustainability of differentiation should be measured with reference both to a single differentiated policy or institution, and to its compatibility with the resilience of the overarching EU governance.
How sustainable particular examples of differentiation might be is linked to constitutional criteria, namely the objectives of the differentiated arrangements and their compatibility with the core constitutional elements of an ever closer Union, as well as to socio-political factors, specifically the support of political elites and citizens for a certain form of differentiation. Possible variations of sustainability are connected to regulatory criteria, namely the scope of the differentiated arrangements, their legal quality and extent of legal commitment; as well as organisational criteria, namely the existence of relevant bodies at different levels of governance that are in charge of different stages of the policy cycle, from agenda-setting to policy evaluation.
As argued by Andreas Eisl and Eulalia Rubio (2022, this Special Issue), most differentiated integration arrangements tend to persist over time once they are created, and the EU is more likely to react to crises through path-dependent and incremental reforms. Particularly in those areas where differentiated integration is a reality, such as the EMU and Schengen, radical shifts towards further integration or complete disintegration are unlikely, while the reaction to crises has usually been more -or a different type ofdifferentiation. Many recent studies have also argued that differentiated integration is likely to further increase in EU foreign policy in the near future (Bassiri Tabrizi 2018; Siddi et al. 2022, this Special Issue). In sectors where member states' practices are highly heterogeneous, such as migration and asylum, differentiation may represent the only available option to start cooperation (Comte 2020).
However, there are also risks of disintegration attached to differentiation, both internal and external. In fact, an over-use of flexibility can create fragmentation among members and between members and non-members, thus jeopardising political unity in the EU, as well as producing a negative impact on the integrity of the EU's internal system of law.
What happened with the Schengen system in connection with recent crises is a case in point. Schengen's flexibility allows members to temporarily reintroduce border controls in the event of a serious threat to public policy and internal security. This provision has been used in a continuously extended and uncoordinated way since the migration governance crisis of 2015-2016 and following the outbreak of Covid-19, thus overstretching Schengen's flexibility. Temporary border checks have been prolonged by member states shifting from one legal basis to another, often with insufficient justification, with disruptive consequences both at the intra-Schengen level and for the entire EU. In fact, they both hampered the correct functionality of the Schengen system and led to instances of infringement of the principle of non-discrimination between EU citizens, thus bearing the risk of an erosion of the acquis in the long term (De Somer et al. 2020).
If we look at the field of external migration policy, differentiated cooperation frameworks with third countries have proliferated since 2015-2016, aimed at reducing migratory pressure at the EU external borders. In particular, the EU and its member states have increasingly resorted to informal differentiation mechanisms -based on non-binding agreements and involving a multitude of actors in intergovernmental and transgovernmental venues -in contrast with legally binding cooperation sustained by supranational procedures. This has resulted in a reduction of the EU's internal legal standards, for example, by watering down the criteria for determining a country as safe in the asylum procedures as a way to facilitate cooperation with countries that have lower human rights safeguards, such as Libya, Morocco, Turkey and the Western Balkans (Okyay et al. 2020). Moreover, the fact that differentiation arrangements such as the EU-Turkey deal or the informal cooperation with Libya are concluded by an unidentified set of actors as explicitly non-binding instruments also make them non-traceable to an EU institution and therefore impossible to adjudicate before the Court of Justice, thus causing a reduction of the EU's rule-of-law standards in external cooperation.
As shown by the cases above, differentiation does not jeopardise per se EU political unity and the normative consistency of European integration. However, the increasing resort to policies associated with intergovernmental venues, soft cooperation mechanisms and informal instruments of differentiation raises significant constitutional challenges, fuelling fragmentation, hampering judicial oversight and eroding human rights safeguards.
The connection of arrangements of differentiated cooperation with the EU institutional framework has been shown to mitigate the risks of fragmentation connected to flexibility and, in the long run, to trigger a process of convergence, as in the case of the European asylum system, where participating states started as early as the 1990s to transfer responsibility to European institutions, harmonise standards and build new auxiliary agencies (Comte 2020). The involvement of EU institutions in differentiation arrangements is also instrumental in providing coherence to various differentiated formats with shared EU goals and decisions. In the case of Treaty-based differentiation arrangements, specific provisions identify the role and mandate of EU institutions. When intergovernmental differentiation initiatives are established outside the EU Treaties, this can be ensured through consultation and cooperation mechanisms, such as in the case of the High Representative in EU foreign policy lead groups, or by ensuring oversight by the European courts as well as the European Parliament.
Moreover, one condition to ensure sustainability of differentiated arrangements in terms of impact on the overall EU governance is to include mechanisms that mitigate potential negative side-effects on non-participating and third countries (Eisl and Rubio 2022, this Special Issue). These mechanisms can take the form of provisions to refrain from discriminatory actions against any member state, voting procedures for important decisions that involve non-participating member states or legal frameworks that facilitate admission of nonparticipating member states. In the case of the Banking Union, mechanisms such as the unityprotection provisions of the Single Resolution Mechanism and the Single Supervisory Mechanism or the double majority voting of the European Banking Agency prevented disintegration risks in the Single Market and helped to generate centripetal forces at the EU level. In fact, they were instrumental in convincing Croatia and Bulgaria to join and in making Denmark and Sweden consider a possible membership (Mack 2020).
Brexit deserves to be treated as a special case, representing a unique instance of differentiated disintegration that questioned the core principles of EU membership as well as what it means to be in or out. The red lines imposed by the EU to guarantee the indivisibility of the four freedoms in the TCA and the Northern Ireland Protocol preserved the integrity of the Single Market. However, it remains to be seen if the complex institutional framework and the binding enforcement and dispute settlement mechanisms introduced to manage the relationship will be able to mitigate political and economic disruption in the face of new crises (Wachowiak and Zuleeg 2022, this Special Issue). 5 For sectors not included in the TCA such as foreign, security and defence policy, an alignment between UK and EU policies might be favoured by the creation of institutional mechanisms allowing the participation of the UK in shaping -but not making -EU decisions (Bond 2020).
The impact of Brexit on internal and external differentiation, and EU governance overall, is a subject of ongoing debate and research. It is too simplistic to conclude that the opt-outs granted to the UK either paved the way for disintegration or were not enough to accommodate divergence. At the same time, if Brexit seems to have triggered centripetal forces in the EU, for example, in the field of economic governance through the adoption of Next Generation EU 6 and in the field of defence with the launch of the Permanent Structured Cooperation, its enabling effect to further integration should not be overstated. In terms of external differentiation, Brexit might suggest a more hard-line approach towards non-EU members, both third countries (no cherry-picked access to the Single Market) and candidate countries (setting the bar for membership even higher). Overall, Brexit seems to have reinforced the assumption that differentiation can help accommodate divergence as long as it does not involve the core elements of EU membership, such as the integrity of the EU's system of law and judicial oversight, as well as its founding principles as laid out in Art. 2 TEU.

Legitimacy and democratic accountability
Accountability can be defined as "a relationship between an actor and a forum, in which the actor has an obligation to explain and to justify his or her conduct, the forum can pose questions and pass judgement, and the actor may face consequences" (Bovens 2007, 450). Closely related to the notion of accountability, and partially depending on it, is the notion of legitimacy. Following Jonas Tallberg and Michael Zürn (2019, 585), legitimacy is defined as the "beliefs within a given constituency or other relevant audience that a political institution's exercise of authority is appropriate". Legitimacy belief depends on several factors, including the procedural standards in decision-making processes and the authority of an institution (Tallberg and Zürn 2019). For supranational institutions such 5 The TCA institutional framework includes a Partnership Council co-chaired at the ministerial level (EU commissioner and UK minister) and a plethora of committees and working groups. 6 Next Generation EU is a €750 billion temporary instrument designed to boost economic and social recovery from the damage caused by the Covid-19 pandemic and financed through bond issuances in the international capital markets by the European Commission on behalf of the EU.
as the EU, the recognition of the authority of an institution is closely tied to both its capacity to deliver effective problem-solving and citizens' general confidence in that institution, thus its socio-political legitimacy (Dellmuth and Tallberg 2015). If we take into consideration the organisational dimension, it is clear how differentiation challenges both the accountability and legitimacy of EU governance, as it creates incongruences between those who take decisions and those who are affected by them (Fossum 2015;Heermann and Leuffen 2020). Indeed, depending on the format of differentiation, citizens and governments might have only marginal or no control over policies that affect them. In the EMU, for instance, the European Parliament represents more citizens than those who are directly affected by Eurozone measures, whereas the Eurogroup and Euro Summits exclude the government representatives of those member states that may be indirectly affected by their decisions (Nguyen 2020).
This incongruence in the democratic constituency between decision-makers and decision recipients is also present in external differentiation in the economic field, where different external agreements with third countries grant access to the Single Market. Nevertheless, these various external economic differentiation regimes include a set of accountability mechanisms, such as joint parliamentary committees that can monitor and require information. Moreover, a number of additional mechanisms have been established in this policy field to complement democratic accountability. For instance, the European Central Bank, in its role of managing the Euro and implementing monetary policy, has only limited accountability channels towards the European Parliament. However, this is supplemented by a narrow specification of the institution's mandate by Art. 282 TFEU and thus by the scrutiny of judicial review (Dawson et al. 2019, 77).
Over the last decade, we have seen a proliferation of new governance arrangements outside the EU Treaties, which have exacerbated problems of accountability. Indeed, the Eurozone crisis led to a strengthening of European executive powers, meaning the European Council, the Euro Summit and the Eurogroup, and the adoption of measures outside the EU legal framework, and thus without the scrutiny of the European Parliament and European Court of Justice. Examples of this are the European Stability Mechanism, where control is provided through the administrative accountability of the Board of Auditors (Nguyen 2020;Mack 2020), or the Treaty on Stability, Coordination and Governance in the Economic and Monetary Union (Papadopoulos 2021;Nguyen 2020). Still, the legitimacy of these emergency measures, introduced in the aftermath of the financial and economic crisis, also depends on their effectiveness and the wider socio-political perception by EU citizens. Whereas these measures reduced EU vulnerabilities, making the Union more effective in facing future crises (Pilati and De Angelis 2020, 18), perceptions of them vary considerably along the North-South and Eurozone membership divides. Citizens from Northern member states are less supportive of deeper economic integration than those from Southern countries, although the latter are nevertheless more sceptical of the benefits of the Euro. At the same time, citizens of countries outside the Eurozone are more sceptical about the potential benefits of the Euro, although there is a significant variation across countries (Stahl 2021, 11-2). Sometimes policy-makers from non-Eurozone members even express fears that further institutionalisation of the EMU could push non-Eurozone countries to the periphery of European integration (Meissner and Tekin 2021, 8, 12). All this creates different constraints that should be carefully taken into consideration when evaluating the impact of differentiation on EU governance.
Turning to differentiated cooperation in the intergovernmental area of foreign, security and defence policy, as mentioned above, differentiation has typically taken place through informal groups of member states leading EU foreign policy (Grevi et al. 2020;Koutrakos 2017;Delreux and Keukeleire 2017). Marco Siddi, Tyyne Karjalainen and Juha Jokela (2022, this Special Issue) have argued that, while missing formal accountability mechanisms, informal groups of member states have usually had a relatively positive impact on the legitimacy of EU foreign policymaking, as they spurred the Union into action on specific issues by generating consensus, especially when their initiatives have been in line with EU values and identity. Nevertheless, there are clear limits to the accountability of these differentiated groups steering EU foreign policy. Indeed, they risk creating exclusive formats, limiting access to information and participation in policy planning and policymaking for the other member states. A key element to mitigate these risks is to facilitate connections to the formal EU foreign and security policy structures, for example, through the participation of the office of the EU High Representative in various differentiation formats and ensuring some level of alignment to the EU's common values and established common positions (Ibid.). At the same time, the greatest risk for the EU's common foreign, security and defence policy seems to be inaction. In fact, only 10 per cent of EU citizens see a national response to international crises as their favourite option. On the contrary, 30 per cent would prefer an EU response to international crises, 17 per cent would favour cooperation among selected countries or an international coalition, 16 per cent cooperation within NATO and 13 per cent would like their country to act within the United Nations (UN) framework. In particular, citizens from smaller EU member states seem to be supportive of a European response to international crises, citizens in Central-Eastern European and Baltic countries are more supportive of a NATO response, while the UN is favoured in Sweden (Stahl 2021, 18-9).
Another case where legitimacy has mainly been a function of effectiveness is EU differentiation in border, asylum and police cooperation (Comte and Lavenex 2022, this Special Issue). As far as differentiation proved effective and succeeded in abolishing internal border controls and integrating the related cooperation in EU structures, it was perceived as effective and legitimate, and free movement of people still enjoys great support by EU citizens (European Commission 2020). Yet, growing external migration challenges have undermined the existing arrangements, ushering in disintegration tendencies. In a recent survey among EU-27 citizens, a relative majority of 39 per cent declared to strongly or somewhat support an exit of their country from the Schengen area (Stahl 2021, 14-5). Despite almost two decades of successful policies in this area, limitations in border, asylum and police cooperation have emerged. In particular, the lack of compliance by member states, together with increasing difficulties in reforming the system, have created a sense of crisis over both the effectiveness and legitimacy of cooperation. All in all, the intermediary status of this policy field has led to a contradictory situation in which neither the European Parliament, nor the national parliaments and governments, have full control; this has undermined both the effectiveness and legitimacy of the policy as such, and confused the publics, which show a limited understanding of these differentiated arrangements (14). As a consequence, the EU has largely externalised the management of migratory challenges through a number of deals with third countries to help reduce pressure at its external borders. Nevertheless, this temporary relaxation has come at a high price. As we noted in the previous section, these deals downgrade EU protection standards by lowering criteria for recognition of the "safe third country" clause of the EU Asylum Procedures Directive 7 and elude the scrutiny of the European Parliament and European Court of Justice.

Conclusion
Our investigation started from the assumption that the traditional regulatory approachaimed at analysing who is bound by which policies and how -is not enough to fully evaluate the impact of differentiation on EU governance and needs to be supplemented with additional elements. The organisational element allows us to grasp the different forms, venues and actors of differentiation so as to determine who participates in the production and implementation of differentiated policies and how. The constitutional element connects the different forms of differentiation to the foundational elements of EU constitutionalism and identity, such as political unity (that is, an ever closer Union) and legal uniformity (that is, rule of law). The socio-political element goes beyond the analysis of differentiation as a policy practice and qualifies it as a policy choice by investigating the preferences of policy-makers and citizens directly or indirectly affected by it.
These elements were used to conduct a comparative analysis of crucial sectors of EU integration, namely economic governance, foreign, security and defence policy, and JHA, including migration, to test different forms of differentiation. The case studies taken into consideration confirmed that differentiation is likely to produce a relative improvement in terms of policymaking and policy implementation across policy sectors and for the EU overall, but its impact is more contested if we take a problem-solving approach and evaluate the results achieved against an ideal scenario. There is no one-size-fits-all solution in terms of institutional set-up, as more stringent or looser forms of differentiation can be best suited for different policy objectives. Clarity of objectives is therefore determinant to identify the right form of governance. Nevertheless, the degree of effectiveness can be influenced by other factors, such as adherence to common European values and identity, and the adaptability of differentiation schemes to evolving circumstances and policy priorities.
Moreover, the assessment of various forms of differentiation against the criteria of sustainability has shown that excessive recourse to flexibility can produce centrifugal dynamics and have a negative impact on EU political unity and normative contingency, thus leading to fragmentation and even disintegration in the long run. Formal differentiation arrangements based on binding agreements and institutional oversight are likely to produce a process of convergence within each policy sector as well as providing coherence among various differentiation formats and with EU goals and decisions. One condition increasing sustainability is to avoid negative effects on non-members, for example, through formal commitment to non-discrimination by members and through formal and informal mechanisms of participation for non-members in the decisionshaping process of differentiated policies and institutions.
Finally, differentiation challenges both the accountability and legitimacy of EU governance, as it creates incongruences between those who take decisions and those who are affected by them, thus breaking the democratic identity between decision-makers and decision recipients. This holds true for both internal and external differentiation. Against this background, the lack of direct democratic control can be mitigated, inter alia, through administrative or judicial accountability mechanisms, or by enabling the selective participation of representatives of the excluded constituencies in the preparatory and implementation phases of the governance cycle. In addition, the case studies clearly show a tendency toward the adoption of (differentiated) measures outside of the EU's legal framework, or of completely informal differentiation to deal with unforeseen circumstances and crises. These measures, while they may offer short-term solutions, can pose serious challenges to the accountability and legitimacy of such differentiated policies, undermining their socio-political acceptability and damaging intra-EU cohesion in the longer term. Maintaining their formal anchoring in the EU legal framework and ensuring scrutiny by EU institutions is fundamental if such differentiated arrangements are to remain legitimate and accountable. At the same time, formal accountability mechanisms need to be complemented by the willingness of European policy-makers and citizens to accept certain differentiated EU policy arrangements as appropriate ways to solve problems.
Based on this analysis, it is possible to conclude that differentiation can be not only compatible with, but also conducive to a more effective, cohesive and democratic EU, if a number of conditions are met. In particular, we found that differentiation can foster EU integration when it is based on formal arrangements, preferably established within the framework of the Treaties, has clear objectives in line with EU's core values and shared interests and includes binding commitments for participating members, a direct link to the EU's institutional framework, information and participation procedures for non-members, as well as a mix of accountability mechanisms (democratic, judicial, administrative). On the contrary, looser form of differentiated cooperation established outside the EU Treaties, combined with the lack of EU institutional anchors, judicial oversight and democratic participation, should be avoided as they can lead to fragmentation and possibly disintegration dynamics, due to excessive flexibility when it comes to the identification of their ultimate goals, establishment of the criteria for participation and adherence to the EU's system of values and norms.