OPINION OF ADVOCATE GENERAL
SHARPSTON
delivered on 30 April 2020 (1)
Case C‑693/18
Procureur de la République
against
Company X,
other parties:
CLCV and Others,
A and Others,
B,
AGLP and Others,
C and Others
(Request for a preliminary ruling from the Vice-President responsible for investigation of the tribunal de grande instance de Paris (Regional Court, Paris, France))
(Reference for a preliminary ruling — Approximation of laws — Regulation (EC) No 715/2007 — Motor vehicles — Pollutant emissions — Defeat device — Program acting on the electronic engine controller — Technologies and strategies to limit the production of pollutant emissions — Diesel engines)
Introduction
1. The reference for a preliminary ruling made in this case by the Vice-President responsible for investigation at the tribunal de grande instance de Paris (Regional Court, Paris, France) concerns the interpretation of Articles 3(10) and 5(2) of Regulation (EC) No 715/2007. (2)
2. In that context, the Court is requested to clarify, for the first time, the meaning of several concepts laid down in that regulation, including ‘defeat device’ and ‘emission control system’. (3)
Legal background
International law
3. Regulation No 83 of the United Nations Economic Commission for Europe (UNECE) — Uniform provisions concerning the approval of vehicles with regard to the emission of pollutants according to engine fuel requirements (4) lays down technical requirements for the approval of motor vehicles. (5)
4. Paragraphs 2.16 to 2.16.3 of that regulation provide:
‘2.16 “Defeat device” means any element of design which senses temperature, vehicle speed, engine rotational speed, transmission gear, manifold vacuum or any other parameter for the purpose of activating, modulating, delaying or deactivating the operation of any part of the emission control system, that reduces the effectiveness of the emission control system under conditions which may reasonably be expected to be encountered in normal vehicle operation and use. Such an element of design may not be considered a defeat device if:
2.16.1. the need for the device is justified in terms of protecting the engine against damage or accident and for safe operation of the vehicle, or
2.16.2. the device does not function beyond the requirements of engine starting, or
2.16.3. conditions are substantially included in the Type I or Type VI test procedures.’
EU law
Decision 97/836/EC
5. Article 1 of Decision 97/836/EC (6) provides as follows:
‘The Community shall accede to the United Nations Economic Commission for Europe Agreement concerning the adoption of uniform technical prescriptions for wheeled vehicles, equipment and parts which can be fitted to and/or be used on wheeled vehicles and the conditions for reciprocal recognition of approvals granted on the basis of these prescriptions, hereinafter referred to as the “Revised Agreement”.
…’
6. Under Article 3(1) thereof, ‘… the Community shall state that its accession will be restricted to implementation of the UN/ECE Regulations listed in Annex II to this Decision’.
7. Regulation No 83 (UNECE) is referred to in Annex II to that decision.
Directive 2007/46/EC
8. Under Articles 34 and 35 of, and Annex IV to, Directive 2007/46/EC, (7) Regulation No 83 (UNECE) is incorporated into the Community vehicle type-approval procedure.
Regulation No 715/2007
9. Recitals 1, 5, 6 and 12 of Regulation No 715/2007 read as follows:
‘(1) … The technical requirements for the type approval of motor vehicles with regard to emissions should therefore be harmonised in order to avoid requirements which vary from one Member State to another and to ensure a high level of environmental protection.
…
(5) Achieving EU air quality objectives requires a continuing effort to reduce vehicle emissions. …
(6) In particular, a considerable reduction in nitrogen oxide emissions from diesel vehicles is necessary to improve air quality and comply with limit values for pollution.
…
(12) Efforts should be continued to implement stricter emission limits, including reduction of carbon dioxide emissions, and to ensure that those limits relate to the actual performance of vehicles when in use.’
10. Article 3 of that regulation, entitled ‘Definitions’, states that:
‘…
(4) “gaseous pollutants” means the exhaust gas emissions of carbon monoxide, oxides of nitrogen, expressed in nitrogen dioxide (NO2) equivalent, and hydrocarbons;
…
(6) “tailpipe emissions” means the emission of gaseous and particulate pollutants;
…
(10) “defeat device” means any element of design which senses temperature, vehicle speed, engine speed (RPM), transmission gear, manifold vacuum or any other parameter for the purpose of activating, modulating, delaying or deactivating the operation of any part of the emission control system, that reduces the effectiveness of the emission control system under conditions which may reasonably be expected to be encountered in normal vehicle operation and use;
…’
11. Article 4(1) and (2) provides:
‘1. Manufacturers shall demonstrate that all new vehicles sold, registered or put into service in the Community are type approved in accordance with this Regulation and its implementing measures. Manufacturers shall also demonstrate that all new replacement pollution control devices requiring type approval which are sold or put into service in the Community are type approved in accordance with this Regulation and its implementing measures.
These obligations include meeting the emission limits set out in Annex I and the implementing measures referred to in Article 5.
2. Manufacturers shall ensure that type approval procedures for verifying conformity of production, durability of pollution control devices and in-service conformity are met.
In addition, the technical measures taken by the manufacturer must be such as to ensure that the tailpipe and evaporative emissions are effectively limited, pursuant to this Regulation, throughout the normal life of the vehicles under normal conditions of use. …’
12. Under Article 5(1) and (2):
‘1. The manufacturer shall equip vehicles so that the components likely to affect emissions are designed, constructed and assembled so as to enable the vehicle, in normal use, to comply with this Regulation and its implementing measures.
2. The use of defeat devices that reduce the effectiveness of emission control systems shall be prohibited. The prohibition shall not apply where:
(a) the need for the device is justified in terms of protecting the engine against damage or accident and for safe operation of the vehicle;
(b) the device does not function beyond the requirements of engine starting;
or
(c) the conditions are substantially included in the test procedures for verifying evaporative emissions and average tailpipe emissions.’
Regulation (EC) No 692/2008
13. Article 1 of Regulation (EC) No 692/2008 (8) provides that the regulation ‘lays down measures for the implementation of Articles 4, 5 and 8 of [Regulation No 715/2007]’.
14. Under Article 2(18) of that regulation, ‘emission control system’, in the context of the OBD (9) system, means ‘the electronic engine management controller and any emission-related component in the exhaust or evaporative system which supplies an input to or receives an output from this controller’.
French law
15. Article L.213-1 of the code de la consommation (Consumer Code) imposes a penal sanction when ‘anyone, whether or not a party to the contract, … has deceived or attempted to deceive a party to the contract, by any means or practice whatsoever, even through a third party, as to either: (1) the nature, kind, origin, essential qualities, composition or content of necessary ingredients of any goods; or (2) the quantity or identity of the things supplied, by supplying a good other than the specific thing that formed the subject matter of the contract; or (3) suitability for use, risks inherent in the use of the good, checks carried out, usages or the precautions to be taken, …’. (10)
16. Article L.213-2 of the Consumer Code provides that the penalty may be increased when the offences committed ‘have had the consequence of rendering use of the goods hazardous to human or animal health’. (11)
17. Under Article L.213-6 of the Consumer Code, legal entities found criminally responsible for the offences defined in Articles L.213-1 and L.213-2 of the same code also incur the penalties provided for in Article 131-39(2) to (9) of the code pénal (Penal Code). This may involve, for example, a prohibition on the activity ‘in the exercise or on the occasion of which the offence was committed’.
The dispute in the main proceedings, the questions referred and the procedure before the Court
18. Company X is a car manufacturer that markets motor vehicles in France. It is apparent from the order for reference that the company placed on the market vehicles equipped with software (‘the software at issue’) which can distort the results of type-approval tests for emissions of gaseous pollutants, such as nitrogen oxides (‘NOx’).
19. During the type-approval process for emissions of gaseous pollutants, vehicles are tested according to a procedure the parameters of which are precisely defined by regulation. These include the speed profile to be followed, temperature and vehicle preconditioning. The speed profile used for the approval test (the New European Driving Cycle, abbreviated to ‘NEDC’) consists of replicating four urban cycles, as well as one extra-urban cycle, in a laboratory (and not under real conditions). That type-approval test is intended, inter alia, to ascertain whether the volume of NOx emitted exceeds the limits imposed by Regulation No 715/2007.
20. The vehicles called into question in this case have an exhaust gas recirculation (EGR) valve.
21. The EGR valve is one of the technologies used by vehicle manufacturers (such as company X) to manage and reduce final NOx emissions. EGR is a system that redirects some of the exhaust gas from a combustion engine to the gas inlet manifold (the engine air supply) in order to reduce final NOx emissions.
22. Specifically, the EGR pollution control system consists of a pipe that returns the exhaust gas (produced by the incomplete combustion of the fuel) to the manifold, with a heat exchanger to cool the burnt gas and an EGR valve that varies the flow of burnt gas back into the manifold.
23. The opening of the EGR valve is controlled by the electronic engine controller. The controller is an on-board computer that commands physical devices within the vehicle. The controller sends instructions to actuators depending on the information provided by its sensors. These actuators control the state of various mechanical parts in the engine. The opening of the EGR valve (on which the volume of gas returned to the manifold, and therefore the effectiveness of the pollution control system, depends) is controlled in real time by the electronic engine controller which, according to the information gathered by various sensors (such as speed and temperature), sends instructions to the EGR valve actuator. How wide that valve is open is therefore determined by the electronic engine controller and, ultimately, by the source code of the software installed on it (that is to say, the software at issue in the present case (12)).
24. In the present case, on 28 September 2015, following publications in the press, the Vice-President of the Regional Council of Île-de-France (France) responsible for transport filed a report with the parquet de Paris (Prosecutor’s Office, Paris, France) regarding the actions of company X in relation to the EGR valve and the software at issue, which was installed on some of its vehicles.
25. On 2 October 2015, the Parquet de Paris (Prosecutor’s Office, Paris) classified those actions as aggravated deception and requested the Office central de lutte contre les atteintes à l’environnement et à la santé publique (Central office for combating damage to the environment and to public health) to carry out an investigation into the conditions in which the vehicles in question had been placed on the French market.
26. At the same time, the Ministère de l’Écologie (Ministry of Ecology, France) asked the Service national des enquêtes (National Investigation Service; ‘SNE’) of the direction générale de la Concurrence, de la Consommation et de la Répression des Fraudes (Directorate-General for Competition, Consumer Affairs and Fraud Control; ‘DGCCRF’) to initiate investigations to determine whether vehicles marketed in France were equipped with the software at issue.
27. The SNE compiled a report, to which were appended the results of the tests carried out by the Union technique de l’automobile, du motocycle et du cycle (Technical Union for Automobiles, Motorcycles and Bicycles; ‘UTAC’), which is the only laboratory authorised to carry out vehicle approval tests in France. Those tests aimed to ascertain whether fraud had been committed. They showed that, in some of company X’s vehicles, NOx emissions were up to 3.6 times higher than the theoretical values recorded when they were approved.
28. Additional tests carried out by the Institut français du pétrole Energies Nouvelles (French Institute for Petroleum and New Energies; ‘IFPEN’) on three vehicles also showed that NOx emissions were specifically reduced when an approval cycle was detected, (13) with the EGR valve open significantly wider.
29. In October 2015, the parquet de Paris (Prosecutor’s Office, Paris) organised a search of the premises of the French subsidiary of company X in order to establish whether there was evidence proving that the subsidiary was aware of the existence of the software at issue.
30. In January 2016, the chairman of the subsidiary was questioned during a hearing where he appeared on a voluntary basis. He stated that he had discovered the existence of the software at issue through the press and did not know how it worked. He nevertheless stated that the software was installed in certain diesel engines, on an (approximate) total of 950 000 vehicles in France, which would be recalled for that software to be updated. He added that he did not acknowledge that the software was deceptive and, according to him, it aimed to ‘bring about an optimal decrease in NOx emissions’.
31. The investigation led to the opening of a judicial investigation, which was assigned to the referring judge and two other investigating judges on 19 February 2016. The prosecutor’s application to commence proceedings states that the alleged offence consists in having ‘since 1 September 2009, … by any means whatsoever, even through a third party, whether or not a party to the contract, deceived purchasers of vehicles fitted with diesel … engines as to the essential qualities of the vehicles and the checks conducted, accompanied by the circumstance that the acts had the result of rendering use of the goods hazardous to human or animal health’, in breach, inter alia, of Articles L.213-1 and L.213-2 of the Consumer Code.
32. At this stage of the criminal proceedings, company X has the status of a witness with legal representation. When it was summoned before the investigating judges on 28 March 2017, company X formally refused to answer the questions that they put to it. Company X also refused to make available information requested by the investigating judges (in particular regarding the methods used to calibrate its engines).
33. In addition, since the opening of the judicial investigation more than 1 200 people have filed civil actions.
34. During those proceedings, an expert was commissioned to analyse the results of the tests carried out by the administrative authority (in other words, the tests carried out by UTAC and IFPEN) and any other technical analyses with a view to describing how the software at issue operated and explaining its effects in terms of increased NOx emissions from vehicles on which it was installed.
35. In his report filed on 26 April 2017 (‘the expert report’), the expert noted that, under the regulations in force, emission control systems must function during normal vehicle operation. It appeared that, under normal driving conditions, in the vehicles examined, the EGR valve was not controlled consistently with the mode observed upon approval. The normal operating mode did not, as the tests carried out by UTAC showed, comply with the regulatory limits on pollution. If the approval cycle had not been detected and the operation of the EGR valve modified, it would not have been possible for the vehicles in question to comply with those limits in normal usage conditions. The effectiveness of the pollution reduction system was therefore reduced in real life.
36. The expert thus concluded that there was a device which detected the approval cycle and modified the operation of the exhaust gas recirculation system (the ‘EGR system’) for the purposes of that approval. The existence of that device resulted in an increase in NOx emissions from vehicles driving under normal conditions. The expert also explained that if the operation of the EGR valve in real traffic had been in line with its operation during the type approval tests, those vehicles would have produced, especially in urban traffic, significantly (around 50%) less NOx, but probably a little more (around 5% more) carbon monoxide, unburned hydrocarbons and carbon dioxide in return. (14) The power of those vehicles would probably have been marginally reduced. Maintenance operations would have been more frequent and expensive due, among other factors, to increased clogging up of the engine.
37. Lastly, the expert stated that the EGR system was a pollution control device in the sense that it was fitted to engines with the sole aim of reducing NOx emissions, that the decrease in its opening reduced the effectiveness of the emission control system and resulted in increased NOx emissions, and that that decrease was observed under normal conditions of vehicle use. By contrast, the reduced opening of the EGR valve led, in practice, to greater engine acceleration and higher performance potential. It also resulted in less clogging up of air intake ducts, valves and the combustion chamber, contributing to engine longevity and reliability.
38. In view of those technical details, the referring judge states that the legal mechanism of the deception — if it were established — would consist in having misled the purchasers of the vehicles concerned as to the essential qualities thereof, namely their non-compliance with Regulation No 715/2007, resulting from the presence in those vehicles of a defeat device as referred to in Article 3(10) and Article 5(2) of that regulation, consisting in programming the electronic engine controller operating the EGR valve to identify the approval cycle so that the activation of the NOx emission control system was increased during that cycle, and not under normal conditions of use.
39. The deception would be aggravated by the circumstance that the use of the vehicles would have endangered human and animal health since diesel engine exhaust gas was classified as a carcinogen by the International Agency for Research on Cancer (IARC) in 2012.
40. The referring judge points out that devices acting on the operation of emission control systems may take different forms. The definition of ‘defeat device’ set out in Article 3 of Regulation No 715/2007 includes several concepts which have not yet been interpreted by the Court.
41. Given that categorisation of the conduct in question as deception, as envisaged in the present proceedings, is based on the categorisation of the software at issue as a ‘defeat device’, the national court considers that it requires clarification as to the scope of the provisions referred to above in order to decide whether company X should be charged and committed for trial following the judicial investigation.
42. In the light of the foregoing, the referring judge decided to refer the following questions to the Court for a preliminary ruling:
‘(1) Interpretation of the concept of “design”
(a) What is covered by the concept of “element of design” in [Article 3(10)] of [Regulation No 715/2007], which defines “defeat device”?
(b) May a program integrated in the engine control calculator or more generally acting on that calculator be considered to be an element of design within the meaning of that article?
(2) Interpretation of the concept of “emission control system”
(a) What is covered by the concept of “emission control system” in [Article 3.10] of [Regulation No 715/2007], which defines “defeat device”?
(b) Does this emission control system include only the technologies and strategies aimed at treating and reducing emissions (in particular [of] NOx) after they have been created, or does it also incorporate the different technologies and strategies that enable the initial production of emissions to be limited, such as [EGR] technology?
(3) Interpretation of the concept of “defeat device”
(a) Is a device that detects any parameter connected with the conduct of the approval procedures provided for in [Regulation No 715/2007], for the purposes of activating or adjusting upwards, during those procedures, the operation of any part of the emission control system, and thus obtaining approval of the vehicle, a “defeat device” within the meaning of [Article 3(10)] of [Regulation No 715/2007]?
(b) If so, is that defeat device prohibited under [Article 5(2)] of [Regulation No 715/2007]?
(c) May a device as described in Question 3(a) be characterised as a “defeat device” if the upwards adjustment of the activation of the emission control system is effective, not only during the approval procedures, but also on specific occasions when the precise conditions detected for the purpose of adjusting the emission control system upwards during those approval procedures are encountered in actual traffic?
(4) Interpretation of the exceptions provided for in Article 5
(a) What is covered by the three exceptions provided for in [Article 5(2)] in Chapter 2 of [Regulation No 715/2007]?
(b) Might the prohibition of the defeat device activating or adjusting upwards the operation of any part of the emission control system specifically during the approval procedures be disregarded for one of the three reasons listed in [Article 5(2)]?
(c) Is slowing down the aging or the clogging-up of the engine among the requirements of “protecting the engine against damage or accident” or of “safe operation of the vehicle” that may justify the presence of a [defeat device] within the meaning of [Article 5(2)(a)]?
43. Written observations were submitted by the French and Italian Governments, by parties to the main proceedings (company X and civil parties A and Others) and by the European Commission.
44. The abovementioned Governments, parties to the main proceedings (company X and civil parties A and Others, B, AGLP and Others and C and Others) and the European Commission also presented oral argument at the hearing on 7 November 2019.
Analysis
Preliminary observations
45. It seems to me appropriate to begin by setting out in greater detail the regulatory framework for the approval of motor vehicles in the European Union.
46. The type-approval of motor vehicles is subject to detailed rules under EU law, based on provisions relating to the establishment and functioning of the internal market (currently Article 114 TFEU).
47. Directive 2007/46 establishes a harmonised framework comprising administrative provisions and general technical requirements for the type-approval of all new motor vehicles (light and heavy) falling within its scope.
48. In the automotive sector, each piece of framework legislation thus establishes the procedure for, and the legal consequences of, vehicle type-approval. The substantive conditions for type-approval are covered in separate regulations.
49. Type-approval operates as follows: the manufacturer submits a vehicle prototype to the competent authorities in order to demonstrate that it meets all the substantive conditions laid down in the regulatory acts referred to in Annex IV to Directive 2007/46. Where type-approval is granted by the competent administrative authority, the manufacturer commences production of the vehicle type in question. Each vehicle produced in conformity with the approved type may be placed on the market without additional testing.
50. In this connection, harmonised technical requirements for emissions are set out in Article 5 of Regulation No 715/2007. The regulation requires manufacturers to equip their vehicles so as to enable the vehicle, in normal use, to comply with that regulation and its implementing measures.
51. Article 5(2) of Regulation No 715/2007 further provides that the use of ‘defeat devices’ (as defined in Article 3(10) that regulation) is prohibited, subject to certain limited exceptions.
52. Therefore, under Article 5 of Regulation No 715/2007, vehicles must be designed not only in such a way that they comply with the limits laid down in the regulation during normal use, but also in such a way that the effectiveness of their emission control system cannot be reduced ‘under conditions which may reasonably be expected to be encountered in normal vehicle operation and use’. (15)
53. For the purposes of applying Article 5 of Regulation No 715/2007, the Commission has set out, in Regulation No 692/2008, specific laboratory test procedures by means of which manufacturers may establish, with a view to obtaining type-approval, that a new type of vehicle complies with the prescribed emission limits.
54. In the present case, the emission measurements prescribed by the test procedure were based on the only driving cycle applicable at the time, namely the NEDC (as defined above). (16) The tests are standardised, laboratory-based tests, not real-world tests — making them more susceptible to strategies to detect and circumvent them.
Admissibility of the questions referred
55. Before proceeding to consider the substance of the questions submitted for the Court’s consideration, it is necessary to establish whether they are admissible.
56. Some of the parties to the main proceedings (A and Others) have submitted in their observations that the categorisation of the software at issue as a ‘defeat device’ is only relevant to the validity of the vehicle approval. Whether that categorisation is established or not, it allegedly has no bearing on the existence of deception as to the ‘essential qualities’ or the checks performed. According to those parties, the Consumer Code penalises deception ‘by any means or practice whatsoever’: it does not therefore matter whether (or not) an infringement of Regulation No 715/2007 is involved. Thus, the questions referred for a preliminary ruling are irrelevant to determining the outcome of the dispute in the main proceedings. (17)
57. Company X likewise submits that those preliminary questions are irrelevant to determining the outcome of the main proceedings. In its view, those proceedings (as set out in the order for reference) consist in determining whether it can be charged with deception and committed for trial on that ground. Company X argues that under French criminal law, the offence of deception requires a combination of a material element and an intentional element, the existence of which is doubtful in the present case. Company X submits that, irrespective of the answers which the Court may give to the questions submitted by the referring judge, it could not in any event be charged (nor, a fortiori, committed for trial).
58. In addition, company X relies on the principle of the legality of offences and penalties: in the absence of a provision expressly criminalising the acts alleged in the present case, it cannot be convicted of deception.
59. Furthermore, company X claims that there is, at this stage, no dispute before the national court as regards the questions submitted, on the ground that it has not yet stated its position, (18) either orally or in writing, on those same questions before the national court. Those questions are merely a request for an advisory opinion on rules of EU law, unrelated to any dispute. In that respect, those same questions are purely hypothetical. Such is the case, in particular, of the fourth question referred, concerning the exceptions set out in Article 5(2) of Regulation No 715/2007, which company X has never hitherto relied on in the proceedings pending before the national court.
60. Moreover, according to company X, the questions referred were not the subject of an exchange of arguments between the parties before they were submitted to the Court, which is contrary to the principle of sound administration of justice.
61. Finally, at the oral hearing, company X also argued that the questions referred were ‘premature’ because the analyses carried out so far (and reflected, inter alia, in the expert report) did not make it possible to define the facts of the dispute with sufficient precision.
62. In my view, the arguments presented by parties A and Others and by company X should be dismissed in their entirety.
63. To begin with, it should be noted that, according to the Court’s settled case-law, by virtue of the cooperation between the Court of Justice and the national courts established by Article 267 TFEU, it is solely for the national court before which the dispute has been brought, and which must assume responsibility for the subsequent judicial decision, to determine, in the light of the particular circumstances of the case, both the need for a preliminary ruling in order to enable it to deliver judgment and the relevance of the questions which it submits to the Court. Consequently, since the questions raised concern the interpretation of EU law, the Court is in principle required to give a ruling. (19)
64. In passing, and although this question has not been raised by the parties to the main proceedings, I would also point out that, according to settled case-law, ‘the judge investigating a criminal matter or the investigating magistrate constitutes a court or tribunal within the meaning of [Article 267 TFEU], appointed to give a ruling, independently and in accordance with the law, in cases coming within the jurisdiction conferred on it by law in proceedings intended to culminate in decisions of a judicial nature’. (20)
65. It follows that questions relating to EU law enjoy a presumption of relevance. The Court may refuse to rule on a question referred by a national court for a preliminary ruling only where it is quite obvious that the interpretation of EU law that is sought bears no relation to the actual facts of the main action or its purpose, where the problem is hypothetical, or where the Court does not have before it the factual or legal material necessary to give a useful answer to the questions submitted to it. (21)
66. In the present case, it is clear from the order for reference that the categorisation of the software at issue (which determines the width of opening of the EGR valve and, as such, the level of final NOx emissions) as a ‘defeat device’ under Article 3(10) and Article 5(2) of Regulation No 715/2007 may have a bearing on whether an offence is found under French criminal law (namely aggravated deception under Articles L.213-1 and L.213-2 of the Consumer Code). The arguments of parties A and Others and of company X as to the futility of the questions submitted must therefore be dismissed since those questions are obviously relevant to the purpose of the main proceedings.
67. While the interpretation of Regulation No 715/2007 is likely to have a bearing on whether an offence is found, the arguments based on French criminal law which company X puts forward (in particular as regards the existence of the material and/or intentional element) cannot, however, succeed. They are irrelevant to the assessment of the admissibility of the questions referred (from the point of view of EU law) and relate to matters which fall exclusively within the jurisdiction of the national court. (22)
68. The arguments relating to the principle of legality of offences and penalties are also irrelevant since the abovementioned regulation governs the type-approval of motor vehicles and does not, as such, lay down rules of criminal law.
69. With regard to the allegedly hypothetical nature of the questions referred, I also entertain serious doubts concerning company X’s submissions.
70. The fact that company X’s procedural strategy before the national court consists of not answering the investigators’ questions, refusing to disclose the source code of the software at issue and denying the deceptive nature of that software does not establish the absence of dispute: on the contrary, I am inclined to think that this strategy reveals the existence of a real controversy of fundamental interest to the outcome of the criminal proceedings pending before the national court.
71. As regards, in particular, the fourth question referred (which concerns the interpretation of the exceptions to the prohibition of defeat devices laid down in Article 5(2) of Regulation No 715/2007), company X claimed that it had never relied on those exceptions before the national court, so the question referred is hypothetical. When questioned on that point by the Court, company X nevertheless admitted, quite clearly, that it did not rule out relying on those exceptions at a later stage, during the proceedings pending before the national court. That admission also seems to me to indicate that those questions are not just ‘a simple request for an advisory opinion’ on purely hypothetical issues.
72. With regard to the absence of an exchange of arguments between the parties before the national court, I would simply point out that Article 267 TFEU does not make a reference to the Court subject to there having been an inter partes hearing in the proceedings in the course of which the national court refers the questions for a preliminary ruling. (23) Company X’s submission on that point must likewise be dismissed.
73. Finally, as regards the allegedly premature nature of the questions referred in the present case, I note that it is for the national court alone to decide when to refer a matter to the Court. (24) I would also point out that company X’s submissions on that point leave me unconvinced. The facts of the dispute in the main proceedings came to light in 2015 and were the subject of several technical analyses (which were themselves the subject of a detailed examination, the findings of which are set out in the expert report). In that regard, it seems to me difficult to argue (as company X did during the hearing) that ‘no specific set of facts has been established at this stage’. On the contrary, I believe that the Court has the necessary factual and legal material to provide a meaningful answer to the questions put to it.
74. Accordingly, in the light of the foregoing considerations, I consider that the questions referred have been established as admissible.
The first question referred
75. By its first question (composed of two sub-questions, which will be examined together), the national court seeks to determine — in essence — whether a program installed on the electronic engine controller or, more generally, acting on that controller can be regarded as an ‘element of design’ within the meaning of Article 3(10) of Regulation No 715/2007.
76. In my view, this question should be answered in the affirmative.
77. The term ‘defeat device’ defined in Article 3(10) of Regulation No 715/2007 means any element of design ‘which senses temperature, vehicle speed, engine speed (RPM), transmission gear, manifold vacuum or any other parameter for the purpose of activating, modulating, delaying or deactivating the operation of any part of the emission control system, that reduces the effectiveness of the emission control system under conditions which may reasonably be expected to be encountered in normal vehicle operation and use’.
78. As the French Government has rightly pointed out, such a definition gives a wide scope to the concept of ‘element of design’. Such an element may consist of both mechanical and computer components controlling the activation of such parts, if it affects the operation of the emission control system and reduces its effectiveness. That is also the position taken by company X in its written submissions.
79. I would clarify that it must be an element that originates from the vehicle manufacturer. In the case of on-board software, it does not matter whether it is pre-installed before the vehicle is sold or downloaded at a later date during an update (imposed or recommended by the manufacturer); by contrast, it cannot be an element installed solely at the initiative of the owner or user of the vehicle, without any link to the manufacturer.
80. The first question referred should therefore be answered as follows: Article 3(10) of Regulation No 715/2007 should be interpreted as meaning that a program installed on the electronic engine controller, or, more generally, acting on that controller may be regarded as an element of design within the meaning of that article, provided that it forms an integral part of that controller.
The second question referred
81. By its second question (composed of two sub-questions, which I shall examine together), the national court asks, in essence, whether Article 3(10) of Regulation No 715/2007 must be interpreted as meaning that the concept of ‘emission control system’ covers exclusively technologies and strategies which reduce emissions ‘downstream’ after they have been produced or whether, on the contrary, that concept also covers technologies and strategies which, like the EGR system, reduce the amount of emissions produced ‘upstream’ in the first place.
82. There are in fact two main categories of techniques enabling manufacturers to optimise the performance of their vehicles in terms of pollutant emissions: first, ‘in-engine’ strategies (such as the EGR system), which consist in minimising the production of gaseous pollutants in the engine itself, and, second, ‘after-treatment’ strategies, which consist in treating the emissions after they have been produced (for example, the NOx absorber catalyst system).
83. Company X advocates a highly restrictive interpretation of the concept of ‘emission control system’. According to it, that concept can only refer to emission-related components which are located in the exhaust system and excludes in-engine strategies. That interpretation is (mainly) based on Article 2(18) of Regulation No 692/2008, according to which, in the context of the OBD system, ‘emission control system’ means ‘the electronic engine management controller and any emission-related component in the exhaust … system which supplies an input to or receives an output from this controller’. (25)
84. I disagree with company X’s view.
85. I note from the outset that the concept of ‘emission control system’ is not defined in Regulation No 715/2007.
86. In order to clarify its scope, reference should be made to the interpretative criteria established by the Court. It will be recalled that, in accordance with settled case-law, when interpreting a provision of EU law, ‘it is necessary to consider not only its wording but also the context in which it occurs and the objectives pursued by the rules of which it is part’. (26)
Literal interpretation
87. Literally, an ‘emissions control system’ is a component of a vehicle designed to control its emissions.
88. In that respect, I note that the EGR system may therefore be presumed to fall within the scope of that concept since it aims to reduce final NOx emissions. The expert report appended to the order for reference clearly states that the EGR system ‘is fitted to engines with the sole aim of reducing NOx emissions’ and that ‘gases circulating through the EGR system are eventually released into the atmosphere’. (27) It is clearly a ‘NOx emission control device’. (28)
89. That is also the position taken by the Commission in its written observations, where it points out (correctly, in my view) that the EGR system is obviously part of an emission control system, since it makes it possible to control the volume of those emissions according to pre-set parameters by reinjecting a greater or lesser volume of exhaust gas into the engine’s intake system. (29)
90. The French and Italian Governments, and the civil parties who expressed their views in written observations or at the hearing, favoured a similar interpretation.
91. Unlike company X, I do not believe that that reading of the concept of ‘emission control system’ is overly broad or that it could include any component of a vehicle that has any impact on the volume of pollutant emissions. At the hearing, company X argued that the electric window-opening system or the air-conditioning system also had an impact on vehicle emissions, although they could not be called ‘emission control systems’. Those examples do not seem apposite to me: the mechanisms referred to by company X are not intended to limit NOx emissions, which is precisely the purpose of the EGR system. (30) It is that very difference in purpose which, in my view, justifies classifying the EGR system as an emission control system. (31)
Contextual interpretation
92. As regards the context, the other provisions of Regulation No 715/2007 should first be examined to see whether they can provide useful additional material for the answer.
93. Article 4(2) of the regulation imposes an obligation on manufacturers to achieve results: the technical measures taken by manufacturers must be such as to ensure that tailpipe emissions are effectively limited. Article 3(6) defines ‘tailpipe emissions’: these are the emission of gaseous and particulate pollutants, without further qualification.
94. Therefore, neither Article 4(2) nor Article 3(6) specifies at which stage of the vehicle’s operation (or by which technical means) these emissions are to be modulated or reduced.
95. Regulation No 715/2007 is technology neutral, in so far as it does not impose any particular technological solution. It only sets a target to be achieved in terms of limiting emissions, which are measured at the outlet of the exhaust pipe.
96. The distinction made by company X between in-engine strategies and after-treatment methods for dealing with exhaust gas is therefore not justified by the wording of Regulation No 715/2007. Moreover, from a factual point of view, that distinction does not seem to make much sense. Whilst company X argued glibly at the hearing that ‘what does not come out of the engine is not an emission’, the fact remains that in practice, as the Commission noted with almost rustic common sense, NOx is still NOx: it always ends up coming out of the exhaust pipe (regardless of whether production of NOx has been initially limited in the engine or it has been treated in the exhaust system). (32)
97. Nor is the Byzantine distinction between ‘upstream’ and ‘downstream’ reduction methods drawn by company X justified in the light of Regulation No 83 (UNECE) and, in particular, paragraph 2.16 thereof. (33) The fact that paragraph 6.5.1.3 of Appendix 1 of Annex 11 to that regulation refers to two after-treatment methods as emission control systems (or anti-pollution systems) does not automatically imply that other methods (such as the EGR system) may not fall within the scope of that concept: these are only isolated examples, not an exhaustive list.
98. What about Regulation No 692/2008, which was relied on by company X both in its written submissions and at the hearing?
99. Let me begin by pointing out that Regulation No 692/2008, adopted by the Commission, (34) is a lower-ranking instrument than Regulation No 715/2007 (issued by the Council and the European Parliament): the former cannot, in principle, limit the latter’s scope. Furthermore, Article 2(18) of Regulation No 692/2008 (cited by company X) applies ‘in the context of the [on-board diagnostic] system’ and is not general in scope.
100. Furthermore, the position taken by company X seems to rest on a misreading of Article 2(18). According to company X, that provision presupposes that an emission control system can only be a component ‘which relates to emissions’ and ‘which is located in the exhaust system’. (35) However, the provision refers to any component (relating to emissions) of the exhaust system, particularly in the French version of the regulation. (36) It is therefore not necessary for this component to be physically located in the exhaust system.
101. Like the Commission, I am inclined to assume that this misreading stems from the English version of Regulation No 692/2008, which refers to any ‘emission-related component in the exhaust … system’. (37) That phrasing (in the English version (38)) does not seem to me to be correct on reading Regulation No 715/2007, the English version of which systematically refers to ‘tailpipe emissions’ (in the French version, this term is translated as ‘émissions au tuyau arrière d’échappement’, emissions at the tailpipe (39)). The terms used clearly reflect the reasoning underpinning Regulation No 715/2007: emissions are always measured at the outlet of the exhaust system, as they always end up being released from that system (before polluting the air). However, the mechanism for controlling emissions does not necessarily have to be located in the exhaust system strictly speaking.
102. Finally, I would also note that Appendix 2 of Annex XI to Regulation No 692/2008 expressly includes the EGR system in its list of emission control systems. That confirms, in so far as may be necessary, that company X’s interpretation of Article 2(18) of that regulation is incorrect.
Teleological interpretation
103. Let me now turn to the aims of Regulation No 715/2007 and their bearing on the interpretation of the term ‘emission control system’.
104. According to recitals 1 and 5 of that regulation, it aims, inter alia, at ensuring a high level of environmental protection, and achieving EU air quality objectives requires a continuing effort to reduce vehicle emissions. Recital 6 states that ‘in particular, a considerable reduction in NOx emissions from diesel vehicles is necessary to improve air quality and comply with limit values for pollution’.
105. Furthermore, it is apparent from Article 4 of the regulation that it aims to limit emissions effectively throughout the normal life of vehicles under normal conditions of use.
106. In the light of the EU legislature’s intention, as expressed very clearly in those provisions, I consider, like the French Government and the Commission, that the concept of ‘emission control system’ should be given a broad construction. Limiting the scope of that concept to exhaust gas after-treatment methods (by excluding in-engine strategies such as the EGR system) would considerably reduce the effectiveness of Regulation No 715/2007. In that respect, the distinction put forward by company X is unwarranted.
Conclusion
107. In the light of the literal, contextual and also teleological interpretation of the provisions of Regulation No 715/2007, I consider that the answer to the second question referred should be as follows: Article 3(10) of Regulation No 715/2007 should be interpreted as meaning that the concept of ‘emission control system’ includes both technologies, strategies, mechanical and computer components which make it possible to reduce emissions (including NOx) ‘upstream’ by limiting the initial production thereof, such as the EGR system, and those which make it possible to treat and reduce them ‘downstream’ after they have been produced.
The third question referred
108. The third question referred includes three sub-questions; I shall deal with the second sub-question when examining the fourth question in the following section of this Opinion.
109. The first and third sub-questions concern — in essence — whether Article 3(10) of Regulation No 715/2007 is to be interpreted as meaning that a device which detects any parameter linked to the conduct of the type-approval procedures prescribed in that regulation, in order to activate or modulate upwards, during those procedures, the operation of any part of the emission control system, and thus to obtain approval of the vehicle, constitutes a ‘defeat device’ within the meaning of this article, even though the upward activation of the operation of the emission control system may also occur from time to time when the exact conditions that trigger such activation occur in real traffic.
110. In my opinion, the answer to this question is ‘yes’.
111. A ‘defeat device’ is an element of design that senses various parameters (temperature, vehicle speed, and so forth) for the purpose of activating, modulating, delaying or deactivating the operation of any part of the emission control system, that reduces the effectiveness of the emission control system under conditions which may reasonably be expected to be encountered in normal vehicle operation and use.
112. Company X puts forward two arguments to refute that categorisation in the present case.
113. By its first argument, it claims that the EGR system is not an emission control system: therefore, an element of design that modulates the operation of the EGR system cannot be characterised as a ‘defeat device’. In view of the answer I propose be given to the second question referred, that argument cannot succeed. I shall not therefore dwell on it any further.
114. Its second argument runs as follows: defeat devices that manipulate upwards the operation of an emission control system in laboratory tests (such as NEDC tests) do not reduce the effectiveness of the emission control system. Only modulation induced during normal use of a vehicle establishes the existence of a defeat device.
115. That argument seems to me both specious and unfounded, both factually and legally.
116. From a factual perspective, it is apparent from the order for reference and the expert report that the EGR system operates in two modes, controlled by the software at issue. When a cycle characteristic of approval testing is detected, the EGR system switches to ‘mode 1’. Otherwise, when it detects that the conditions characteristic of approval testing are absent, the system selects ‘mode 0’.
117. In mode 1, the EGR valve is wide open and allows the vehicle to meet regulatory limits for NOx emissions. Conversely, in mode 0 (which is, in practice, the mode that prevails in real driving conditions), the EGR valve is not completely deactivated, but it is not so wide open. Such modulation results in much higher NOx emissions than those in the test phase (40) and, in any case, in a result which does not comply with the limits laid down in Regulation No 715/2007. (41)
118. As the Commission and the French Government have rightly pointed out, it is therefore clear that the device in question ‘modulates’ the operation of part of the emission control system, since it varies the level of emissions by switching from one mode to another, depending on the detection of various predefined parameters.
119. In other words, by defaulting to mode 0 in normal conditions of vehicle use, the process has the effect of modulating the effectiveness of the emission control system downwards. The fact that this process only results in increased activation when the conditions specific to approval testing are detected does not affect that finding.
120. From a legal point of view, the inevitable conclusion is that the argument advanced by company X is not in keeping with the wording, the context or the purpose of Article 3(10) of Regulation No 715/2007 either.
121. As the Commission has rightly pointed out, that provision is not based on a dichotomy between the testing phase for the type-approval of a vehicle and the subsequent period of normal use of the vehicle. The type-approval of motor vehicles is based on test procedures which must, as far as possible, correspond to what the future normal use of the vehicle will be after it is placed on the market. The test is intended to reproduce ex ante the actual driving conditions that the vehicle will encounter in normal use.
122. The NEDC test method is, in a manner of speaking, a theoretical trip which represents (in a stylised manner) the ‘conditions which may reasonably be expected to be encountered in normal vehicle operation and use’. Partial or complete deactivation of an emission control system, programmed to occur systematically outside that theoretical trip, inevitably leads to a reduction in the effectiveness of the system in normal conditions of use. A violation of Article 5(2) of Regulation No 715/2007 unavoidably results from that artificial deactivation.
123. There is therefore no legislative basis for the position advocated by company X.
124. In addition, from the point of view of context, reference should also be made to Article 4(2) of Regulation No 715/2007, which lays down the obligation to ensure that tailpipe emissions are effectively limited throughout the normal life of vehicles under normal conditions of use. (42) A device whose sole purpose is to ensure compliance with regulatory limits during laboratory tests would de facto run counter to that obligation. The expert report agrees with this conclusion: in the expert’s opinion, ‘according to the spirit of the law, emission control systems must be operational during normal vehicle operation’. (43) Accepting company X’s premiss would be tantamount to ‘conceding that the legislation does not intend to reduce pollutant emissions in daily life … but merely to submit technologies to progress checks’. (44) Evidently, such an interpretation would not be in line with the systematic aim of Regulation No 715/2007.
125. The expert’s remarks bring me naturally to the teleological aspect of the question: in view of the objectives pursued by Regulation No 715/2007 (as set out in points 104 and 105 of this Opinion), there can be no doubt that the position taken by company X would unjustifiably restrict the effectiveness of Regulation No 715/2007 and the prohibition of the devices referred to in Article 3(10) and Article 5(2) of that regulation.
126. Finally, I agree with the Commission and the French Government that it is irrelevant whether the upward modulation of the operation of the emission control system can take place from time to time during normal vehicle use. As the Commission has observed, the chances of such a coincidence occurring are infinitesimal (given the specific characteristics of the NEDC test). Compliance by the vehicle with the limits laid down in Regulation No 715/2007 must be the rule during normal use and not an exception linked to the accidental association of similar conditions to the approval tests.
127. In the light of the foregoing, I consider that the answer to the third question referred should be as follows: Article 3(10) of Regulation No 715/2007 should be interpreted as meaning that a device which detects any parameter linked to the conduct of the approval procedures provided for in that regulation, for the purposes of activating or modulating upwards, during those procedures, the operation of any part of the emission control system, and thus obtaining approval of the vehicle, constitutes a ‘defeat device’ within the meaning of that article, even if upward modulation of the operation of that emission control system may also occur from time to time when the exact conditions that trigger it occur by chance under normal vehicle operating conditions.
The fourth question referred
128. By sub-question (b) of the third question, and by the fourth question (itself made up of three sub-questions), the national court questions the lawfulness of a defeat device such as that at issue in the main proceedings and, in particular, whether such a device may fall within the scope of one of the exceptions to the prohibition laid down in Article 5(2) of Regulation No 715/2007. More specifically, the national court seeks to determine whether slowing down the ageing or clogging up of the engine can be included amongst the requirements of ‘protecting the engine against damage or accident’ or of ‘safe operation of the vehicle’ within the meaning of Article 5(2)(a) of that regulation and thus warrant the presence of that defeat device.
129. I recall that the use of defeat devices that reduce the effectiveness of emission control systems is prohibited and that there are only three exceptions to that general prohibition, namely where: (a) the need for the device is justified in terms of protecting the engine against damage or accident and for the safe operation of the vehicle; (b) the device does not operate beyond the requirements of engine starting; or (c) the conditions are substantially included in the test procedures for verifying evaporative emissions and average tailpipe emissions.
130. The exceptions laid down in Article 5(2)(b) and (c) are plainly inapplicable in the present case, in view of the facts set out above.
131. By contrast, it is necessary to analyse the exception under Article 5(2)(a) of Regulation No 715/2007, which allows the presence of a defeat device to be justified where it is needed to protect the engine against damage or accidents and to ensure the safe operation of the vehicle.
132. It is important to note that, according to settled case-law, exceptions are to be interpreted strictly so that general rules are not negated. (45) In that respect, the interpretation of exceptions may not go beyond the cases explicitly envisaged by the provision in question. (46)
133. I must therefore dismiss at the outset company X’s contention the exception in question must be given a ‘broad’ interpretation or application. (47)
134. In the present case, the terms ‘accident’ and ‘damage’ must be interpreted. In my view, the scope of these terms can be elucidated through a literal and teleological reading thereof.
135. As regards their literal interpretation, I note that it is generally accepted that the French term ‘accident’ refers to an unforeseen, sudden event that results in damage or danger, such as injury or death. (48) The French term ‘dégât’ refers to damage generally resulting from a violent or sudden cause. (49) The terms ‘accident’ and ‘damage’ used in the English version of Regulation No 715/2007 do not, in my opinion, gainsay that meaning. (50)
136. Where the wording of a provision of EU law is clear and precise, it should be adhered to. (51)
137. A defeat device can therefore be justified under Article 5(2)(a) of Regulation No 715/2007 only if it is necessary to protect the engine against sudden damage.
138. In my view, therefore, the Italian Government’s broad interpretation to the effect that the concept of ‘damage’ should be extended to include wear and tear, loss of efficiency, depreciation in the vehicle’s value as a result of aging and the progressive clogging up of its engine, should be dismissed.
139. As the Commission and the French Government have rightly pointed out, ageing and clogging up of the engine or a component thereof are the inexorable result of normal use of the vehicle. These are ordinary, predictable effects of the progressive accumulation of contaminants in the engine over the normal life of the vehicle under normal conditions of use, which can be mitigated by long-term regular, scheduled maintenance. They are a separate matter from accidents, damage or threats to the safe operation of the vehicle.
140. I now come to the teleological aspect of the question. In my opinion, in view of the objectives of Regulation No 715/2007 and, in particular, the objectives of protecting the environment and improving air quality in the EU, (52) the broad interpretation advocated by the Italian Government is in no way justified.
141. That interpretation would negate the general rule (in other words, the prohibition of defeat devices that reduce the effectiveness of emission control systems).
142. It is the responsibility of vehicle manufacturers to ensure that vehicles comply with the limits set by emissions legislation throughout their normal operation (53) and that vehicles operate safely within those limits. While it cannot be ruled out that the operation of an emission control system may adversely affect (in the long term) the durability or reliability of the engine, that circumstance does not justify deactivating the emission control system during normal vehicle operation under normal conditions of use for the sole purpose of protecting the engine against aging or progressive clogging up.
143. In other words, the interpretation put forward (in the present case) by the Italian Government cannot be accepted, since it deprives the prohibition of its effectiveness and manifestly runs counter to the intention of the EU legislature, which is to bring about a reduction in pollutant emissions by setting limits which must be complied with during the normal use of every vehicle placed on the market.
144. Moreover, such an interpretation would de facto imply that economic interests (such as preservation of the resale value of the vehicle) take precedence over public health. (54) That outcome would be contrary to both the letter and the spirit of Regulation No 715/2007. (55)
145. What are the implications for the present case?
146. In my opinion, only the immediate risk of damage which affects the reliability of the engine and creates a specific hazard when the vehicle is driven is capable of justifying the presence of a defeat device.
147. It is for the national court — the sole arbiter of the facts — to establish whether the device at issue in the main proceedings falls within the scope of the exception analysed above. (56)
148. However, it seems appropriate to note that, according to the expert report, the EGR system ‘is not destructive to the engine’. (57) The system is nevertheless likely to degrade the engine’s performance in use and accelerate clogging up, which can make maintenance operations ‘more frequent and more expensive’. (58) In the light of that observation, it seems to me that the defeat device in question is not necessary to protect the engine against accidents or damage and to ensure the safe operation of the vehicle.
149. The answer to sub-question (b) of the third question and to the fourth question should therefore be as follows: Article 5(2)(a) of Regulation No 715/2007 should be interpreted as meaning that the objective of slowing down the aging or clogging up of the engine does not justify use of a defeat device within the meaning of that provision.
Conclusion
150. In the light of the foregoing considerations, I propose that the Court give the following answer to the questions referred for a preliminary ruling by the Vice-President responsible for investigation of the tribunal de grande instance de Paris (Regional Court, Paris, France):
(1) First question referred
Article 3(10) of Regulation (EC) No 715/2007 of the European Parliament and of the Council of 20 June 2007 on type approval of motor vehicles with respect to emissions from light passenger and commercial vehicles (Euro 5 and Euro 6) and on access to vehicle repair and maintenance information should be interpreted as meaning that a program installed on or, more generally, acting on the electronic engine controller may be considered as an element of design within the meaning of that article, provided that it forms an integral part of that controller.
(2) Second question referred
Article 3(10) of Regulation No 715/2007 should be interpreted as meaning that the concept of ‘emission control system’ covers both technologies and strategies as well as mechanical or computer components that enable the initial production of emissions (including nitrogen oxides) to be limited ‘upstream’, such as an exhaust gas recirculation system, and those which enable emissions to be treated and reduced ‘downstream’ after they have been produced.
(3) Third question referred
Article 3(10) of Regulation No 715/2007 should be interpreted as meaning that a device which detects any parameter connected with the conduct of the approval procedures provided for in that regulation for the purposes of activating or modulating upwards, during those procedures, the operation of any part of the emission control system, and thus obtaining approval of the vehicle, constitutes a ‘defeat device’ within the meaning of that article, even where the upward modulation of the operation of that emission control system may also occur from time to time, when the precise conditions that trigger it occur by chance under normal vehicle operating conditions.
(4) Fourth question referred
Article 5(2)(a) of Regulation No 715/2007 should be interpreted as meaning that the objective of slowing down the ageing or clogging up of the engine does not justify the use of a defeat device within the meaning of that article.