A theory of external judicial politics: the ECJ as cautious gatekeeper in external relations

Abstract

Scholars have extensively studied how the European Court of Justice (ECJ) interacts with Member State courts. The ECJ’s behaviour vis-à-vis international tribunals remains, however, underexplored despite its salience for EU global actorness. The ECJ does at times condone and at other times reject cooperation with international tribunals in that it either authorises or prohibits EU and Member State participation in relevant regimes. What drives ECJ behaviour? While intuitive, European law fails to fully account for it. This study draws on models of bounded discretion to explain ECJ behaviour in external judicial politics. It argues that two factors – namely jurisdictional overlap between the European legal order and international tribunals as well as the centrality of these tribunals in global governance – decisively influence the preferences of the ECJ, Member States, the European Commission and Parliament and thus delimit the range of politically viable rulings and shape ECJ behaviour.

Keywords:

• European Integration
• judicial politics
• ECJ
• international tribunals
• autonomy

The role of the European Court of Justice (ECJ) in post-Brexit relations has emerged as a highly salient issue in current talks between the United Kingdom and the European Union (EU). The British government insists that it cannot accept the ECJ as a neutral court with jurisdiction over Northern Ireland – still part of the Single Market regime – and demands the creation of a novel tribunal under international law to resolve post-Brexit disputes. EU negotiators, in turn, refuse to review the Withdrawal Agreement and the ECJ’s powers. This deep-rooted disagreement has produced heated exchanges and threats to suspend the agreements governing post-Brexit cooperation and trade. While commentary abounds, it is noteworthy that few observers have paid attention to the ECJ’s views on this matter.

The ECJ has indeed been playing a key role in defining the EU’s relationship with international tribunals and regimes and is likely to have a decisive say over any post-Brexit dispute settlement mechanism. Since the 1970s, the ECJ has been regularly prompted to rule on the ability of the EU and Member States under European law to participate in international tribunals and related regimes. These rulings, while few in number, are of considerable political salience in that they shape EU external relations and constitutionalise the EU as global governance actor. What is more, they suggest that the ECJ holds an ambivalent stance towards international tribunals. Whereas the ECJ has predominantly pursued a cooperative approach vis-à-vis Member State judiciaries to co-opt them for European Integration (see Alter 2001; Burley and Mattli 1993; Sweet and Brunell 2012), it oscillates between cooperative and non-cooperative approaches towards international tribunals. The ECJ, for instance, condoned the EU’s participation in the novel Investment Court System (ICS) and the Dispute Settlement Body (DSB) of the World Trade Organisation (WTO) and thereby accepted the need to jointly manage with international tribunals potential norm conflicts arising between European and international law. The ECJ, however, blocked the EU’s accession inter alia to the European Court for Human Rights (ECtHR), the Court of the European Economic Area and the European and Community Patent Court and thereby shielded the European legal order and itself from judicial interreference. Little is known about these ‘external’ judicial politics of the ECJ. What informs ECJ decisions to opt for or against cooperative approaches?

It is intuitive to turn to European law to explain ECJ jurisprudence. After all, the ECJ’s mandate is to authoritatively interpret European law. Upon closer scrutiny though, a legalistic explanation cannot fully account for ECJ behaviour. First, European primary law – apart from laying out a basic allocation of competences in external relations – remains almost entirely silent on the relationship between the EU, the Member States and international tribunals. The ECJ instead invokes the so-called ‘autonomy’ of the European legal order as the legal basis to evaluate whether the EU or her Member States can participate in an international tribunal and regime (Lenaerts 2019). This doctrine, as Alter (2001: 2) shows, is not enshrined in the European Treaties but the ECJ asserted it in a fundamentally political process through its caselaw. Second, the ECJ has even invoked this doctrine to override the very few explicit Treaty provisions dealing with the EU’s relationship with international tribunals and regimes. In Opinion 2/13, the ECJ indeed referred to it to block the EU’s accession to the ECtHR explicitly mandated in Art. 6.2 TEU of the Treaty of Lisbon (CJEU, 2014). Third, the ECJ’s recourse to its doctrine of autonomy has been widely criticised as arbitrary and at times manifestly self-serving and aimed at minimising the influence of international competitor tribunals (de Witte 2010; Odermatt 2016). Lawyers criticise that it is impossible to identify legal-institutional properties of international tribunals that are compatible or incompatible with European law (Kassoti and Odermatt 2020; Mavroidis and Cantore 2018). In sum, European law – even in the eyes of lawyers – is insufficient to fully explain ECJ behaviour.

This study draws on models of bounded discretion to explain ECJ behaviour in external judicial politics (Carrubba and Gabel 2014; Larsson and Naurin 2016). These models imply that the ECJ seeks to promote its own policy agenda – namely the advancement of European Integration and consolidation of its judicial powers vis-à-vis domestic and international tribunals – through its jurisprudence yet is constrained by its legal-political environment. These constraints take the form of legal norms and interpretation techniques that delimit the range of legally admissible rulings as well as of varying preference constellations among Member States, European Commission and Parliament that delimit a subset of politically viable rulings. It is within this legally politically defined space that the ECJ can rule and engage in agency slack without facing the risk of grave reputational damages, non-compliance or political override (Carrubba and Gabel 2014; Kelemen 2012). Taking into consideration the vagueness of European law in this domain, constellations of stakeholder preferences should play the predominant role in shaping ECJ behaviour. This study argues that two factors in particular influence these preference constellations in external judicial politics: First, jurisdictional overlap – understood as geographical and substantive congruence – between the European legal order and international regimes and tribunals should determine the risk of norm conflict and judicial competition and therefore notably influence ECJ preferences on cooperation. High overlap should cause ECJ opposition to participation. Second, the centrality of international tribunals and regimes within global regime complexes should determine how strongly notably the Member States, Commission and Parliament feel about participation. Stakeholders should cohesively push for participation in regimes and tribunals of high centrality thus limiting the ECJ’s leeway. In sum, this study advances the argument that regime properties shape constellations of stakeholder preferences, which then influence ECJ behaviour in external judicial politics.

The study tests this model through case-oriented Qualitative Comparative Analysis (QCA) of ECJ jurisprudence, relevant legal literature, an assessment of regime properties and preference variation in relevant ECJ proceedings. The next sections review the judicial politics literature and develop the theory in greater detail. The fourth section discusses research design and methodological choices. The last sections present the empirics and conclude.

Taking stock – the ECJ and judicial cooperation

Why does the ECJ cooperate with other courts? Scholars of European judicial politics have extensively studied cooperation between the ECJ and Member State courts. Rationalist, neo-realist and legalistic research suggests that Member State preferences determine ECJ behaviour including cooperation with national judiciaries (Garret and Weingast 1993; Garrett 1995). Neo-functionalist and historical institutionalist research, in turn, see the ECJ as causally independent actor that drives cooperation with national judiciaries to advance European Integration and to consolidate its judicial powers (Alter 2001; Blauberger and Schmidt 2017; Burley and Mattli 1993; Pavone and Kelemen 2019; Sweet and Brunell 1998; Weiler 1994). Other scholars, finally, adopt a bottom-up perspective and caution that national courts – rather than the ECJ or Member State governments – initiate preliminary ruling procedures and drive cooperation. The propensity of national courts to seek preliminary rulings is seen to depend on economic factors (Sweet and Brunell 1998), legal cultures and institutions (Alter 2009, 1996; Kelemen and Pavone 2018; Pavone and Kelemen 2019), opportunistic and strategic motivations (Leijon 2021) and the rule of law in Member States (Blauberger and Kelemen 2017).

While the literature on European judicial politics offers insights on cooperation between the ECJ and Member State courts, it remains silent on cooperation with international tribunals. Schimmelfennig’s (2006) study on how the ECJ cites ECtHR jurisprudence to fight off challenges from national high courts constitutes a rare exception. While it does not propose a theory of external judicial politics, it implies that the ECJ cooperates with international tribunals to consolidate the European legal order and its judicial independence. This vision of the ECJ echoes neo-functionalist and historical institutionalist accounts of internal judicial politics as well as research on transnational judicial communication (see Slaughter 1994; Voeten 2010). The latter suggests that international courts cross-reference rulings of other courts – a form of judicial cooperation – to strengthen their position vis-à-vis states and to increase their influence. The following sections draw on insights from these literatures to develop a theory of ECJ behaviour in external judicial politics.

A theory of external judicial politics

To theorise external judicial politics, it is important to recognise that cooperation in internal and external judicial politics are processes of different nature. Internal cooperation is a legal-administrative process in which the ECJ and Member State courts directly interact to jointly resolve legal questions. External cooperation, in turn, plays out through ECJ rulings on the constitutionality of EU and Member State participation in international tribunals and regimes. The ECJ and international tribunals do not directly interact, but the ECJ acts as gatekeeper for international cooperation. Through its jurisprudence, it either enables or blocks the EU and Member States from participating in international tribunals and regimes and thereby allows or prevents interactions between European and international jurisprudence and law. It follows from this insight that a theory of ECJ behaviour in external judicial politics needs to explain how the ECJ arrives at its rulings.

This study draws to that end on models ‘bounded discretion’ of international courts (Carrubba and Gabel 2014; Larsson and Naurin 2016). These models stipulate that international courts like the ECJ are not mere administrators of justice but seek to leave an imprint on policy. The judicial discretion of international courts and their ability to shape policies are seen to be limited by law as well as the constellation of preferences of contracting states and key stakeholders. The ECJ’s ability to promote its agenda should be high, if the law is vague and stakeholders – such as the Commission, Parliament and Member States – are either supportive, disinterested or divided over the ECJ’s agenda hindering political override through law-making. The ECJ’s ability to promote its agenda should be limited, in turn, if the law leaves little interpretative leeway or stakeholders are united in opposition to the ECJ’s agenda making political override a credible scenario. Most scholars of European judicial politics accept this fundamental assumption of models of ‘bounded discretion’. Alter indeed observes (2014: 338) that it would be surprising if the ECJ systematically ignored Member State preferences. Despite this consensus though, scholarly disagreement persists over the scope and determinants of courts’ discretion due to epistemological challenges in identifying the considerations that inform ECJ decisions (Carrubba and Gabel 2014; Larsson and Naurin 2016; Sweet and Brunell 2012).

Models of bounded discretion offer a valuable approach to theorise ECJ behaviour in external judicial politics. As European law provides for a vast range of legally acceptable rulings in external judicial politics, variation in the constellation of stakeholder preferences delimiting the range of politically viable rulings should be particularly important in shaping ECJ behaviour. What then determines ECJ, European Commission, Parliament and Member State preferences and thus drives variation in preference constellations on cooperation with international tribunals and regimes? To answer this question, this section first defines the fundamental interests of these actors in external judicial politics, secondly discusses how these translate into context-specific preferences regarding specific tribunals and thirdly lays out how they affect overall preference constellations constraining or enabling the ECJ:

• The ECJ is seen to pursue two fundamental interests: to consolidate the European legal order and thereby advance European Integration (Burley and Mattli 1993; Weiler 1994); and to preserve its judicial independence (Alter 2009). While these interests explain the ECJ’s cooperative attitude vis-à-vis Member State judiciaries, they also explain why the ECJ should be cautious regrading cooperation with international tribunals and regimes. EU and Member State participation in international tribunals may promote the constitutionalisation of the EU as a global actor, yet may trigger norm conflicts between European and international law that result in judicial competition between the ECJ and international tribunals.

• The European Commission and Parliament are seen to hold and promote similar interests: to advance European Integration; and to consolidate their political powers where appropriate in view of subsidiarity. Importantly, they are seen as keen promoters of greater EU involvement in external relations. Hence, the Commission and Parliament are likely to share ECJ concerns about the autonomy of the European legal order, but to equally attach great importance to the constitutionalisation of the EU as a global governance actor.

• The interests of the Member States are more complex to model in that they are committed to furthering their national welfare and only indirectly to European Integration to the extent that it supports the former (Hooghe and Marks 2009). Regarding external judicial politics, Member State governments should focus on pursuing their functional needs in international affairs inter alia through participation in international tribunals and regimes that resolve interstate collective action problems and afford less attention to the autonomy of the European legal order unless obvious significant risks for the EU and its Single Market occur.

Fundamental interests form the basis for the formation of context-specific actionable preferences. Preference formation is indeed the process through which actors analyse their environment that conditions the pursuit of their fundamental interests to arrive at actionable preferences. This study argues that two exogenous factors decisively – yet not exclusively – shape ECJ, European Commission, Parliament and Member state preferences and thus overall preference constellations defining the range of politically viable ECJ rulings in external judicial politics: 1) jurisdictional overlap of the European legal order and international tribunals and regimes; 2) the centrality of international tribunals and regimes for global governance.

Jurisdictional overlap: Jurisdictional overlap between the European legal order and international tribunals and regimes should notably shape ECJ preferences on external cooperation and to a lesser degree Commission, Parliament and Member State preferences. Jurisdictional overlap refers to overlap in terms of substantive law and regime membership. Overlap is high, if the European legal order and international tribunals and regimes regulate similar issue areas and have a similar Eurocentric membership. Overlap is limited, in turn, if the European legal order and international tribunals and regimes regulate different issue areas and/or EU parties represent only a fraction of regime membership.

Jurisdictional overlap matters as it affects the probability that norm conflicts and judicial competition arise between European and international law and between ECJ and international tribunals. Norm conflicts occur when EU and Member States face contradictory obligations under European and international law. Judicial competition, in turn, arises when international tribunals interpret European law as fact or relevant international law to resolve disputes and challenge the ECJ’s judicial monopoly. Both phenomena challenge the autonomy of the European legal order including the ECJ’s judicial powers (de Witte 2010; Lenaerts 2019; Odermatt 2016) and thus go against the ECJ’s fundamental interests in judicial politics.

It follows that the ECJ should turn more critical of EU and Member State participation in international tribunals and regimes as jurisdictional overlap increases. Greater overlap results in a higher number of substantive issues and interstate relations being co-regulated and co-adjudicated by both the European legal order and international regimes. While the ECJ has been fairly successful in deflecting equivalent challenges from Member State constitutions and courts by asserting the supremacy and direct effect of European law (Alter 2009; Weiler 1994), it cannot easily mitigate norm conflicts and judicial competition through doctrinal tools in that international law is a priori binding on the EU, the ECJ and Member States. As overlap grows, the ECJ should thus become more hesitant to condone participation and cooperation with international tribunals and regimes. The European Commission, Parliament and Member States may share concerns over judicial competition but generally attach less importance to it.

Regime centrality: The centrality of international tribunals and regimes, in turn, should notably shape the preferences of the European Commission, Parliament and Member States on cooperation. Regime centrality refers to the significance of tribunals and regimes within so-called global regime complexes governing international affairs (Raustiala and Victor 2004). Most global governance domains do not come under the purview of a single international tribunal or regime, but various tribunals and regimes co-exist and co-govern. While membership in certain core tribunals and regimes is essential for states to address international collective action problems and exert influence (e.g. WTO), other tribunals and regimes play secondary roles and are substitutable (e.g. UNCTAD). Regime centrality, in other words, describes how functionally significant tribunal and regime membership are seen to be.

Regime centrality should determine how strongly the European Commission, Parliament and the Member States care about participation and thereby about cooperation with international tribunals and regimes. As discussed above, the European Commission, Parliament and Member States should attach greater importance to the ability of the EU and Member States to participate in global governance than the ECJ to consolidate their power and to enhance welfare through the resolution of interstate collective action problems. Hence, they should be more or less willing to entertain ECJ concerns and jurisprudence hindering participation in tribunals and regimes depending on regime centrality. They should be little inclined to accept ECJ concerns regarding participation in tribunals and regimes of high centrality manifesting itself in homogenous support for participation. They may be more accepting of ECJ concerns or indeed less interested regarding tribunals and regimes of low centrality. Due to idiosyncratic stakeholder properties – for instance variation in Member States’ trade dependence or attachment to Human Rights – one should expect some variation in the perception of regime centrality and thus preferences on tribunal and regime participation.

Last, it is important to spell out how jurisdictional overlap and regime centrality may shape overall preference constellations. After all, models of bounded discretion stipulate that the ECJ’s ability to pursue its agenda in external judicial politics hinges on the constellation of stakeholder preferences and more precisely whether stakeholders are interested, divided or united against or in favour of the ECJ’s preferences (Figure 1). As theorised above, the Commission, Parliament and Member States should fairly cohesively push for participation in tribunals and regimes that are of high centrality yet may show uninterested or divided over participation in tribunals and regimes of low regime centrality. The ECJ, in turn, should be highly concerned about participation in tribunals and regimes that exhibit high jurisdictional overlap and less concerned about participation in tribunals and regimes of limited overlap. Hence, high regime centrality in combination with low jurisdictional overlap should result in a preference configuration conducive to cooperative ECJ rulings, whereas low regime centrality and high jurisdictional overlap should be least likely to produce cooperative rulings. This theoretical reasoning is enshrined in below hypotheses while a legalistic explanation is enshrined in a counter-hypothesis:

Figure 1. A model of bounded ECJ discretion in external judicial politics.

H_1.1: The ECJ should adopt a cooperative approach vis-à-vis international tribunals, if tribunals and regimes exhibit low jurisdictional overlap and/or high regime centrality.

H_1.2: The ECJ should adopt a non-cooperative approach vis-à-vis international tribunals, if tribunals and regimes exhibit high jurisdictional overlap and/or low regime centrality.

C₁: Cooperation and non-cooperation between the ECJ and international tribunals and regimes exclusively reflect legal considerations.

Research design and methodological strategy

How to test empirically the validity of these hypotheses and causal argument? In a first step, it is necessary to define the universe of external judicial politics. According to above definition, external judicial politics play out when the ECJ gets asked to rule on participation and cooperation with international judicialized tribunals that provide third-party dispute resolution. For the sake of parsimony, this study thus disregards the political dynamics that lead (or not) to these requests as well as caselaw on diplomatic dispute settlement mechanisms. ECJ records suggest that the ECJ received ten requests for Opinions and preliminary rulings concerning participation in relevant tribunals (de Witte 2010; Lenaerts 2019; Odermatt 2016; Vajda 2019). It is in these ten cases that here-theorised dynamics and factors should play out.

In a second step, it is necessary to clarify the research strategy to assess these ten cases in view of the hypotheses and theoretical model. Models of ‘bounded discretion’ come with the epistemological challenge that it is impossible to directly observe ECJ preferences and decision-making. The ECJ uses law as a ‘mask and shield’ (Burley and Mattli 1993) to conceal political motives in its jurisprudence. Testing the validity of models of ‘bounded discretion’ is thus a complex and approximative endeavour. To address this challenge, the study uses case-oriented Qualitative Comparative Analysis (QCA) (Ragin 2008; Thomann and Maggetti 2020). QCA is a set-theoretic approach that consists of mapping co-occurrences and set-relations between explanatory factors and outcomes through Boolean truth tables to identify sufficient and necessary conditions for an outcome of interest to materialise. It is predominantly used to analyse and compare small and intermediate numbers of cases (5>N > 70) and is a powerful tool to uncover conjectural causation. Thomann and Maggetti (2020) further distinguish between case-oriented and condition-oriented QCA. Whereas condition-oriented studies compare intermediate (or large) numbers of cases and seek to produce findings of external validity, case-oriented studies compare few cases yet enrich truth tables with in-depth case analyses. Case-oriented studies, in other words, import classic qualitative methods – such as process tracing – into QCA to generate findings of high internal validity.

In accordance with this case-oriented approach, this study conducts four tests to build confidence in the internal validity of the hypotheses and theoretical model: First, it offers a historical overview of external judicial politics to provide a ‘thick’ understanding of the relevant cases including their setting, actor preferences, regime properties and ECJ rulings (Table 1). Second, it scrutinises the validity of legalistic explanations of ECJ behaviour (C₂) through a survey of legal scholarship. Third, it conducts a QCA based on a truth table to evaluate whether regime properties and ECJ rulings cooccur as theorised in hypotheses H_1.1 and H_1.2. Last, the study takes stock of Commission, Parliament and Member State preference constellations – as recorded in their submissions to ECJ proceedings – as an intervening variable that transmits tribunal and regime properties into varying ranges of politically viable rulings and ECJ behaviour. It thus qualitatively emulates large-n studies building on models of ‘bounded discretion’ , which sidestep epistemological problems arising from the unobservability of ECJ preferences by identifying correlations between constellations of stakeholder preferences and ECJ rulings (see Larsson and Naurin 2016). In combination, these four tests should allow rejecting or building confidence in the hypotheses, counter-hypothesis and here-developed model of external judicial politics.

Table 1. Summary table.

		Opinion 1/76 (Fund Tribunal)	Opinion 1/91 (EEA Court)	Opinion 1/94 (WTO DSB)	Opinion 2/94 (ECtHR I)	MOX Plant (UNCLOS)	Opinion 1/09 (ECPC)	Opinion 2/13 (ECtHR II)	Achmea (Intra-EU ISDS)	Opinion 1/17 (ICS)	Komstroy (intra-EU ISDS)
European case law identifier		ECLI:EU:C:1977:63	ECLI:EU:C:1991:490	ECLI:EU:C:1994:384	ECLI:EU:C:1996:140	C-459/03	ECLI:EU:C:2011:123	ECLI:EU:C:2014:2454	C-284/16	ECLI:EU:C:2019:341	C-741/19
Time period		1977	1992	1994	1996	2006	2009	2014	2019	2019	2021
Core question put to ECJ		Commission: Is the agreement on the European laying-up Fund compatible with European law?	Commission: Is the proposed EEA Court compatible with European law?	Commission: Is the EU competent to conclude GATS and TRIPS Agreements including provisions on the WTO Dispute Settlement Body?	Council of Ministers: Does the EU hold the necessary competences to accede to the ECHR and ECtHR?	Commission: Is Ireland’s resort to UNCLOS arbitration compatible with European law?	Commission: Is the ECPC compatible with European law?	Commission: Is ECtHR membership compatible with European law?	German Federal Court of Justice: Is intra-EU investment arbitration compatible with European law?	Belgian government: Is the Investment Court System compatible with European law?	Paris Courts of Appeal: Is intra-EU investment arbitration under the Energy Charter Treaty compatible with European law?
Independ. variables	Geographic overlap	0.90 (Eurocentric)	0.90 (Eurocentric)	0.12 (Global)	0.57 (Eurocentric)	0.15 (Global)	0.72 (Eurocentric)	0.57 (Eurocentric)	1.00 (EU-centric)	0.15 (Global)	0.52 (Eurocentric)
	Substantive overlap	Low	High	High	Low	High	High	High	High	High	High
	Regime centrality	Secondary regime	Secondary regime	Core regime	Secondary regime	Secondary regime	Secondary regime	Secondary regime	Secondary regime	Core regime	Secondary regime
Intervening variable	Preference constellation of key actors	Minor interest & divisions:	Minor divisions:	Unanimous support for participation:	Minor divisions:	Minor interest & divisions:	Strong interest & divisions:	Support for participation:	Strong interest & divisions:	Unanimous support for participation:	No formal position on substance:
		Council is mostly concerned about the EU’s external competences; DK and UK voice concerns over the Fund Tribunal’s impact on the European legal order.	BE and UK deem EEA Court compatible with European law; most Member States, Commission, Council of Ministers and Parliament remain silent on compatibility; only ES deems EEA Court incompatible.	Council of Ministers and Commission consider GATS, TRIPS and thus DSU compatible with European law; disagreements only over exclusive or mixed ratification.	AU, BE, FI, DE. IT, GR, SE support accession; FR, PT, ES, IE, UK question need and EU competence to accede.	IE and SE deem UNCLOS arbitration compatible; Commission deems it incompatible; no further submissions	CZ, DK, DE, EE, NL, PL, PT, RO, SI, FI, SE and UK consider ECPC compatible with European law; BE, FR, Parliament and Commission suggest minor amendments; IE, GE, ES, IT, CY, LI and LU consider ECPC incompatible with European law.	Council of Ministers, Commission and Parliament deem ECtHR membership generally compatible with European law; some however acknowledge the potential for legal-judicial tensions.	AU, DE, FR, NL, FI deem intra-EU investment arbitration compatible; CZ, EE, GE, ES, IT, CY, LV, HU, PL, RO, SK and Commission deem it incompatible.	Council of Ministers and Commission deem ICS compatible with European law; BE voices some concerns mostly to clarify legal situation.	Council of Ministers, FR, FL, HU, SE deem ECJ not competent to render preliminary decision in that core dispute only concerns non-EU parties.
Dependent variable	ECJ ruling	Non-cooperative	Non-cooperative	Cooperative	Non-cooperative	Non-cooperative	Non-cooperative	Non-cooperative	Non-cooperative	Cooperative	Non-cooperative
Theoretical evaluation of findings		Support for H1.2	Support for H1.2	Support for H1.1	Support for H1.2	Support for H1.2	Support for H1.2	Support for H1.2	Support for H1.2	Support for H1.1	Support for H1.2

Evaluating the importance of regime overlap and centrality: Hypotheses H_1.1 and H_1.2 stipulate that jurisdictional overlap between the European legal order and the relevant international tribunal and regime as well as regime centrality affect ECJ jurisprudence. Do tribunal and regime properties and ECJ rulings co-vary as expected? And to what extent do different properties qualify as sufficient or necessary conditions for cooperative or non-cooperative ECJ rulings in the given set of ECJ cases?.

It is further necessary to operationalise the independent variables. Jurisdictional overlap encompasses both geographical and substantive-legal overlap. Geographical overlap is measured through an index (G) that records the share of EU parties (EU and/or Member States) in the total membership (M) of the relevant international regime (G = EU/M). If the index is higher than 0.5, then more than half of the regime members are EU parties thus qualifying as Eurocentric. If the index is below 0.5, then the majority are third countries qualifying as global. Substantive legal overlap is deemed high/low depending to the extent that European and international law regulate the same issue areas and are thus likely to create norm conflicts. In line with Goldstein et al. (2000), it is useful to conceptualise substantive overlap in terms of sectorial scope and legal precision. European law, for instance, was vague and only touched on certain aspects of fundamental and Human Rights in the early 1990s resulting in low substantive overlap with the ECHR. Since the 2000s though, European law has come to encompass all aspects of this domain in great precision thus resulting in high substantive overlap. Last, the study distinguishes between core and secondary regimes of global governance to capture regime centrality. Core regimes are defined as essential for sectorial global governance and membership as key for global actorness. Secondary regimes, in turn, may be important yet are non-vital and substitutable for global governance. The WTO and the international investment regime for instance have been cornerstones of global trade and investment governance for six decades. Any economy that strives to take part in the modern world economy has acceded to the WTO and concluded so-called bilateral investment treaties (BITs). The ECHR, in turn, is a prominent regime for Human Rights protection yet it cannot be seen as non-substitutable (see EU Charter of Fundamental Rights) or key to global Human Rights governance (see UN) thus qualifying as a secondary regime. Finally and importantly, regime properties may evolve over time making it necessary to record their properties at the time of relevant ECJ ruling.

Empirical analysis

Historical overview: The following paragraphs provide a historical account of the evolution of external judicial politics to contextualise ECJ decisions before moving on to explicitly testing the hypotheses and counter-hypothesis.

External judicial politics have their roots in the founding years of European Integration. The EU was created as a classic intergovernmental regime based on conventional international treaties and law (Weiler 1994). In 1963, however, the ECJ started asserting in Van Gend en Loos that the founding Treaties had created a novel legal order that is ‘autonomous’ both from the domestic legal orders of the Member States and from international law (Figure 2). This doctrine, which the ECJ has been elaborating since then, is in many regards similar to the concept of sovereignty. ECJ judge Vajda (2019) defines it along three dimensions: 1) normative, 2) institutional and 3) jurisdictional autonomy. Normative autonomy implies that the European legal order is an autonomous source of norms and public authority that exists independently of the international and national legal orders. It manifests itself, on the one hand, in the primacy and direct effect of European law vis-à-vis Member State law. It implies, on the other hand, that the European Treaties are not part of conventional international law but akin to a sui generis constitution shielded from international law interference. Institutional autonomy, in turn, refers to the division of competences between the European Institutions and thus the EU and Member States. Jurisdictional autonomy, finally, refers to the judicial hierarchy of the European legal order that comes to the fore in the ECJ’s monopoly to authoritatively interpret European law and the prerogative of Member State courts to request preliminary rulings.

Figure 2. Ideal-type representation of internal and external judicial politics.

In the first decades of European Integration, the consolidation of the internal autonomy of the European legal order vis-à-vis Member State law and courts stood at the forefront of European judicial politics (Alter 2001; Burley and Mattli 1993). The ECJ pursued a strategy of judicial openness vis-à-vis Member State courts to co-opt them (Figure 2). This strategy showed successful yet recent challenges by Member State governments and courts caution that this work continues (Blauberger and Kelemen 2017). The consolidation of the external autonomy of the European legal order vis-à-vis international law and tribunals, in turn, first surfaced in the late 1970s and gained traction since 1990s due to the growing legalisation and judicialization of international affairs (Alter et al.2019; Goldstein et al.2000) and the EU’s greater involvement in external relations. These developments amplified the potential for norm conflict and judicial competition between international law and tribunals as well as European law and the ECJ. To protect the external autonomy, the ECJ resorted to a complex strategy of cautious gatekeeping opting at times for and against cooperation with international tribunals (Table 1).

The ECJ first ruled on participation and cooperation with an international tribunal in 1977. In Opinion 1/76 (ECLI:EU:C:1977:63), the Commission asked the ECJ to assess whether the draft agreement creating the European Laying-Up Fund for Inland Waterway Vessels between the EU, the Member States and Switzerland was compatible with European law. The agreement was meant to set up an international tribunal and fund to manage surplus carrying capacity for goods on the Rhine and Moselle rivers. The ECJ ruled that the EU was competent to conclude the agreement though found – echoing concerns of the Danish and British government – the so-called Fund Tribunal to undermine the jurisdictional autonomy of the European legal order. The Fund Tribunal would arguably undermine the ECJ’s judicial monopoly to interpret European law regarding transport policy.

The external autonomy of the European legal order again surfaced in ECJ jurisprudence in 1991 when negotiations between the EU and members of the European Free Trade Area (EFTA) on the creation of the European Economic Area ended. In the light of EU-internal legal discussions on the draft EEA Agreement, the Commission asked the ECJ to assess whether the planned EEA Agreement and notably the EEA Court were compatible with European law. The EEA Court should resolve disputes among EEA members – including EU Member States – based on the Acquis Communautaire and carried the risk of partially replacing the ECJ at the top of the European judicial hierarchy. Hence, the ECJ found in Opinion 1/91 (ECLI:EU:C:1991:490) that the planned EEA Court was incompatible in that it would challenge the ECJ’s monopoly to interpret European law (jurisdictional autonomy) and endanger the coherence of European law. The EU and EFTA states consequently redesigned – and effectively marginalised – the EEA Court, which the ECJ then found to be in compliance with European law in its follow-up assessment (ECLI:EU:C:1992:189).

The next episode in external judicial politics occurred with the conclusion of the Uruguay Round and creation of the World Trade Organisation (WTO) in 1994. In the light of long-standing disagreements over competences in external economic relations, the Commission asked the ECJ to evaluate whether the EU was competent to conclude the WTO Agreement notably regarding its annexes the General Agreement on Trade in Services (GATS) and the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPs). This request indirectly raised the question whether the planned Dispute Settlement Body (DSB) was compatible with European law in that the GATS and TRIPs Agreement both foresee dispute resolution through the DSB. The ECJ ruled in Opinion 1/94 (ECLI:EU:C:1994:384) that both agreements had to undergo mixed ratification yet remained silent on the compatibility of the DSB with the autonomy of the European legal order. The ECJ’s silence puzzles lawyers until today. Many assume that the ECJ sidestepped questions of external autonomy due to the centrality of the WTO DSB for global trade and the EU as a trading power and project.

External judicial politics came again to the fore in the Mox Plant case (C-459/03) in 2001. This dispute between Ireland and the United Kingdom concerned the pollution of the Irish Sea with radioactive substances from the British nuclear facility in Sellafield. After years of discussions and mediation attempts, Ireland submitted the dispute to ad hoc arbitration under the United Nations Convention on the Law of the Seas (UNCLOS). The European Commission held that the dispute predominantly concerned European environmental law and that Ireland was breaching Art. 344 TFEU, which requires Member States not to submit intra-EU disputes based on European law to other tribunals than foreseen under the European Treaties. In 2006, the ECJ ruled in favour of the Commission in that Ireland’s recourse to UNCLOS undermined the jurisdictional autonomy of the European legal order. In consequence, the UNCLOS tribunal resolved itself and Ireland took the dispute to the ECJ.

In the early 2000s, the EU, its Member States and neighbouring countries set out to create a European patent system under the umbrella of the European Patent Organisation (EPO) that would grant patent protection across 38 jurisdictions across the continent. To rationalise dispute resolution and prevent judicial venue shopping, the negotiating parties agreed to establish a European and Community Patent Court (ECPC) with compulsory jurisdiction. Anticipating concerns over the external autonomy of the European legal order, the negotiating parties agreed that the ECPC should be obliged to request and follow preliminary rulings of the ECJ on questions touching on European law. To dispel doubts, the Commission asked the ECJ for an Opinion. In 2009, the ECJ delivered a negative verdict in Opinion 1/09 (ECLI:EU:C:2011:123) on two grounds: First, the ECPC could not be forced to request preliminary rulings and to comply with them thus compromising normative autonomy. Second and more importantly, the proposal would deprive ordinary Member State courts to hear patent-related disputes and to make preliminary ruling requests thus undermining jurisdictional autonomy. The ECJ stressed that Member State governments could not sign away this quasi-constitutional prerogative of Member State courts in an international treaty. Following Opinion 1/09, the ECPC proposal was fundamentally revised to accommodate ECJ concerns – notably by embedding this tribunal in the European legal order to allow ECJ purview – yet ratification remains uncertain.

The most controversial episode of external judicial politics concerned the EU’s accession to the European Convention on Human Rights (ECHR) and her Court (ECtHR). The question first arose in Opinion 2/94 (ECLI:EU:C:1996:140) due to the growing entanglement of European law with Human Rights when the Council of Ministers sought on Opinion on whether the EU held sufficient competences to start accession negotiations with the ECHR (Schimmelfennig 2006). The Commission, Parliament and Member States showed broad though not unanimous support for EU accession at this point. The ECJ, nonetheless, denied this question in 1996. The Member States consequently decided to strengthen the EU’s competences regarding Human Rights notably through the adoption of the Charter of Fundamental Rights of the EU in 2000, which emulates the ECHR. The Treaty of Lisbon (2009) then legally mandated the EU’s accession to the ECHR in Art. 6.2 TEU. In 2010, the EU started accession negotiations and – conscious of the ECJ’s weariness – involved ECJ representatives. Under ECJ guidance, negotiators included substantive and procedural provisions in the accession protocol to avoid inadvertently challenging the external autonomy of the EU’s legal order. Notably, they designed a co-respondent mechanism to allow the EU and Member States to jointly respond and internally determine legal responsibility in ECtHR proceedings to protect institutional autonomy. Once finalised, the Commission requested an Opinion on the compatibility of the draft agreement with European law. To the surprise of academics, policy-makers and even the advocate general, the ECJ rendered a negative verdict in 2014. It observed that ‘…the EU and its institutions, including the Court of Justice, would be subject to… decisions and the judgments of the ECtHR…’ which could ‘call into question the Court’s [ECJ] findings in relation to […] EU law’ (ECLI:EU:C:2014:2454). It concluded that despite all precautions the ECtHR could – in theory – still find itself in the position to judge on the allocation of competences (institutional autonomy) and interferer with the EU’s judicial hierarchy and ECJ’s powers (jurisdictional autonomy). To comply, the EU reluctantly stopped the accession process. Opinion 2/13 sent shockwaves through the EU. While some expert observers and Member State representatives acknowledged the risk of the ECtHR interfering with ECJ powers, most commentators chided the ECJ as a rogue agent overruling the explicit will of the Member States enshrined in primary law.

The last episodes of external judicial politics concerned the global investment regimes and so-called Investor-to-State Dispute Settlement (ISDS). ISDS as enshrined in thousands of bilateral investment treaties (BITs) enables foreign investors to sidestep potentially biased courts in host states in case of expropriation to claim monetary compensation. Old capital-exporting Member States have been at the forefront of negotiating BITs with ISDS. Notably in the 1990s, they concluded hundreds of BITs with their Central and Eastern European neighbours. As these countries gradually joined the EU, these BITs became so-called intra-EU BITs among Member States creating an international law regime governing intra-EU foreign direct investment within the Single Market. For decades, the Commission has been demanding Member States to terminate intra-EU BITs in that they distort competition within the Single Market. This long-standing dispute entered its decisive stage in 2016 in the context of the arbitration proceedings Achmea v Slovakia, when the arbitration tribunal ruled that Slovakia had to pay €22 million in compensation to the Dutch insurance company Achmea for breach of the Dutch-Czechoslovakian BIT. Slovakia challenged the arbitral award in Germany, where the arbitration tribunal had its seat, arguing that the relevant intra-EU BIT had ceased to be in force with Slovakia’s EU accession and the tribunal lacked jurisdiction. The German Federal Court of Justice requested from the ECJ a preliminary ruling on the compatibility of intra-EU investment arbitration with European law. In 2019, the ECJ published its verdict (C-284/16). It held that intra-EU arbitration tribunals may have to interpret European law to resolve disputes yet – as they are no ‘Member State courts or tribunals’ in the sense of Art. 267 TFEU – are barred from seeking preliminary rulings potentially undermining the coherence of European law (normative autonomy) and the ECJ’s judicial monopoly (jurisdictional autonomy). To comply with the ECJ ruling, the Member States reluctantly agreed to terminate intra-EU BITs in May 2020 to remove the legal basis for intra-EU arbitration. They did not, however, terminate their membership in the Energy Charter Treaty (ECT). The ECT is an agreement among 54 parties including the EU, all EU Member States (except Italy) and East European and Central Asian countries. It contains ISDS provisions that have been invoked some 80 times in intra-EU disputes. The Achmea ruling triggered questions over the compatibility of intra-EU ISDS under the ECT with European law. In 2021, following a request from the Paris Court of Appeal, the ECJ ruled in Komstroy (C-741/19) that intra-EU ISDS under the ECT – much like under intra-EU BITs – was incompatible with the European legal order and ordered the EU and Member States to stop intra-EU arbitration under the ECT.

The Achmea judgement threw up questions about the general compatibility of ISDS – and indeed the EU’s newly developed Investment Court System (ICS) – with European law. Due to societal backlashes against conventional ISDS in the context of trade talks with USA, the EU decided to reform and replace ISDS with its new ICS in future trade and investment negotiations. While the ICS differs in many regards from conventional ISDS, it also exhibits certain commonalities. To address these concerns heads on, the Belgian government asked the ECJ for Opinion 1/17 (ECLI:EU:C:2019:341) on the compatibility of the ICS with European law. Following the Achmea ruling, experts expected another negative verdict and started holding funeral eulogies on the end of the EU’s ability to enter into modern trade and investment agreements. Since Opinion 2/13, frustration with the ECJ had grown and many feared that the ECJ – in its perceived quest to ring-fence its powers – would torpedo EU efforts to shape global economic governance and assume international leadership. When the ECJ delivered a positive verdict in 2019 in Opinion 2/17, observers expressed great relief and noted that the ECJ had come to its senses in that had gone back on some of its most contested arguments in previous decisions.

Evaluating legalistic explanations: Having given a historical overview of external judicial politics, the next section tests the intuitive counter-hypotheses C₁, which stipulates that only legal considerations guided ECJ decisions. This study cannot conduct a legal analysis of the complex cases presented above – a vast legal literature offers detailed assessments – but instead surveys this literature to assess whether the legal community sees the ECJ as faithfully interpreting European law or rogue agent.

A remarkably clear picture emerges from this survey. Many scholars of European law chide the ECJ for frequently adopting extreme positions with the barely veiled purpose to protect its judicial powers against encroachment from international tribunals and law through its jurisprudence and to the detriment of the EU’s ability to participate in global governance. This political agenda of the ECJ manifests itself in an – at times – allegedly arbitrary application of its doctrine. Kassoti and Odermatt (2020: 7) indeed note that the concept of external autonomy remains nebulous and constitutes ‘…a gate to the outside world… The ECJ is the gatekeeper. The principle of autonomy can be left open to allow interaction with the international legal order or it can be closed. That decision lies with the gatekeeper’. Mavroidis and Cantore (2018: 1) add that the ECJ has ‘…failed to establish clear criteria to identify ex ante the circumstances…’ that constitute a challenge to the EU’s external autonomy thereby circumscribing the EU’s ability to play a proactive role in global governance. De Witte’s (2010: 150) critique of the ECJ is even more pronounced. He chides the ECJ in that ‘…the autonomy of the EC legal system is put forward as a rhetorical shield to help to protect the Court’s own exclusive jurisdiction in Community law matters…’ against competitor tribunals. Koutrakos, lastly, observes that the ECJ has at times advanced the ‘most orthodox reading of the orthodoxy of European law’ (2019: 7) with regard to the European legal order’s external autonomy going against assessments of General Advocates and the European legal community.

This fervent critique motivated various ECJ members to intervene in academic discussions to explain the court’s reasoning and manage mounting frustration. ECJ judges Lenaerts (2019), Rosas (2017) and Vajda (2019) for instance wrote lengthy justifications to defend ECJ jurisprudence against criticisms of arbitrariness and opportunism. They in essence stress that the ECJ is entrusted to protect the constitutional features of the European legal order against short-sighted political actions yet they have not been able to fully deflect above concerns. In sum, the legal community assumes that both legal and political-opportunistic motives fuel ECJ behaviour thus challenging counter-hypotheses C₁.

Evaluating the importance of jurisdictional overlap and regime centrality: QCA uses Boolean truth tables to map logically possible and empirically observed combinations of dichotomously coded independent and dependent variables (Ragin 2008; Thomann and Maggetti 2020). Truth tables facilitate data description and the identification of conjectural causation. The findings reported in the truth table (Table 2) confirm hypothesis H_1.1 and H_1.2: First, jurisdictional overlap, as theorised, promotes and qualifies as necessary condition for non-cooperative ECJ rulings in the given set of ECJ rulings. Absent high regime centrality, jurisdictional overlap always coincides with non-cooperation. Second, geographical overlap and substantive overlap – analysed separately – only qualify as sufficient conditions in that non-cooperative rulings may occur, if only one of these conditions is present. Third, as theorised, high regime centrality qualifies as necessary condition for cooperative ECJ rulings. High regime centrality counteracts jurisdictional overlap, whereas low regime centrality coincides with non-cooperative rulings. In sum, the ECJ largely rejected cooperation with Eurocentric regimes of secondary importance that exhibit high overlap with the European legal order, yet mostly condoned cooperation with global core regimes even when exhibiting substantive overlap.

Table 2. Crisp-set truth table.

Geographic overlap	Substantive overlap	Regime centrality	Outcome	Count	Case name
1	1	0	0	5	Opinion 1/91; Opinion 1/09; Opinion 2/13; Achmea; Komstroy
1	0	0	0	2	Opinion 1/76; Opinion 2/94;
0	1	0	0	1	Mox Plant
1	1	1	0	0
0	0	0	0	0
0	0	1	0	0
1	0	1	0	0
0	1	1	0	0
0	0	1	1	0
0	1	1	1	2	Opinion 1/94; Opinion 1/17
1	1	1	1	0
0	0	0	1	0
1	0	0	1	0
0	1	0	1	0
0	0	1	1	0
1	1	0	1	0

Note: Geographical overlap is coded as 1, if labelled Eurocentric and 0 if labelled global in Table 1; substantive overlap is coded as 1, if reported as high and 0 if reported as low in Table 1; regime centrality is coded as 1, if it is reported as a core regime and 0 if it is reported as a secondary regime in Table 1; outcome is reported as 1 if ECJ decision is cooperative and 0 if it is non-cooperative.

Assessing the causal mechanism: The preceding analysis suggests that regime properties and ECJ rulings overall co-vary as expected. To scrutinise the underlying model of bounded ECJ discretion, it is helpful to further unpack the empirics and evaluate – as far as epistemologically possible – how regime properties co-vary with Commission, Parliament, Member State and ECJ preferences on participation in international tribunals and regimes. Stakeholder preferences are modelled as intervening variable that transmit variation in regime properties into variation in the range of politically viable ECJ rulings and thus ECJ behaviour. While it is impossible to observe ECJ preferences apart from its rulings, it is possible to scope Commission, Parliament and Member State preferences through their submissions to ECJ proceedings (see Larsson and Naurin 2016) (Table 1). Several observations can be made on this basis.

First, the Commission, Parliament and Member States were cohesive and supportive of participation and thus cooperation with international tribunals embedded in core regimes. In the context of Opinion 1/94 and 1/17, they unanimously signalled their expectation of EU participation in that both regimes are considered bedrocks of global economic governance. As expected, the ECJ adopted a cooperative approach in both instances. Second, the Commission, Parliament and Member States were often disinterested, divided or at least hesitant regarding participation and cooperation with Eurocentric secondary regimes (see Opinions 2/94, 1/09, 2/13 and Achmea). Manifest preference heterogeneity or hesitant support reflected diverging appreciations of the legal risks for the autonomy of the European legal order emanating from regime participation (see Opinions 1/76, 1/91, 2/94, 2/13) or heterogenous financial stakes (see Achmea and Komstroy). In the context of Opinion 2/13 on the ECtHR, the Council, for instance, formally supported EU participation yet it became clear during the ECJ proceedings that certain Member States acknowledged the potentially detrimental impact of EU participation on the external autonomy of the European legal order. In these instances, the ECJ indeed rendered, as theorised, non-cooperative rulings arguably as hesitation, disagreement and disinterest hinder political override and increase the ECJ’s room of manoeuvre.

Conclusion

Why does the ECJ at times condone and then again reject cooperation with international tribunals and regimes? This study has argued that legalistic explanatory approaches cannot fully account for ECJ behaviour in external judicial politics (C₁). Instead, it theorised that international regime properties shape stakeholder preferences, which in turn influence the range of politically viable rulings and ECJ behaviour (H_1.1 & H_1.2). The empirical analysis largely supported this theoretical argument. The study’s key contribution lies in drawing attention to and theorising external judicial politics. Until now, political scientists have afforded no attention to the complex relationship between the ECJ and international tribunals despite its growing salience in an increasingly interconnected, legalised and judicialized world. As this study shows, external judicial politics decisively influence the EU’s global actorness beyond questions of external competence allocations. Current discussions between the EU and the United Kingdom over future bilateral dispute resolution and the ECJ’s powers in Northern Ireland are indeed likely to bring external judicial politics further into the limelight. To fully understand the role of judicial politics and the ECJ in European Integration, it is thus necessary to broaden the analytical focus and move away from an inward-looking research focus. Looking ahead, political scientist may want to further explore when and why requests for Opinions and preliminary rulings on participation international tribunals get filled and thus enable the ECJ to rule. This study disregarded this question for the sake of parsimony, despite its importance for a thorough understanding of external judicial politics and the ECJ’s role in constitutionalising the EU as a global actor.

Acknowledgements

I am very grateful for suggestions and feedback on this manuscript by the peer reviewers as well as Alexander Katsaitis, Moritz Schmoll and Jürgen Basedow.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Notes on contributors

J. Robert Basedow

J. Robert Basedow is Assistant Professor for International Political Economy at the European Institute of the London School of Economics and Political Science. His research focuses on international trade and investment policy, politics of international law and dispute settlement mechanisms.