1.This article examines how Swedish courts justify decisions not to refer questions to the Court of Justice of the European Union (CJEU). It is based on an analysis of 668 non-referral decisions across all court types and levels. By identifying the reasons provided, this article evaluates the explanations judges consider legitimate for nonreferral, and compares these with the legal obligation to give reasons. The findings reveal a nuanced picture of how Swedish courts approach reasons-giving within the preliminary ruling procedure (PRP), highlighting both commendable and questionable practices. Courts of last instance predominantly rely on the CILFIT exceptions, particularly the irrelevance of EU law, followed by acte éclairé, while acte clair is used only infrequently. Lower courts commonly resort to vague formulations such as “no reason to refer”. Such minimal reasoning may reflect reluctance to engage with the PRP, thought it may also stem from legalistic considerations. In any event, a few more words explaining why no questions were sent could go a long way in building trust in how the Swedish judiciary handle the PRP, both from the perspective of the parties and the Union legal order.s. 35
INTRODUCTION
1.The preliminary ruling procedure (PRP) is widely recognized as the cornerstone of the EU legal order, crucial for ensuring the uniform interpretation of Union law. Yet, its activation depends entirely on national judges. Despite its central role, we know surprisingly little about the cases where domestic judges choose not to refer questions to the Court of Justice of the European Union (CJEU). One way to investigate this legal phenomenon is by analysing the justification behind decisions not to refer. These explanations reveal two key insights. First, they shed light on what judges regard as legally valid grounds for withholding questions, whether in response to party requests or ex officio. Second, courts adjudicating at last resort are required to justify decisions not to refer. This allows their reasoning to be assessed in light of the legal obligation to give reasons.s. 35
2.Although the duty to provide reasons is most clearly established for courts of last instance, it is lower courts that handle the majority of cases. The landscape of judicial reasoning in this context would be incomplete without including these courts. Furthermore, it has been noted that courts of last instance “[do] not seem to be the most acute problem for effective judicial protection. The weak link … lies where lower national courts can discretionarily and without (effective) scrutiny refrain from making use of the preliminary reference procedure”.s. 36
3.While all national courts share the overarching responsibility to uphold effective judicial protection under EU law, it is important to distinguish between their respective obligations for the PRP. Courts adjudicating at last instance are legally required to refer questions and must provide explicit reasons when they choose not to. In contrast, lower courts may refer and face less stringent requirements to justify their decision.s. 37
4.This article investigates how Swedish courts have justified their decisions not to refer questions to the CJEU between 2013 and 2021, based on data that captures when referral is mentioned explicitly in judgments. It maps and categorizes the type of reasons provided, identifying recurring patterns and evaluating the legal soundness of the reasons-giving practice. In doing so, it provides insight into how the PRP functions at the national level and why some questions never reach the Luxembourg court.s. 37
5.Sweden represents a suitable case study for two reasons. To begin with, it was the first Member State to introduce legislation requiring apex courts to give reasons for non-referral, following an infringement procedure. Additionally, Sweden offers near-complete access to data where a party has requested a preliminary ruling or where judges have raised the issue on their own initiative, but no questions were sent. This type of access is rare throughout the Union, where data is often limited to last instance courts.s. 37
6.The mainstream view is that Nordic judges are reluctant to ask questions to the CJEU. Wind argues that Nordic legal traditions of weak judicial review and majoritarianism make courts cautious about requesting preliminary rulings. Sweden stands out in that it was the subject to an infringement proceeding concerning the conduct of its courts of last instance. According to the Commission, although the supreme courts were bound by the obligation to refer, they had hidden behind formal admissibility assessments and thereby disregarded Article 267(3) TFEU, by failing to request preliminary rulings when refusing leave to appeal. Also, the lack of reasoning for the non-referral decisions, when declaring the case inadmissible, made it “impossible for the individual to verify whether the obligation under [Article 267(3)] has been complied with”, and “… not possible for the Commission as guardian of the Treaties to control that Article [267(3)] is complied with”.s. 37
7.To avoid being brought before the CJEU and the risk of a finding of a treaty violation, the Swedish Parliament enacted legislation requiring courts of last instance to provide reasons when deciding not to refer. By this move, the infringement action was dropped, and it was hoped that not only would the national legislation increase the transparency, but also encourage more referrals from Sweden’s apex courts. However, in an interview with Swedish Radio, the former Supreme Court President Bo Svensson remarked, “…it does not matter if the rules are written down. The Supreme Court will continue to act as it always has”. Former Minister of Justice Thomas Bodström agreed, adding that “there won’t be any change in how the Supreme Court acts. But we might avoid criticism on this issue”.s. 38
8.The upfront statements to the Swedish media reflect the alleged reluctance of Swedish courts to engage with the PRP. More than two decades after the Commission’s 2004 infringement action, and the subsequent introduction of a requirement for courts of last instance to give reasons for non-referral, it remains to be seen how the Swedish judiciary provides such reasons.s. 38
DATA AND METHOD
1.The dataset for this study consists of 668 cases in which Swedish judges explicitly decide not to refer questions to the CJEU, most often following a request from one of the parties. To examine the practice of reasons-giving, it is important to identify instances where either a party requested the national court to refer, or where the judges explicitly decided against referral on their own initiative.s. 38
2.The data relied upon were collected from Norstedts Juridik’s database JUNO that includes most of the national court judgments from 2013 to present; however, this study limits the data collection period to 2021. A set of keywords was used to identify relevant cases, and subsequently false positives were removed, and identical cases collapsed to one representative observation.s. 38
3.Although the dataset offers a novel and nearly complete view of the reasonsgiving practice in a Member State within the context of the PRP, it is not without limitations. There is a gap regarding cases where applications for leave to appeal were denied by courts of last instance. These decisions are difficult to systematically capture because they are not indexed on requests for preliminary rulings. There are clear indications that such decisions tend to be less thoroughly reasoned than those where leave to appeal is granted. Consequently, the bulk of data from the two supreme courts derives from judgments, while decisions made during the leave to appeal stage remain largely unrepresented.s. 39
4.The coding of types of requests was conducted in two stages: categories were constructed during an initial review of all cases and subsequently refined and recorded a second time.s. 39
5.Theil highlights how “legal language is open-textured and responsive to subtle shifts that can be easily missed or open to competing interpretations”. Classifying how national judges give reasons is challenging. I constructed the categories from the lines of reasoning identified in the judgments, and attempted to stay as close as possible to the type of reasoning the text conveyed. Although this arguably achieves a high level of validity, by capturing the explicit form of reasoning presented, it remains impossible to be entirely certain that what was inferred from the legal reasoning aligns with what the individual judges had in mind.s. 39
6.The main challenge, however, is one of reliability. This is common in legal research, as legal assessment often involves terms that are ambiguous and open to different interpretations, as described in the quote by Theil. Having said that, most statements were relatively straightforward to classify, while a few were difficult to comprehend. This is a finding in itself. Twelve observations were excluded owing to challenges in makes sense of the reasoning provided.s. 39
7.Furthermore, to accurately capture how courts provide reasons in the context of non-referral, it is insufficient to extract only the specific string referencing the decision not to refer. Hence, the entire judgment must be examined, as relevant reasoning may be implicit or elaborated elsewhere in the merits. In light of the European Convention on Human Rights (ECHR), the quality of the party’s request largely steers what is expected of the national judges in terms of the level of detail required in the reasoning. Accordingly, a brief, onesentence justification may not be problematic where the request itself is poorly substantiated and the decision not to refer is accompanied by further reasoning in other parts of the merits.s. 40
8.The resulting typologies provide a useful framework for making sense of the practices of national courts. They are structured at two levels: a general category (for example, acte éclairé), and a set of corresponding sub-categories, such as explicit reference to a CJEU case, implicit reference elsewhere in the merits, or no clear reference.s. 40
9.Central to the analysis are the typologies that emerge, as they represent the general patterns of reasons provided, which in turn can be assessed in relation to the legal requirements for reasons-giving. This helps identify commendable examples and the lines of reasoning that should be avoided. In contrast, the ambition is not to assess the ‘correctness’ of each and every reasoning in the context of the cases. Similarly, this article does not focus on whether Swedish courts of last instance correctly apply the CILFIT exceptions. Instead, it distinguishes between the formal act of providing reasons and the substantive adherence to the legal obligation to refer, although this distinction is not always clearcut. While the reasoning may formally satisfy the requirement to provide reasons, explaining why no questions were sent in light of the CILFIT criteria, it may nonetheless fall short of, for example, that the correct interpretation of an EU act is beyond any reasonable doubt. This allows for criticising, for example, courts of last instance providing reasons outside the CILFIT criteria, but not the actual application of Article 267 of the Treaty on the Functioning of the European Union (TFEU).s. 40
THE FINDINGS
1.3.1 How Swedish courts give reasons for non-referral Table 1 (see below) outlines the landscape of judicial reasoning within the PRP from the Swedish courts, distinguishing between different court levels under Article 267 TFEU. Courts adjudicating at last instance are generally required to refer questions to the CJEU and must provide reasons when invoking exceptions to this obligation. These exceptions – based on the CILFIT criteria, irrelevance, acte éclairé, or acte clair – require a clear demonstration of how they apply to the case at hand. In contrast, lower courts, while also obliged to refer in cases involving doubts about the validity of Union law, otherwise enjoy the broadest discretion in deciding whether to refer. Their reasons-giving requirements are less clearly defined, having a lower bar than apex courts. The types of reasoning illustrated in Table 1 range from those firmly rooted in EU law to others that rest on uncertain footing. The following parts evaluate the practice of reasons-giving across each category. 3.2 Irrelevancy Article 267(2) TFEU specifies that lower courts may refer “if it considers that a decision on the question is necessary to enable it to give judgment”. Similarly, under Article 267(3) TFEU and the exceptions developed in the case law of the CJEU, last instance courts are exempted from their obligation to refer if the interpretation of EU law is irrelevant to the outcome of the case. Advocate General Emiliou specifies that “‘necessity’ must be understood as the ability of the issue to influence the outcome of the case (put very simply: who wins, who loses and why).” There is a presumption of relevance for questions referred, although it can be rebutted for those that are hypothetical.s. 41
2.Irrelevance of Union law is frequently cited by Swedish courts as a reason for non-referral. It accounts for 28% of all observations and stands out as the most common justification among courts of last instance, where it comprises 40% of all stated reasons. This ground is fully legitimate under EU law, and the reasons provided do not have to be more complicated than this. Its prevalence suggests that Swedish courts often, in response to a party request, highlight that EU law is irrelevant to the outcome and hence that there is no need for a preliminary ruling. When national courts resolve the case without addressing EU law, referral becomes moot.s. 43
3.There are several commendable examples that demonstrate this approach clearly. For instance: “[t]he issue raised in the case concerning the interpretation of Article 132(1)(m) of the VAT Directive is not relevant to the assessment of the case. There is no reason to request a preliminary ruling”, “[t]he administrative court of appeal considers that the question of EU law is irrelevant to the outcome of the case …”, or “[g]iven the outcome of the case, there are no grounds to examine the request to obtain a preliminary ruling …”.s. 43
4.Although this exemption is valid, it risks becoming an easy escape from the obligation to refer if approached too lightly. Anthony already warned in 1966 that the most convenient way for courts to sidestep the preliminary ruling procedure would be to ‘circuminterpret’ EU law and fall back on national law instead. Hence, there may be some legal gymnastics in play within this category to avoid referral. At the same time, it might also be that the claims of the parties tend to be abundant.s. 43
5.However, not all observations in this category clarify whether the court is applying the intended assessment of irrelevance, i.e. determining that EU law is irrelevant to the outcome – or merely using the term “unnecessary” in a more casual sense, implying that a preliminary ruling is unnecessary even though EU law is still relevant. Craig and de Búrca highlight how “… Article 267 TFEU does not provide that the reference must be necessary, but that a decision on the question is necessary to enable the national court to give judgment”.s. 43
6.The observations belonging to the category of “unnecessary to refer” typically take the form of statements such as: “a preliminary ruling is not necessary to decide the case” or other vague formulations such as: “the administrative court notes in this regard that, while it is indeed possible to request a preliminary ruling …, it considers that the question in the case is not of such a nature that it is necessary to request a preliminary ruling”.s. 43
7.In this instance, EU law may indeed be relevant to the outcome, yet the court’s reasoning does not make that assessment transparent. While such a line of reasoning is uncommon among courts of last instance, it occurs more frequently in lower courts. Nonetheless, courts of last resort tread on thin ice when they assert, for example, that “the interpretative questions in the case do not concern EU law in such a way that the Labour Court is required to request a preliminary ruling”, which is ambiguous. 3.3 Acte éclairé and acte clair Among the more common justifications for non-referral by courts of last instance are the core CILFIT exceptions, particularly acte éclairé and, to a lesser extent, acte clair, alongside irrelevance. This is unsurprising, as apex courts are required to anchor non-referral in one of these three established exceptions. In contrast, lower courts more sporadically rely on existing case law as an explicit reason not to refer, or assert that the interpretation of a Union act is clear.s. 44
8.A straightforward reason for non-referral is citing existing CJEU case law that addresses the same legal questions or points of law, and applying it to the facts at hand. This approach appeared in 31% of observations for apex courts and 9% for lower courts. National courts are expected to apply the case law of the CJEU and, although they are always welcomed to ask the Court to update its jurisprudence, it makes sense not to refer in these scenarios. However, the level of detail varies. Ideally, courts should explicitly reference the relevant CJEU judgment and explain how it is applied, thereby showing that the CJEU has already provided the answers. The next best approach is to make an implicit reference, that is, to state that relevant CJEU case law exists without specifying which, in relation to the decision not to refer, while subsequently applying CJEU case law elsewhere in the merits. Lastly, in a few observations, national courts state acte éclairé without touching further on the topic, not even implicitly.s. 44
9.Hence, acte éclairé is not always relied upon with the rigour required of courts of last resort, but there are good examples to highlight, such as the following: In total, there were fewer explicit references to CJEU case law by courts of last instance (12 in total), compared to 33 implicit references, and four observations containing no clear reference.s. 44
10.Acte clair is without doubt the CILFIT exception that has generated the most discussion of the three in the literature. When it was introduced, it was warned that it would provide an easy escape route for courts of last instance to avoid asking questions to the CJEU. To mitigate this risk, the CJEU set the threshold very high, making it exceedingly difficult for national apex courts to satisfy. Hence, the exception was slightly reformulated in Consorzio Italian management, but not by much. To conclude that the correct interpretation of Union law is so obvious as to leave no room for any reasonable doubt, apex courts must go through a number of steps. They must assess that the interpretation is “… equally obvious to the other courts or tribunals of last instance of the Member States and to the Court of Justice”, compare language versions, bear in mind that EU law contains autonomous concepts, and consider the context as well as the telos of the provision and Union law as a whole.s. 45
11.When relied upon by the Swedish courts, they seldom make it explicit that they have performed the checklist outlined by the CJEU. In this regard, there were only two cases in which courts of last instance made it clear that they had consulted language versions other than the Swedish one. Hence, the last instance courts should do a better job of communicating that they approach this exception with the required rigour and explain ‘why’ the EU law in question is so obvious beyond any reasonable doubt, viewed through the lens of the CILFIT criteria. As things stand, the results indicate that acte clair is not handled with the diligence needed. At the same time, the courts of last instance rarely rely on acte clair alone: this occurred in only nine cases, not counting the observations comparing language versions. Otherwise, it is used by the two apex courts in combination with references to existing CJEU case law.s. 45
12.The Swedish Supreme Court seems fond of the statement that “the content of Union law is clear or has been clarified”. It is usually possible in these observations to connect the dots through references to CJEU case law in the merits, but the statement itself seems low effort and does not really specify why these two CILFIT exceptions were applicable. In this regard, there is strong reason to endorse Bernitz’s position, who, in light of the legal requirements, has called for the apex courts to abandon this practice and instead “always state in their reasoning the grounds on which they have found the EU legal position to be so clear or clarified that there is no room for reasonable doubt”.s. 46
13.The practice of other courts also includes references to two, or all three, CILFIT exceptions. The Patent and Market Court of Appeal for example, highlights how “EU law is clear or has already been clarified through the case law of the [CJEU]. It is therefore not necessary to request a preliminary ruling in order to decide the case”. The observations combining acte éclairé and acte clair are not problematic per se, although the acte clair doctrine is supposed to represent the last exception when EU law is relevant, and there is no prior judgment from the CJEU. The main issue however, lies in the failure to explicitly reference the CJEU ruling relied upon, to explain how that judgment applies to the case, and to articulate the steps of the acte clair test when concluding that the interpretation of EU law leaves no room for reasonable doubt. To highlight when this is done effectively, “whether the coaching position in question falls within the scope of these provisions can be determined based on the wording of the provisions and the case-law of the [CJEU] …” where the court also compares the Swedish and English language versions, and explicitly references the relevant CJEU judgment.s. 46
14.The main area identified for improvement is the tendency to rely on generalised statements that fail to explain how the CILFIT criteria were applied in the specific case. Superficial, checkbox-style references to the acte éclairé or acte clair exceptions do not meet the legal requirement as set out in Consorzio Italian Management, to demonstrate their applicability.s. 46
15.However, it is not only courts adjudicating at last instance that explain nonreferral by reference to existing CJEU judgments or to the clarity of Union law beyond any reasonable doubt. Lower courts, which are not bound by the same strict standards, also do so. Broberg and Fenger warn against setting the bar too high with them leaning into the CILFIT criteria, designed for courts at last instance. There are a few examples in which lower courts explicitly reference CILFIT, but the main issue does not appear to be that lower courts are setting the bar at the same level as courts of last instance. There are not many examples where the lower and intermediate courts explicitly address the test established by the CJEU for acte clair, with only three observations comparing language versions out of 44. A typical reasoning establishes that “the administrative court finds that the meaning of Union law regarding the issue of R&D services, as governed by Directive 2004/18/EG, is not so unclear as to warrant a request for a preliminary ruling …”.s. 47
16.It would be good practice also for lower courts to explain how the interpretation of EU law leaves no room for reasonable doubt. Without requesting a preliminary ruling, they risk misapplying EU law. In this way, concluding acte clair without communicating how that conclusion was reached is suspect, and can be perceived as arbitrary in light of well-argued requests from the parties to the contrary. 3.4 EU Law is valid An additional point of interest concerns what is largely missing from the general trends observed. The landscape of reasoning shows that nearly all observations concern the interpretation of EU law, while issues of validity are almost completely absent. All national courts have an obligation to refer, when doubting the validity of Union law, and it may therefore be expected that such issues are referred and, as a result, do not appear in the non-referral data. At the same time, it is legally acceptable for national courts to disagree with the parties when deeming EU law to be valid, and applying it without a reference.s. 47
17.In the dataset, there are two observations in which courts explicitly decline to refer while providing reasons to the validity of Union law, while there are five cases in total in which parties challenged the validity of Union law. An example from the first type is reflected well with reference to Foto-Frost and IATA and ELFAA, in which the national court concludes that: By contrast, another observation, though legally acceptable in its reference to CJEU case law, appears to conflate the facultative approach to questions of interpretation and the obligation for questions of validity. It states that: “according to Article 267 … the administrative court may request a preliminary ruling from the [CJEU] if it is necessary in order to assess the validity of the Commission’s decision”. The applicant had clearly specified that it challenged the validity of the Commission’s decision, whereas the national court appears to approach the matter as one of interpretation.s. 48
18.Two main takeaways follow. First, there may at times be a fine line between questions of interpretation and validity. Second, observations addressing validity are rare when compared to the overall range of reasons provided for nonreferral. 3.5 Other reasoning not clearly anchored in Article 267 TFEU While the previous categories represent legitimate grounds for non-referral, albeit sometimes reflecting weak reasoning, the “other” category includes typologies that are questionable per se, even when the specific reasoning may be extensive. These observations come mostly from lower courts, which account for 58% of the reasons provided by them, compared to 18% for courts of last instance.s. 48
19.The dominant sub-category is the vague assertion that the court “finds no reason to refer”, appearing in 199 observations, of which 195 originate from lower courts. This typically involves uninformative statements such as “the administrative court of appeal finds no reason to request a preliminary ruling”.s. 48
20.On its own, this type of reasoning offers little insight beyond acknowledging that a party requested a referral. It fails to explain why the court declined to refer, leaving both parties and the public to guess at the reasoning behind the decision. Yet, consulting the judgments as a whole, several observations include implicit reasoning elsewhere in the merits, which helps clarify the decision not to refer. The less common category “in light of the above, no reason to refer” is similar in that judges conclude that there is no reason to refer in light of what has been stated already in the case. In some observations, the limited – close to nothing – explanation may be understandable when party requests are vague, lack specific questions or arguments for why referral is necessary. However, the point should be made that judges ought to be transparent and briefly explain why no referral was needed even when party engagement is low. Moreover, this group also includes cases where requests were well-substantiated. Generally, this type of reasoning warrants criticism.s. 49
21.Even courts of last instance rely on this approach in four observations,“finding no reason to refer”. Although rare, the Land and Environmental Court of Appeal and the Migration Court of Appeal for example states that “… there are no grounds to request a preliminary ruling from the [CJEU]”, “… there is no reason to request a preliminary ruling from the [CJEU]”, without offering further explanations. I consider these to fall short of the standard expected from courts adjudicating at last instance.s. 49
22.Another form of reasoning is the hybrid of “no such question raised + no doubts/unnecessary”, which can be located at all court levels. Article 267 TFEU provides that “[w]here such a question [interpretation of the treaties, or interpretation or validity of EU acts] is raised before [a national court] …”, then that court may request a preliminary ruling, or for courts of last instance, shall request a preliminary ruling. AG Ćapeta interprets the notion of a question being raised as equivalent to its relevance to the outcome of the case. However, a more common approach is to divide it into two steps, first, whether EU law has been raised, either by the court or by the parties, and second, whether it is relevant to the outcome of the case. Generally, this category consists of observations that are difficult to grasp and share common characteristic, justifying their treatment as a separate category.s. 49
23.The formulations are vague, such as “… the case has not given rise to any question that casts doubt on the interpretation of Union law and would necessitate a preliminary ruling in order to decide the matter”. Such phrasing is muddled and unclear and it seems to be a standard formula copied across judgments. Further confusion arises when “no question raised” is combined with assertions that the interpretation of EU law is beyond any reasonable doubt. Either a question of EU law is raised, or it is not – there is no middle ground in which no question is raised and that there is simultaneously no unclarity for the question of Union law.s. 49
24.Lower- and intermediate courts also found the reason not to refer within the national legal system, mostly by pointing to conclusions from apex courts that these had previously opted not to refer. As an example: “in light of the Supreme Administrative Court’s position in the judgment HFD 2014 ref. 14, the administrative court of appeal finds no grounds to request a preliminary ruling …”. Although this practice makes sense from a national judicial hierarchy perspective, it is important to remember that, in matters of EU law, the CJEU is the court to follow. National courts must make independent decisions on whether to refer, regardless of the position of apex courts. It is a better practice to reference the specific reasoning provided by the court of last instance, for example the same CJEU case-law, rather than simply stating that referral is unnecessary based on a higher court’s decision. This is reflected in the following observation: “[A]lthough it is the CJEU and not the Administrative Supreme Court that serves as the ultimate interpreter of EU law, the Court of Appeal’s position is that the CJEU’s case law [referenced by the Swedish Supreme Court] … was sufficiently clear …”.s. 50
25.Moreover, there are 13 cases where no reason is provided at all after requests from parties. While the initial reaction may be one of concern over arbitrariness, it should be noted that courts of last instance do not have to refer, based on strictly formal “grounds of inadmissibility specific to the procedure before that national court or tribunal, subject to compliance with the principles of equivalence and effectiveness”. The four observations from the apex courts either dismiss the case, or remand it to the lower court. These seem not to raise too much concern, but for clarity, it would be advisable to explicitly reject the party claim for referral as is done in other similar observations.s. 50
26.There are also a few cases where lower courts limit the reasoning to state that they have no duty to refer and that the EU law in question needs no clarification from the CJEU. For example, This line of reasoning is problematic. The direct actions available for individuals are limited and it is notoriously difficult to be granted standing under Article 263(4) TFEU. Hence, the idea is that national courts, and the PRP, provide an indirect route to Luxembourg in the “complete system of legal remedies”. If this route is cut off by national courts along the lines of reasoning above, it signals potential issues for effective judicial protection.s. 50
CONCLUSION
1.The findings present a nuanced picture of how Swedish courts approach reasons-giving under the PRP. Rather than revealing a uniform pattern, the practice reflects both commendable and questionable examples.s. 51
2.At the level of courts of last instance, judges typically rely on the CILFIT exceptions, as expected. However, the depth and clarity of the reasoning vary considerably, bearing in mind the CJEU’s requirement that courts of last instance must ‘show’ how one of the three exceptions applies.s. 51
3.Among the exceptions, irrelevancy, acte éclairé, and acte clair, the most frequently cited by courts of last instance is the irrelevance of Union law to the outcome of the case. By contrast, acte clair is rarely invoked, either alone, or in conjunction with acte éclairé. This is a noteworthy finding, as acte clair is often portrayed as the most problematic exception. When relied upon, it requires careful substantiation due to its inherent risk of arbitrariness.s. 51
4.Irrelevancy, by contrast, accounts for 40% of all observations at the level of courts of last instance. National courts may only refer questions that are relevant to the resolution of the case. At the same time, this exception can easily be used to hide from the duty to refer.s. 51
5.Lower courts similarly justify non-referral on the basis of irrelevance, suggesting that EU law is not pertinent to the resolution of the case. For a large part however, they rely on vague formulations such as “no reason to refer”, which provides little to no insight unless implicitly supported elsewhere in the judgment.s. 52
6.These practices can be read in different ways. Taken together, the broader picture shows that Swedish courts rely on a range of justifications for non-referral. Some are fully legitimate, such as the irrelevance of Union law or reliance on existing CJEU case law. Others, however, raise concerns about the quality of judicial reasoning and, potentially, the proper application of Article 267 TFEU and gaps for effective judicial protection.s. 52
7.Instances of minimalistic reasoning, often found in lower courts, particularly when not accompanied by implicit clarifications in other parts of the merits, could be read as fitting the narrative of ‘the reluctant Nordics’, as they leave the underlying reasoning for non-referral unclear. At the same time, it is equally plausible that such practices reflect other potential explanations, such as the facultative nature of Article 267(2) TFEU, the weakness of the parties’ requests, or even time constraints.s. 52
8.Similarly, when courts of last instance invoke acte clair without explaining how that criterion is met, this may suggest a reluctance to engage with the PRP. However, when they rely on existing CJEU case law, which is the more common of the two justifications, it becomes more difficult to criticise them on normative grounds for non-referral. Similarly, the justification that EU law is irrelevant can be read in two ways: as a legally sound conclusion, or as a fallback on national law that masks an underlying reluctance. Generally, the practice of giving reasons can potentially be explained by a variety of factors, rather than solely by the narrative of the ‘reluctant Nordics’.s. 52
9.Regardless of the underlying motives, Sweden was subject to an infringement proceeding more than 20 years ago concerning the alleged reluctance of the two Supreme Courts. The solution identified was to require courts of last instance to justify decisions of non-referral, an obligation that today also follows from both the ECHR and EU law. This would, it was hoped, put an end to non-referral decisions made in the dark.s. 52
10.This concern echoes Potter’s observation that “[i]f there is any truth in the aphorism that justice must not only be done but seen done, then a decision without reasons given must be regarded as undesirable, because it must be suspect since it may be arbitrary”. In this light, the often minimalistic reasoning provided by Swedish courts in non-referral decisions risks undermining confidence in the proper application of the PRP and may fail to provide parties with sufficient justification for rejecting their requests. It must be emphasised that individuals have a right to a reasoned decision of non-referral, at least in relation to courts of last instance.s. 52
11.A few additional words from the Swedish courts could go a long way in building trust in how they handle the PRP. This could remove suspicions of reluctance to engage in dialogue with the CJEU, and instead show that many are the acceptable situations for not engaging in dialogue with the Luxembourg court.s. 53