Institutional Challenges for External Differentiated Integration: The Case of the EEA

This working paper analyses the institutional challenges related to external differentiated integration in the European Economic Area (EEA). It focuses mainly on the formulation of EEA-relevant EU legislation and its incorporation into the EEA Agreement. The paper shows that over the past 25 years various institutional arrangements have been added to the initial institutional framework of the EEA in order to increase and maintain substantive integration. However, the European Union (EU) has been consistent in protecting the autonomy of its decision making which is why the EEA EFTA States have far-reaching access to EU policy making but never the right to vote. The EEA EFTA States therefore insist on separate EEA decision making whenever possible. This has given them a surprisingly large amount of room for manoeuvre for instance, by deliberately delaying the incorporation of politically sensitive acts, and by making EEA-specific adaptations to EU acts. On the other hand, they were also forced to introduce simplified procedures for EEA decision making in order to cope better with the high legislative dynamics of the EU. These procedures give priority to the efficacy of the EEA over the decision-making autonomy of the EEA EFTA States by establishing a more or less automatic rule transfer from the EU to the EEA. Thanks to its far-reaching functional and institutional integration the EEA provides a good example for the analysis of the legal and political feasibility of external differentiated integration. Above all, the results of the empirical analysis demonstrate how difficult it is to reconcile the integration reservations of non-Member States with the principles of the EU in an institutional framework designed to ensure the long-term good functioning of their relations.

1 Introduction * 'No integration without representation!' The dictum that has been used by Rittberger (2007) to describe the nexus between European integration and parliamentary democracy does also refer to a fundamental challenge of external differentiated integration. Differentiated integration can be defined as the incongruence between 'the territorial extension of European Union (EU) membership and EU rule validity' (Holzinger and Schimmelfennig 2012: 292). If at least one EU Member State is legally excluded from an EU rule, this is called internal differentiated integration (Duttle et al. 2017). By contrast, external differentiated integration describes the situation where EU rules are also binding on a non-Member State. The distinction between internal and external differentiated integration may blur if a non-Member State receives specific opt-outs within the functional scope of its legal relations with the EU (Frommelt 2020a). In this paper, however, those specific types and mechanisms of differentiated integration are not relevant.
External differentiated integration is based on a formal and legally binding agreement between the EU and the respective non-Member State. To date there are various models of external differentiated integration in place such as the European Economic Area (EEA), the association agreements of the EFTA States to Schengen and Dublin, the EU-Swiss bilateral relations, the EU-Turkey customs union, the EU's relations with the small-sized states of Andorra, Monaco and San Marino, the Energy Community, as well as the association agreements with the Eastern European Countries in the European Neighbourhood Policy (ENP). These models provide to 'varying degrees access to and participation in the EU's internal market as well as an institutionalized relationship involving extensive reciprocal rights and obligations, approximation to or adoption of the acquis, policy cooperation and integration' (Gstöhl and Phinnemore 2019b: 175).
In external differentiated integration there is usually a mismatch between the territorial scope of EU policies and the composition of the legislature as non-Member States have no right to vote in the EU decision-making process, but EU law still forms the basis of their legal relations with the EU. To overcome this mismatch there are various ways how non-Member States can participate in EU policy making and there are also different ways how EU law becomes part of the legal relations between the EU and a non-Member State. It is thus not surprising that an increasing number of studies compares the functional scope and institutionalisation of the various models of external differentiated integration Lavenex 2011;Lazowski 2014;Gstöhl 2015;Gstöhl and Phinnemore 2019a). However, such comparative studies cannot depict the details of the institutional relations between the EU and a non-Member State, and thus the numerous institutional and practical challenges that arise in the administration of external differentiated integration.
Addressing this gap in the literature, this working paper questions the legal and political challenges of external differentiated integration by analysing the institutional arrangements of the EEA and its changes over time. The EEA Agreement provides for substantial functional and institutional integration between the EU and the participating EFTA States Iceland, Liechtenstein and Norway. Due to its broad functional scope and elaborated institutional framework the EEA is widely seen as the most far-reaching model of integration of non-Member States with the EU (Tobler et al. 2010). The institutional set-up of the EEA can best be described as a two-pillar structure, with the EFTA institutions matching those on the EU side. It shall ensure that within the functional scope of the EEA the same legal obligations apply to the EU States and the EEA EFTA States. both about making an account and about holding to account'. With regard to external differentiated integration, Fossum (2015: 13) argues that the 'EU's democratic coping mechanisms are configurated along EU membership lines, not along the territorial reach of EU policies'. This means that the more closely a state is affiliated to the EU without being a member, the higher the incongruence and lower the accountability. In other words, the more 'closely associated a non-member is to the EU, the more the relationship resembles a hegemonic one, even though the EU was not designed as a hegemon' (Fossum 2016: 2).
Similar to Fossum, Eriksen (2018: 1) argues that 'under the conditions of complex interdependence and economic integration, political differentiation can undermine the fundamental conditions for democratic self-rule'. By being excluded from EU decision making but still obliged to adopt EU rules, non-Member States experience 'the individual loss of sovereignty without being compensated with participatory rights in EU decision making' (ibid.: 16). Eriksen describes therefore the association of the EEA EFTA States as 'self-incurred dominance ' (ibid.: 11). Moreover, the participation of non-Member States in EU policy making suffers from 'a bureaucratization of there domestic governance process as the executive bureaucracy implements relevant EU legislation' (Schimmelfennig and Winzen 2020: 182). Further democratic issues for non-Member States are limited transparency on the real level of integration and a lack of political debate on the goals of the public policies resulting from their integration with the EU (Frommelt 2017).
The existing literature is indeed very critical about the EEA's overall democratic legitimacy. Because there is no joint decision-making mechanism for EEA rules and the EFTA States did not obtain a right of initiative of new EEA rules, there is a 'substantial loss of "operational sovereignty" for the EFTA countries without offering them satisfactory "voice opportunities" in return' (Gstöhl 1994: 333). Scholars have characterised the EEA therefore as 'legalized hegemony' (Pederson 1994), 'fax democracy' (Eliassen and Sitter 2004), 'semi-colonial' (Tovias 2006) and a 'hierarchical setting, in which EEA EFTA members have subordinated themselves to "foreign rule" by the EU' (Lavenex et al. 2009: 818).
These descriptions are contrasted by the fact that the EEA Agreement enjoys high public support in all three EEA EFTA States. For instance, only 5 percent of the Liechtenstein people call the EEA Agreement a bad agreement for Liechtenstein (Frommelt 2020b). At 17 percent this share is slightly higher in Norway, but still very low (Sverdrup et al. 2019). Moreover, people in Norway and Liechtenstein do not think that EU accession would give them significantly more influence in EU policy making. It is not clear whether this support is driven by instrumental cost-benefit calculation or whether it is based on a deeper confidence in the EEA and its institutions. However, it shows that people's assessment of the EEA is unlikely to be solely based on its democratic qualities. Moreover, many experts overlook the fact that although EEA law is based on EU law, there is still an EEA decision-making process that allows the EEA EFTA States to delay the incorporation of new EEA-relevant EU legislation into the EEA Agreement and to make EEA-specific adaptations.
According to Tallberg and Zürn (2019: 10), the 'principal institutional sources of legitimacy […] are the procedures through which IOs [international organisations] make decisions and the performance of IOs'. Moreover, they distinguish between 'democratic' and 'purposive' qualities of the two dimensions procedure and performance. Qualities are democratic 'when they give expression to or promote core values of the democratic process', and 'purposive when they serve or promote shared ends ' (ibid.: 11). Patterns such as 'participation, accountability, deliberation or transparency' are important to assess the democratic qualities of a procedure. By contrast, the purposive qualities of procedures are 'expert advice, efficiency and legality'. The democratic qualities of performance are the protection of rights and democratic processes, whereas the purposive qualities of performance are 'problem solving, collective welfare gains and distributive fairness' (ibid.: 12).
In this paper I argue that the democratic qualities of the EEA can be fed from two different sources. First, the access that the EEA EFTA States have to the policy making of EEA-relevant EU legislation Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers enables them to shape the content of an EEA-relevant EU act before it is formally adopted by the EU. The more extensive this access and the more equal the EEA EFTA States are treated with the EU Member States, the less the democratic qualities of the EEA will differ from those of the EU. However, the inclusion of the EEA EFTA States in EU policy making is limited by the fact that the EU wants to protect the autonomy of its decision making.
Second, to become legally binding for the EEA EFTA States, an EEA-relevant EU act adopted by the EU has to be formally incorporated into the EEA Agreement by a unanimous decision of the representatives of the EEA EFTA States and the EU in the EEA Joint Committee. The EEA Joint Committee can also make EEA-specific adaptations to an EU act. The democratic qualities of the EEA therefore directly correspond with the procedures for EEA decision making: The higher the autonomy of the EEA EFTA States in EEA decision making, the higher the democratic qualities of the EEA. High autonomy of the EEA EFTA States, however, could endanger the integrity of EU law by allowing the EEA EFTA States to opt out of certain legal obligations that result from their participation in the EU's internal market.
The purposive qualities of an international organisation are also closely linked with the functional integration that this organisation provides. In the EEA, functional integration enables the EEA EFTA States to participate in the EU's single market. This participation is based on the principle of homogeneity. Homogenous does not necessarily mean 'the same' (Baur 2019: 35) but still requires that new EEA-relevant EU legislation is continuously incorporated into the EEA Agreement as 'closely as possibly' after its adoption by the EU (Article 102 EEA Agreement). The purposive qualities of the EEA can therefore be based on the efficiency and legality of its procedures, as well as its ability to guarantee the EEA EFTA States access to the EU's single market. In the EEA context, efficiency and legality can best be operationalised by the time required to incorporate a new EU act into the EEA Agreement (Frommelt 2017). The faster the incorporation of new EU acts into the EEA Agreement, the more efficient the EEA decision-making procedure will be and the better the EEA will function.
To analyse the legitimacy of the EEA, this working paper will describe the different procedures for policy shaping and decision making in the EEA. Before this, the following chapter briefly describes the negotiations on the EEA Agreement and how they were determined by the EU's wish to protect the autonomy of its decision making and legal order and the EEA EFTA States' reservations about political integration.

The creation of the EEA
On 2 May 1992, after 'prolonged and exhaustive' negotiations (Gstöhl 1994: 336), representatives of the then European Communities (EC) and its 12 Member States, together with representatives of the seven Member States of the European Free Trade Association (EFTA), signed the Agreement on the European Economic Area (EEA). After Switzerland opted out of the EEA as a result of a popular vote, the EEA Agreement entered into force on 1 January 1994 for the European Union (EU) and five EFTA States: Austria, Finland, Iceland, Norway and Sweden. The turbulent beginning to the EEA continued when Austria, Finland and Sweden joined the EU on 1 January 1995, leaving Iceland and Norway as the only two EFTA States that were also Contracting Parties to the EEA. On 1 May 1995, Liechtenstein finally joined the EEA after adaptations to the EEA Agreement and the Swiss-Liechtenstein Customs Union Treaty were approved by a popular vote.
By removing all barriers to the free movement of goods, services, capital and persons, the EEA's functional scope is similar to the EU's internal market. In addition, the EEA Agreement includes horizontal and flanking policies, e.g. environmental protection, where cooperation is necessary to ensure a level playing field across the EEA. The EEA Agreement extends the EU's internal market to the EFTA States, but does not cover the Common Agricultural Policy, the Common Fisheries Policy, the Customs Electronic copy available at: https://ssrn.com/abstract=3712440 The signing of the EEA Agreement was the result of a process triggered by the wish to create a common economic space comprising both the EU and EFTA (Rye 2015;Gstöhl 1994). Although bilateral free trade agreements for industrial goods had been in place since the early 1970s, and several sectoral agreements followed, the completion of the EU's internal market 'raised concerns of marginalisation and of being left behind' in the EFTA States (Leuffen et al. 2013: 118). A key moment on the way to the formation of the EEA was the Luxembourg summit in 1984, when the intention of forming a common European economic area was first expressed. The scope of the EEA Agreement was a crucial issue, in particular as the EFTA States wanted to have some unlimited derogations from the EU acquis (Gstöhl 1994: 342-344). There were also different views on a financial contribution by the EFTA States to the economic and social cohesion of the EU. In the end, however, the search for an institutional framework for the envisaged association turned out to be the biggest challenge.
The EFTA States aimed to avoid political integration that would pool their sovereignty by qualified majority decisions or supranational institutions. They also strove to have a say in EU decision making for EEA-relevant policies. By contrast, the EU was eager to prevent the EFTA States 'free riding' and 'cherry picking' from the internal market. It was not willing to jeopardise either the integrity of its legal order or the autonomy of its decision-making processes.
This became clear in the speech made by EC Commissioner Willy De Clercq at the EC-EFTA ministerial meeting in Interlaken in 1987. De Clercq mentioned the three principles that determine the EU's relationship with non-Member States to this day: The first principle states that 'Community integration comes first'. Hence, the EU's relations with third countries should not prevent or slow down the progress of integration within the EU itself. In other words, against the background of already complex and time-consuming decision making in the EU, the EU did not want to 'complicate that process even further by establishing formal institutional arrangements with non-Member States' (De Clercq 1987). The second Interlaken principle is therefore that the EU's 'decision-making authority must be preserved'. Finally, in his speech, De Clercq refers to the different interests of EU Member States, and that a decision can only be achieved 'if an overall and balanced solution is found'. His third principle therefore is that the relations between the EU and a non-Member State are defined by a 'balance of benefits and obligations'. These principles were repeated in Jacques Delors' famous speech to the EP in January 1989, which initiated the decisive period of the EEA negotiations.
Whereas a common decision-making process was never realistic, the negotiators of the EEA Agreement initially agreed on an institutional framework comprising a common EEA Court and an EEA Surveillance Authority. However, the European Commission requested an opinion from the European Court of Justice (ECJ) on the compatibility of the proposed EEA Agreement with the existing Treaties of the European Communities. On 14 December 1991, the ECJ adopted its famous Opinion 1/91, according to which the creation of an independent EEA Court composed of judges from the ECJ and the EFTA States and functionally integrated with the ECJ was incompatible with EU law, since it would be 'likely adversely to affect the allocation of responsibilities defined in the Treaties and the autonomy of the Community legal order, respect for which must be assured exclusively by the Court of Justice pursuant to Article 164 of the EEC Treaty'.
The ECJ has confirmed its opinion on several occasions (see Müller 2020). Moreover, the substance of the Interlaken principles can also be found in the European Council's guidelines for Brexit negotiations. The European Council (2017) 'reiterates that any agreement with the United Kingdom will have to be based on a balance of rights and obligations, and ensure a level playing field'. A non-member of the Union 'that does not live up to the same obligations as a member, cannot have the same rights and enjoy the same benefits as a member'. The guidelines of the European Council also state that the EU 'will preserve its autonomy as regards its decision making as well as the role of the Court of Justice of the European Union'. Another principle of the Brexit negotiations that was not explicitly mentioned Robert Schuman Centre for Advanced Studies Working Papers in the Interlaken speech by Willy De Clercq but still formed an essential part of the EEA negotiations is the goal of 'preserving the integrity of the Single Market' by excluding 'participation based on a sectorby-sector approach'.
The comparison of the principles of the EEA negotiations and the Brexit negotiations indicates a certain consistency of the EU's relations with non-Member States. The remainder of the working paper therefore considers how these principles have been implemented in practice and thus how consistent the institutional arrangements of the EEA are with its underlying principles. The next chapter thus briefly describes the institutional framework of the EEA.

The institutional framework of the EEA
The EU largely prevailed in the negotiations on the EEA (Gstöhl 1994: 355). The institutional set-up of the EEA can best be described as a two-pillar structure, with the EFTA institutions matching those on the EU side. This two-pillar structure encompasses EEA decision making as well as the supervision and judicial control of EEA law. The EU and EFTA sides therefore manage their internal matters independently, but there are various informal and formal mechanisms to ensure cooperation between the EU and EFTA institutions.

EEA Council
In the EEA Council, the EU is represented by the rotating Presidency of the Council of the EU and by representatives of the European External Action Service (EEAS) and the European Commission, while the EEA EFTA States are represented by their Ministers for Foreign or European Affairs. Representatives of the EFTA Secretariat and the EFTA Surveillance Authority (ESA) attend EEA Council meetings as observers. The Presidency of the EEA Council alternates between the EU and the EFTA side. The EEA Council meets twice a year and provides political impetus for the development of the EEA Agreement and the guidelines for the EEA Joint Committee.
The EEA Council is the highest political body in the EEA. In recent years it has increasingly been used as a forum for political discussion. However, the EEA Council has not yet reached the political relevance originally envisaged by the EFTA States. This can be explained, inter alia, by the fact that the actual scope for political impetus to shape the further development of the EEA is limited by the fundamental imbalance of power between the EU and EEA EFTA States. As a result, the EEA EFTA States address challenges to the functioning of the EEA Agreement at a technical rather than political level, and thus in the EEA Joint Committee rather than in the EEA Council.

EEA Joint Committee
The EEA Joint Committee is composed of representatives of the EEAS, the three EEA EFTA States (usually at ambassadorial level) and ESA as observer. It meets approximately six to eight times a year and takes further decisions through correspondence about twice a year. Decisions in the EEA Joint Committee are taken by consensus between the EU on the one hand and the EEA EFTA States on the other. The EEA EFTA States must speak with one voice, which means that they have to reach an agreement between themselves before formal negotiations with the EU take place. The obligation to speak with one voice is often seen as one of the greatest challenges of the EEA Agreement, as the EEA EFTA States have to agree on a common position, even when their preferences and capabilities in respect of adopting and implementing EU law may differ. On the other hand, the "single voice principle" is likely to increase the homogeneity of EU and EEA law, as it forces the EEA EFTA States to cooperate, and restricts the demand for country-specific opt-outs by the EEA EFTA States from the outset.
The EEA Joint Committee is responsible for the day-to-day management of the EEA Agreement and for decisions on the incorporation of new or amended EU legislation into the EEA Agreement. The EEA Electronic copy available at: https://ssrn.com/abstract=3712440

Other joint EEA bodies
In contrast to the EEA Council and the EEA Joint Committee, the roles of the EEA Joint Parliamentary Committee and the EEA Consultative Committee are purely advisory. The EEA Joint Parliamentary Committee is composed equally of members of the EP and members of the national parliaments of the EEA EFTA States, while the EEA Consultative Committee is composed of representatives of the social partners of the EU and EEA EFTA States. The two bodies strengthen the dialogue between the EU and EFTA pillars, with the aim of improving the understanding of the special nature of the EEA Agreement and its Contracting Parties. The two institutions are not directly involved in the EEA decision-making process, but monitor and review EEA-relevant developments by drawing up relevant reports and resolutions.

EFTA Standing Committee
The Standing Committee of the EFTA States is based on a separate international agreement between the EEA EFTA States. Therefore, neither the EU nor Switzerland are parties to the Agreement on the Standing Committee of the EFTA States. The main purpose of the Standing Committee is to draw up decisions that will later be taken by the EEA Council or the EEA Joint Committee. The necessity of such a body results from the obligation of the EEA EFTA States to speak with one voice in the EEA Council and in the EEA Joint Committee. However, the Standing Committee also exercises various administrative and executive functions in connection with the EFTA pillar of the EEA. In particular, it is responsible for decisions on the relationship between the three EEA EFTA States.
The Standing Committee consists of the Ambassadors of Iceland, Liechtenstein and Norway, as well as observers from Switzerland and ESA. It normally meets the day before the meeting of the EEA Joint Committee in order to adopt a common position before that meeting. In general, the Standing Committee decides unanimously. In certain situations, however, the Standing Committee can take decisions and make recommendations by majority vote. For example, in the area of financial services, the possibility of majority voting was introduced for the inclusion of the European Supervisory Authorities in the EEA Agreement (see Fredriksen andJónsdottír 2018: 1096). The possibility of majority decision making shows that there are different modes of governance in the EEA, some of which go beyond the principle of purely intergovernmental cooperation.

EFTA Subcommittees
The Standing Committee has five subcommittees dealing with the free movement of goods, capital and services, persons, horizontal and flanking policies, and legal and institutional issues. Sub-committees I-IV have met jointly since 2009. Their task is to prepare for the incorporation of new EU legislation into the EEA Agreement. Thus, Subcommittees I-IV must also approve a draft decision of the EEA Joint Committee before it can be passed to the EU. The meetings of Subcommittees I-IV are usually followed by a joint meeting with the EU in order to discuss draft decisions and find common solutions. Subcommittee V is composed of legal experts from the foreign ministries of Norway and Iceland and the EEA Coordination Unit of Liechtenstein. It is primarily concerned with the analysis and evaluation of legal and institutional issues. In addition to the subcommittees, there are several working groups for the individual regulatory areas. These working groups are composed of experts of the EEA EFTA States' Robert Schuman Centre for Advanced Studies Working Papers national administrations. The working groups make a fundamental contribution to the EEA incorporation process by providing practical and legal analysis of each new EU legislative act.

EFTA Secretariat
The EFTA Secretariat is not part of the EEA's two-pillar structure, but is an EFTA authority that dates back to the foundation of the EFTA. The EFTA Secretariat has offices in Brussels, Geneva and Luxembourg. The Secretary General is based in Geneva. Nevertheless, the EEA accounts for a large part of the work of the EFTA Secretariat based in Brussels. This includes, in particular, assisting the EEA EFTA States in the incorporation of new EU law into the EEA Agreement, and in so-called 'decision shaping' as well as coordinating the cooperation with the relevant EU institutions and bodies. In addition, the EFTA Secretariat provides the public with various information on the processes and institutions of the EEA and the development of law in the EEA. In simple terms, the EFTA Secretariat has a coordinating function vis-à-vis the EEA EFTA States and the EU, provides the institutions and actors of the EEA with expertise, and takes care of the public image of the EEA. Although the EFTA Secretariat is not part of the two-pillar structure of the EEA and is not mentioned elsewhere in the EEA Agreement, it occupies an outstanding position for the way the Agreement operates.

EEA Surveillance Mechanism
The monitoring mechanism of the EFTA pillar consists of the ESA, located in Brussels, and the EFTA Court, located in Luxembourg. As in the case of the Standing Committee, the tasks and powers of the ESA and the EFTA Court are governed by a separate agreement between the EEA EFTA states, the socalled Surveillance and Court Agreement (SCA). In contrast to the Standing Committee, however, the establishment of the ESA and the EFTA Court was provided for in the EEA Agreement itself (see Article 108 EEA Agreement).
The EFTA Court deals with infringement procedures brought by the ESA against an EEA EFTA State concerning the implementation, application or interpretation of EEA law. At the request of a court of an EEA EFTA State, the EFTA Court also issues opinions on the interpretation of EEA law in the EEA EFTA States, has jurisdiction to settle disputes between two or more EEA EFTA States, and hear complaints about decisions of the ESA. There is an ongoing formal and informal exchange between the ESA and the European Commission as well as between the EFTA Court and the CJEU.
ESA and the EFTA Court are responsible for ensuring that the EEA EFTA States fulfil their obligations under the EEA Agreement. The surveillance mechanism is very similar to that of the EU. But there are also some differences. For example, the members of the ESA College are appointed solely by the EEA EFTA governments and do not have to be approved by a parliament, as in the case of the European Commission. Moreover, most College members have close ties to politics in the EEA EFTA States by having served various years as civil servant or diplomat. Some experts see this as threat of the independency of ESA. There is also no provision in the EEA Agreement, which gives the EFTA Court the right, at the ESA's request, to impose a lump sum or penalty payment for non-compliance with a judgment of the EFTA Court. Furthermore, judgments and opinions of the EFTA Court are not meant to be binding. In practice, however, they have a 'persuasive authority', i.e. in the vast majority of cases they are complied with, just like rulings of the CJEU. Finally, we should also mention the small size of the EFTA Court, with only three members, and the small number of cases compared to the CJEU, which is due not only to the fact that there are only three EEA EFTA states, but also to the fact that their courts less frequently seek an opinion from the EFTA Court by comparison with the average EU State.
In summary, the EEA's institutional structure consists of two independent pillars and some common institutions. This model, agreed by the EU and EFTA, formally preserves the decision-making autonomy of both sides, whilst at the same time creating the institutional framework for joint efforts in order to secure a homogeneous and dynamic economic area. On the other hand, the following chapters of this Electronic copy available at: https://ssrn.com/abstract=3712440 Institutional Challenges for External Differentiated Integration: the Case of the EEA European University Institute 9 paper will show that the EU's functional and institutional integration relevant to the EEA is highly dynamic and far-reaching, which often makes it impossible to allocate decision-making powers in the two-pillar structure in a way that is consistent with the basic principles of the EEA. If such an allocation of decision-making powers is not possible within the EEA's two-pillar structure, this requires specific arrangements to include the EEA EFTA States in the EU policy making. Gstöhl and Phinnemore (2019b: 179) summarise the institutional requirements for a privileged partnership between the EU and non-Member States in four questions: '(1) How to keep agreements upto-date in light of relevant new EU law;

Policy shaping
(2) how to monitor partners' compliance; (3) how to ensure the uniform interpretation of agreements in line with the EU law from which they are derived; and (4) how to settle disputes between the parties'. In other words, the institutions of external differentiated integration have to address the entire process from policy formulation to the settlement of potential disputes that are linked with a specific policy (Lavenex and Schimmelfennig 2009;Lavenex and Krizic 2019). This paper mainly focuses on the stages of policy formulation and decision making. Under socalled 'decision shaping', the EEA EFTA States are involved in the creation of new EU legislation from an early stage. This chapter first describes the types of policy shaping and its legal framework. Afterwards, it focuses on additional institutional arrangements that have become necessary in order to ensure an adequate representation of the EEA EFTA States in EU policy making. Finally, the chapter analyses the strategies of policy shaping and analyses the legal limitations of non-Member States participation in EU policy making. The chapter shows that the EEA EFTA States' access to the EU policy making varies considerably across the different policies of the EEA, but that the EU has always been able to preserve its decision-making autonomy because it has never given the EEA EFTA States the right to vote.

Types of policy shaping in the EEA
Decision shaping describes 'the process of contributing to and influencing policy proposals up until they are formally adopted' (EFTA Secretariat 2009: 20). In a wider sense, decision shaping can also be described as 'the phase of preparatory work undertaken by the European Commission to draw up new legislative proposals. The EEA Agreement contains provisions for input from the EFTA side at various stages before new legislation is adopted' (EFTA Secretariat 2019). The term itself is not included in the EEA Agreement, nor is it an official EU concept. Nevertheless, it has become the established term to describe the various ways in which the EEA EFTA States contribute to EU policy making despite not having the right to vote. Hence, decision shaping contrasts with decision making on new legal acts, which is left to the EU Member States. Moreover, the term 'decision shaping' might be too narrow as the inclusion of the EEA EFTA States in EU policy making 'is not limited to influencing individual decisions, but rather EU policy of relevance to the EEA more broadly' (EFTA Secretariat 2009: 20). Hence, 'policy shaping' might be a more accurate term to fully capture the EEA EFTA States' access to EU policy making.
There are at least four different ways for the EEA EFTA States to shape EU policies (Frommelt 2017;Baur 2020): (1) The EEA EFTA States can second national experts (SNE) to the European Commission or other EU institutions. The aim of such a secondment is to supply the EU institutions with expertise but also to increase and spread knowledge of the EU, the EEA and their member states. In general, there are around 50 national experts of the EEA EFTA States seconded to the EU. Most of them are Norwegians.
(2) The EEA EFTA States can submit comments and written contributions to the relevant services of the European Commission and the EP in order to express their views on a specific policy. Such comments can address any kind of legislative or strategic documents of the EU but mostly refer to a white paper or to the proposal of a new EU act. On average the EEA EFTA States have submitted 10 comments per year since 2001 (Karlsen 2020). (3) The EEA EFTA States may participate in EU Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers committees and expert groups as well as specified EU bodies such as the board of supervisors of an EU agency, albeit without the right to vote. This is actually the only way for the EEA EFTA States to be formally represented in EU policy making. However, as will be shown in detail below, the participation is limited to specific stages of EU policy making and may even face further constrains. (4) Where the EEA EFTA States do not have formal access to EU policy making, they may use other forms of influence such as lobbying. Figure 1 illustrates the policy shaping by the EEA EFTA States. The EEA EFTA States have access to EU policy making mainly in the preparatory phase of an EEA-relevant EU act. The EEA Agreement ensures that the national experts of the EEA EFTA States can participate in expert groups of the European Commission on equal terms with the national experts of the EU Member States but without the right to vote. As soon as an EU act is handed over to the EU legislators, the EEA EFTA States are no longer formally represented in EU policy making. This "representation gap" can only be compensated by a political dialoguefor instance within the EEA Council or the EEA Joint Committeeor by lobbying. By contrast, the EEA EFTA States are represented throughout the main steps of the policymaking process for delegated EU acts (Article 290 TFEU) or implementing EU acts (Article 291 TFEU) adopted by the European Commission but again without the right to vote. The same applies for decisions taken by EU agencies in which the EEA EFTA States participate.

Figure 1: Participation of the EEA EFTA States in EU policy making
Source: author's own compilation

Legal framework of policy shaping
The legal framework of the EEA EFTA States' access to EU policy making is based on Articles 79 to 81 of the EEA Agreement (cooperation outside the four freedoms) as well as Articles 99 to 101 (institutional provisions). Article 79 of the EEA Agreement states that the Contracting Parties shall strengthen the dialogue with each other in order to achieve closer cooperation. To this end, the Contracting Parties should also exchange views on framework programmes, special programmes or other EU actions outside the four freedoms. Article 80 stipulates that the EEA EFTA States may participate in the relevant programmes. Finally, Article 81 regulates the participation of the EEA EFTA Electronic copy available at: https://ssrn.com/abstract=3712440 Institutional Challenges for External Differentiated Integration: the Case of the EEA European University Institute 11 States in the committees assisting the European Commission in the implementation and development of such programmes. The status of the EEA EFTA States in a committee is determined by their formal participation in the related programme and the financial contributions linked to it. Furthermore, Article 81 EEA states that institutions, undertakings, organisations and nationals of the EEA EFTA States shall have the same rights and obligations under such EU programmes as institutions, undertakings, organisations and nationals of EU Member States.
The most relevant provisions for decision shaping are set out in Articles 99 to 101. Article 99 requires the European Commission to seek the advice of experts from the EEA EFTA States when drafting new legislation within the functional scope of the EEA Agreement, at the same time and to the same extent that it does with experts from EU Member States. During the phase preceding the decision of the Council of the EU, the Contracting Parties may consult 'each other in the EEA Joint Committee at the significant moments at the request of one of them' in order to establish a 'continuous information and consultation process' (Article 99 EEA). Finally, Article 99 states that the Contracting Parties 'shall cooperate in good faith during the information and consultation phase with a view to facilitating, at the end of the process, the decision taking in the EEA Joint Committee'. This wording shows that the EEA EFTA States' access to EU policy making must also ensure the functionality of the EEA Agreement.
Article 100 governs the participation of the EEA EFTA States in the EU comitology system. Comitology committees assist the European Commission in the exercise of its enforcement powers. Article 100 provides that the EEA EFTA State experts shall be consulted on the same basis as experts of the EU Member States when drafting measures. Overall, the participation of the EEA EFTA States should be as broad as possible, which essentially corresponds to full involvement except for the right to vote. Finally, Article 101 refers to further committees that could be important for the functioning of the EEA Agreement and are therefore open to the EEA EFTA States. These committees are listed in Protocol 37 to the EEA Agreement.

Additional arrangements for policy shaping
Protocol 1 to the EEA Agreement stipulates how EU acts incorporated into the EEA Agreement shall be applied. Due to Protocol 1, the Contracting Parties to the EEA Agreement do not have to make recurrent adaptations for EEA-specific provisions to EU acts when incorporating them into the EEA Agreement. For instance, if an EU act sets up a new committee, there is generally no need for an EEAspecific adaptation because Protocol 1 ensures that the EEA EFTA States gain access to that committee as soon as the EU act is incorporated into the EEA Agreement.
Despite Protocol 1, however, there are numerous provisions governing the EEA EFTA States' access to EU policy making in the Annexes and Protocols to the EEA Agreement, based on EEA-specific adaptations that were agreed in the EEA Joint Committee when those EU acts were incorporated into the Agreement. For instance, such ad hoc rules for EU policy shaping by the EEA EFTA States were necessary in order to provide the EEA EFTA States with access to the various EU agencies and their bodies. As those bodies are not reflected in the EEA's two-pillar structure, the provisions set out in the EEA Agreement were not sufficient to ensure EEA EFTA participation. Instead, the conditions of the EEA EFTA States' access to those bodies had to be agreed when incorporating the founding acts of those EU bodies into the EEA Agreement (Frommelt 2017: 60-63).
Textbox 1 lists some examples of ad hoc rules on the EEA EFTA States' participation in EU policy making. It shows that the wording of such adaptations may differ. However, all adaptations have in common that they grant the EEA EFTA States access to EU policy making but without the right to vote. The difference between observer status (example 2) and full participation without the right to vote (example 1) is not clearly defined. Interview evidence suggests that as an observer the EEA EFTA States not have a mandate to issue opinions and have limited access to documents. Example 3 gives the EEA EFTA States particularly far-reaching access as the adaptation text also refers to internal committees and panels within the functional scope of the EEA Agreement. However, the EEA EFTA States can only Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers be represented in bodies where all EU Member States are represented. Taking the example of the European Banking Authority, this means that the EEA EFTA States can participate in the Board of Supervisors but not in the Management Board. Another example of far-reaching inclusion of the EEA EFTA States in EU policy making is provided by the European Data Protection Board, where the EEA EFTA States fully participate but for the right to vote and to stand for election as chair or deputy chairs of the Board. An interesting point is here that positions of the supervisory authorities of the EEA EFTA States are recorded separately by the Board. The EEA EFTA States have never doubted the EEA relevance of the BEREC Regulation. However, the regulation raised challenges for the two-pillar system since it was not clear how the supervisory governance framework established therein would function in the EEA. The main questions were whether BEREC should be competent to issue opinions concerning the EEA EFTA States or whether there should be a separate body in the EFTA pillar. Moreover, the BEREC Regulation contained provisions limiting the status of the EEA EFTA States on the Management Board of BEREC as mere observers (32009R1211, Article 4, para 1). From the perspective of the EEA EFTA States, this was not in line with the institutional and procedural framework of the EEA Agreement. As a result, the EEA EFTA States proposed to include an adaptation text setting out that the EEA EFTA States shall have the right Electronic copy available at: https://ssrn.com/abstract=3712440 According to the Norwegian Government (Norway 2012: 11) long-term lobbying efforts vis-à-vis EU institutions and Member States shall enhance the credibility of Norway and provide a solid basis for Norway to have an influence on EU policy making. The experience of Norway and Iceland relating to maritime affairs is particularly appreciated. The EEA EFTA States may also be invited to participate in the Council's informal meetings, organised by the respective Council Presidency prior to the ordinary biannual meeting of the European Council. Still, in contrast to the Schengen association, the EEA Agreement does not ensure any kind of formal access to Council meetings or its working groups (Baur 2019). Moreover, the actual degree of the EEA EFTA States' cooperation with the Council strongly depends on the rotating presidency (EEA Review Committee 2012: Chapter 9.3.3).
A common pattern of the EEA EFTA States' reports on their engagement in EU policy making is their lack of capacity. Although the three EEA EFTA States differ greatly in size, compared to the EU Member States they are all small states. As a result, the EEA EFTA States do not have the capacity to participate in all of the committees and expert groups to which they have access, and thus have to define their own political priorities. Intra-EFTA collaboration is therefore particularly important for the two smaller EEA EFTA States, Iceland and Liechtenstein (Jónsdottír 2013: 48;Frommelt 2015), and is addressed in the procedures for the incorporation of EU acts into the EEA Agreement according to which 'EEA EFTA experts attending such Commission meetings inform experts of other EEA EFTA States and the Secretariat about EEA-relevant acts under preparation by the Commission' (Decision of the Standing Committee of the EFTA States No 1/2014/SC).
Although the EEA Agreement ensures full access to expert groups of the European Commission and comitology committees, the representatives of the EEA EFTA States may still face restrictions. For instance, some EEA EFTA experts have reported to the EFTA Secretariat (2002: 14) that they were asked to leave a committee meeting before voting took place. This was mainly the case when nongovernmental stakeholders from the EU States were participating in a committee and the representatives of the EEA EFTA States were asked to leave the meeting together with them. On the other hand, various EEA EFTA experts have reported that they do not face such restrictions in the committee they attend (ibid.: 14). Several other issues are regularly addressed by experts of the EEA EFTA States, such as restricted access to documents, late invitations and document transmission, unfavourable seating plans or restrictions on their speaking time. Such restrictions, however, are probably the exception rather than Robert Schuman Centre for Advanced Studies Working Papers the rule, and there seems to be no structural discrimination of the EEA EFTA States in EU committees (Frommelt 2017: 69). Interview evidence also suggests that the chair of a committee plays an important role in ensuring that the involvement of experts from the EEA EFTA States is equal to that of their EU counterparts (ibid.: 70). Finally, it is important that most of the committees to which the EEA EFTA States have access generally decide by consensus rather than by majority or unanimous vote. Figure 2 shows that after the new comitology system of the Lisbon treaty was implemented in 2013 the big majority of the EU acts incorporated into the EEA Agreement were delegated or implementing acts. By contrast, between 2013 and 2018 only 8 percent of the incorporated EU acts were legal acts adopted by the Council or the Council and the EP jointly. This shows how important for the EEA EFTA States their access to the comitology system of the EU is. Nevertheless, it cannot compensate for the democratic costs that result from the limited access to the policy making of the EU Council and the EP as the comitology is determined by the technical expertise of the executive bureaucracy whereas the main goals and standards of public policies are set by the EU legislators.  2009: 26) active European politics, meaning vigorous decision shaping, remains mainly a way to gather information rather than to exert political influence (EEA Review Committee 2012: Chapter 9). In other words, the representation of the EEA EFTA States in EU policy making does not increase their autonomy but the overall functioning of the EEA.

Legal limits of policy shaping
Regulation (EC) No 1211/2009, the old BEREC Regulation, explicitly excluded the participation of third countries. But can an EU act also enable full representation of third countries in EU bodies? This question was raised mainly in the context of the EEA EFTA States' participation in EU agencies that can enact binding decisions vis-à-vis the EEA EFTA States. Literally speaking, participation without voting rights is not participation (Bekkedal 2019: 416). As a result, Bekkedal argues that 'from the perspective of the naïve Norwegian (…) the most integrated outsider must be allowed to participate fully in the EU agencies, with voting rights' if they also accept all competences of such an EU body and contribute financially.
This demand corresponds with the basic idea of Heermann and Leuffen (2020: 14)  be designed and governed with respect to a unified set of democratic principles in order to maintain the EU's overall legitimacy'. According to Heermann and Leuffen, these democratic principles are 'autonomy', 'accountability' and 'political equality'. To implement these principles, any 'mismatch between the differentiated territorial scope of public policies and the composition of legislatures' has to be avoided (ibid.: 13). Although Heermann and Leuffen focus on internal differentiated integration, the principles of autonomy, accountability and equality might be extended to external differentiated integration as well if the non-Member State does indeed strive for full policy-specific integration and does not insist on specific institutional and procedural arrangements.
According to Müller (2020), full representation of the EEA EFTA States in EU policy making is not possible under the current legal framework of the EU. Müller's argumentation is based on CJEU Opinion 1/91, which declared the initial institutional framework of the EEA incompatible with the principles of the EU, in particular the EU's wish to protect the autonomy of its legal order and decision making. Müller argues that the CJEU has confirmed the importance of the autonomy of the EU's legal order in various judgments, in particular Opinion 2/13 on the accession of the EU to the European Convention for the Protection of Human Rights and Fundamental Freedoms. Consequently, the principle that the EU's full decision-making power must be preserved in the EU's relations with non-Member States remains valid. Nonetheless, taking into account how diverse EU governance has become by the proliferation of new bodies and modes, it would at least be worth considering whether for distinct bodies from the EU institutions, such as EU agencies and where decisions are taken on the basis of a simple majority of the voting members, full representation of third countries should become possible in the future. This is on the condition that the EEA EFTA States refrain from any special arrangements within the policy field governed by the respective EU agency.
In summary, this chapter has shown that there are various ways in which the EEA EFTA States can access the making of EEA-relevant EU legislation. However, their formal representation is only legally guaranteed in the various expert groups and comitology committees of the European Commission as well as the bodies of the EEA-relevant EU agencies. Their lack of access to the EU Council and EP severely limits the EEA EFTA States' influence on new EEA-relevant EU legislation. On the other hand, empirical analyses of the legislative dynamics of the EEA show that less than 10 percent of the EU legislation incorporated into the EEA Agreement was adopted by the EU Council or the EU Council and the EP. In other words, the majority of EU legislation incorporated into the EEA Agreement was adopted by the European Commission and thus prepared by bodies to which the EEA EFTA States had access. Nevertheless, the involvement of EEA EFTA States in EU policy making cannot guarantee the same democratic quality as EU policy making does for the EU Member States, nor can it fully legitimise the dynamic incorporation of new EEA-relevant EU legislation into the EEA Agreement. Against this background, the following two chapters focus on the different procedures for incorporating EEArelevant EU legislation into the EEA Agreement. They show that the more the EEA EFTA States are subordinated to EU decision making, the more relevant their representation becomes in EU policy making.

EEA Decision making
The principle of homogeneity requires that new EEA-relevant EU legislation is continuously incorporated into the EEA Agreement. Since the EEA Agreement entered into force on 1 January 1994, more than 10,500 EU acts have been incorporated into its Annexes and Protocols (EFTA Secretariat 2020a). This high number of incorporated EU acts confirms the dynamic nature of the EEA Agreement. Likewise, it shows how important efficient procedures are in order to cope with the dynamics of EU legislation. This chapter presents the different procedures of EEA decision making. It then describes the potential outcomes of EEA decision making before analysing the level of autonomy of the EEA EFTA States in EEA decision making.
Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers

Standard procedure of EEA decision making
When the EEA Agreement entered into force there was only a single procedure foreseen for EEA decision making. The so-called "standard procedure" takes into account that the 'administration and management of the EEA is shared between the EU and the EEA EFTA States in a two-pillar structure' (EFTA Secretariat 2020b). Substantive decisions relating to the EEA Agreement and its operation are therefore 'a joint venture and are taken by joint EEA bodies (…) established by the EEA Agreement and consisting of representatives both from the EU side and the EEA EFTA States' (ibid.). The EEA EFTA States 'have not transferred any legislative competences to the joint EEA bodies' (ibid.). Hence, the standard procedure for EEA decision making does not formally constrain their legislative sovereignty.
The five phases of the standard procedure are: 1) The early assessment of EEA challenges relating to a proposed EU act; 2) the formal assessment of the final EU act upon its publication in the Official Journal of the European Union; 3) the drafting of the Joint Committee Decision (JCD) to incorporate an EU act; 4) the formal adoption of the JCD; and 5) the national ratification of the JCD if one or more of the EEA EFTA States needs to fulfil a particular constitutional requirement.
Phase 1 takes place before an EU act is formally adopted by the EU. It is thus not an incremental part of EEA decision making. However, such an early assessment is likely to reduce the time required to incorporate the EU act into the EEA Agreement. This applies in particular for implementing acts and delegated acts (Isaksen 2020: 60) where the EEA EFTA States benefit from their far-reaching participation in EU policy making.
In phase 2 and 3 each EU act has to undergo a 'thorough process in several steps involving the EFTA Secretariat, the EFTA States, the European Commission and the EEAS, and in certain cases also the Council of the EU' (ibid.: 58). The purpose of this process is the assessment of the EEA relevance of an EU act, its need for adaptations, the compatibility of such adaptations with the principle of homogeneity, and the necessity of parliamentary ratification. These assessments are often executed in parallel in the EFTA and EU pillar. In the EFTA pillar, the EFTA Secretariat coordinates the exchange between the national experts of the EEA EFTA States represented in the EFTA Working Groups, EFTA Subcommittees and EFTA Standing Committee. On the EU side, the incorporation process is coordinated by EEAS. The rules of procedure for deciding on EU position in the EEA Joint Committee are laid down by Regulation (EC) No 2894/94. If an EU act requires 'more than mere technical adjustments' those adaptations have to be adopted by the Council based on proposal from the European Commission (Article 1(3)). The EP has to be consulted (Article 3(2)). In practice, there is also an 'interservice consultation' in the EU pillar involving the relevant Commission Directorates-General (DGs) as well as the Commission's Legal Service (Isaksen 2020: 63).
The incorporation of a new EU legal act takes place by a formal decision of the EEA Joint Committee. In Case E-6/01 CIBA the EFTA Court has described a decision of the EEA Joint Committee as a 'simplified form of an international agreement between the Community and its Member States on the one hand, and the EFTA States party to the EEA Agreement on the other'. The role of the EEA Joint Committee can therefore not be restricted to adopting the relevant EU acts into the EEA. Instead, the EEA Joint Committee can within the 'boundaries of the EEA Agreement' make adaptations to EU acts when necessary and can also issue joint statements.
As the EEA EFTA States have to speak with one voice in the EEA Joint Committee, they need to agree on a common position. In other words, any kinds of adaptations or opt-outs require the agreement of all EEA EFTA states and must be agreed with the EU as well. For the Liechtenstein government (2020: 37) the "speaking with one voice principle" is one of the biggest challenges of the EEA as one EEA EFTA State can seriously delay the incorporation process for all three EEA EFTA States. From an integrationist perspective, however, the "speaking with one voice principle" is necessary to avoid cherry-picking by the EEA EFTA States.
Electronic copy available at: https://ssrn.com/abstract=3712440  (Neier 2020;Frommelt 2017). Finally, due to their dualistic legal orders, Iceland and Norway have to transpose an EU act into national legislation before it can have effect within the domestic legal system (Isaksen 2020: 67). This does not apply to Liechtenstein that has a monistic legal order. However, it should be noted that the question of direct applicability in the EEA is still unclear (Bekkedal 2020). Figure 3 illustrates the standard procedure for EEA decision making starting with the adoption of an EEA-relevant EU act by the EU legislators and ending with the decision of the EEA Joint Committee to incorporate the new EU act into the EEA Agreement.

Further procedures for EEA decision making
The standard procedure applies to all EU acts that are not subject to the "simplified" or "fast-track" procedure. The simplified procedure was introduced in 2001 for veterinary acts concerning imports from third countries and was subsequently extended to other veterinary acts. It derogates from the general rule that all EU acts have to be incorporated into the EEA Agreement by way of a JCD before becoming applicable in the EEA EFTA States. According to the simplified procedure in the veterinary field, the EEA EFTA States 'simultaneously take measures corresponding to those taken by the EU' (Standing Committee 2014: point 2.2.3). This means that the EEA EFTA States are obliged to implement and apply EU acts in the same manner and within the same deadlines as those applicable in the EU Member States. Four times a year the EEA Joint Committee is invited to take note of the EU acts that are applicable to EEA EFTA States based on the simplified procedure. However, this is mainly for transparency reasons and does not give the EEA Joint Committee any substantive role in the procedure. A special variation of the simplified procedure was also established by the incorporation of Regulation (EU) 2016/679 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data (General Data Protection Regulation, GDPR).
In 2014, the EEA EFTA States agreed on the introduction of the so-called "fast-track procedure". This followed increasing pressure by the EU on the EEA EFTA States to reduce delays when incorporating new EU secondary law into the EEA Agreement. The fast-track procedure applies to EU acts that, by their nature: (1) do not raise any EEA horizontal challenges; (2) do not need any adaptations; and (3) do not entail any national constitutional requirements to be fulfilled (EFTA Secretariat 2016: 6). It can be summarised as follows: based on the criteria defined by the Standing Committee, the EFTA Secretariat identifies EU acts suitable for the fast-track procedure; if the EEA EFTA State experts do not oppose the choice of procedure by the EFTA Secretariat, the EFTA Secretariat prepares a draft JCD, which is first approved by the relevant EFTA subcommittee and then by the EEA Joint Committee.
While the simplified procedure has been imposed by the EU, the fast-track procedure was invented by the EEA EFTA States and does not involve the EU pillar. The fast-track procedure is designed to ensure the swift incorporation of new EU secondary law into the EEA Agreement. However, in contrast to the simplified procedure, where EU acts are incorporated without a JCD and without the formal involvement of the EFTA Subcommittees, the fast-track procedure includes the same steps and players as the standard procedure but with stricter deadlines. Moreover, the EEA EFTA States decide whether an EU act is subject to the fast-track procedure or the standard procedure. To further increase the speed of EEA decision making, the EEA EFTA States may even waive the requirement to send out a standard sheet in certain policy areas. Figure 4 shows that approximately 50 percent of the EU acts adopted by the EU in the years 2015 to 2018 that later on were incorporated into the EEA Agreement, were incorporated by the fast-track procedure. In 2018 the EU institutions adopted 528 legal acts that were considered as EEA relevant and therefore transferred into EEA decision making: 211 EU acts were subject to the standard procedure, 53 to the simplified procedure and 264 to the fast-track procedure. The fast-track procedure can thus be seen as the most important procedure in terms of the number of EU acts incorporated into the EEA Agreement.
Electronic copy available at: https://ssrn.com/abstract=3712440 The application of the measures contained in an implementing act can be discontinued by each EEA EFTA State individually if an agreement on the formal incorporation into the EEA Agreement of the implementing act could not be reached in the EEA Joint Committee within twelve months of the entry into force of that implementing act. As a result, the other Contracting Parties to the EEA Agreement shall restrict or prohibit the free flow of personal data to an EFTA State that does not apply the measures contained in an implementing act. In other words, the simplified procedure for adequacy decisions deviates from the "single voice principle" in the EEA as the threat of suspension in case of nonincorporation applies not to all EEA EFTA States but to the EEA EFTA State unwilling or unable to incorporate the EU act into the EEA Agreement.
Considering the broad functional scope of the EEA and the different characteristics of EU secondary law, the need for different procedures is not surprising. Nevertheless, neither the simplified procedure in the veterinary field, the simplified procedure for adequacy decisions in the field of data protection nor the fast-track procedure were foreseen in the EEA Agreement, but were introduced to ensure the continued homogeneity in highly interdependent and sensitive areas and to increase the capacity of the EEA's institutional framework by avoiding time-consuming assessments of purely technical EU acts. The simplified procedure in the veterinary field thereby seriously constrains the EEA EFTA States' legislative sovereignty in the sense that there is an automatic policy transfer from the EU to the EEA. Likewise, the fast-track procedure reduces the function of EEA decision making to a purely formal process. These restrictions on the EEA EFTA States' decision-making autonomy are partly compensated Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers for by the EEA EFTA States' increased representation in EU policy making through their participation in the EU committees that adopt the EU acts subject to simplified and fast-track procedure. However, this participation is still without the right to vote, and thus cannot overcome the mismatch between the territorial scope of an EU act and the composition of its authors. The introduction of the new procedures can therefore be explained by the weighting of purposive qualities of legitimacy such efficiency and legal security over democratic qualities like autonomy, accountability and transparency.

Dynamic but not automatic
Decision making in the EEA is based on an already adopted EU act. Its expected outcome is the timely and completed incorporation of a new EEA-relevant EU act into the EEA Agreement. Nevertheless, EEA decision making remains a distinct process that can lead to different outcomes.
(1) The EEA EFTA States reject the incorporation of a new EU act into the EEA Agreement in the EEA Joint Committee. In this event the procedure set out in Article 102 of the EEA Agreement may be invoked. This stipulates that if a conciliation procedure has failed to resolve the disagreement, the parts of the EEA Agreement directly affected by the EU act in question are suspended.
(2) At least one of the three EEA EFTA States fails to ratify a JCD and thus the incorporation of an EU act into the EEA Agreement. Again, the procedure set out in Article 102 EEA may be invoked. (3) The EEA EFTA States and the EU agree on the exclusion of an EU act initially marked as EEA relevant from EEA decision making. This EU act will then no longer be treated as EEA relevant. (4) The incorporation of a new EEA-relevant EU act into the EEA Agreement can be delayed. Such a delay may result from different preferences between the EU and EEA EFTA States or between the EEA EFTA States themselves. In many cases, however, delayed incorporation is due to the high complexity of the EEA's institutional framework and is therefore inherent in EEA decision making. (5). The EEA EFTA States and the EU can adopt specific adaptations to a new EU act incorporated into the EEA Agreement. Such an adaptation 'implies that the relevant provisions of that act shall be read in a specific way for the purposes of the EEA Agreement' (EFTA Bulletin 2019: 38).
To date, the EEA EFTA States have never rejected the incorporation of a new EEA-relevant EU act. Likewise, the EEA EFTA States have never failed to fulfil the constitutional requirements that they indicated when incorporating an EU act into the EEA Agreement. By contrast, between 1994 and 2015, the EEA EFTA States excluded more than 1,250 EU acts from the EEA decision-making process. However, most of these were purely technical acts in the field of veterinary matters. Others had to be excluded because they were no longer applicable in the EU due to the lengthy EEA decision making. Only very few EU acts were excluded due to different assessments of the EEA relevance of an EU act by the EU and the EEA EFTA States (Frommelt 2017: 146-150).
Empirical analysis of the time required to incorporate new EU law into the EEA Agreement shows that this incorporation is often seriously delayed (Frommelt 2017). As a result, the EEA EFTA States and the EU Member States have had to comply with an EU act at the same time in less than 20 percent of the EU acts incorporated into the EEA Agreement. A very prominent example of a delayed incorporation is Directive 2008/6/EC with regard to the full accomplishment of the internal market of Community postal services (Third Postal Directive, 32008L0006). In April 2011, following a decision of the labour party's national meeting, the Norwegian Prime Minster Jens Stoltenberg, stated that the Norwegian government did not intend to incorporate the directive into the EEA Agreement (Norway 2012: 25). However, Directive 2008/6/EC was never officially excluded from incorporation and, thus far, the EU has not invoked Article 102 of the EEA Agreement. Indeed, shortly after its election in October 2013, the Norwegian government with Prime Minster Erna Solberg announced that it would 'lift Norway's reservation vis-a-vis the incorporation of the Third Postal Directive into the EEA Agreement' (European Commission 2013). Nonetheless, in June 2020, the incorporation of the Third Postal Directive was still pending.
Electronic copy available at: https://ssrn.com/abstract=3712440 Institutional Challenges for External Differentiated Integration: the Case of the EEA European University Institute 21 Figure 5 shows how the time required to incorporate an EU act into the EEA Agreement has changed over time and differs between the standard procedure and the fast-track procedure. It is calculated as the number of days between the adoption of an EU act by the EU and the date of the decision of the EEA Joint Committee to incorporate this EU act into the EEA Agreement (and thus not its entry into force). For EU acts adopted by the EU in 1994 the median time to incorporation was 1297 days. The high values for EU acts adopted between 1994 and 1997 can be explained by the fact that many EU acts related to Annex I to the EEA Agreement (veterinary and phytosanitary matters) were initially not treated as EEA relevant by the EEA EFTA States. As a result, Annex I was not amended for various years. After a review of this practice in 1998 the contracting parties decided to incorporate into the EEA Agreement all outstanding EU acts. Figure 5 confirms that the median time required for the incorporation of an EU act is much lower for EU acts incorporated by the fast-track procedure than for EU acts incorporated by the standard procedure. For EU acts adopted by the EU in the year 2018 the median time to incorporation was 158 days for EU acts subject to the fast-track procedure compared to 549 days for EU acts subject to the standard procedure. The median time for all EU acts was 218 days which is slightly lower than the median time to incorporation before the introduction of the fast-track procedure (245.5 days).

Figure 5: Time required to incorporate EU acts into the EEA Agreement
Source: author's own calculation based on the Eurlex database Empirical analysis also shows that the number of country-specific or EEA-specific exemptions is higher in the EFTA pillar of the EEA than in the EU pillar. For instance, between 1994 and 2015 nearly 8 percent of EEA-relevant EU acts were incorporated into the EEA Agreement with substantive amendments that changed the scope of application of the act within the EFTA pillar of the EEA. For instance, such adaptations exclude the applications of provisions of an EU act that do not fall within the scope of the EEA Agreement like provisions that govern the EU's relations with third countries (e. g. Directive 2014/59/EU, JCD 21/2018) or that define certain infringements as criminal offences (e. g. Directive 2005/35/EC, JCD 65/2009). Frommelt (2020a) also shows that there is higher supply of country-specific opt-outs in the EEA than in the EU. This applies in particular to the two small states Iceland and Liechtenstein while the demand for country-specific opt-outs by Norway is mostly only realised if such opt-outs were also provided to EU Member States, and thus are not EEA-specific optouts. Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers In summary, the empirical analysis of EEA decision making demonstrates that although there is dynamic incorporation of new EU law into the EEA Agreement, there is at least for the standard procedure not an automatic transfer of policies from the EU to the EEA. As EEA decision making is based on EU law, the EEA EFTA States will always be rule takers but there still is an EEA decision making on these rules that formally preserves the decision-making autonomy of the EEA EFTA States.

Threat of suspension
In the event of disagreement between the EU and the EEA EFTA States on the incorporation of a new EU act into the EEA Agreement or in case of a seriously delayed incorporation, the procedure set out in Article 102 EEA may be invoked. This stipulates that the parts of the EEA Agreement directly affected by the EU act in question are suspended if a conciliation procedure has failed to resolve the disagreement. However, the Contracting Parties are required 'to examine all further possibilities to maintain the good functioning' of the EEA (Article 102(4)). Moreover, Article 102(6) states that 'rights and obligations which individuals and economic operators have already acquired' under the EEA Agreement shall remain in the case of a suspension.
To date, Article 102 has only been formally invoked twice, and in both instances the six-month conciliation meant that a suspension was averted (see Frommelt 2017: 60-62). Taking into account that the EEA EFTA States are highly dependent on access to the single market, Pelkmans and Böhler (2013: 53) call the Article 102 procedure and the possible suspension of substantial parts of the EEA Agreement the 'nuclear option' with 'the practical effect of bypassing the formal maintenance of sovereignty (on EEA issues) for the EEA-3 countries [the EEA EFTA States] and forcing them to accept a de facto dependence on EU decision making'. The 'EU's material leverage stems from the attraction of its single market' (Lavenex 2014: 889) and the high dependence of the EEA EFTA States on the access to the single market.
However, the practical relevance of Article 102 may be overestimated. So far, the EU has sought to address disagreements on the incorporation of new EU acquis into the EEA Agreement through discussions (principally at a technical level) to convince the EEA EFTA States to take corrective action, although this may cause 'lengthy negotiations and unproductive situations of public political controversy' (European Commission 2012: 8-9). The reticence towards invoking Article 102 may therefore result from the EU's commitment to compromise and reach a consensus rather than resort to hard bargaining. Indeed, theories of international relations suggest that in the long run legitimacy is a much cheaper means to secure compliance than coercion (Lindblom 1977;Tallberg and Zürn 2019: 2). In addition, the EU's reticence to invoke the procedure to suspend parts of the EEA Agreement can be explained by the ambiguous wording of Article 102 (Baur 2016b: 74). For instance, Article 102 does not specify whether a suspension would apply to just the policy area in question or to the entire policy field. It is also unclear whether the Council (and hence the EU Member States) or the EEAS is competent to invoke Article 102. Finally, the EU's reticence to invoke Article 102 could be interpreted as the result of the little relevance of the EEA EFTA States for the overall functioning of the EU's single market (Frommelt 2019).
To sum up, the procedure foreseen in Article 102 remains an important 'deterrent' in order to prevent cherry-picking and free-riding by the EEA EFTA States (European Commission 2012: 9). Due to its ambiguity, however, it has so far not seriously limited the autonomy of the EEA EFTA States in EEA decision making. As long as the EEA EFTA States do not oppose the incorporation of an EU act in principle, there is not much that the EU can do against a delayed incorporation. This might also be explained by the fact that in the past 25 years, the EEA EFTA States cleverly highlighted their overall commitment to the good functioning of the EEA whenever they feared political pressure from the EU regarding the incorporation of a specific EU act. Frommelt (2017: 238) observes an inherent conflict between legitimacy deriving from the EEA's output as a homogenous and dynamic economic area, and legitimacy shaped by the input of the EEA EFTA States' domestic political institutions. In other words, any political assessment of new EEA-relevant EU legislation by the national parliaments or business associations of the EEA EFTA States is likely to delay the incorporation of those EU acts into the EEA Agreement. This means that an increase in the input legitimacy by more participative and elaborate EEA decision making is likely to cause temporarily differentiated rules for the EEA EFTA States and the EU States and thus a decrease in the EEA's output legitimacy.

Parliamentary procedure
According to Article 103 of the EEA Agreement, a JCD can be binding on a Contracting Party only after the fulfilment of constitutional requirements. In practice, this means that the EEA EFTA States can indicate constitutional requirements to a JCD, which then requires formal ratification by the national parliament of the respective EEA EFTA State. Members of parliament of an EEA EFTA State, however, can only approve a JCD. They cannot make any changes or adaptations, neither to the JCD nor the EU act itself.
Each EEA EFTA State has its own criteria to follow before launching a procedure in accordance with Article 103. In Liechtenstein, for instance, the EEA/Schengen Committee of the Liechtenstein parliament (EWR/Schengen-Kommission, LGBl. 2013.009, Article 69) meets a few days prior to the meeting of the EEA Joint Committee in order to decide whether a procedure in accordance with Article 103 is necessary for a specific EU act. The parliamentary EEA/Schengen Committee makes its decision based on a recommendation from the EEA Coordination Unit of the Liechtenstein government. The recommendation itself is based on an expert opinion of Liechtenstein's Constitutional Court (StGH 1995/14, Liechtensteinische Entscheidungssammlung (LES) 1996. Parliament's approval of a JCD is only necessary if the corresponding JCD changes domestic law or has financial consequences. By contrast, delegated and implementing EU acts, EU acts with a specific addressee, non-binding EU acts and EU acts that simply amend or consolidate other EU acts do not require approval by the Liechtenstein parliament.
The Constitutional Court also stated that the Committee's decision has to reflect the efficiency and proper working of the EEA Agreement. Finally, it highlighted the necessity of close cooperation between government and parliament to ensure the democratic legitimacy of the EEA policy-making process. These specifications give all players certain room for interpretation as to whether the stipulation of constitutional requirements is necessary or not. If Liechtenstein has pointed out the need to fulfil its own constitutional requirements to a JCD, parliament has to approve the respective JCD in a plenary meeting.
Iceland and Norway apply similar criteria when constitutional requirements are stipulated. However, whereas the Liechtenstein parliament is mostly involved only a short time before a JCD is finally adopted, the parliaments of Iceland (Althingi) and Norway (Storting) may also be involved at an earlier stage. In 2010, Althingi revised its rules on the parliamentary procedure for matters concerning EEA in order to strengthen the parliamentary impact on EEA law. According to the new rules any EU acts of which the implementation into Icelandic law requires amendments to statutory laws 'must be submitted for review three times to the parliament before being finally incorporated into the EEA Agreement' (Einarsdóttir 2019: 11): First, such an EU act has to be submitted to the Foreign Affairs Committee of Althingi when the EEA EFTA States are assessing its EEA relevance and its need for adaptations. The Committee can then involve other parliamentary committees such as the Constitutional and Supervisory Committee. Second, an EU act has to be submitted to the Committee when the draft of the JCD is finalised. Before each meeting of the EEA Joint Committee, the Minister of Foreign Affairs therefore presents the Foreign Affairs Committee the EU acts to be adopted by the EEA Joint Committee. Third, an EU act has to be submitted to Althingi after the JCD has been adopted in order to request Althingi to lift the constitutional requirements.
Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers The first phase of consultation in particular led to serious delays in the incorporation of new EU legislation into the EEA Agreement. As a result, the rules introduced in 2010 were revised again in 2018. However, the rules still foresee that the Foreign Affairs Committee is consulted at the preparation stage of a JCD.
Between 1994 and 2019, constitutional requirements were indicated for 546 JCDs, which is 12 percent of all JCDs adopted. Iceland indicated 331 constitutional requirements, Liechtenstein 240 and Norway 317. On average it takes 283 days until Althingi lifts a constitutional requirement, compared to 179 days in Liechtenstein and 186 days in Norway. The fact that 450 days after the adoption of a JCD, ratification of 25 percent of the JCD to which Iceland has indicated constitutional requirements is still pending, shows that Article 103 of the EEA Agreement can again seriously delay the incorporation of new EU legislation into the EEA Agreement. This confirms the inherent conflict of the EEA between the EEA EFTA States' political wish and constitutional obligation to protect their legislative autonomy on the one hand, and the need to ensure efficient administration of the EEA Agreement in order to preserve the homogeneity of EEA law on the other hand.
In summary, the analysis of EEA decision making has shown that new procedures had to be introduced within the EFTA pillar, but also between the EFTA and EU pillars, in order to cope with the legislative dynamics of the EU. Those new procedures do not necessarily deprive the EEA EFTA States of their final legislative powers, but still constrain their decision-making autonomy by establishing more efficient and mostly also hierarchical procedures for EEA decision making. Indeed, efficiency has turned out to be the main challenge of EEA decision making. The very fact that a specific EEA decision-making process exists makes it difficult to ensure that a new EEA-relevant EU act enters into force simultaneously for the EU Member States and the EEA EFTA States. If the EEA relevance of such an EU act has to be assessed first, or if the EEA EFTA States and the EU have to negotiate on EEA-specific adaptations before the EU act can be incorporated into the EEA Agreement, it seems inevitable that incorporation will be delayed. Delayed incorporation can endanger the integrity of EU law if the EEA EFTA States benefit from access to the EU's internal market even though they have not yet incorporated a new EU legal act and therefore are not legally bound by this act. The practice of EEA decision making thus shows that the requirement for the dynamic incorporation of new EU law led to a certain softening of the basic principles of the EEA on both sides. While the EEA EFTA States had to accept restrictions on their decision-making autonomy in order to make the EEA Agreement more efficient, they were also able to exploit the complexity of the EEA decision-making procedures to deliberately delay the incorporation of new EU law, which ultimately undermined the integrity of EU lawat least on a selective and temporary basis.
In the next section, the institutional challenges of the EEA will be assessed based on the EEA EFTA States' participation in EU agencies.

Participation in EU agencies
A special case of EEA decision making can be seen in the EU agencies and other EU bodies that can take legally binding decisions of EEA relevance. These institutions are separate legal entities from the EU institutions and carry out certain tasks under EU law. Many of these tasks are also EEA relevant, and therefore the founding legal EU act establishing these bodies should be incorporated into the EEA Agreement. As there were hardly any EU agencies in place at the time of conclusion of the EEA Agreement, not a single provision governing the relationship between the EEA EFTA States and EU agencies was inserted into the Agreement. Accordingly, there are also no institutions in the EFTA pillar that correspond to the EU agencies. The legal framework of an EU agency for the purpose of the EEA Agreement must therefore be regulated in the JCD incorporating the founding act of that agency into the EEA Agreement. This chapter first introduces four different models of institutional arrangements for the EEA EFTA States' participation in EU agencies, and then presents some empirical cases for each model.

The four models of participation
At the time of writing, the EEA EFTA States participate in 17 EU agencies and other decentralised EU bodies on the basis of the EEA Agreement. The conditions for the participation of the EEA EFTA States differ for each EU body, as do the powers of these institutions vis-à-vis the EEA EFTA States. A challenge for the two-pillar structure occurs in particular when these EU agencies or decentralised bodies have the competence to take binding decisions. According to the logic of the two-pillar structure, this competence vis-à-vis the EEA EFTA States must be located within the EFTA pillar. Also, the judicial review of such binding decisions must be reflected in the EFTA pillar. However, this might undermine the homogeneity of the EEA, as EU agencies are characterised by a particularly high level of expertise and degree of independence. Hence, creating an equivalent to an EU agency in the EFTA pillar would require a lot of resources and still it would be highly unlikely that such an EFTA body would gain the same authority as the EU body.
By focusing on EU agencies or decentralised bodies that have the competence to take binding decisions, this paper distinguishes four different models for the allocation of power to enact legally binding decisions of an EU body in the EEA (see Bekkedal 2019 for a very similar classification): 1) The decision-making power of an EU body is delegated to the national authorities of the EEA EFTA States; 2) decision-making power is delegated to EFTA Standing Committee; 3) decision-making power is delegated to ESA; and 4) decision-making power remains at the EU body. In a nutshell, the decisionmaking power of the supranational EU body can for the purpose of the EEA EFTA States be exercised by a national body, an EFTA body or an EU body. The difference between models two and three is that the EFTA Standing Committee is a purely intergovernmental body, whereas ESA is a quasisupranational body.
Other conceivable options would be that the EEA EFTA States establish an independent agency for the EFTA pillar, or that they participate in the EU agency on the basis of a separate international agreement beyond the EEA Agreement. Since these two models have not been taken into consideration thus far, they will not be analysed in this paper.
The four models have a different impact on the legislative autonomy of the EEA EFTA States. The formal responsibility for taking binding decisions is, however, only one side of the coin. It is also important to consider whether the decision-making power is conditional in the sense that there are formal restrictions in place on the decision-making power of a national authority or ESA. Again, there are at least four different ways to condition decision-making power in the EFTA pillar: 1) The body taking binding decisions addressed to the EEA EFTA States has to enact such decisions within a clearly defined and often very tight timeframe, starting with the corresponding decision of the EU body; 2) decisions addressed to the EEA EFTA States have to be adopted on the basis of drafts prepared by the EU agency at its own initiative; 3) the power to taking binding decisions addressed to the EEA EFTA States is enacted under the surveillance mechanism of the EEA giving ESA the possibility to initiate an infringement procedure if an EEA EFTA State breaches EEA law; and 4) in the case of disagreement between the bodies of the EFTA and EU pillars on a specific decision, the EEA dispute settlement mechanism can be invoked. It is possible that all conditions are in place at the same time, reducing the formal autonomy of the EEA EFTA States and their national authorities to a rather symbolic feature.
From the perspective of the EU, one can argue that the more the decision making vis-à-vis the EEA EFTA States is tied to an EU body, the more likely it is that the integrity of EU law will be preserved. By contrast, from the perspective of the EEA EFTA States, their representation in EU policy making becomes more important the more an EU body constrains their decision-making power. The EEA EFTA States' access to the policy making of an EU agency mostly follows the same template by including full participation in all committees without the right to vote (see above). It is therefore not necessary to describe these arrangements in detail. The following sections thus focus on the four different models of decision making. Robert Schuman Centre for Advanced Studies Working Papers

Models within the two-pillar structure
An example for the delegation of decision-making power to the national authorities of the EEA EFTA States (model 1) is provided by Regulation (EC) No 726/2004, which sets out Community procedures for the authorisation and supervision of medicinal products for human and veterinary use and for the establishment of a European Medicines Agency (EMA). Regulation (EC) No 726/2004 was incorporated into the EEA Agreement by JCD 61/2009, which included an EEA-specific adaptation stipulating that the EEA EFTA States must adopt corresponding decisions on approval of medicinal products 'simultaneously and within 30 days of the adoption of the decision' by the EU. In this way, the EEA EFTA States were able to avoid a formal transfer of their decision-making power to the EU. Since the decision-making power of the national authorities of the EEA EFTA States is linked to the decision of the competent EU body and constrained by a clearly defined period, material homogeneity is likely to be maintained. The adaptation text also states that the EEA Joint Committee shall be informed about such decisions and shall periodically publish lists of such decisions in the EEA Supplement to the Official Journal of the European Union. In addition, ESA shall monitor implementation of the decision and apply mutatis mutandis in the event of 'disagreement between the Contracting Parties on the interpretation of these provisions [...] in the sense of Part VII of the Agreement', which means that the affected parts of the EEA Agreement could be suspended if no agreement is reached. Similar arrangements have been made for the authorisation of chemicals (REACH, Regulation (EC) No 1907/2006JCD 25/2008) and also for novel foods (Regulation (EC) No 285/97; JCD 147/2015).
To date, there is no empirical evidence of the delegation of the decision-making power of an EU agency to the EFTA Standing Committee (model 2). However, interview evidences suggest that this model has been taken into consideration when negotiating the EEA EFTA States' participation in EU agencies. Indeed, there are various EU acts where the competences of the European Commission are carried out by the EFTA Standing Committee. Moreover, the adaptations text for Regulation (EC) No 1907/2006 and some other regulations states that 'as regards the EFTA States, the Agency shall, as and when appropriate, assist the EFTA Surveillance Authority or the Standing Committee, as the case may be, in the performance of their respective tasks' (JCD 25/2008).
A very prominent example of the difficulties in incorporating EU agencies into the EEA Agreement is the EEA EFTA States' participation in the European System of Financial Supervision (ESFS), which is composed of the European Systemic Risk Board (ESRB) and the three European Supervisory Authorities: the European Banking Authority (EBA), the European Insurance and Occupational Pensions Authority (EIOPA) and the European Securities and Markets Authority (ESMA). These authorities have the competence to adopt individual binding decisions addressed to competent national authorities and financial market institutions or participants. The European Supervisory Authorities are empowered to: (i) take decisions to temporarily restrict or even prohibit certain financial market activities if they jeopardise the orderly functioning of the EU financial market; (ii) take decisions requiring financial market participants to implement and comply with the relevant EU requirements; (iii) take emergency measures; and (iv) settle disagreements between competent authorities.
In October 2014, the Finance Ministers of the three EEA EFTA States and the EU Member States agreed on the broad lines of incorporation of the European Supervisory Authorities into the EEA Agreement. In September 2016, after more than five years of intensive negotiations, the Contracting Parties finally incorporated the founding acts of the European Supervisory Authorities into the EEA Agreement. However, this required several EEA-specific adaptations (more than 50 for each authority; see JCDs 199/2016;200/2016;201/2016) and an amendment to the Surveillance and Court Agreement (SCA). These adaptations and amendments provide that all decisions in the framework of the European Supervisory Authorities shall be taken by ESA on the basis of drafts prepared by the relevant authority on its own initiative or at the request of ESA (model 3). In other words, although ESA will act in complete independence according to the adaptation text, it will in fact only 'rubber-stamp' the draft decisions submitted to it by the relevant European Supervisory Authority (see Fredriksen and Franklin 2015: 679). This "docking" of the EFTA pillar to the EU was thought to be necessary to counteract the Electronic copy available at: https://ssrn.com/abstract=3712440 A similar solution was agreed when Regulation (EU) No 713/2009 establishing an Agency for the Cooperation of Energy Regulators (ACER) was incorporated into the EEA Agreement in 2017 and entered into force in September 2019, after the Icelandic parliament finally ratified the JCD (Reuters 2019). Here too, ESA was given the competence to adopt a decision addressed to the national authorities of the EEA EFTA States based on a draft prepared by ACER (JCD 93/2017). Moreover, the adaptation text explicitly states that such a decision can be challenged before the EFTA Court. It should be noted that such a clause is not included in the JCDs for the European Supervisory Authorities, which in the end makes no difference as the right of appeal to the EFTA Court is already enshrined in the SCA. However, in contrast to the European Supervisory Authorities, the adaptation texts for ACER give the national authorities of the EEA EFTA States the option to request that ESA reconsider its decision. ESA shall forward such a request to the EU agency, which 'shall consider preparing a new draft for the EFTA Surveillance Authority and reply without undue delay' (JCD 93/2017). In the case of a disagreement between ESA and the EU agency, the matter can be referred to the EEA Joint Committee to deal with it in accordance with Article 111 of the EEA Agreement, which foresees that the contracting parties can request the CJEU to give a final ruling.

One-pillar model
No two-pillar solution was agreed for the incorporation of the General Data Protection Regulation (EU) 2016/679 (GDPR), but rather a one-pillar approach (model 4). By JCD 154/2018, the EEA EFTA States and the EU agreed that the European Data Protection Board (EDPB) can also issue binding decisions vis-à-vis the national supervisory authorities of the EEA EFTA States. In addition, this 'assimilation model' (Bekkedal 2019: 408)which actually requires subordination rather than just assimilationestablishes a somewhat simplified procedure for the implementation of adequacy decisions enacted by the European Commission (see above). In a joint declaration to the JCD, the Contracting Parties declare that the procedure agreed for the incorporation of the GDPR does not create a precedent for future adaptations of EU acts. This may be true from a legal perspective, but political pressure may lead to the opposite direction being taken in the future.
Already in the case of Regulation (EC) No 1592/2002 establishing common rules in the field of civil aviation and establishing a European Aviation Safety Agency (EASA), the EEA-specific adaptation of JCD 179/2004 gives the EU agency the competence to take decisions that are directly applicable throughout the EEA. This model ensures the homogeneity of EEA law but bypasses the decision-making powers of the EEA EFTA States and EFTA institutions. Such a deviation from the two-pillar structure of the EEA was only acceptable to the EEA EFTA States because decisions by EASA are of a very technical nature and require a high level of expertise. The model also applies to Regulation ( The question of whether the EU or EFTA pillar has the power to take legally binding decisions is closely related to the question as to where judicial review of those decisions takes place. In the accompanying documents to Regulation (EU) No 2016/679 submitted to parliament, the Liechtenstein government (2018: 16-17) notes that any natural or legal person in the EEA EFTA States has the right to bring an action for annulment of a decision of the EDPB before the ECJ. However, if a court in an EEA EFTA State has to interpret Regulation (EU) No 2016/679, it still has to request an advisory opinion from the EFTA Court and not the ECJ. The fact that the EFTA Court can be asked to assess the Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers possible annulment of a decision made by an EU body places it in the difficult position of having to assess the possible outcome of a hypothetical action for annulment before the ECJ. Referring to a similar case regarding the EMA, Fredriksen and Franklin (2015: 677) note an 'apparent lack of effective judicial protection' as long as the EEA EFTA States 'have no way to get the ECJ to rule on the legality of the underlying EU decision'.
This chapter has shown that the EEA EFTA States' decision-making power is seriously constrained when participating in an EU agency that can take legally binding decisions. This raises the question of whether the EEA EFTA States should be fully represented in the policy making of those agencies (see above). As mentioned above, this has not been possible thus far but might be worth considering in an increasingly differentiated Europe. On the other hand, this chapter has also shown that even in a onepillar model like the EEA EFTA States' participation in the EDPB, some specific institutional arrangements remain in place to take account of the EEA EFTA States' status as non-Member States and their reservations about political integration. Such arrangements may justify why the EEA EFTA States were not given real decision-making power within the EU agency.
Given the different approaches so far, it remains to be seen how the relationship between the EEA EFTA States and EU bodies will be resolved in the future. At present, EU legal acts concerning the new European Labour Authority (Regulation (EU) 2019/1149), the European Union Agency for Railways (Regulation (EU) 2016/796) and the BEREC Regulation (Regulation (EU) 2018/1971) are to be incorporated into the EEA Agreement.
The EEA EFTA States' participation in EU agencies shows again that the institutional framework of the EEA was not sufficient to ensure the proper functioning of the EEA Agreement. The new institutional arrangements have in common that they give priority to the efficacy of the EEA over the decision-making autonomy of the EEA EFTA States, while the decision-making autonomy of the EEA EFTA States is particularly constrained if the institutional arrangements deviate from the EEA's twopillar structure by leaving the competence for legally binding decisions vis-à-vis the EEA EFTA States within the EU pillar. As there is no template for allocating the decision-making power of decentralised EU bodies in the EEA's two-pillar structure, the institutional arrangements have to be negotiated separately for each upcoming EU body. This piecemeal approach begets the risk of increasing complexity and inconsistency. The following chapter therefore analyses how the nature of the EEA Agreement as an initially purely intergovernmental agreement has changed over time.

Modes of governance in the EEA
In EU studies, Lindberg and Scheingold (1970) were the first scholars to provide a systematic categorisation of when and how the EU can claim authoritative decision-making power. Their work has been updated by several scholars (e.g. Börzel 2005;Leuffen et al. 2013: 11). The main aim of such a categorisation is to measure the extent to which Member States have pooled their sovereignty in a given policy area, for instance by moving from unanimous to majority voting or by delegating their decisionmaking authority to supranational institutions, such as the European Commission and the EP, and involving the jurisdiction of the CJEU.
Thus far, no systematic categorisation of decision making in the EEA has been produced. This can be explained by the fact that the EFTA States opposed political integration. As a result, the EEA Agreement was designed as an agreement under public international law. This means that the decision making foreseen in the main part of the EEA Agreement is strictly intergovernmental and there are no formal requirements for the EEA EFTA States to pool their sovereignty. Since the main part of the EEA Agreement has not been updated so far, the basic principles of the EEA Agreement are also supposed to have stayed the same. However, due to the various rules on EEA decision making that have been anchored in the Annexes to the EEA Agreement when incorporating new EU acts with specific Electronic copy available at: https://ssrn.com/abstract=3712440 The decision-making autonomy of the EEA EFTA States, and thus their degree of political integration, varies across the different policies of the EEA and ranges from full autonomy to full delegation to EU institutions. The decision-making autonomy of the EEA EFTA States remains unaffected if decision making in the EEA is based on intergovernmental coordination between the EEA EFTA States. Such cooperation takes place in the EEA's flanking policies. The EEA EFTA States' participation in an EU programme covered by the EEA's functional scope is mostly based on a JCD to incorporate the EU act setting up a new EU programme into the EEA Agreement. Each of the three EEA EFTA States can veto the incorporation of such an EU act. However, the special quality of decision making in the flanking policies is that the EEA EFTA States do not have to 'speak with a single voice'. Hence, if an EEA EFTA State opts out of an EU programme incorporated into the EEA Agreement, this does not trigger the Article 102 procedure. For instance, although Liechtenstein's parliament rejected the principality's participation in the Horizon 2020 Framework Programme for Research and Innovation, this did not prevent the incorporation of the relevant EU act into the EEA Agreement; the Contracting Parties simply added an opt-out clause for Liechtenstein to JCD 109/2014. By contrast, according to the standard EEA procedure, any adaptations or opt-outs require the agreement of all the EEA EFTA States as well as that of the EU. Considering the economic dependence of the EEA EFTA States on non-discriminatory access to the EU's internal market, the threat of restricted market access puts the intergovernmental coordination between the EFTA and EU pillars under a "shadow of hierarchy" and makes it legally and political constrained. Although the EEA EFTA States have not formally pooled their sovereignty, in the daily administration of the EEA Agreement they face considerable pressure to align with EU decisions. This pressure is even higher as the EEA EFTA States are forced to speak to the EU with a single voice.
To monitor and enforce the EEA Agreement, ESA is competent to adopt certain decisions. These decisions can refer to the notification of specific norms and technical standards or professional diplomas, but can also include the competence to issue fines or legally binding decisions in emergency situations. When considering the policy-making power of ESA, two modes of governance need to be distinguished. First, there is "quasi-supranational centralisation" within the EFTA pillar, meaning that ESA has exclusive decision-making power over the EEA EFTA States and only limited coordination with the European Commission. Second, there is quasi-supranational centralisation across the EU and the EFTA pillars, where ESA makes its decisions based on draft proposals from the European Commission.
This distinction between centralisation within the EFTA pillar and centralisation across the EFTA and EU pillars is important. In the first case, any transfer of decision-making autonomy is legitimised by the fact that it is transferred from the EEA EFTA States to a body that consists of national representatives of the EEA EFTA States. By contrast, in the second case, the decision-making autonomy of ESA is constrained by the obligation to cooperate closely with the European Commission and/or other EU bodies. This means that the EEA EFTA States are not fully represented in the body drafting the decisions that are legally binding on them. This is the case, for instance, regarding the European Supervisory Authorities, where ESA takes decisions based on drafts prepared by the respective authority.
The governance of the EEA can also deviate fully from the EEA's two-pillar structure. In the field of aviation security, EASA has the exclusive competence to make decisions that directly apply throughout the EEA. Put differently, the EEA EFTA States have fully delegated their decision-making power to a supranational EU body. This is also the case for the EDPB established by Regulation (EU) No 2016/679 (see above). Hence, decisions of the EDPB can have a direct and binding effect on national supervisory authorities in the EEA EFTA States. Table 1 summarises the different decision-making mechanisms in the EEA.
Electronic copy available at: https://ssrn.com/abstract=3712440 Robert Schuman Centre for Advanced Studies Working Papers The myriad of decision-making rules in the EEA results from the need for closer interaction between the EU and the EEA EFTA pillar and the commitment to a homogenous and dynamic economic area. It also reflects the advanced and often very technical integration within the policies covered by the EEA Agreement. Arguably, the evolution of the different governance modes of the EEA mainly ensures the proper dynamic functioning of the EEA. On the face of it, the deepening of the EEA is thus based on a functionalist logic that results from the high level of integration of the EEA EFTA states and their desire to maintain access to the EU's internal market. However, the deepening of the EFTA pillar of the EEA can also be driven by the EEA EFTA bodies, domestic stakeholders, and the EEA EFTA States themselves. For instance, in the context of aviation safety, the deepening of the EEA can be explained by the common interest of the EEA EFTA States' governments in having safe aircrafts. The EEA EFTA States have been unwilling and unable to provide the necessary expertise to assume the tasks of the European Aviation Safety Agency. In the case of the European Supervisory Authorities the deepening is mainly the result of the active engagement of companies and entrepreneurs in the financial services sector who feared the potential loss of their access to the EU market for financial services if EU acts relating to the European Supervisory Authorities were not incorporated into the EEA Agreement. For this reason, commercial stakeholders lobbied the EEA EFTA governments to ensure swift incorporation based on a pragmatic institutional solution.
Another approach to explain the institutional dynamics of the EEA is provided by the theory of historical institutionalism (Gstöhl and Phinnemore 2019b). The EEA EFTA states are locked into the EEA's two-pillar structure because the price of exiting the EEA Agreement would be very high due to a lack of alternative models of integration. Path dependence (Pierson 1996) thus works in favour of integration within the EEA's two-pillar structure even if this integration has unintended consequences such as an increasing delegation of decision-making power to EFTA and EU bodies. This deepening is supported by the relative power asymmetry with the EU resulting from the EEA EFTA States' small size and their dependence on access to the internal market.
Further research would be necessary to quantify the evolution of the competences of the EU and EFTA bodies in relation to the EEA EFTA States and compare this with the initial reservations of the EEA EFTA States regarding such transfer of competences. In the absence of research in this area, this chapter explicitly refrains from characterising the EEA Agreement as "intergovernmental" or "supranational". Instead, it highlights the ambiguity and complexity of the extent of centralisation in the EEA, as well as the persistence of different policy-specific modes of governance that range from strictly intergovernmental cooperation to subordination to supranational EU bodies. Moreover, it shows that the extent of transfer of state power in the EFTA pillar but also from the EFTA to the EU pillar of the EEA has increased over time. As a result, there is no doubt that the political integration of the EEA Agreement goes far beyond what is usual for an agreement under public international law and also far beyond what was initially intended by the EEA EFTA States.

Discussion
External differentiated integration allows the projection of EU legal rules beyond formal EU membership providing a legal order not only for the EU but for the European region (Schimmelfennig and Winzen 2020: 179). One of the significant challenges of external differentiated integration is the lack of congruence of rule makers and rule takers (Eriksen 2018;Schimmelfennig and Winzen 2020). The EEA is no exception here. It requires the continuous incorporation of new EEA-relevant EU rules into the EEA Agreement while the EEA EFTA States' participation in EU policy making is mutually constrained. This paper has shown, however, that the classification of the EEA EFTA States as rule takers does not fully capture the complexity of the institutional arrangements of the EEA and the democratic challenges resulting from these arrangements.
The EEA Agreement is based on EU law, and there is no common decision making on EEA-relevant EU legislation between the EEA EFTA States and the EU. However, empirical analysis of standard EEA decision making shows that the EEA EFTA States did not completely cede their legislative sovereignty, and there is also no overshadowing dominance by the EU. The standard procedure of EEA decision making gives the EEA EFTA States the possibility at least to delay the incorporation of new EU rules into the EEA Agreement. Iceland and Liechtenstein in particular also benefit from a substantial number of opt-outs, which were mostly agreed when incorporating those EU acts into the EEA Agreement. Finally, there are various EEA-specific adaptations to EU acts incorporated into the EEA Agreement that take into account the EEA's specific institutional framework and limited functional scope. By contrast, at least thus far, the EU has been rather cautious in using its superior bargaining power to prevent such delays. One explanation for this can be that the suspension mechanism foreseen in Article 102 of the EEA Agreement is likely to be impractical to invoke on a regular basis. Another explanation is that, overall, the functioning of the EEA has always been good, and there was also a credible commitment by the EEA EFTA States to the EEA and its principles. Hence, it would not have been in the interest of the EU to endanger the proper functioning of the EEA and its good relations with the EEA EFTA States just because of some single issues.
While the standard procedure for EEA decision making has given the EEA EFTA States a surprisingly large amount of room for manoeuvre, with the simplified procedure in the veterinary field, the simplified procedure for adequacy decisions in the field of data protection and the fast-track procedure for EU acts without horizontal challenges, the EEA EFTA States have introduced a more or less automatic rule transfer from the EU to the EEA. The same applies to the EEA EFTA States' participation in EU agencies that can take legally binding decisions vis-à-vis the participating states.
The introduction of such institutional arrangements was often controversial and subject to lengthy negotiations. The underlying conflict was thereby the same as in the negotiations of the EEA Agreement back in the 1990s. While the EU wants to preserve the integrity of its legal order and decision making, the EEA EFTA States do not want to cede legislative power to the European institutions. As long as this conflict continues, EU legal acts, which not only regulate policy issues but also assign certain competences to the EU institutions, are likely to trigger new institutional challenges for the EEA. We can therefore argue that the negotiations on the institutional design of external differentiated integration are never finished due to the dynamics of EU law.
The new institutional arrangements have disclosed the limits of the EEA's two-pillar structure and its nature as an agreement under public international law. This has also raised concerns about the compatibility of the institutional arrangements agreed with the EU with the national constitutions of Iceland and Norway. In both Iceland and Norway, the national constitution constrains the transfer of state power to an international organisation. In the case of Iceland, such a transfer is only possible 'if it is delimited, well defined and not overly onerous to any Icelandic parties' (Einarsdóttir 2019 based on the report of the committee appointed by the Minister of Foreign Affairs on the Constitution and the EEA Agreement, 6 July 1992). Due to this principle, the incorporation of the founding acts of the Robert Schuman Centre for Advanced Studies Working Papers European Supervisory Authorities as well as other EU agencies was highly controversial among both politicians and lawyers in Iceland (Einarsdóttir 2019).
In Norway, any arrangements must be in conformity with the requirements that flow from Article 115 of the Norwegian Constitution, according to which the transfer of legislative power to organisations of which Norway is not a member is not possible (Bekkedal 2019;Holmøyvik 2015). Since the EU insists on the autonomy of its decision making, and therefore does not allow full participation by the EEA EFTA States, this could be seen as a legal limit of external differentiated integration rooted in the domestic level. However, based on a three-fourths majority by the national parliament, Norway can transfer competences to an EFTA body that can be vested with identical powers to the EU body. This solution has already been applied by the Norwegian government to ensure the incorporation of the European Supervisory Authorities into the EEA Agreement. The handling of those constitutional restrictions in Iceland and Norway remains inconsistent (Holmøyvik 2015), and has therefore not prevented the EEA EFTA States' subordination to EU bodies such as EASA or the EDPB. Indeed, thus far, the EEA EFTA States have shown a remarkable level of pragmatism and flexibility in order to maintain their level of functional integration. Nonetheless, the EEA EFTA States' participation in EU agencies shows that, in particular in dualistic countries, external differentiated integration can also be constrained by national constitutional law. This paper has also shown that EEA decision making and the EEA EFTA States' access to EU policy making are interdependent. The more constrained the decision-making autonomy of the EEA EFTA States, the more important their involvement in EU policy making. If we simply focus on the efficacy of the EEA, it is without doubt that far-reaching representation of the EEA EFTA States in EU policy making is favourable, as it is likely to reduce the number of procedural steps and the actors involved in EEA decision making.
While the division of decision-making power between the two pillars of the EEA Agreement is often highly controversial, negotiations on the institutional arrangements for access for the EEA EFTA States are usually rather straightforward; at least if the EU does not define additional restrictions for the access to its institutions in the respective EU act. This can be explained by the fact that the EU has always been consistent in protecting the autonomy of its decision making, which ruled out the right to vote for the EEA EFTA States from the outset of any negotiations. As a result, equal participation by the EEA EFTA States and EU Member States in EU bodies is not possible. Therefore, the involvement of EEA EFTA States in EU policy making cannot fully legitimise the dynamic incorporation of new EEA-relevant EU legislation into the EEA Agreement.
To become rule makers instead of rule takers, full participation by the EEA EFTA States in EU policy making would be particularly important in EU agencies, as well as in the EU committees preparing the EU acts to be incorporated into the EEA Agreement by simplified procedures. However, this would also require that the EEA EFTA States accept the pooling of sovereignty that comes with decision making that is no longer based on unanimity. It is unclear whether the EEA EFTA States would indeed be willing to agree to this kind of integration, as it would bring the EEA EFTA States very close to EU membership.
Schimmelfennig and Winzen (2020: 182) argue that EEA EFTA States 'certainly have to pay the democratic costs of being decision-takers in the EU's market and flaking policies' but also highlight that 'this arrangement has not been imposed but negotiated in response to the democratic choice to avoid full membership'. With regard to the newly introduced institutional arrangements described in this paper this might not be fully correct as those arrangements were indeed more or less imposed by the EU as they were the only option for the EEA EFTA States if they wished to maintain their existing level of integration. However, this does not change the fact that the EEA EFTA States could opt for an EU membership in order to avoid the institutional constrains of the EEA.
On the other hand, this paper has also shown that the EEA EFTA States' limited access to EU policy making makes the administration of external differentiated integration more complex and may even threaten the integrity of EU law if the incorporation of new EEA-relevant EU legislation is delayed, or Electronic copy available at: https://ssrn.com/abstract=3712440 if decisions taken by the national authorities of the EEA EFTA States deviate from the guidelines of an EU agency. It is likely that for most EU agencies the participation of the EEA EFTA States would have been agreed much faster by combining the requirement of full participation in EU agency with the offer of full voting rights. Likewise, the legality would be higher and the democratic cost for the EEA EFTA States lower. This raises the question of the extent to which the EU should enable external differentiated

Conclusions
The EEA Agreement is the most far-reaching and institutionalised agreement between the EU and a non-Member State. It is thus often seen as 'the institutional benchmark' for a privileged partnership with the EU (Baur 2019: 27). Its main features are the so-called 'two-pillar structure' for institutional integration and the 'homogeneity principle' for substantive functional integration. Due to its dynamic nature, the EEA also requires the continuous incorporation of new EEA-relevant EU legislation into the EEA Agreement.
Since the EEA is an institutional benchmark for external differentiated integration the EEA is well placed to analyse the institutional challenges that come with external differentiated integration. This paper has shown that the initial institutional integration of the EEA was not sufficient to ensure its intended functional integration. To cope with the high legislative dynamics of the EU, new procedures for EEA decision making were necessary. Their main purpose was no longer to ensure the EEA EFTA States' autonomy but to increase the EEA's efficiency, and thus to maintain its proper functioning. Likewise, new institutional arrangements had to be found to allow the EEA EFTA States' participation in EU bodies such as agencies that were not foreseen by the EEA's two-pillar structure. These arrangements may deviate from the two-pillar structure by placing the EEA EFTA States directly under the influence of the EU pillar. They also change the nature of the EEA Agreement, which can therefore no longer be regarded as a fully intergovernmental agreement.
Undoubtedly, the new arrangements and procedures have improved the relevant processes in the EEA. Overall, however, they have increased the complexity of the EEA. According to Zahariadis (2013: 812) institutional complexity such as that which is found in the EEA can have four implications: First, it raises administrative costs; second, it generates more complexity; third, it gives rise to political conflict; and fourth, it safeguards diversity. In the case of the EEA, not all of these implications can be seen since the overall functioning of the EEA is still good. Nevertheless, due to its complex institutional setup, the EEA cannot be seen as an ideal model for external differentiated integration (Sverdrup 2011). Indeed, the current negotiations with the European microstates Andorra, San Marino and Monaco (Maiani 2019), as well as the draft for an institutional agreement between the EU and Switzerland (Swiss government 2020), show that the future of external differentiated integration is more likely bilateral and less institutionalised.
On the other hand, most of the institutional challenges presented in this paper also apply to any other form of advanced external differentiated integration as they emerge from the institutional requirements of functional integration in the EU and are thus not specific to the EEA's institutional framework. This of course raises the question of whether gradual integration of non-Member States into the EU is even possible or whether a clear dividing line between Member States and non-Member States is needed? The only such dividing line so far is that non-Member States do not have the right to vote in the EU decision-making process. However, the example of the participation of the EEA EFTA States in EU agencies shows that this dividing line also poses various problems that make external differentiated integration more difficult. The search for an effective and democratic institutional design for external differentiated integration will continue.
Electronic copy available at: https://ssrn.com/abstract=3712440 The European Commission supports the EUI through the European Union budget. This publication reflects the views only of the author(s), and the Commission cannot be held responsible for any use which may be made of the information contained therein.