The CJEU’s give-and-give relationship with executive actors in times of crisis

Abstract

In a crisis, ordinary rules must often give way to a more expedient approach. Such emergency competences tend to favour executive decision-making over legislative procedures. In a European Union (EU) shaken by successive crises, this situation risks leading to permanent competence creep. While considerable attention has been devoted to the impact of crisis on legal and political decision-making within the Union, the position of the Court of Justice (CJEU) – and its impact on the distribution of powers within the EU – has been less researched. This Article fills the gap by exploring how the Court reviews the exercise of power in times of crisis by executive actors at the Union and Member State levels. Using migration law as a case study, it qualitatively and quantitatively examines how the CJEU has responded to crisis both in its scrutiny of measures of containment, and through its adjudication of migration cases in general before and after the acute phase of the 2015 refugee crisis. The Article shows that the crisis has led the CJEU to take a more lenient approach towards the executive powers at both the Union and the Member State level. It argues that this effectively amounts to a withdrawal from the judicial control function and enables an expansion of executive power that is likely to have effects lasting beyond any given emergency.

1. Introduction

Crisis appears to have become the new normal for the European Union (EU).¹ Since, as the proverb proclaims, desperate times call for desperate measures, this has consequences for the exercise of public power, as well as for the methods used to scrutinise it. In a crisis, normal rules must often give way to more expedient procedures, in order to facilitate a rapid response to an urgent situation. Such an approach, while perhaps necessary to avoid or contain damage in an immediate crisis, supplants the ordinary laws and undermines their associated safeguards. In particular, these emergency measures tend to be taken by the executive on the basis of formal or – more typically in the EU’s case – informal emergency competences that often minimise the participation of the legislature in the decision-making process.² As crisis becomes a semi-permanent state of affairs (as it has arguably become the case in the EU), there is a danger that this will lead to transfer of power – de jure or de facto – from elected representatives of the people to less accountable executive actors,³ increasing the power of the executive superstructure and exacerbating the ‘democratic deficit’ for which the Union has already been criticised.⁴

Clearly, there are good reasons for executive empowerment in times of crisis, particularly as regards the need for quick and decisive action. Nevertheless, research has shown that the invocation of emergency powers often opens up the door to democratic backsliding.⁵ Following the theory of tripartite separation of powers, judicial review is a given part of the guarantees against such developments.⁶ At the same time, the CJEU is widely regarded as a pragmatic and politically sensitive actor. A driver of European integration, certainly⁷ – indeed, perhaps an ‘activist’ in its pursuit⁸ – the Court is ultimately keenly aware of the Realpolitik surrounding its judgements and their implementation.⁹ This situation raises doubts as to how far the Court can be trusted to form the vanguard of democracy or individual liberties in times of crisis.

The legal nature of EU emergency measures both facilitates and obstructs judicial intervention. On the one hand, the lack of comprehensive or consistent regulation of a state of emergency means that crisis measures often entail bending rules or circumventing them. This renders the executive vulnerable to charges of illegality.¹⁰ On the other hand, measures of an unprecedented, unregulated, and informal nature may not fall readily under the review powers of the Court – potentially leaving it without jurisdiction, and would-be claimants without effective access to judicial remedies.¹¹

Yet, while significant scholarly attention has been devoted to the impact of crisis on legal and political decision-making within the EU,¹² the contribution of the Court to these decision-making processes has not drawn similar scrutiny. Against this background, the present paper explores how the CJEU reacts to crisis.¹³ In particular, it sets out to answer two questions. First, how intensely does the Court scrutinise the exercise of executive emergency competences? While a high standard of review ought arguably to be upheld as a necessary counterpoint to the risk of executive power creep, previous literature has questioned the feasibility – both practically and theoretically – of upholding such a standard.¹⁴ Second, does crisis prompt the Court to be restrained in its assessment of substantive legal issues? A more cautious judicial approach might be expected in response to greater political controversy,¹⁵ even if, from a strictly legal positivist standpoint, no change in case law is supposed to occur insofar as the crisis does not affect the content of substantive law.

Using migration law as a case study, this Article undertakes a qualitative and quantitative examination of how the CJEU responds to crisis both in its scrutiny of measures of containment taken at Union level and in its assessment of migration cases in general before and after the acute phase of the so-called refugee crisis in the late summer and autumn of 2015. It finds that not only did the Court take a lenient approach in its review of Union crisis measures, but this judicial reticence has continued to characterise the Court’s outlook on migration law in general since the crisis. It argues that this effectively amounts to a withdrawal from the judicial control function, enabling an expansion of executive power that is likely to have effects lasting beyond any given emergency. Rather than the give-and-take implied by the principles of institutional balance and separated powers, the relationship between the Court and the executive appears to be one where the Court gives and keeps on giving.

The argument proceeds as follows. Section 2 briefly introduces the EU’s emergency competences in the area of migration law, particularly as applied in the 2015 refugee crisis. Section 3 analyses the Court’s judicial review of the exercise of Union emergency competences, arguing that the Court has interpreted these competences broadly (A) as well as allowing the Council to cut corners in the decision-making process (B) and that these interpretations are likely to have an impact lasting beyond any given crisis situation (C). Section 4 presents the results of an empirical examination of the Court’s migration case law in the years before and after the acute refugee crisis. After some methodological remarks (A), it demonstrates that the crisis has prompted the Court to be more restrained (B) while at the same time shifting its trust away from Member State executive actors and towards its judicial colleagues (C). On this basis, Section 5 concludes that the Court’s reticence gives rise to largely unsupervised executive authority at the Union level and authority under limited, decentralised judicial supervision in the Member States. Section 6 places the conclusions in a wider context and points to future research avenues.

2. Emergency competences in EU migration law

Emergency competences in migration law are explicitly provided for in Article 78(3) of the Treaty on the Functioning of the EU (TFEU), which takes as its point of departure that an emergency arises when the Union faces a sudden inflow of migrants.¹⁶ This in itself is a relatively narrow conception of an emergency: it does not include emergencies in general, or even migration-related emergencies, but only emergencies of a certain kind. It is also worth noting that an emergency qualifies as such mainly for administrative reasons. It is not the plight of refugees as such or the conditions at their place of origin that constitute the crisis, but rather the structural pressure that immigration places on the administrative systems of the Union and its Member States. So construed, a crisis to some extent arises due to matters of law and administration at the place of destination. In the case of the 2015 crisis, the Dublin regime arguably worsened the burden borne by Italy and Greece, by preventing migrants from lawfully proceeding to other Member States.¹⁷

Article 78(3) states that, in a situation of emergency, the ordinary legislative procedure – which according to Article 78(2) normally applies in the field of migration – may be set aside. Instead, the Council may adopt, upon receiving a proposal from the Commission and only after consulting with the European Parliament, provisional measures that will apply in the emergency. Resorting to the emergency competence, therefore, has two major effects: strengthening the executive power, and bypassing the Parliament as a co-legislator.¹⁸ The Council decides by qualified majority on measures pursuant to Article 78(3) TFEU, unless it opts to deviate from the Commission’s proposal, in which case unanimity is required according to Article 293 TFEU. The Article 78(3) TFEU competence has been triggered only once thus far: in 2015, when the Council authorised the reallocation of refugees from the heavily burdened entry states of Italy and Greece to other Member States at a further remove from the Union’s external borders.¹⁹ It has also formed the basis for a Commission proposal for temporary measures at the Union border with Belarus.²⁰ (This proposal, however, has not yet been adopted.)

In addition, certain acts of secondary law contain similar powers of derogation addressed to the Member States. An example of this is Article 18 of the Returns Directive,²¹ according to which Member States may – in situations where an exceptionally large number of immigrants are to be returned – provide for longer times for judicial review and for a departure from ordinarily prescribed conditions of detention for persons to be returned.

Mention may also be made here of Article 72 TFEU. Although not specifically designed for emergencies, this Article declares that Member States retain power over matters affecting the law and order or internal security within each Member State – issues that naturally gain in prominence during a crisis. At least potentially, therefore, this provision implies a broadening of Member State powers of derogation from Union law in times of crisis, with a (temporary) transfer of competence from the Union to the national level. However, as will be further discussed in section 3 below, the Article is to be interpreted strictly.²²

3. Judicial review of emergency competence in Union law

A. Substantive review: the scope of emergency competence

On three occasions, the Court of Justice has had reason to examine the emergency competence conferred on the Commission and the Council by Article 78(3) TFEU: first in its July 2016 ruling in Jafari ²³; and thereafter in two cases, the first of which concerned the validity²⁴ and the second the enforcement²⁵ of the refugee relocation scheme.²⁶ Of these, the Jafari ruling is of less interest, as it dealt with Article 78(3) TFEU only incidentally. The second ruling, which followed a validity action brought by Hungary and Slovakia and supported by Poland, is most relevant for our purposes here, and it will be analysed thoroughly in the following sections. The infringement proceedings against Poland, Hungary, and Czechia – in which the emergency competence was largely undisputed – will be discussed more briefly.

As has been noted above, Article 78(3) is an exception. On this basis alone, one might expect it to be interpreted restrictively. This is especially so since, as the Court established early on in its reasoning in Slovakia and Hungary v Council, the competence set out in Article 78(3) is a non-legislative competence that allows for the derogation from legislative acts – a state of affairs that, one might imagine, underscores the need for restrictiveness further. Yet, the Court took a markedly expansive approach on the scope of the emergency competence conferred on the Council and the Commission.

The non-legislative nature of the competence, which is largely uncontroversial, follows directly from the formal definition of legislative acts in Article 289 TFEU, in conjunction with the fact that Article 78(3) does not prescribe any legislative procedure.²⁷ Although the Court itself designated the competence only negatively – as ‘non-legislative’²⁸ – the residual approach to the definition of the three branches implies that the adoption of legal measures on such a basis is an executive function.²⁹ This classification is supported by the modesty of the manner in which the Parliament – the most direct representative of the people – is included in the decision-making process under Article 78(3). As for the power to derogate from secondary legislation on this legal basis, which had been questioned by the applicant governments, the Court based its conclusion on two arguments. First, relying on the wording of Article 78(3), it observed that said Article does not define the nature of the measures to be adopted; and so, does not limit itself to implementing or accompanying acts.³⁰ In itself, however, this provides no positive support for a more extensive interpretation of the competence in question which also enables the executive to override provisions adopted by the legislature. In order to arrive at that conclusion, the Court instead cited the ‘overall scheme and objective’ of the provision, especially in relation to the ordinary legislative competence for the common asylum system laid down in Article 78(2). In particular, it observed, a restrictive interpretation that fails to include the right to derogate from legislation would ‘significantly reduce the effectiveness’ of the emergency competence.³¹ The conclusive argument for the existence of an executive power to derogate from legislative acts was thus the effectiveness of that very power. By contrast other concerns, such as the hierarchy of legal norms and the wider participation guaranteed by the legislative procedure – not to mention the effectiveness of the ordinary legislative competence laid down in Article 78(2) – were left entirely outside the Court’s reasoning.

The Court did observe that any derogations must be limited both in time and in substance.³² In its subsequent examination of these limits, however, it set the bar low, repeatedly emphasising the discretion of the Council as the designated decision-maker under the procedure set out in Article 78(3). In particular, it held that said Article confers ‘broad discretion’ on the Council regarding both the type of measures to be taken and the length of the period during which they shall apply. Moreover, the Court noted, the inclusion of mechanisms for dealing with ‘possible developments of the situation’ does not mean that the executive has stepped outside its competence.³³ The Court declined to furnish any guidance on how long a measure can apply while still qualifying as temporary. This reticence gives rise to the question of how long a situation can be said to constitute the ‘development’ of a (previous) emergency, such that the executive can lawfully handle it by dispensing from legislative provisions. While the Court did conduct a review of the 24-month applicability of the relocation scheme, it did so only very lightly, and it concluded that the Council had not ‘manifestly exceeded the bounds of its discretion’.³⁴ The Court’s reluctance to impose a stricter test – which, in casu, the chosen applicability period may well have passed³⁵ – could be taken to suggest that almost any definite term would suffice to pass the ‘provisional’ threshold. Thus, the discretion enjoyed by the executive to derogate from secondary EU law in times of crisis must be described as very broad.

An alternative approach to derogation from EU law can be seen in the joined infringement proceedings against Hungary, Poland, and Czechia for failing to carry out their obligations under the relocation scheme. In this case, in a clear parallel with its interpretation of Article 78(3), the Court accepted in principle that Member States were entitled, on the basis of Article 72 TFEU, to derogate from obligations under EU law where necessary in order to guarantee law and order and internal security. In contrast with its judgement in Slovakia and Hungary v Council, however, it underlined that Article 72 TFEU must be interpreted strictly, and that derogation is permissible only when a Member State has proven the necessity for such derogation.³⁶ This effectively brings the question within the scope of EU law, and thus ultimately within the Court’s own competence. Furthermore, the Court noted, since the Council had made provision within the relocation scheme itself for exceptions on grounds of security to the duty to accept relocated refugees, there could be no right to derogate from the legal instrument for such a reason.³⁷ Instead, it held that the Member States must be accorded a wide margin of discretion when applying the statutory exceptions laid down in the EU legal acts.³⁸ Such discretion is limited, however, to the assessment of individual cases; and consistent, objective, robust, and specific evidence must be cited in support of the decision in question.³⁹ One might accordingly question whether what remains really qualifies as a wide margin of discretion.

Yet another approach to the applicability of emergency competences was taken in the recent Landkreis Gifhorn case, in which the Court ruled on the limits of Member State emergency competence under Article 18 of the Returns Directive.⁴⁰ In this case, as in that of Commission v Hungary, Poland, and Czechia, the Court held that a Member State’s decision to activate the emergency clause is subject to judicial review by the courts of that Member State in each individual case that comes before them.⁴¹ The Court made clear that this review is not limited to the application of the exceptional powers in the case at hand, but must also address the question of whether the invocation of emergency competences is justified.⁴² Furthermore, the national court must take into account any circumstances cited by the parties, as well as all other circumstances it considers relevant. Thus, the decision of a Member State executive or legislature⁴³ to invoke an emergency competence is to be subjected continually to a full judicial review by national courts competent to engage in ex officio fact-finding, running an inherent risk of contradictory rulings that would render the emergency measures at the same time both applicable and inapplicable. This again forms a contrast with the ‘broad discretion’ that the Court accorded the EU institutions which are engaged in ‘complex assessments’ and in political choices under the emergency competence set out in Article 78(3).⁴⁴

While there is nothing to necessitate an identical approach to the emergency competences set out in the three Articles discussed above, it is easy – in light of Commission v Hungary, Poland, and the Czech Republic and Landkreis Gifhorn – to see alternative lines of argument that would have entailed a stricter circumscribing of executive emergency powers. The scope of the ‘provisional’ measures permitted could easily have been set more narrowly than just that they be limited in time; for instance, the Court could have required the Council to demonstrate that the chosen duration was indeed necessary. It could also have set the standard of review higher than ’manifest infringement’; eg, by requiring more than a single sentence in the preamble as justification. Similarly, it could have ruled that provisional actions by the executive should be limited to what is necessary to address the immediate emergency – with or without requiring the Council to demonstrate the necessity for such actions – while leaving contingency measures for possible developments to be handled in accordance with the normal legislative procedure set out in Article 78(2).

The Court’s generosity vis-à-vis the Commission and the Council regarding actions taken by them under Article 78(3) was also evident in its interpretation of the criteria for exercising the powers in question – in particular, the criterion that the emergency be attributable to a sudden inflow of migrants. In this regard too, the Court accorded the institutions broad discretion. The requirement of suddenness, the Court found, does not preclude the possibility that an inflow of migrants does form part of a longer-term development. The defining characteristic is instead whether the inflow is such as to make the ‘normal functioning of the EU common asylum system impossible’.⁴⁵ The Court then cited statistical data supporting the (undisputed) fact that the inflow of migrants into Greece and Italy in the summer and early autumn of 2015 was indeed massive, in both absolute and relative terms. The Council, it therefore concluded, had made no ‘manifest error of assessment’ in finding the inflow to be ‘sudden’.⁴⁶

Again, the approach taken in Landkreis Gifhorn is different. The Court made the existence of an emergency situation (in the meaning of Article 18 of the Returns Directive) conditional on whether the increasing inflow of migrants was unforeseeable, or if it was otherwise excusable that the Member State in question had failed to take sufficient structural measures to cope with the rising inflow.⁴⁷ This stands in sharp contrast to the Court’s interpretation of Article 78(3) in Slovakia and Hungary v Council, where it accepted the classification of inflow of refugees as sudden if it has made the normal functioning of the asylum system impossible – apparently without consideration of whether structural adaptations could feasibly have been undertaken. This too illustrates how the Court has consistently opted for the broadest interpretation of the scope of the Council’s competence under Article 78(3), even as it has taken a more restrictive approach to similar provisions in other legal acts.

In sum, the Court’s attitude vis-à-vis the political institutions under Article 78(3) TFEU must be characterised as highly deferential. As the analysis has shown, this stance is not the only possible interpretation by the Treaty, but represents an interpretative choice on the part of the Court, possibly suggestive of how the Court perceives of its own role in times of crisis: as a supporting actor to the Council and Commission.

B. Procedural review: the decision-making process

Regarding the decision-making process too, the Court took a generous view vis-à-vis the executive, represented in this case by the Council and the Commission. It would appear to be uncontroversial that the European Council cannot, under its power to define strategic guidelines for the area of freedom, security, and justice, alter the institutional decision-making rules of the Council.⁴⁸ The Court’s refusal to take into account the politically sensitive character of the issues at stake in this regard appears sound as a matter of the rule of law. It is namely in connection with the handling of politically sensitive issues that the guarantee afforded by constitutional decision-making procedures – that all relevant actors are included – is most relevant.

More remarkable is the Court’s leniency regarding the executive’s repeated revisions of the initial proposal. This concerns both the relative positions of the Commission and the Council, as well as the degree of involvement from the Parliament.

As noted earlier, Article 78(3) requires only that the Parliament be consulted. The Court, while reaffirming that effective participation of the Parliament is a ‘fundamental democratic principle’ and an ‘essential element in the institutional balance’,⁴⁹ held that once it has been consulted, the obligation to consult it again only arises in the case of amendments to the initial proposal that involve a difference ‘in essence’. The Court held that such amendments had indeed been made: the original proposal identified Hungary as a frontline state and a beneficiary of the relocation scheme, alongside Italy and Greece; whereas the final decision provided for relocation only from the latter two Member States. The Court accepted, however, that the provision of oral information, given by the Council president at an extraordinary sitting of the Parliament – on the day before the latter adopted a resolution expressing its support for the proposal – meant that the Parliament ‘must necessarily have taken account of’ the changes.⁵⁰ This conclusion, which De Witte and Tsourdi characterise as ‘rather light-hearted’,⁵¹ was apparently not shaken by the fact that the Parliament was still referring in that resolution to relocation measures for the benefit of Italy, Greece, and Hungary; or that it had explicitly requested to be notified and consulted again in the case of amendments.⁵² Nor did the Court comment on the fact that, formally speaking, the amendments announced by the Council president in the Parliament before the vote appear to have been adopted only after it.⁵³ This nonchalant treatment of the (already minimised) role of the Parliament in the decision-making process is particularly hard to understand in view of the fact that, as more than one commentator has pointed out,⁵⁴ it would have been quite possible for the CJEU to take a stand on this point without necessarily invalidating the measures in question.

Furthermore, Article 293 TFEU requires the Council either to follow the proposal as put forward by the Commission – whether originally of after amendments – or else to depart from it by unanimous decision. The Council had not acted unanimously, so the question arose as to whether the alterations to the original proposal had actually been proposed by the Commission. In this regard, the Court had already ruled in 1994 ruled that amendments do not have to be made in writing.⁵⁵ This appears to be common practice.⁵⁶ However, citing a greater need for flexibility in a crisis situation, the Court appeared to go a step further, holding that amendments can be made not just orally but also implicitly. This would seem to be the import of its conclusion that the Commission, merely by participating in the adoption process, can be considered to have amended a proposal if its participation ‘clearly shows that it has approved the amended proposal’.⁵⁷ In the case at hand, the Commission argued that the Council meetings where the amendments had been introduced had been attended by two Commission representatives, who were empowered to approve amendments. Neither the Commission nor the Council, however, appears to have presented any evidence to the Court that these two Commission representatives had taken any actions at any point to exercise that power – aside from simply taking part in the meeting. Indeed, at a (slightly) earlier point in the judgement, the Court explicitly noted that ‘the Council made other amendments to the Commission’s initial proposal following the Parliament’s adoption of that legislative resolution’⁵⁸ – a statement that appears to contradict the Court’s later conclusion (although it can be explained away, perhaps, as a sloppy way of noting that the amendments had been made at Council meetings). The effect of this virtually all-encompassing understanding of what constitutes action by the Commission in the decision-making process is to concentrate power further in the hands of the Council, and is difficult to reconcile with the wording of Article 193 TFEU.

Interestingly, the Court also noted that the Commission itself did not consider its power of initiative to have been undermined in this case.⁵⁹ The Commission is generally known for fiercely guarding its prerogatives, so this admission is remarkable in itself. It cannot be excluded that the Commission genuinely agreed that its representatives can be presumed, when they participate in Council meetings, not only to endorse but indeed to have themselves proposed any amendments against which they do not protest. Alternatively, the Commission’s position can be understood as a pragmatic choice. That is, it did not wish the relocation scheme to be invalidated, so it colluded with the Council – even though the latter had not, strictly speaking, acted either unanimously or upon a proposal from the Commission.

Either way, it is highly questionable that this represents a relevant legal consideration for the Court of Justice. In their analysis of Slovakia and Hungary v Council, De Witte and Tsourdi note how rare it is for the CJEU to disagree with the other EU institutions when they are agreed among themselves.⁶⁰ They observe as well that the Parliament’s feeling ‘no need to intervene in the present case to defend its institutional prerogatives’ may be part of the reason for the Court’s easy dismissal of that institutions right to be consulted again.⁶¹ However, while this may be true as an empirical matter, it remains problematic from a constitutional perspective. The rules on majority decision-making – and particularly those that require unanimity – function as protection for the minority. Whether the other institutions insist upon it or not – and perhaps particularly when they do not – the role of the Court should be to uphold the distinction between the power of proposal and that of decision, which the Treaty drafters placed in different bodies.⁶² Instead, the Court allowed the institutions to reinvent their functions ad hoc in order to achieve a politically desired solution, thereby effectively granting an exemption from the decision-making rules laid down by the Treaties, and permitting a minority that should have been heard to be outvoted. If the Court not only refrains from intervening against such collusion, but also can be expected (as De Witte and Tsourdi imply) to refrain from so doing, then one might wonder who shall guard the integrity of the democratic decision-making process, in time of crisis or not.

C. Long-term effects

The rulings analysed above are likely to affect longer-term relations between the Court and other EU institutions, well beyond the current crisis.⁶³ The most obvious effects concern the application of the Article 78(3) competence in possible future crises. The above analysis has demonstrated that the Court has expanded the competence in four ways: first, there is no fixed upper limit for what constitutes a ‘provisional’ measure; second, an inflow of migrants can be seen as ‘sudden’ even if it is predictable; third, the Parliament may be said to have been consulted even if essential amendments are presented only informally; and fourth, the requirement of unanimity in the Council effectively applies only when it has shut Commission representatives out of meetings or else intends to adopt a version which the Commission actively opposes. One might reasonably expect, therefore, that the threshold for the Commission and the Council to use the Article 78(3) competence, instead of the ordinary legislative competence prescribed by Article 78(2) TFEU, will be lower in the future. This serves to increase the power of the Council, and to a lesser extent that of the Commission as well; and it weakens the democratic influence of the Parliament and the voters it represents.

Beyond this particular legal basis, furthermore, it is worth noting that the Court treats the fact of ongoing crisis in many respects as a reinforcing factor, but not as a sine qua non. Discussing the criterion of suddenness in Article 78(3) TFEU, for example, the Court remarked – in a separate paragraph and without obviously or explicitly limiting its statement to the interpretation of that Article – that ‘EU institutions must be allowed broad discretion when they adopt measures in areas which entail choices, in particular of a political nature, on their part and complex assessments’.⁶⁴ Since almost all measures taken by the Council or the Commission entail political choices, at least to some extent, this statement would seem to have a wide potential reach.

Furthermore, regarding the forms for Commission amendments to proposals under Article 293 TFEU, the Court first cited previous case law on the need for a degree of flexibility.⁶⁵ Only thereafter did it remark that this flexibility ‘must, a fortiori, prevail in the case of the procedure for adopting an act on the basis of Article 78(3) TFEU’, due to the specific circumstances under which the latter Article is intended to be applied.⁶⁶ It is unclear whether, moving forward, the acceptance of implicit amendments under Article 293 TFEU – an Article that applies to Union decision-making in general – will be considered an example of the ‘a fortiori’ flexibility required in times of crisis, or simply accepted as part of the flexibility that is ‘characteristic’ of the adaptation of EU legal acts in general. The wording of the judgement suggests the latter.

Similarly, in responding to the objection that the relocation scheme constituted a disproportionate action, the Court initially observed broadly that the institutions must be granted a wide discretion when making complex assessments of a political character. It is only when the measures chosen are ‘manifestly inappropriate’ that they are to be declared invalid.⁶⁷ Only thereafter did the Court state that these principles ‘also’ apply in the field of migration, and ‘in particular’ when decisions are being taken under Article 78(3). These formulations all leave the road open for further expansion in future case law, well beyond a given crisis situation.

4. Migration law at the CJEU before and after the crisis

A. Methodological remarks

Emergency competences are just one facet of the EU’s crisis response. While the existence and extent of emergency regimes vary between Member States, the Union’s response consists to a great extent of adaptations in the ordinary legal framework.⁶⁸ In order to assess the effect of the migration crisis on the Court’s outlook more generally, I conducted an empirical examination of its case law in this field in the years immediately preceding and following the crisis. I used original data from the SepaRope Judiciary Dataset for the analysis.⁶⁹ The dataset includes all Court of Justice cases decided between 1 January 2010 and 31 December 2020 which are classified in the Court’s Curia database as falling under the heading of ‘Area of Freedom, Security and Justice’ and one of the following three subcategories: ‘Immigration Policy’, ‘Border Checks’, and ‘Asylum Policy’. This comes to 142 judgements in all. Of these, the vast majority are preliminary reference cases, with only four having been brought before the Court in other types of proceedings (two through the infringement procedure under Article 258 TFEU and two as validity challenges under Article 263 TFEU). Six cases gave rise to questions concening the validity of EU legal acts. Thus, observations based on this data mainly concern the Court’s relation to executive or legislative power at the national level.

Each judgement in the dataset is subdivided further into different legal issues, reflecting the distinguishable legal problems brought before the Court.⁷⁰ The dataset comprises 283 legal issues in all. Of these, 96 belong to 52 unique judgements delivered by the Court in the five and a half years preceding the acute phase of the crisis – that is, between 1 January 2010 and the summer judicial recess of 2015. The remaining 187 legal issues featured in 90 unique judgements delivered during a five-and-a-half-year period during and after the height of the refugee crisis (1 September 2015 to 31 December 2020). Unsurprisingly, these figures in themselves illustrate how migration became a more contentious area with the onset of the crisis.

B. Increased restraint after the crisis

A comparison of the Court’s case law before and after the height of the 2015 refugee crisis suggests that the crisis heralded a new and more cautious outlook for the Court – one marked by greater restraint vis-à-vis executive decision-makers.

For each legal issue identified, the dataset records whether the Court’s ruling implies that a public actor has committed a violation of EU law, and if so which branch of government that actor belongs to.⁷¹ Such violations are not limited to formal infringement actions brought by the Commission under Article 258 TFEU – as observed above, such actions are very uncommon in the dataset – but they include all rulings of the CJEU which imply that a public body has acted in a manner inconsistent with its obligations under EU law. If increased politicisation of the policy area renders the Court more cautious, this should be visible in a drop in the relative number of such rulings, indicating that the Court has become less willing to provoke adverse reactions. By contrast, if the risk of formal or informal executive competence creep due to crisis prompts the Court to heighten its scrutiny in order to safeguard the constitutional division of powers, then the opposite effect should be evident: an increase in the relative number of findings that EU law has been violated. In fact, as we have seen in section 3, the Court has adopted the former approach in its review of Union emergency competences.

As Figure 1 shows, the relative frequency of violations of EU law established by the Court on the part of a Union or (more commonly) Member State actor fell drastically during the height of the refugee crisis and the five years that followed, as compared with the previous five and a half years. From 2012 through 2014, infringements were found in about six out of every ten legal issues annually, as compared with between two and four out of ten after the crisis. The average frequency of violations was 51 per cent among the totality of issues decided prior to the outbreak of the crisis, and 29 per cent among those decided after it. The effect is clear and dramatic.

Figure 1. Issues where a violation of EU law by a Member State or EU executive or legislative body was found, expressed as a fraction of the total number of issues per year. Numbers indicate the absolute number of issues in each category.

These figures strongly suggest that the Court responds with restraint not just when faced with a crisis or in its immediate aftermath, but in the longer run following it besides. This is particularly remarkable considering that the number of Member State infringements, in the Commission’s assessment, increased from 2014 on.⁷²

A similar pattern was found when examining the legal issues contained in the SepaRope Judiciary Dataset that gave rise to questions concerning the compatibility of legal acts by the Union or its Member States with higher-ranking EU norms.⁷³ In the case of legal issues decided before the acute phase of the crisis, these questions were decided mainly in favour of the higher-ranking EU legal rule, with the measure under review being upheld in only 46 per cent of the issues. After the outbreak of the crisis, by contrast, the reviewed measure was upheld in a majority of the issues (58 per cent).

One group of Member States defied the pattern; in cases concerning Member States that joined the Union in the 2004 enlargement or later, violations of EU law were still found in 44 per cent of the issues decided after the crisis outbreak (to be compared to the overall average of 29 per cent in all issues decided during the same period). In practice, this observation is driven by the Visegrad group. As several of these states were among those most vocally opposed to taking in refugees, this is not surprising. It is also possible that the concurrent rule of law crisis in the same Member States affected the Court’s outlook.

These results suggest that the occurrence of a crisis has both immediate and long-term effects on the Court’s outlook, by pushing it towards greater caution and restraint. This finding also fits in well with the low standard of review applied by the Court to Union emergency measures, as explored in section 3 above. It furthermore demonstrates that the restraint shown by the Court in that case law is not limited to Union institutions, as most of the actors found to be at fault in the dataset examined here operated at Member State level.

Legal factors, such as legislative reform, cannot explain the suddenness of the change shown in Figure 1 or its correlation with the refugee crisis. Most of the cases decided in the months or even years after the crisis were pending before the Court prior to the onset of the crisis. They thus concerned events that predated the crisis, and they were governed by legal acts in force at the time and thus adopted well before the dramatic increase in the number of incoming refugees that coincided with the new approach taken by the CJEU. Legal changes prompted by the crisis should have had observable effects on the Court’s rulings until some time after the event, when the measures had entered into force and cases governed by the new legislation had made their way into court. Nor had there been any legislative reforms in the years leading up to the crisis that could explain a sudden fall in violations in late 2015. Moreover, no potentially paradigm-shifting rulings capable of explaining the pattern are identified in the literature. Indeed, since the cases examined here span the whole area of migration law, it is unlikely that one or even a few important precedents delivered in early or mid-2015 could have had the dramatic effects observed in the dataset.

Instead, the findings are best explained as an instance of strategic restraint on the part of the Court– ie, a politically (broadly understood) or at least non-legally motivated avoidance of controversy. This kind of judicial behaviour is frequently ascribed to courts in political science scholarship,⁷⁴ and previous research has convincingly demonstrated its operation in connection with rulings handed down by the Court in the face of a clear risk of legislative override.⁷⁵ In the specific context of the EU refugee crisis, Goldner Lang has conceptualised this behaviour as ‘judicial passivism’, understood as a judicial strategy in which ‘the Court consciously decides not to use its judicial power where it could.’⁷⁶ The stronger political charge of Court rulings in such a situation renders this explanation plausible, as does the lack of any legally convincing explanations. Furthermore, such behaviour matches a broader pattern of the Court as an actor keenly aware of its relationship to important political actors and anxious to avoid conflict with them, particularly on potentially controversial matters.⁷⁷

Understood in this way, the sharp decline in findings of Member State violations of EU law during and after the peak of the migration crisis indicates that, while the Court is reluctant (as shown in section 3 above) to take the same broad view on emergency competences at the national level as at the Union level, it still actively avoids entering into conflict with Member States in policy areas affected by crisis. This more pragmatic approach avoids laying down constitutional precedents that may be difficult to contain later.⁷⁸ Nevertheless, such rulings may have lasting effects on the substantive legal field.⁷⁹ As the data explored here demonstrates, such lasting effects do indeed arise. The present analysis cannot tell us whether such effects are the result of lenient precedents set during the crisis period and subsequently upheld, or whether instead they indicate continued caution on the part of the Court even as a given crisis abates.

C. Shifts in judicial trust

A second effect that can be observed in the dataset when comparing the Court’s behaviour before and after the crisis is what appears to be a shift away from trusting executive actors and towards trusting the Court’s judicial colleagues at the national level. Each legal issue in the dataset is coded for whether the Court expresses deference towards another actor.⁸⁰ Deference is understood as accepting another actor’s exercise of discretion, or leaving final assessment on the matter for another actor at a later stage of adjudication. For each issue where deference is detected, it is graded on a three-grade scale as minimal, partial, or full. It is also noted to whom the Court defers.

Analysis of the data shows that the Court remained about as likely to defer to other actors after the crisis as before it. A small difference can be detected in the degree of deference: minimal deference grew relatively more common after the crisis (rising from 29 to 55 per cent), at the expense of partial deference (which fell from 71 to 43 per cent). Full deference was very uncommon during both periods. While crisis thus does not appear to have had any major effect on the Court’s assertiveness or its willingness to share final decision-making power, the figures indicate that the Court became somewhat less willing to defer to other actors during and after the crisis, as compared with before. This pattern is particularly clear as regards southern and central European Member States, where the rate of partial deference was 80 per cent or higher before the crisis, but fell to only around 35 per cent during and after it. The ready interpretation of these figures is that the Court lost trust in the discretion of political actors in the Member States under most strain because of the crisis.

To whom, then, does the Court defer? A clear pattern on this point is shown in Figure 2. Before the crisis, the Court deferred final decisions in roughly equal measure to executive and to judicial actors (42 per cent and 50 per cent, respectively, of the cases where it deferred to any other actor at all). After the crisis, by contrast, it virtually stopped deferring to any actors outside the judiciary, with its deference to executive actors dropping to 14 per cent of the cases. This development occurred in equal measure in all Member States.

Figure 2. Issues on which the CJEU showed deference to an executive, legislative, or judicial actor, respectively, expressed as a fraction of the total number of issues on which it showed deference was expressed. N = 24 pre-crisis and 42 post-crisis.

Two explanations for this change would appear to be possible. The first, mirroring the strategic account offered above for the Court’s increased restraint during the crisis and its aftermath, suggests that the Court’s trust shifted at the time of crisis. The argument here is that the CJEU, while reluctant for strategic reasons to put itself on a collision course with the executive power in an area perceived to be in crisis, lost (or withdrew) its trust in executive discretion at the national level. Understood in this fashion, deferring less to the executive can be seen as a way to retain a measure of judicial control. This interpretation also finds some support in the increased incidence of minimal deference after the crisis. Crisis, in this narrative, made the Court less inclined to trust in the good judgement of the executive, while at the same time making it shrewder in choosing its battles. This indicates a difference between the vertical and horizontal relationships of the Court, as no similar reduction can be seen in its willingness to defer to executive discretion at the Union level.

Another and more legal explanation for this shift in the Court’s deference must be recognised, however. In practice, deference to a judicial actor here entails deference to national courts. In preliminary reference cases (which, as noted in section A above, form the bulk of the dataset), such deference can be expected as a consequence of the division of competences between the CJEU and the referring courts.⁸¹ While this division of competences has not been reformed, it is quite possible and perhaps even likely that the character of the cases changed after the crisis.⁸² Due to the relatively small number of issues where the Court showed deference, the results from breaking down this development by year are highly uncertain, and they should be approached with great caution. Nevertheless, the figures suggest that the shift in the Court’s ‘deference partners’ may have been somewhat less abrupt than that in its judicial restraint. If these indications are correct, the shift may have been triggered not by the outbreak of crisis in itself, but instead by the inflow of crisis-related cases. Such an increase can be expected to occur only with some delay after the actual events (since it takes time for cases to reach the Court), and such crisis-related cases may conceivably have traits that distinguish them from other migration cases and which justify showing greater deference to the referring court.

5. Two faces of judicial restraint

The question posed in this Article is how crisis affects the Court’s relationship to other branches of power and in particular to executive actors, who are the most likely ones to see their powers grow in times of crisis. Using the 2015 refugee crisis as a case study, I have explored this issue from two complementary perspectives: doctrinal and empirical. Both perspectives lead to similar results, namely that the crisis led the Court to adopt a strategy of judicial restraint or tolerance vis-à-vis Union and Member States’ executives. The analysis has also shown that this crisis-induced restraint has long-term effects on the policy area as a whole. At the same time, the empirical analysis showed that national executive actors were granted less discretion after the crisis than before. Only in relation to a small number of central and eastern European Member States, which for various reasons might have been considered at particular risk of evading their Union law obligations, did the Court retain a firmer stance after the outbreak of the crisis.

Taken together, these findings support three conclusions regarding the Court’s approach to crisis. First, crisis does affect the Court’s outlook on a legal area – and not just on the specific provisions for coping with emergencies, but also on its own role in the area more generally. Second, the Court’s reaction to crisis is one of deference and withdrawal. Rather than stepping up in the aftermath of an emergency in order to guard against a creeping expansion and possible abuse of executive power, the Court appears to respond by aligning its role with that of the executive. The result is a united front of public power in the face of crisis, but a decrease in judicial control of executive power. Both of these effects appear to extend beyond the immediate crisis, resulting in a lasting reallocation of powers both between the Union and Member State levels and between the executive, the legislative, and the judiciary.

Third, these effects appear to play out somewhat differently vis-à-vis actors at the Union and Member State levels, respectively. In cases involving the direct review of emergency competences, the generosity shown by the Court to the Union executive does not extend to the Member States. The greater leeway enjoyed by the Member States in times of crisis shows up instead in the Court’s assessment of substantive legal issues in the area affected by crisis, and it is checked by a shift in the Court’s trust from executive to judicial actors. Distinguishing between the intra-Union relations between the Court and other institutions (the horizontal dimension) and the relations between the Court and national actors (the vertical dimension), the findings indicate that the CJEU in the case of both relationships favours an interpretation of law that serves to increase executive competences. In the horizontal dimension, however, it also trusts the executive with extensive discretion, whereas in vertical situations it prefers to retain the legality assessment within the judiciary. Perhaps the two approaches can be described thusly: unsupervised authority for the Union executive, and authority under supervision for executive actors at the Member State level.⁸³

The CJEU’s attitude towards the Council’s and the Commission’s exercise of Treaty emergency competences increases the scope for political decision-makers to exercise discretion that is not or only superficially subjected to judicial supervision. In relation other institutions at the Union level, the Court’s case law radiates far-reaching trust – both in their assessments of the design of the specific measures under scrutiny and, more importantly, in their willingness and ability to exercise the same degree of sound judgement in the future. Had the Court feared the possibility of a future power grab, it would likely have set a more ambitious standard of review as a precedent for upcoming cases.

Where the CJEU’s relation to actors at the Member State level is concerned, a more complex picture emerges. As can be seen from the contrast between the rulings in Slovakia and Hungary v Council and Landkreis Gifhorn, the Court has not been prepared to extend the same margin of appreciation to national actors; and the quantitative analysis demonstrates a lower level of trust in Member State executives after the crisis. At the same time, it shows that the Court has nonetheless extended an olive branch, in the shape of a more forgiving ‘everyday’ application of migration law following the crisis. These findings can be reconciled as both indicating a relationship of permissiveness without trust; executive authority under judicial supervision. This suggests that the Court attempts to avoid controversy while at the same time exercising damage control; recognising the strain caused by crisis, the Court avoids unnecessary conflict with Member State executives by holding them to account only in a relatively small number of cases – presumably those where a violation is clearest or most troubling – while retaining within the judiciary the power to scrutinise executive (and legislative) measures on a continual case-by-case basis.⁸⁴ At the same time, the Court’s case law entails a move towards decentralised judicial review, in which power is retained within the judiciary but decentralised to the courts of the Member State.

6. Outlook and future research

While it would be unfair to claim that crises have been celebrated in EU law, there is a persistent narrative – dating back from Jean Monnet’s vision of a Europe ‘built through crises’⁸⁵ – of how one crisis after another has provided the impetus for further European integration.⁸⁶ As this Article has demonstrated, however, the effect of crisis is to lower not just the bar of political viability, but also that of judicial acceptance. Crisis leads, that is, not only to further European integration but to integration driven by other branches of government than those foreseen by the ordinary decision-making procedures – and thereby quite possibly to a different kind of integration. This should prompt us to reflection on the ‘flexible’ character of current EU decision-making in general and crisis management in particular – especially since the succession of crises plaguing the Union includes a rule-of-law crisis.⁸⁷

The conclusions in this paper are based on a case study of the 2015 migration crisis and therefore do not allow for far-reaching general conclusions. Future research should broaden the perspective to other crises that have shaken the EU in order to determine whether this behaviour is specific to the migration crisis, or typical of the CJEU’s crisis response in general. Furthermore, especially in light of the Court’s tendency to defer final assessment to national courts, it would be worthwhile examining to what extent national courts, particularly those at the apex of national judicial hierarchies, are more willing to exercise meaningful judicial review.

Nevertheless, the findings give cause for concern. The principle of institutional balance – which in EU constitutional law is the substitute for (or perhaps an adapted version of) the principle of separation of powers⁸⁸ – is generally defined as the requirement that all EU institutions act within their powers.⁸⁹ One might wonder, however, whether alongside this demand that they not exceed their prerogatives, such actors also have an obligation to use the powers conferred upon them to their full extent – at least in relation to each other.⁹⁰ When institutions start colluding, namely, they abandon their functions and undermine the separation of powers.⁹¹ In other words, the balance can be upset as surely when institutions fail to exercise their powers as when they seize new ones. The Court has proven itself to be a reliable ally of its institutional colleagues in their hour of need, but a functioning balance of powers should resemble less a friendly alliance than a jealous rivalry. If that rivalry is laid to rest, democratic rules are set aside.