OPINION OF ADVOCATE GENERAL
ĆAPETA
delivered on 6 July 2023(1)
Case C‑122/22 P
Dyson Ltd,
Dyson Technology Ltd,
Dyson Operations Pte Ltd,
Dyson Manufacturing Sdn Bhd,
Dyson Spain, SLU,
Dyson Austria GmbH,
Dyson sp. z o.o.,
Dyson Ireland Ltd,
Dyson GmbH,
Dyson SAS,
Dyson Srl,
Dyson Sweden AB,
Dyson Denmark ApS,
Dyson Finland Oy,
Dyson BV
v
European Commission
(Appeal – Energy – Directive 2010/30/EU – Indication by labelling and standard product information of the consumption of energy and other resources by energy-related products – Commission Delegated Regulation (EU) No 665/2013 – Energy labelling of vacuum cleaners – Testing method – Non-contractual liability of the European Union – Sufficiently serious breach of EU law)
I. Introduction
1. This case is the latest step in the judicial saga between the vacuum cleaner manufacturer Dyson Ltd, as well as the other appellants that are part of the same group of companies (which I will refer to collectively as ‘Dyson’), and the European Commission relating to the testing methods for the energy labelling of vacuum cleaners under EU law. (2)
2. It concerns an appeal brought by Dyson against the judgment of the General Court of 8 December 2021, Dyson and Others v Commission (T‑127/19, EU:T:2021:870; ‘the judgment under appeal’).
3. By that judgment, the General Court dismissed Dyson’s action for non-contractual liability of the European Union, brought under Article 268 TFEU and the second paragraph of Article 340 TFEU. Dyson claimed to have sustained damage as a result of the Commission’s adoption of Delegated Regulation (EU) No 665/2013 concerning the energy labelling of vacuum cleaners. (3) The General Court held that the condition concerning the existence of a sufficiently serious breach of EU law had not been met.
4. This case therefore concerns the concept of sufficiently serious breach, which is a condition for both the liability of the European Union and that of the Member States in damages for breach of EU law. It requires the Court to clarify the method for assessing the seriousness of the breach. More specifically, the present case raises the following important questions. Does it really matter whether the EU institution concerned – in this case, the Commission – had a discretion to act in the relevant field or whether it was left with very limited or no discretion in that regard? How can it be determined whether the EU institution had a discretion? How should the Court use the factors set out in the case-law which might excuse the breach in question?
II. Background
A. The background to this case in a nutshell
5. Before venturing into the complex background to this case in more detail, a short introductory explanation is in order. By Delegated Regulation 665/2013, the Commission, among other things, determined the method for testing the energy efficiency of vacuum cleaners, which is necessary for awarding an energy efficiency label to such appliances. The method chosen was based on testing vacuum cleaners at the beginning of their initial use, when their receptacles just start filling with dust. I will refer to this method as ‘the empty bag test’.
6. In Dyson v Commission, (4) Dyson brought an action for annulment before the General Court, claiming that that method did not show the energy efficiency of vacuum cleaners ‘during use’, as required by the basic legislation (Directive 2010/30/EU (5)) which Delegated Regulation 665/2013 supplemented. After the initial judgment of the General Court was set aside on appeal by the Court of Justice, (6) which then referred the case back to the General Court, that latter Court annulled Delegated Regulation 665/2013. (7) The General Court considered that the empty bag test could not show the energy efficiency of vacuum cleaners ‘during use’. For that, it was necessary that the receptacles were already filled to a certain level.
7. It is further necessary to explain why Dyson was interested in challenging the empty bag test. Dyson produces and sells cyclonic or bagless vacuum cleaners. (8) In contrast to conventional bagged vacuum cleaners (9) which generate suction to pull dust and dirt off the floor and into a disposable bag, bagless vacuum cleaners, as their name suggests, do not use bags. A cyclonic vacuum cleaner uses centrifugal force to separate dust and debris from the air, thus eliminating the need for disposable bags.
8. This distinction is important when it comes to the testing methods of the energy performance of vacuum cleaners at the heart of this case. It is my understanding that the problem is as follows: when a bagged vacuum cleaner uses suction to pick up dust, that affects its performance; when the bag or receptacle is empty, air flows relatively easily through the machine’s filter and pores in the bag so suction is maintained; however, as the vacuum cleaner is used, the bag and the filter become clogged with dust, and there is a reduction in airflow. As a result, the vacuum cleaner’s suction power, cleaning performance and energy efficiency decline, as the vacuum cleaner consumes more power to combat clogging and to compensate for the loss of suction. In contrast, cyclonic vacuum cleaners do not have bags, so no clogging occurs. Suction power can be maintained whether the receptacle is filled or not. In other words, cyclonic vacuum cleaners perform just as well full as they do empty. That finding is not challenged by any of the parties to these proceedings.
9. In a nutshell, the crux of the matter lies in the fact that, as claimed by Dyson in this case, testing with an empty bag represents a vacuum cleaner’s maximum possible dust pick-up performance and thus conceals the fact that once the consumer uses the vacuum cleaner, bag clogging and suction loss occur for most traditional bagged vacuum cleaners. Therefore, cleaning performance and energy consumption of vacuum cleaners vary when tested using an empty bag versus when the bag is ‘dust-loaded’.
10. Even if the empty bag test as a ground for energy labelling was abandoned after the General Court’s annulment judgment, Dyson considered that it had suffered damages during the period in which Delegated Regulation 665/2013 was in force, that is, prior to its annulment. Dyson therefore brought an action for damages before the General Court.
11. Before discussing the proceedings before the General Court (D) and the procedure before the Court of Justice (E), I will explain, to the extent necessary to analyse Dyson’s claims, the legislation relevant to this case (B), along with the previous judgments of the Court of Justice and the General Court in this saga (C).
B. EU energy labelling and Delegated Regulation 665/2013
1. Overview of EU energy labelling and ecodesign legislation
12. Generally speaking, EU energy labelling legislation provides information to consumers about the energy consumption and environmental performance of products to help them make informed decisions. It goes hand-in-hand with EU ecodesign legislation, (10) which sets minimum energy efficiency requirements for product design in order to improve environmental performance. Together, EU energy labelling and ecodesign legislation help meet the European Union’s aim of preventing and mitigating climate change, (11) and such legislation is estimated to save consumers in excess of EUR 250 billion per year. (12)
13. EU energy labelling and ecodesign rules are currently in force for a number of products, (13) such as televisions, dishwashers, ovens, refrigerators, washing machines and tumble dryers. EU energy labelling provides information to consumers on the energy efficiency of the particular product, using a comparative colour scale from A (most efficient) to G (least efficient). Apparently, about 85% of consumers recognise the EU energy label, and use it in their purchasing decisions. (14)
2. The legislative framework of Directive 2010/30 and Delegated Regulation 665/2013
14. EU energy labelling can be traced back to Directive 79/530/EEC, (15) which was later replaced by Directive 92/75/EEC. (16) However, it was not until nearly 20 years later – with the adoption of Directive 2010/30 – that vacuum cleaners came to be included in this framework.
15. As indicated by Article 1(1) thereof, Directive 2010/30 established a framework for the harmonisation of national measures on end-user information, particularly by means of labelling and standard product information, on the consumption of energy and where relevant of other essential resources during use, thereby allowing end-users to choose more efficient products. (17)
16. As evidenced by the fact that the expression ‘during use’ is mentioned 15 times in Directive 2010/30, (18) that directive concerns energy labelling intended to reflect energy consumption while the products are being used. Indeed, this issue had been discussed during the decision-making process leading to the adoption of that directive, (19) and it was in fact the Commission, in its proposal for that directive, which introduced language to that effect in several provisions. (20)
17. It is important to note for the purposes of the present case that, at the relevant time, Directive 2010/30 established a general framework, defining the policy objectives and the roles and responsibilities of, inter alia, the Commission, the Member States, manufacturers and traders. Pursuant to Article 10 of that directive, the Commission was empowered to adopt delegated acts (within the meaning of Article 290 TFEU), setting out the energy labelling requirements for specific product groups, including vacuum cleaners.
18. The third subparagraph of Article 10(1) of Directive 2010/30 read as follows:
‘Provisions in delegated acts regarding information provided on the label and in the fiche on the consumption of energy and other essential resources during use shall enable end-users to make better informed purchasing decisions and shall enable market surveillance authorities to verify whether products comply with the information provided.’ (21)
19. Directive 2010/30 has since been replaced by Regulation (EU) 2017/1369. (22) That regulation maintains essentially the same scope as Directive 2010/30, but modifies and enhances some of its provisions to clarify and update their content, taking into account the technological progress achieved in recent years in respect of the energy efficiency of products. (23) It emphasises that testing methods should reflect real-life usage. (24)
20. Unlike other products, vacuum cleaners were not subject to EU energy labelling rules until Directive 2010/30, nor to EU ecodesign rules until Directive 2009/125/EC, (25) and on the basis of an extensive impact assessment, it was decided that the Commission would adopt the relevant regulations in tandem. (26)
21. Delegated Regulation 665/2013 established labelling requirements for most standard types of vacuum cleaners. (27) It obliged suppliers and dealers to comply with the labelling requirements for vacuum cleaners with effect from 1 September 2014. (28) The information to be included on such labelling had to be obtained by reliable, accurate and reproducible testing methods, as set out in Annex VI to Delegated Regulation 665/2013. (29) Paragraph 1 of that annex referred for that purpose to harmonised standards which were published in the Official Journal of the European Union.
22. In that respect, the Commission published references to the European Committee for Electrotechnical Standardisation (‘Cenelec’) Standard EN 60312-1:2013 (‘the Cenelec standard’), which concerns the methods for measuring the performance of vacuum cleaners for household use. However, section 5.9 of the Cenelec standard, which concerned a dust-loaded testing method, was not included. (30) I will refer to this method as ‘the section 5.9 dust-loaded test’.
23. As a consequence of this exclusion, for the purposes of applying Annex VI to Delegated Regulation 665/2013, the harmonised standard for calculating the annual energy consumption of vacuum cleaners was based on a testing method using an empty receptacle (in accordance with sections 4.5 and 5.3 of the Cenelec standard), and not on a testing method using a dust-loaded receptacle (pursuant to section 5.9 of the Cenelec standard).
24. Therefore, only the empty bag test could be used for measuring the energy efficiency of vacuum cleaners under Delegated Regulation 665/2013. (31)
C. Events leading to the proceedings before the General Court
1. Annulment proceedings against Delegated Regulation 665/2013
25. On 7 October 2013, Dyson Ltd brought an action before the General Court seeking the annulment of Delegated Regulation 665/2013 on the basis of Article 263 TFEU. In support of its action, Dyson put forward three pleas in law, alleging, first, lack of competence on the part of the Commission; second, failure to provide a statement of reasons for that regulation; and, third, infringement of the principle of equal treatment.
26. By judgment of 11 November 2015, Dyson v Commission (T‑544/13, EU:T:2015:836), the General Court dismissed that action, rejecting those pleas in law as unfounded. Dyson appealed against that judgment.
27. By judgment of 11 May 2017, Dyson v Commission (C‑44/16 P, EU:C:2017:357; ‘the Court of Justice’s annulment judgment’), the Court of Justice set aside that judgment in so far as the General Court had rejected the pleas alleging the Commission’s lack of competence and infringement of the principle of equal treatment.
28. Importantly, the Court of Justice found that the requirement, as follows from Article 1 and the third subparagraph of Article 10(1) of Directive 2010/30 that the information supplied to consumers must reflect energy consumption when the machine is in use, is an essential element of that directive, (32) within the meaning of Article 290 TFEU. Accordingly, it must be laid down in the basic legislation and cannot be delegated. Moreover, the Court of Justice held that, contrary to the General Court’s findings, to understand the expression ‘during use’ contained in the third subparagraph of Article 10(1) of Directive 2010/30 as referring to the actual conditions of use is not an ‘excessively broad’ interpretation of that provision, but the very meaning of that specification. (33)
29. Consequently, the Court of Justice ruled, in paragraph 68 of that judgment, that the Commission was ‘obliged, in order not to disregard an essential element of Directive 2010/30, to adopt in [Delegated Regulation 665/2013] a method of calculation which makes it possible to measure the energy performance of vacuum cleaners in conditions as close as possible to actual conditions of use, requiring the vacuum cleaner’s receptacle to be filled to a certain level, having regard nevertheless to the requirements concerning the scientific validity of the results obtained and to the accuracy of the information supplied to consumers, as mentioned in particular in recital 5 and Article 5(b) of the directive’. (34)
30. Upon referral back, by judgment of 8 November 2018, Dyson v Commission (T‑544/13 RENV, EU:T:2018:761; ‘the General Court’s annulment judgment on referral back’), the General Court annulled Delegated Regulation 665/2013.
31. In that judgment, (35) the General Court considered that it was apparent from paragraph 68 of the Court of Justice’s annulment judgment that, in order for the method adopted by the Commission to comply with the essential elements of Directive 2010/30, two cumulative conditions must be met. First, in order to measure the energy performance of vacuum cleaners in conditions as close as possible to actual conditions of use, a vacuum cleaner’s receptacle must be filled to a certain level. Second, the method adopted must satisfy certain requirements concerning the scientific validity of the results obtained and the accuracy of the information supplied to consumers. It was apparent from Article 7 of Delegated Regulation 665/2013 and the whole of the case file that the Commission adopted a method for calculating the energy performance of vacuum cleaners based on an empty receptacle. Therefore, the first condition was not met, which was sufficient for the General Court to conclude that the Commission had failed to have regard to an essential element of Directive 2010/30, without it having to rule on whether that method met the second condition.
32. In addition, the General Court pointed out that, if no method of calculation carried out with a receptacle filled to a certain level met the requirements concerning the scientific validity of the results obtained and the accuracy of the information supplied to consumers, it remained open to the Commission to exercise its right of legislative initiative to propose an amendment of Directive 2010/30. (36)
33. The General Court therefore struck down Delegated Regulation 665/2013 in its entirety, finding it unnecessary to rule on the third plea alleging infringement of the principle of equal treatment. (37)
2. Preliminary ruling proceedings concerning the interpretation of Delegated Regulation 665/2013
34. In the meantime, by judgment of 25 July 2018, Dyson (C‑632/16, EU:C:2018:599), the Court of Justice gave a preliminary ruling in response to the referral by a Belgian court of questions relating to the interpretation of Delegated Regulation 665/2013, as well as Directive 2005/29/EC on unfair business-to-consumer commercial practices. (38) The request had been made in the context of a dispute between Dyson Ltd and Dyson BV, on the one hand, and BSH Home Appliances NV (‘BSH’), on the other, concerning several labels describing the energy consumption of conventional (bagged) vacuum cleaners marketed by BSH under the Siemens and Bosch trade marks, including the EU energy label required by Delegated Regulation 665/2013. Dyson Ltd and Dyson BV argued, inter alia, that the use of those labels by BSH without specifying that they reflect results of tests carried out with an empty bag constituted an unfair commercial practice capable of misleading consumers.
35. In its judgment, (39) the Court held, in particular, that Directive 2010/30 and Delegated Regulation 665/2013 are to be interpreted as meaning that no information relating to the conditions under which the energy efficiency of vacuum cleaners was tested could be added to the EU energy label.
D. Proceedings before the General Court and the judgment under appeal
36. On 21 February 2019, Dyson brought an action for damages before the General Court, based on non-contractual liability of the European Union, pursuant to Article 268 TFEU and the second paragraph of Article 340 TFEU.
37. Dyson claimed the sum of EUR 176.1 million, including interest, or, in the alternative, the sum of EUR 127.1 million, including interest, and submitted that the Commission should be ordered to pay the costs. The Commission contended that the action should be dismissed and that Dyson should be ordered to pay the costs.
38. By the judgment under appeal, the General Court held that Dyson had failed to satisfy one of the main conditions for establishing EU non-contractual liability, namely the existence of a sufficiently serious breach, in respect of four alleged infringements of EU law on the part of the Commission. It therefore dismissed the action.
39. First, with respect to the infringement of Article 10(1) of Directive 2010/30, the General Court found that the Commission did not have any discretion allowing it to exceed the powers conferred on it by Directive 2010/30 (paragraph 36 of the judgment under appeal). Nevertheless, the General Court held that a lack of discretion was not sufficient to establish a sufficiently serious breach and that it was necessary to take into consideration certain other factors as set out in the case-law (paragraphs 37 and 38 of the judgment under appeal).
40. The General Court held that, taking into account the Court of Justice’s annulment judgment and the General Court’s annulment judgment on referral back, the application of Article 10(1) of Directive 2010/30 to the specific case of vacuum cleaners gave rise to certain differences of assessment, indicative of difficulties of interpretation with regard to the clarity and precision of that provision and, more generally, of Directive 2010/30 as a whole (paragraphs 40 to 45 of the judgment under appeal).
41. The General Court also held, with regard to the complexity of the situation to be regulated and whether the error made was inexcusable or intentional, that the Commission was entitled to take the view, without manifestly and gravely disregarding the limits on its discretion, that the section 5.9 dust-loaded test was not capable of guaranteeing the scientific validity and accuracy of the information supplied to consumers and to opt instead for the empty bag test which was scientifically sound. The General Court further observed that Article 11(1) of Directive 2010/30 limited the delegation of powers to the Commission for five years, so the Commission could not postpone the adoption of energy labelling rules until Cenelec adopted a dust-loaded testing method. Moreover, neither the European Parliament nor the Council opposed the adoption of Delegated Regulation 665/2013. Those circumstances served to confirm, according to the General Court, that the Commission’s infringement of Article 10(1) of Directive 2010/30 was excusable (paragraphs 94 to 97 of the judgment under appeal).
42. The General Court concluded that, due to those factors, as well as the technical complexity of the problems to be resolved and difficulties in the application and interpretation of the relevant texts, an administrative authority exercising ordinary care and diligence could take the view that it was at risk by deciding to use the dust-loaded testing method rather than the empty receptacle testing method. Accordingly, the Commission demonstrated conduct that could be expected from such an administrative authority and did not manifestly and gravely disregard the limits on its discretion (paragraph 97 of the judgment under appeal).
43. Furthermore, the General Court, relying on those findings, held that the Commission, by adopting the empty bag test, did not manifestly and gravely disregard the limits placed on its discretion or commit a sufficiently serious breach of the other three alleged infringements of EU law, namely the principle of equal treatment (paragraphs 105 to 112 of the judgment under appeal), the principle of sound administration and the duty to act diligently (paragraphs 116 to 119 of the judgment under appeal) and the right to pursue a trade or business (paragraphs 123 to 131 of the judgment under appeal).
E. Procedure before the Court of Justice
44. By an appeal lodged on 18 February 2022, Dyson requests that the Court set aside the judgment under appeal in its entirety, rule that the Commission committed a sufficiently serious breach of EU law and refer the case back to the General Court. Dyson also requests that the Court order the Commission to pay the costs.
45. In its response lodged on 22 June 2022, the Commission requests that the Court dismiss the appeal and order Dyson to pay the costs.
46. Dyson and the Commission also lodged a reply and a rejoinder on 22 August 2022 and 30 September 2022, respectively.
47. A hearing was held on 20 April 2023 at which Dyson and the Commission presented oral argument.
III. Analysis
48. Dyson raises seven grounds of appeal. The first four grounds of appeal concern the Commission’s alleged infringement of Article 10(1) of Directive 2010/30, with the first ground alleging mischaracterisation of Dyson’s pleas and the second, third and fourth grounds alleging misapplication of the legal test for a sufficiently serious breach as follows: the second ground is based on misapplication of the method for assessing the seriousness of the breach as regards the Commission’s lack of discretion; the third ground is based on an incorrect assessment of the legal test for a sufficiently serious breach as regards interpretative difficulties and the clarity and precision of the rule breached; and the fourth ground is based on an incorrect assessment of that legal test as regards regulatory complexity. The fifth, sixth and seventh grounds of appeal allege misapplication of the legal test for a sufficiently serious breach in connection with the alleged infringements of the principle of equal treatment, the principle of sound administration and the right to pursue a trade or business, respectively.
49. I will begin by addressing the first ground of appeal. In my view, the rest of the assessment depends on determining what was exactly the breach that Dyson claimed the Commission committed. The question whether the General Court erred in law in assessing whether that breach was sufficiently serious therefore hinges on that issue. As I am of the view that the General Court mischaracterised Dyson’s claim, many of the arguments by which it found that the breach was not sufficiently serious are irrelevant. I will, nevertheless, assess the other six grounds of appeal, alleging misapplication of the method for assessing the seriousness of the breach.
A. The first ground of appeal
50. By the first ground of appeal, Dyson complains, in essence, that the General Court mischaracterised its plea. Whereas it claimed that the Commission committed a sufficiently serious breach by adopting the empty bag test, the General Court, in paragraphs 52 to 82 of the judgment under appeal, instead wrongly addressed whether the Commission was entitled to reject the section 5.9 dust-loaded test. (40)
51. It is true that, in the judgment under appeal, the General Court devoted considerable attention to assessing whether the Commission was correct to discard the section 5.9 dust-loaded test. That assessment is relevant only if the claim to be addressed is that the Commission wrongly rejected that method. However, that was not Dyson’s claim. Its claim was not that the Commission wrongly excluded one possible dust-loaded testing method. Rather, Dyson claimed that the Commission committed a sufficiently serious breach by choosing the empty bag test.
52. This distinction is important, as it determines the level of discretion that the Commission enjoyed, which, in turn, is important for addressing Dyson’s other pleas before the General Court. Dyson does not deny that the Commission has discretion to decide on what type of dust-loaded testing method it adopts. On the contrary, according to Dyson, choosing the empty bag test was not within the discretion conferred on the Commission under the basic legislation.
53. The Commission argues that it was legally relevant for the General Court to evaluate whether the only other recognised and available testing method at the time of the adoption of Delegated Regulation 665/2013, namely the section 5.9 dust-loaded test, was a scientifically accurate alternative. When that regulation was being drafted, the choice, as determined by the Commission in its discretion in complex policy-making, was between empty bag and partially loaded testing methods, and nothing else; no partially loaded testing method, including the section 5.9 dust-loaded test, was scientifically reliable. Therefore, the Commission argues that it could be excused for adopting what was then the only scientifically reliable testing method – the empty bag test.
54. I cannot agree with the Commission. The unavailability of a scientifically viable partially loaded testing method does not give the Commission leeway to adopt the empty bag test. Therefore, the choice could not have been between the empty bag test and a partially loaded testing method. The choice was only between different partially loaded testing methods or no test at all.
55. In the General Court’s annulment judgment on referral back, the General Court considered that the use of the empty bag test required the amendment of the basic legislation. (41) At the hearing, however, the Commission explained that proposing a legislative amendment to Article 10 of Directive 2010/30 would have been an illogical and difficult course of action. Not only would such an amendment have taken a long time to prepare, but also it would have left Delegated Regulation 665/2013 out of sync with the ecodesign rules, since the EU energy labelling and ecodesign legal frameworks normally work hand-in-hand and reinforce each other with regard to energy-related products.
56. Inasmuch as this might be true, it cannot justify the choice of a testing method which is not possible under EU law, that is, the empty bag test.
57. Therefore, I consider that the first ground of appeal alleging that the General Court mischaracterised Dyson’s plea is well founded.
B. The second ground of appeal
58. By the second ground of appeal, Dyson complains, in essence, that the General Court wrongly found, in paragraphs 37 and 38 of the judgment under appeal, that the infringement of a non-discretionary requirement was not itself sufficiently serious, given the particular context of this case. According to Dyson, the ‘during use’ requirement was an essential element of Directive 2010/30, by which the EU legislature intended to limit the discretion of the Commission. The Commission was aware of the misleading nature of the empty bag test based on its own expert report and the stakeholder consultation responses, and that that test risked harm to the environment and precluded manufacturers from informing consumers of such risks. As Dyson emphasised at the hearing, the General Court found that the Commission had no discretion, so that should have been the end of the matter. (42)
59. The Commission submits that the General Court correctly applied the legal test for a sufficiently serious breach in relation to the Commission’s discretion. (43) As the Commission argued at the hearing, the Court of Justice should endorse the case-law of the General Court that there is no automatic link between the lack of discretion and the finding of a sufficiently serious breach and that the other factors set out in the case-law should be taken into account. The Commission further claimed at the hearing that while reduced or no discretion leads to the finding of a sufficiently serious breach where there is a clear and precise obligation under EU law, that is not the situation in this case.
60. Is the mere infringement of a rule of EU law that leaves no discretion to the EU institution concerned sufficient for the breach of that rule to be considered sufficiently serious, or should the Court nevertheless take into account other factors that may excuse that breach?
61. To my mind, the answer to that question does not follow clearly from the case-law. Equally important, the case-law does not give any guidance as to how to determine when an institution has discretion. Over the years, certain Advocates General (44) and scholarly literature (45) have pointed to the inconclusiveness of the case-law and have questioned the relevance of discretion as a criterion, since the assessment whether a breach is sufficiently serious depends on other factors in any event.
62. My analysis of the existing case-law suggests that discretion has a role in determining whether a breach of EU law can be characterised as sufficiently serious, but that that role is not decisive. In that light, by holding, in paragraphs 37 and 38 of the judgment under appeal, that the lack of discretion is not sufficient to justify the conclusion that there has been a sufficiently serious breach and that it is still necessary to verify whether there might be factors which can excuse the breach, the General Court did not err in law.
63. In what follows, I will briefly discuss relevant case-law relating to the liability of the European Union and that of the Member States. (46)
1. Pertinent case-law on the liability of the European Union and that of the Member States
64. As is well known, since the judgment in Bergaderm, (47) the case-law relating to the liability of the European Union and that of the Member States for damages resulting from a breach of EU law has developed harmoniously. Relying on its earlier judgment concerning State liability in Brasserie, (48) the Court recalled in Bergaderm that the conditions under which the European Union and the Member States may incur liability for damage caused to individuals by a breach of EU law should not differ. This is because the protection of the rights which individuals derive from EU law cannot vary depending on whether an EU or national authority is responsible for the damage. (49)
65. That means that the case-law concerning the conditions for establishing EU non-contractual liability, including a sufficiently serious breach, has developed in line with that concerning State liability. Thus, in Bergaderm, the Court repeated the formula which seemed to be clearly established in the earlier case-law on State liability, according to which the assessment of the seriousness of the breach depends on whether or not the Member State or EU institution enjoyed a discretion when allegedly acting in breach of EU law. Where the institution concerned has no discretion or only considerably reduced discretion, the mere infringement of EU law is considered to be sufficient to establish the existence of a sufficiently serious breach. (50)
66. That distinction between the existence and the non-existence of discretion followed from two cases which were formative for State liability. In the first case in which the Court established the principle of State liability, namely Francovich, (51) the Court did not even mention a sufficiently serious breach as a requirement for liability. (52) However, the circumstances of that case related to a situation in which a Member State had acted in an area of EU law in which it had no discretion. In that case, the breach consisted of the failure of a Member State to transpose a directive in time. While Member States can usually choose the measures by which they transpose a directive into their national law, they do not have any choice as to whether they will transpose a directive at all. In such a case, it is clear that the mere infringement of EU law (non-transposition of a directive) results in a Member State’s liability in relation to an individual who has suffered damage due to such a failure to transpose.
67. The situation would be different if the Member State had transposed a directive, but had done so incorrectly. Member States do enjoy a choice as to how to transpose a directive, so that that error might, in some circumstances, be excusable and would not result in liability. The first case in which the Court had the opportunity to decide how liability arises in a situation in which a Member State has been left with choices did not, however, concern the transposition of a directive, but rather a breach of a Treaty provision. In Brasserie, (53) the two Member States concerned had committed a breach of EU law by choosing to enact domestic legislative measures (regarding the definition of beer and ship registers, respectively), which still infringed the internal market rules as set out in the Treaty. Those Member States were in a situation in which they enjoyed a discretion as to how to regulate. Nevertheless, their discretion in choosing the appropriate measures was limited by those Treaty rules. Those circumstances resulted in the Court finding that where a Member State acts in a field in which it has a wide discretion, liability can arise only if the breach is sufficiently serious.
68. The reason why EU or Member State institutions do not simply incur liability for any breach of EU law, but only for one which is sufficiently serious, stems from the need to strike a balance between the protection of individuals against unlawful conduct on the part of the (EU or Member State) institutions, on the one hand, and the leeway that must be accorded to those institutions in order not to paralyse action by them, on the other. (54)
69. In the judgment in Brasserie, (55) the Court held that the decisive test for finding whether a breach is sufficiently serious in a situation in which a Member State enjoys a discretion is whether it manifestly and gravely disregarded the limits on its discretion. The same test has also been used since the judgment in Bergaderm to determine the seriousness of a breach committed by EU institutions. (56)
70. It was in relation to assessing whether a Member State manifestly and gravely disregarded the limits placed on its discretion that the Court, in Brasserie, (57) listed certain factors which may be taken into account. (58) Thus, for instance, on the one hand, the obscure or imprecise nature of the rule breached or the fact that the position taken by an EU institution may have contributed towards the omission might excuse the breach. On the other hand, the fact that the breach was deliberate, or that the breach persisted despite a judgment finding the infringement in question or settled case-law of the Court on the matter, were factors which would lead to the conclusion that the breach was sufficiently serious. In any case, factors such as those listed in the judgment in Brasserie are not exhaustive, as the case-law seems to suggest. (59)
71. It is important to observe that, in the judgment in Brasserie, the factors which may excuse the breach related to a situation in which the Member States concerned enjoyed a wide discretion.
72. Thus, at the time when the Court chose to align EU and State liability in the judgment in Bergaderm, it seemed that there were clear rules for determining the seriousness of the breach. If an EU institution had no discretion or only a very limited discretion, the mere infringement of EU law would, in itself, be sufficiently serious. If, on the contrary, the EU institution enjoyed a discretion, the breach would be sufficiently serious only if that institution manifestly and gravely disregarded the limits placed on its discretion. Depending on the circumstances of the particular case, the EU Courts were to take into consideration different factors, in order to determine whether the breach could or could not be excused.
73. Turning to the present case, if the rules set out in the case-law concerning the assessment of a sufficiently serious breach were indeed so clear cut, this would lead to the conclusion that the General Court erred in law, as it took into account several factors even though it found that the Commission had no discretion to choose the empty bag test.
74. However, the case-law, as it developed after Bergaderm, does not allow the conclusion that the distinction between the existence and the non-existence of discretion and the relationship between discretion and the seriousness of the breach is so clear cut.
75. In the section that follows, I will offer a classification and a brief analysis of that case-law.
2. Classification of the case-law after the judgment in Bergaderm
76. The case-law following the judgment in Bergaderm can be classified into four groups for the purposes of the present case: first, cases in which the EU Courts considered that there was no discretion and therefore established a sufficiently serious breach; second, cases in which the EU Courts considered that there was no discretion, or only a limited discretion, but after examining different factors, found that the breach was not sufficiently serious; third, cases in which the EU Courts found that there was a discretion but, after examining different factors, concluded that the breach was sufficiently serious; and fourth, cases in which the EU Courts found that there was a discretion and, after examining different factors, concluded that the breach was not sufficiently serious.
(a) The first group of cases
77. First, there is case-law in which the EU Courts found that the institution that had committed the breach had no discretion and that therefore liability automatically followed. There are such cases in relation to both EU (60) and State liability. (61)
78. Interestingly, some of those cases concerning State liability were not cases in which Member States had failed to transpose a directive in time, but rather cases involving the incorrect transposition of a directive, a situation in which Member States usually have a choice. Nevertheless, the Court has considered that, in those cases, Member States were not left with any discretion. The important factor for the Court was the clarity and precision of the rule breached, which left no doubt as to the result to be achieved by the Member States.
79. Applying such reasoning to the case at hand, it would be possible to conclude, as the General Court did, that the Commission did not have any discretion to adopt the empty bag test because the rule in the basic legislation requiring the testing of vacuum cleaners ‘during use’ was clear. However, the case-law relating to the first group suggests that liability should, contrary to what the General Court concluded, follow automatically.
(b) The second group of cases
80. There is also case-law that points in a different direction. The second group of cases concerns situations in which the EU Courts considered that there was no discretion or only a limited discretion in relation to the rule of EU law that had been breached, but nevertheless assessed other factors and found that the breach could be excused. Therefore, no liability arose. While such examples are difficult to find when it comes to State liability, (62) they are not uncommon in the case-law relating to EU liability. (63)
81. For example, the judgment in Holcim (Deutschland) v Commission (64) concerned the reimbursement of bank guarantee charges that the applicant had paid in lieu of fines imposed by a Commission decision in the field of competition law which was later annulled. The General Court held that the Commission’s discretion was reduced in those circumstances and thus the infringement, namely the insufficiency of the evidence adduced in support of the applicant’s impugned practices, was capable of establishing the existence of a sufficiently serious breach. However, taking account of the complexity of the case and the difficulties in applying the Treaty provisions in matters relating to cartels, the General Court held that the breach was not sufficiently serious. On appeal, (65) the Court of Justice upheld the General Court’s approach. (66)
82. Applying that line of reasoning to the present case, it may be concluded that the General Court did not err in law when considering that it may assess additional factors, even after finding that the Commission had no discretion to adopt the empty bag test. Those additional factors could excuse the Commission’s breach, in which case it would not be deemed sufficiently serious to incur liability.
(c) The third and fourth groups of cases
83. The third group of cases concerns those in which the EU Courts considered that the institution had a discretion in relation to the rule of EU law that had been breached and, after taking account of different factors, concluded that the breach was sufficiently serious. Examples can be found in the context of both EU (67) and State liability. (68)
84. For instance, the judgment in Ombudsman v Staelen, (69) involved alleged breaches by the European Ombudsman of the duty to act diligently in connection with the handling of a complaint. The Court of Justice held that the General Court had erred in law by ruling in general terms that a mere breach of the duty of diligence amounted to a sufficiently serious breach; as apparent from the case-law, the Ombudsman enjoys a wide discretion. The Court of Justice found, however, that the Ombudsman had made inexcusable errors and had manifestly and gravely disregarded the limits on its discretion in the conduct of its investigation and thus had committed three sufficiently serious breaches of EU law. (70)
85. In the fourth group of cases, the existence of discretion and the assessment of the different factors resulted in a finding by the EU Courts that the breach was not sufficiently serious. Here, too, the case-law contains examples relating to EU (71) and State liability. (72)
86. For example, the judgment in SGL Carbon and Others v Commission (73) involved the Commission’s allegedly unlawful classification of a certain substance pursuant to the relevant EU regulation. The Court of Justice upheld the General Court’s assessment of the Commission’s conduct as excusable due to the complexity of classifying a substance and the difficulty of interpreting the summation method rule.
87. Applying that line of reasoning to the present case, it may be concluded that the General Court did not err in law by assessing additional factors if the Commission enjoyed a discretion.
88. Did the Commission have a discretion?
89. If the present case is compared to the Brasserie case, the answer to that question would seem to be ‘yes’. That case concerned an area of EU law in which the Member States concerned could make regulatory choices, but a higher rule of EU law imposed limits on those choices. More specifically, EU law excluded discriminatory rules from the choices available to the national legislature of those Member States. In the present case, the Commission adopted a regulation on the energy efficiency of vacuum cleaners as a delegated act, and thus in an area of EU law in which the EU legislature left it to the Commission to make regulatory choices. However, such choices were limited by the impossibility to adopt testing methods other than those ‘during use’. Consequently, this could lead to the conclusion that the Commission exceeded the limits placed on its discretion, and that the General Court was right to assess other factors in order to determine whether disregard for those limits was grave and manifest.
90. However, the same situation could also be described as involving no discretion. From this perspective, the Commission was in fact implementing a mandate imposed by basic legislation (similar to when Member States transpose directives) and then breached a clear rule which did not leave it any discretion.
91. Given that the same situation can be characterised as involving either the existence or the non-existence of discretion, in my view, it seems preferable to adopt the approach suggested by certain Advocates General (as mentioned above in point 61 of this Opinion) that discretion is an important factor in determining whether the breach is sufficiently serious, but it is not decisive. In each case, it is necessary to establish whether the institution concerned had a discretion. However, if there was no discretion, it still does not follow automatically that the breach was sufficiently serious. Depending on the circumstances of each particular case, other factors can be taken into account and might lead to the exclusion of liability. Based on those considerations, regardless of whether the situation in the present case is characterised as one where there was discretion but the limits placed on that discretion were exceeded, or as one involving no discretion, the additional factors can be taken into account in order to determine whether the breach was sufficiently serious.
92. Accordingly, I consider that the General Court did not err in law by finding in paragraphs 37 and 38 of the judgment under appeal that it was necessary to take into account the additional factors. The second ground of appeal should, therefore, be rejected.
93. However, in my view, the General Court erred in law in assessing those additional factors. That leads me to the third and fourth grounds of appeal.
C. The third and fourth grounds of appeal
94. As indicated above (see points 41 and 42 of this Opinion), the General Court considered that the Commission’s breach of Article 10(1) of Directive 2010/30 was excusable. The General Court took into account the complexity of the situation to be regulated and the difficulties in the interpretation and application of the relevant texts. It found that the Commission did not manifestly and gravely disregard the limits on its discretion in choosing the empty bag test rather than the section 5.9 dust-loaded test.
95. By the third and fourth grounds of appeal, Dyson claims, in essence, that the General Court could not lawfully conclude that the Commission’s breach of Article 10(1) of Directive 2010/30 could be excused by interpretative difficulties and regulatory complexity. It was clear that the test to be applied should show energy efficiency ‘during use’ and that the purpose of that rule was to ensure that consumers receive reliable information on the energy efficiency of vacuum cleaners. While choosing an appropriate scientific test might be technically complex, that was not the case with regard to the rule that the Commission could not use a test which did not measure the performance of vacuum cleaners ‘during use’.
96. The Commission argues, in essence, that, in adopting the empty bag test, it considered that it was using a testing method that was as close as possible to real usage while also being reliable, accurate and reproducible, so as to enable end-users to make better informed purchasing decisions. According to the Commission, it was entirely reasonable for it to exercise a degree of caution when defining testing methods and opting to weigh the scientific validity of such methods as a decisive factor when making policy decisions. Therefore, even though the Court of Justice and the General Court subsequently found in the annulment proceedings that the Commission was wrong in adopting this approach (see points 27 to 33 of this Opinion), its breach may be excused and for that reason is not sufficiently serious.
1. Admissibility
97. At the hearing, the Commission argued that the third and fourth grounds of appeal are inadmissible, as those grounds call into question the General Court’s assessment of the facts.
98. I disagree with the Commission.
99. According to the case-law, while the assessment of facts and whether they are complex in an action for EU non-contractual liability is, in principle, for the General Court, (74) the Court of Justice can review the legal characterisation of the facts and the legal conclusions drawn from them in determining whether there has been a sufficiently serious breach. (75)
100. In the present case, as Dyson argued at the hearing, the third and fourth grounds of appeal do not seek a re-examination of the General Court’s findings of fact, but are intended, in essence, to challenge the legal characterisation applied by the General Court on the basis of which it decided that the Commission did not commit a sufficiently serious breach. Dyson specifically seeks to challenge the legal reasoning of the General Court that the breach committed by the Commission involved interpretative difficulties and regulatory complexity, which led the General Court to find that the Commission did not manifestly and gravely disregard the limits on its discretion. That constitutes a point of law subject to review by the Court of Justice on appeal.
101. Therefore, I consider that the third and fourth grounds of appeal are admissible.
2. Substance
102. In my view, the General Court erred in law in finding that the Commission’s breach of Article 10(1) of Directive 2010/30 could be excused by interpretative difficulties or regulatory complexity.
103. The question whether a breach of a rule of EU law may be excused is to be assessed by taking into consideration the circumstances as they existed at the time of the adoption of the decision by which the breach was committed. (76) It is, thus, necessary to assess whether the General Court erred in finding that interpretative difficulties and regulatory complexity could have excused the Commission at the time when it adopted Delegated Regulation 665/2013.
104. In my view, it cannot be accepted in the circumstances of the present case that the Commission, as a ‘good’ administrator exercising ordinary care and diligence, (77) could consider that it was justified in adopting a testing method that misleads consumers about the energy efficiency of vacuum cleaners simply because that was the only testing method available at the time.
105. First, the Commission has not claimed that it was unclear that the testing method to be adopted had to measure the performance of vacuum cleaners ‘during use’. However, its problem was that, at the relevant time, a scientifically viable dust-loaded test did not exist. (78) The Commission therefore decided to use the empty bag test.
106. At the hearing, the Commission stated that, at least at the relevant time, it was still possible to consider that testing vacuum cleaners at the beginning of their use (with empty bags) represented testing ‘during use’. It was only later, in the Court of Justice’s annulment judgment, that the Court explained that ‘during use’ means measuring the energy performance of vacuum cleaners in conditions as close as possible to actual conditions of use, requiring the vacuum cleaner’s receptacle to be filled to a certain level. (79)
107. Can such an excuse be accepted?
108. Indeed, using purely textual interpretation and concentrating on those two words exclusively, one may conclude that the beginning of the use is also ‘during use’.
109. However, at the relevant time, the Commission was aware, as it stated at the hearing, that there was a decline in energy performance with bagged vacuum cleaners. In particular, as indicated by Dyson, this is demonstrated by the Commission’s expert report, (80) along with information provided by consumer organisations (81) and Dyson itself (82) during the stakeholder consultations which led to the adoption of Delegated Regulation 665/2013.
110. The Commission was thus aware, at the relevant time, that the empty bag test could not achieve the objective of Directive 2010/30 to inform consumers about the energy efficiency of vacuum cleaners and enable them to buy more energy efficient ones. Quite the contrary, the Commission could not have been unaware that such a test is misleading for consumers.
111. No good administrator should be allowed to disregard the purpose of the legal provisions that it is implementing. If interpreted in the context of the objectives of Directive 2010/30, the wording ‘during use’ could not, according to the information that the Commission had, be understood as allowing testing with empty bags.
112. In my view, the fact that, at the relevant time, the Parliament or the Council did not oppose Delegated Regulation 665/2013, as they could have done under Article 12 of Directive 2010/30, cannot be used as evidence of an unclear mandate.
113. I do not find, therefore, that interpretative difficulties could excuse the breach committed by the Commission.
114. As far as regulatory complexity is concerned, the Commission explained that it had to balance the need to apply a testing method that reflected actual conditions of use with the requirement that such a method be scientifically accurate and reliable. However, the Commission could not include in the balancing exercise a test which gave misleading results to consumers. There was no regulatory complexity associated with the conclusion that the empty bag test was not an option.
115. This is sufficient to conclude that, notwithstanding the possible technical complexities of finding the appropriate test, it was clear, at the relevant time, that the one test that the Commission chose to use could not have been used. It is, therefore, difficult to see how the Commission, acting as a good administrator, could be excused for not recognising that.
116. Finally, the Commission has argued that, pursuant to Article 11(1) of Directive 2010/30, it was required to adopt a delegated regulation within a period of five years. Therefore, the Commission was right to consider that it was better to adopt the empty bag test than no test at all.
117. In response to that argument, I should first point out that the Commission’s understanding of the five-year period referred to in Article 11(1) of Directive 2010/30 is incorrect. That provision did not impose a final deadline by which the Commission had to adopt a testing method at all costs. It determined only the initial period of delegation upon the expiry of which the Commission was required to make a report. That period was automatically to be extended for another five years if the Parliament or the Council did not revoke the delegation. Moreover, even if the five-year period were a deadline for the Commission, that could not excuse it for opting for a test that was not in conformity with the delegation, as it misleads consumers.
118. In sum, if Dyson’s claim is appropriately characterised (see my analysis of the first ground of appeal) as claiming that the Commission committed a sufficiently serious breach by adopting the empty bag test, it is pretty straightforward that neither interpretative difficulties nor regulatory complexity could excuse the Commission for adopting that test.
119. Therefore, I consider that the third and fourth grounds of appeal are well founded and that the Commission committed a sufficiently serious breach of Article 10(1) of Directive 2010/30.
D. The fifth, sixth and seventh grounds of appeal
120. By the remaining three grounds of appeal, Dyson complains that the General Court erred in law by finding that the breach committed by the Commission was not sufficiently serious in relation to the alleged infringements of the principle of equal treatment (the fifth ground of appeal), the principle of sound administration and the duty to act diligently (the sixth ground of appeal) and the right to pursue a trade or business (the seventh ground of appeal).
121. As seen in point 43 of this Opinion, the General Court relied on its reasoning in respect of the alleged infringement of Article 10(1) of Directive 2010/30 in finding that there was no sufficiently serious breach in respect of those other alleged infringements of EU law. As I have already considered in my analysis of the first, third and fourth grounds of appeal that the General Court mischaracterised Dyson’s plea and erred in law in finding that the Commission’s breach of Article 10(1) of Directive 2010/30 was not sufficiently serious, the General Court’s findings as to whether those other alleged infringements of EU law are sufficiently serious are also vitiated by the same errors of law.
122. Therefore, I consider that the fifth, sixth and seventh grounds of appeal are well founded.
IV. Conclusion
123. In the light of the foregoing, I propose that the Court of Justice should:
– find that the first, third, fourth, fifth, sixth and seventh grounds of appeal are well founded;
– set aside the judgment of the General Court of 8 December 2021, Dyson and Others v Commission (T‑127/19, EU:T:2021:870);
– find that the breach of Article 10(1) of Directive 2010/30/EU of the European Parliament and of the Council of 19 May 2010 on the indication by labelling and standard product information of the consumption of energy and other resources by energy-related products committed by the Commission is sufficiently serious;
– refer the case back to the General Court;
– reserve the decision on costs.