Various Claimants v Mercedes-Benz Group AG & Ors

This is the first judgment resolving issues of substantive dispute between the parties, as opposed to case management and procedural issues, in the Pan-NOx Emissions Group litigation. The issues arise principally in the context of the Lead GLO, in which the Defendants are the Mercedes Benz Group AG & Others (‘the Mercedes Defendants’). However, those who consider that resolution of the issues may indirectly affect them have been given permission to make non-duplicative submissions. As such, Volkswagen AG & Others (‘VW’), Ford Motor Company & Others (‘Ford’) and Nissan Motor Company & Others (‘Nissan’) took part in the hearing and made written and oral submissions (each to a greater or lesser extent). I shall refer to these as other Defendant Manufacturers, and this may include those other Defendants in the wider GLO litigation as context demands. By reason of the procedure adopted, determinations of fact and law in the context of the Mercedes litigation will be binding, insofar as relevant and applicable, across the Pan-NOx Emissions litigation.

The core allegations are that Mercedes, and the other Defendant Manufacturers, sold cars that did not comply with the requirements of the applicable regulatory regime because they contained software that cheated or subverted the emissions control regime in Regulation 715/2007 (‘the Emissions Regulations’) and/or its successors. There are various different alleged methods of achieving this outcome, but for the purposes of this judgment they are referred to generically as ‘prohibited defeat devices’ (‘PDDs’). This alleged non-compliance then lies at the heart of a number of different claims brought in a number of different ways.

The German Federal Motor Transport Authority, the Kraftfahrt-Bundesamt (‘KBA’) made 4 different types of decision, or administrative act, which are relevant to the dispute between the parties (‘the KBA Decisions’). The precise nature and effect of the KBA Decisions are in dispute, but in broad terms:

The preliminary issues to be determined are as follows:

There is no dispute as to the answer to the second issue as between the Claimants and the Mercedes Defendants: it is common ground between their respective experts, and as such the parties, that in the event of a successful appeal against a KBA Decision, the KBA Decision will be annulled with retrospective effect. The answer to issue 2 is therefore ‘No’. Mercedes have appealed each of the Recall Decisions before the German Administrative Courts and those appeals are still pending. In its written submissions, the VW Defendants contend, in respect of this issue, that the determination of the German appellate court (as to the existence or otherwise of a PDD) is binding on the parties. VW did not serve evidence of German Law or provide any detailed submissions on this point. I deal with this issue briefly in paragraph 185 below, finding that the answer accords with the agreed position between the Claimants and the Mercedes Defendants which reflects the common ground between the experts.

The preliminary issue is to be determined against the following assumption (‘the Assumption’):

“The Court shall assume for the purposes of the KBA Trial that the KBA were properly advised as to the nature and existence of any defeat devices and/or updates which were the subject of such decisions”.

The context of this assumption is that, as part of its pleaded case, the Claimants allege that Type Approval was obtained by knowing misrepresentations (see paragraph 159 of the Generic Particulars of Claim, for example). Thus, even if the Claimants lose the preliminary issue based on the Assumption, there remains a challenge to the validity of the Type Approvals.

In approaching the preliminary issue, it is necessary to consider the following sub-questions, in respect of each category of KBA Decision:

During the course of the hearing, the Mercedes Defendants brought an application under Section CPR r.31.22 seeking that I disapply the standard CPR provisions to certain parts of the Sample Decisions which it is said are confidential. During the course of the hearing, certain interested parties also applied for copies of the Sample Decisions which are the subject of determination in this preliminary issue. Pending a full opportunity for submissions to be made, the pro tem position ordered was that the usual rule under CPR r.31.22 does not apply to those passages identified by Mercedes, which were therefore also redacted in copies provided to the interested parties. It is necessary for me to consider whether this temporary position shall remain. It had initially been anticipated that that determination would form part of this judgment. However, in circumstances where this judgment was available to be handed down before the cycle of submissions on the CPR r.31.22 and related issues had closed, the position pro tem shall remain. I shall address the final position in a short supplementary judgment in due course. For the avoidance of doubt, no part of the Sample Decisions quoted in this judgment is confidential.

Questions of German Law are, before me, questions of fact. In respect of German Law issues, I heard evidence from Professor Ekkehard Hofmann for the Claimants, who had also given evidence in Crossley 1 , and Dr Dieter Neumann for the Mercedes Defendants. Both witnesses’ areas of specialisation are in the public law arena: Professor Hofmann has been a Professor of Public Law since 2009; Dr Neumann, who is in private practice, focuses on public commercial law which he describes as administrative and commercial law, with familiarity with complex administrative law and administrative procedural issues.

In general terms, I do not doubt that both Professor Hofmann and Dr Neumann approached the giving of their oral evidence in accordance with their duties to the Court. Both, on the whole, gave answers in cross-examination frankly, whether or not those answers assisted their clients.

It is right, as submitted by the Claimants, that the written evidence of Dr Neumann purported on more than one occasion to rely upon academic sources which, upon investigation, were not wholly appropriate citations in that they either did not provide the relevant support or, indeed, said the opposite. I do not doubt that this was the product of carelessness or oversight rather than a deliberate attempt to mislead the court. Whilst this was undoubtedly of some concern, it does not lead me to reject his evidence in total. Nevertheless, it has been necessary to exercise some caution when considering the extent to which, where his evidence conflicted with that of Professor Hofmann, his view should be preferred, particularly in relation to the question of the application of Sections 20 and 22 of the Straßenverkehrs-Zulassungs-Ordnung (‘StVZO’) in respect of which the most obviously incorrect citation related, and in respect of which Dr Neumann’s opinion was left unsupported - indeed, contradicted.

One particular oddity of the evidence as it transpired is that, coming from their public law perspectives, both experts seemed in general agreement that they preferred – in a normative sense – the approach taken by the lower Administrative Courts when construing the tenor of the KBA decisions, which supported what might be described as a more expansive view of the regulatory content and binding effect of such decisions. This has meant that, in large part, the Mercedes Defendants pray in aid the evidence of the Claimants’ expert, Professor Hofmann, rather than seek to suggest his evidence should be rejected. However, the preliminary issue focusses not on what the law should be, or might be following constitutional challenges in the German Courts which have not yet taken place, but what the law actually is, at least as at the date of this judgment, in the context of litigation in which the Claimants are seeking damages in a civil law action in the Courts of England and Wales. I return to this further below.

It is helpful to set out a general framework of the German legal system, drawn from the helpful joint statement from the experts.

Pursuant to Article 95 para. 1 of the Basic Law (‘GG’), the German court system is divided into branches, which include Ordinary Courts and Administrative Courts.

Civil Courts are part of the Ordinary Courts. Their jurisdiction derives from Section 13 of the Courts Constitution Act (Gerichtsverfassungsgesetz, “GVG”), and they have jurisdiction over the civil disputes, family matters and non-contentious matters (civil matters) as well as criminal matters for which neither the competence of administrative authorities nor the jurisdiction of the Administrative Courts has been established and for which no special courts have been created or permitted by provisions of Federal Law. This results in administrative and Civil Courts deciding on different legal areas and this jurisdiction is exclusive. This means that an administrative court may not, under any circumstances, decide on a legal action before a Civil Court.

An administrative act is a form of action of an executive body, which includes the KBA. An administrative act requires a statutory basis that must result from a formal law. For the legal assessment of an administrative act, a distinction must be made between legality, effectiveness and nullity. The legality concerns the question of whether the requirements for state action specified by the statutory basis are met. For the effectiveness of an administrative act, it is essentially sufficient that it is announced to the addressee and the parties involved and is not void. The nullity of an administrative act within the meaning of Section 44 VwVfG is to be assumed if it contains particularly serious deficiencies.

The Federal Constitutional Court (‘BVerfG’) is the Constitutional Court at the Federal level. It is itself a constitutional institution (Verfassungsorgan) and, as the highest court of the German legal system, it stands out from the courts of appeal. It decides on the interpretation of constitutional law in cases governed by Article 93 GG, but is not competent as a so-called "super revision court" for the interpretation of other law. This means that in such cases the BVerfG only decides whether a court's decision violates constitutional provisions and not whether the court has correctly applied the rest of the law.

The decisions based on constitutional complaints, in which the BVerfG declares a law to be compatible or incompatible with the constitution (the Basic Law) or null and void, have the status of law themselves pursuant to Section 31 para. 1 of the Federal Constitutional Court Act (hereinafter: ‘BVerfGG’), so that the decisions of the BVerfG must be followed by other courts. They have formal de jure binding effect on inferior courts, therefore. The highest German court of administrative law is the Federal Administrative Court (Bundesverwaltungsgericht, ‘BVerwG’). As a Federal Court, its precedents are de facto but not strictly legally binding towards the courts subordinate to it.

Precedents are decisions in which a legal question that arises again in another case has already been decided. It is not the final decision in an individual case, but only the answer given by the court in the grounds of the judgment to a legal question which arises in the same way in the case to be decided by a lower court that is binding.

The equivalent of the BVerwG for civil law is the Federal Court of Justice (Bundesgerichtshof, ‘BGH’). The lower Civil Courts are equally bound by the precedents of the BGH, as explained above for the BVerwG: they are again, therefore, de facto but not strictly legally binding. Whilst Professor Hofmann agreed with the proposition that arguments could be made to a lower Civil Court to try to persuade it not to follow BGH jurisprudence, even without that BGH jurisprudence having been contradicted by the Administrative Courts, both experts agreed that it is very unusual for the lower courts to deviate from the precedents of the Federal Courts, despite the fundamental freedom of the lower courts in their decisions. If the lower courts were to rule differently from the precedents, it would be highly likely that the disadvantaged party would ultimately appeal to the Federal Court again. It is no doubt for this reason that Dr Neumann was prepared to accept that, in certain important respects explored more fully below, the decisions of the BGH represented ‘settled jurisprudence’.

The joint statement sets out a number of agreed principles in respect of the proper approach under German Law to the issues of whether, and in what respects, an administrative act is binding. I set these out below.

An administrative act is binding on third parties according to the prevailing opinion in administrative law case law and literature if the following requirements / requirements pursuant to Sections 43, 44 VwVfG are met, in other words, if:

It is agreed that each of the Type Approvals, Recall Decisions and Mandatory Update Decisions satisfied these criteria. There is a dispute, however, about whether the Voluntary Update Decisions are null and void. Professor Hofmann also posits additional criteria, relating to not-yet incontestable, or unimpeachable, decisions. Resolution of this dispute is only relevant if the binding (or potentially binding) element of the Decision includes a determination of the presence or absence of PDDs.

The experts agree that the binding part of an administrative act is interpreted according to the reasonable recipient’s perspective. The decisive factor, according to the established case law of the BGH and BVerwG, is therefore the declared will of the public authority, as the recipient could understand it in an objective assessment. The main focus is on the ‘tenor’ (or operative part) of the administrative act. In order to determine the content of the binding part of an administrative act, the following aspects must also be taken into account:

The formula generally held by the BVerwG is:

“The regulatory content must be determined in accordance with the rules developed for Sections 133 and 157 BGB. The interpretation of an administrative act is not based on the subjective ideas of the addressee or the issuing public authority. In accordance with the interpretation rule of Section 133 BGB, the decisive factor is the declared will as the addressee could understand it in an objective assessment. Even for the interpretation of an administrative act, only such circumstances are to be taken into account that were recognisable to the addressee upon receipt of the declaration of intent. It is not the inner will but the objectively declared will that is decisive as it is understood by the addressee.”

It is also agreed between the experts that a decision of an administrative authority can have declaratory effect, if, on an objective interpretation of the decision, the administrative authority so intended.

There are four sample Type Approval Decisions: those dated 8 June 2012 (e1*2007/46*0928*00) (‘TAD1’) and 23 April 2015 (e1*2007/46*0457*13) (‘TAD2’) are EC Type Approvals of a type of complete vehicle. Those dated 2 September 2016 (e1*2001/116*0480*22)(‘TAD3’) and 24 February 2017 (e1*2001/116*0480*23)(‘TAD4’) are extensions of Type Approval of a type of complete vehicle. They all consist of a certificate accompanied by annexures and enclosures. Both types of certificates identify matters such as the make, type, commercial name, category of vehicle, the name and address of the manufacturer, the name and address of the assembly plant(s). They then state (using identical or materially similar words):

“The undersigned hereby certifies the accuracy of the manufacturer's description in the attached information document of the vehicle(s) described above ((a) sample(s) having been selected by the E C type-approval authority and submitted by the manufacturer as prototype(s) of the vehicle type) and that the attached test results are applicable to the vehicle type.

1.

For complete vehicles:

The vehicle type meets the technical requirements of all the relevant regulatory acts as prescribed in Annex IV to Directive 2007/46/EC [or 70/156/EEC in relation to e1*2001/116*0480*23 dated 24 February 2017].

2.

…

3.

Type approval is granted.”

There are then various attachments which include an ‘Information Package’ and ‘Test Results’.

TAD3 and TAD4 include what are called collateral clauses. TAD3 contained the following wording, for example:

“This type-approval is granted under the resolutory condition, that from the date of approval until 30.11.2016 the following documentation will be submitted to Kraftfahrt- Bundesamt:

1.

Description of all measures to reduce emissions within the engine and with aftertreatment system

2.

Description of the base emission strategy that is active throughout the speed and load operating range of the vehicle unless an auxiliary emission strategy is activated;

3.

Description of the auxiliary emission strategy that becomes active and replaces or modifies a base emission strategy for a specific purpose and in response to a specific set of ambient or operating conditions and only remains operational as long as those conditions exist.”

The experts agree that the sample “Type Approval Decisions” are to be regarded as administrative acts of the Federal Motor Vehicle Office and that they have a binding effect on the Civil Courts across legal proceedings. They both recognise that the tenor in the four sample “Type Approval Decisions” consists, as a minimum, in the granting of the Type Approval (“Type approval is granted”). Dr Neumann’s view was that the tenor also extended to the words, ‘The vehicle type meets the technical requirements of all the relevant separate Directives as prescribed…’ and that, even if the tenor were restricted, the latter words would have to be taken into account in interpreting the tenor.

Dr Neumann’s evidence was that, in his view and in accordance with the approach of the German Administrative Courts, from the wording of the tenor of the sample “Type Approval Decisions”, there was a comprehensive approval effect, which in particular includes the finding that the KBA has not determined a violation of the ban on the use of PDDs – in other words, an implicit declaration as to the absence of PDDs. This is because regulatory requirements pursuant to the Framework Directive pursuant to which Type Approval is granted includes the Emissions Regulation, which expressly prohibits PDDs. Article 5(2) of the Emissions Regulation states ‘The use of defeat devices that reduce the effectiveness of emission control systems shall be prohibited.’ It is therefore said that either taken in isolation, or coupled with the explicit statement that that vehicle meets the regulatory requirements, the ‘tenor’ of the Type Approval decision amounts to an operative, binding determination or declaration under German Law that no PDDs exist.

The Claimants’ position is that:

In the BGH decision of 26^th June 2023, the court was concerned with a claim brought against VW.

EC Type Approval had been granted for a car with a diesel engine of the EA288. The complainant had financed the purchase price by means of a loan, and required the respondent to provide compensation as if he had not concluded the purchase agreement. The complainant argued that the car contained a prohibited defeat device. The claim was being brought pursuant to Sections 823 and 826 of the German Civil Code. The Regional Court of Osnabrück dismissed the claim, and the Higher Regional Court of Oldenburg did not admit the appeal.

The appeal succeeded in the BGH. Paragraph 12 of the decision states, in the context of the claim under Section 826 (which is a claim dependent upon “intentional… damage… in a manner that violates common decency”):

The BGH came to a similar conclusion in a claim against Mercedes on 8 August 2023 in which the court (in the context of a Section 826 claim) stated that EC Type Approval is not issued with respect to a particular vehicle nor to a group of specific vehicles in the sense of manufactured vehicles of a particular series, but merely approves a type of vehicle which conforms with the information contained in the certification. The implications for the facts of the operative part of an EC Type Approval (including eventual modifications) goes no further than the establishment of the regularity of the vehicle type put forward for assessment by the relevant certifying authority.

In their joint report, the experts agreed that the BGH did not address the question of whether the binding effect of a Type Approval can also preclude a claim for damages under Section 823 (a claim effectively aligned with a claim for tortious damages in English law). However, it was Professor Hofmann’s view under cross-examination that the Type Approval (non-) application to individual vehicles issue would be common to claims under both Civil Code sections. As such, he considered that there would be a ‘pretty good’ argument to apply the case law referred to under Section 826 to
Section 823 claims. No reason was put to Professor Hofmann, nor advanced in submission, as to why the determination of the BGH in respect of the irrelevance of a Type Approval decision to individual vehicles, by reason of the limited construction of its tenor, would not be as applicable to a tortious claim for damages under Section 823 as to one for immoral damages under Section 826.

Dr Neumann also clearly considered that a claim for damages in a Civil Court would follow the reasoning, and that this would be the case whether the claim was brought under Sections 823 or 826:

For what it is worth, this was also the analysis adopted by Mercedes’ (German) Lawyers, White & Case, in their constitutional complaint to the BVerwG (which has not yet been admitted for consideration), describing the decision of the BGH as one which has a de facto irreversible impact on claims for damages being brought against them in German courts. This would not, of course, be the case if the Mercedes Defendants’ analysis of the limited effect of the decision advanced before me is correct.

Although not a section of the decision explored in evidence, the view that Type Approval is not relevant to a claim for damages under Section 823 as well as Section 826, does appear consistent with the BGH’s own analysis in respect of the section 823 claim. In paragraph 34 of the decision, the BGH held:

“It must be assumed that the Respondent has issued an incorrect certificate of conformity. A certificate of conformity is incorrect if the motor vehicle concerned is, pursuant Art. 5 II Regulation (EC) 715/2007, equipped with a prohibited defeat device because the certificate then shows an actual non-existent conformity of the specific motor vehicle pursuant to Art. 5 II Regulation (EC) 715/2007. In this case, the content of the underlying EC Type Approval is not relevant” (my emphasis).

Dr Neumann confirmed in evidence, and I accept, that these decisions of the BGH have been followed in the Higher Regional Courts. It was in the context of pointing out that, notwithstanding this, there was administrative law literature which took a different view that Dr Neumann placed support on an article by Professor Schröder from 2017. Dr Neumann confirmed that he took the views of this academic seriously. In respect of the scope of EC Type Approval, Professor Schröder stated, after a number of paragraphs of analysis from first principles:

“It is therefore not convincing to assume that the EC type-approval is merely a variant of the classic general operating licence with regard to its regulatory content and that it is structurally comparable with the latter. In fact, EC type-approval does not relate to the individual vehicles in series production at all, but only to the type. Only this view takes account of the fact that the individual vehicles obtain their marketability in the internal market directly through the certificate of conformity and only indirectly through the EC type-approval.”

Dr Neumann fairly admitted that, far from supporting his own view, Professor Schröder in this part of the article supported the analysis adopted by the BGH. The following day Dr Neumann sought to provide an explanation of his reliance on Professor’s Schröder’s views by pointing to a separate passage which, in fairness to Dr Neumann, his report had specifically identified by way of footnote. However, the sentence upon which he relied (‘It can be inferred from this that an EC type-approval must exist as an administrative act; only then can the certificate of conformity transfer its positive effects to the individual vehicle’) was contained in a section of the article considering the validity of the Certificate of Conformity, not the binding effect of EC Type Approval on the individual car. I do not accept that, as suggested by Dr Neumann, Professor Schröder’s article was internally inconsistent, nor that the sentence he relied upon (when read in its proper context) provided any support for the view he expressed in his report.

There can be no doubt, in my judgment, that in the context of a civil law claim for damages, which would be determined by the German Civil Courts, the proper applicable German Law is that the tenor of a Type Approval does not extend to any relevant binding determination about the presence or absence of PDDs within an individual vehicle.

In her closing submissions, Ms Davies KC, on behalf of the Mercedes Defendants, urged upon the Court that this cannot be right because it would undermine the EC Type Approval system “if civil courts or other public authorities could substantially deviate from unimpeachable decisions of the Member State competent authorities”: words taken from Professor Hofmann’s evidence before Waksman J in Crossley 1. In addition, Ms Davies KC repeated Dr Neumann’s view that marketability of vehicles throughout the EU would therefore seriously be impaired. There are two answers to these related submissions. First, this is not, in fact, a submission that in anyway undermines the proper analysis of the effect of Type Approval under German law carried out above. It is, at best, a plea to what the law should be, but that is – even if correct - not the issue I have to resolve. Second, there is no evidence that the suggested inevitable systemic collapse of the harmonised framework for vehicle registration and sale as a result of the approach taken by the BGH has happened or is even remotely likely to happen. When asked what the actual effect has been of German BGH decisions, and many subsequent decisions at lower court level following this approach (which permits courts to determine whether defeat devices actually exist in the context of an award of damages against a manufacturer, notwithstanding the continuing existence and validity of a vehicle’s Type Approval for the purposes of its sale, registration and continued use on the roads) on the harmonised framework, Mr Kennelly KC, for VW, was unable to help, other than at a theoretical level.

Ms Davies KC rightly points out that the BGH decision of 26 June 2023 has been appealed to the Federal Constitutional Court. But that complaint has not yet been admitted by the court, let alone heard or determined. It is not pending in any immediate sense at all. Whilst it is common ground that such a decision could affect the present approach to the binding effect of Type Approval decisions adopted by the BGH, this fact is entirely irrelevant to the state of law as it stands today, which is what I have to decide.

It is in this context that the Mercedes Defendants rely upon the fact that both of the experts were, from their broadly public law perspectives, of the view that in the context of the Administrative Courts the BGH may have got the issue wrong. Professor Hofmann regarded the approach of the Administrative Courts and the Civil Courts as irreconcilable. It is not at all obvious to me why that is necessarily the case: it is readily conceivable that the approach to consideration of the binding effect of a document in the context of its annulment under public law might be conceptually different to the approach taken in the context of civil law suits affecting private law rights.

I do not need to speculate, however. The question is whether it is appropriate for me to consider the approach of the BGH as properly reflecting German Law as to the binding effect of Type Approval in the context of the present litigation. The answer is clear: the BGH approach is the relevant one for the purposes of determining the preliminary issue before me. The division between the roles of the administrative court and the Civil Courts is unambiguous. As confirmed by Dr Neumann in the context of the 8 August 2023 decision, although the administrative court may have priority in deciding whether to annul a decision, that does not in any way preclude the Civil Court from continuing with a claim for damages. Both experts were of the view that, if a claim was brought today for damages akin to the claim before this Court in the German Civil Courts (the only German courts competent to hear such a claim), that claim would not be precluded by the existence of a Type Approval, because a Type Approval (or modified Type Approval) does not represent any binding decision as to the presence or absence of PDDs in specific vehicles in German Law as applied by those courts. If the decision of the KBA granting Type Approval is not binding as to the presence or absence of PDDs in a civil claim brought under German Law, it obviously cannot be binding under English Law.

It follows that the answer to Preliminary Issue 1 in relation to Type Approval Decisions is ‘No’, and that is the case for any relevant time period.

In light of the foregoing, it is not strictly necessary to consider this issue, but the Mercedes Defendants contend that the answer may have ramifications for future arguments in these proceedings and, in deference to this, I will go on to consider the matter.

Mr Campbell KC relies upon paragraphs 12 and 13 of the 26 June BGH decision in support of his contention that the tenor of a KBA Decision is limited, and does not extend to a declaration as to the presence or absence of a PDD. Paragraph 13 relevantly states (emphasis added):

“It follows from Article 9 I 1a, first sentence a) of Directive 2007/46/EC that the competent public authority shall not grant EC Type Approvalin respect of a specific vehicle or in respect of a group of specific vehicles in the sense of the vehicles produced in a particular series, but shall only approve a vehicle type which conforms to the information in the description. Accordingly, the binding effect of an EC Type Approval cannot extend beyond the information in the description … For example, in the event of a failure to disclose a defeat device, the EC Type Approval, which does not cover a corresponding software without a description thereof, may be maintained, and the competent public authority may limit itself to ensuring conformity by means of other measures less serious than the cancellation of the EC Type Approval. This could include, for example, ancillary provisions. However, a change in the description by the vehicle manufacturer, as is conceivable in the sense of later information on defeat devices, may also result in the need for a new Type Approval.”

In this respect, the Claimants submit that the Administrative Courts are aligned with the view of the BGH. They rely upon the Schleswig Administrative Court decision of 17^th January 2024. This was a decision relating to a claim brought by the DUH, an environmental and health-relevant consumer protection body in Germany. The Defendant was the KBA. VW was among the manufacturers joined as interested parties. At paragraph 374, the Court stated as follows (emphasis added):

“According to the law applicable at the time of Type Approval, the joined parties were not obliged to specify engine control software in the information package that influences the vehicle types with regard to emission behaviour. However, the manufacturers had to provide the approval authority with information on the operation of the exhaust gas recirculation system (EGR), including its operation at low temperatures (Art. 3 Para. 9 of the implementing regulation). As far as is known, the joined parties have not disclosed any defeat devices. It was only after it became known that various manufacturers in the USA and the European Union were violating the ban on the use of defeat devices on a large scale that the obligations to cooperate were tightened. Even at the time of Type Approval, however, it was a prerequisite for its issue that vehicle types did not contain any unauthorised defeat devices. The Type Approvals remained valid even though the vehicles were actually fitted with an unauthorised defeat device. Contrary to a legal opinion previously held by the defendant, the existing unauthorised defeat device was also not tacitly approved. This is because the approval only legalises the object to the extent that it has been applied for and a positive decision has been made by the authorities. This means that originally only the approval for a type without (unauthorised) defeat devices was granted. However, if the authority is aware of a defeat device, incorrectly assumes a permissible exception to fundamental ban and expressly approves it, this constitutes an effective but unlawful Type Approval …”

Thus, the administrative court was drawing a distinction between the position where the existence of a defeat device is, and is not, known by the KBA on the basis of the information provided to it for the purposes of Type Approval determinations. Where no defeat device is identified in the information, the approval will not extend to a determination that no defeat devices (whether permissible or not) exist. However, where a defeat device is identified in the information provided, but the KBA wrongly conclude that it is a permissible exception to the prohibition and approve the type, the implication is that the tenor of such a decision will extend to a declaration as to the absence of PDDs in the vehicle. Even in this case, of course, in light of the inapplicability of the Type Approval to an individual vehicle, as set out above, this would preserve the effectiveness of the Type Approval decision as part of the regulatory regime, but would not prevent a claim for damages.

Dr Neumann’s view is that the Schleswig Court is wrong in its approach. In support of this, he relies upon the principle of investigation found in Section 24 of the German Code of Administrative Court Procedure (‘VwVfG’). This states that a public authority (which would include the KBA) shall investigate the facts ex officio, shall determine the scope of investigations and is not bound by the submissions and the requests for evidence of the parties involved. I was not persuaded by Professor Hofmann’s view in relation to the potential relevance of this section that where an authority had an obligation to follow a specific procedure during an investigation, this general duty would be precluded; or, indeed, that there is (even if this was right) any relevant preclusionary procedure in the context of the KBA’s approval process. From the existence of this general principle comes Dr Neumann’s conclusion that since the KBA could only grant Type Approval if satisfied that the vehicle type did not contain PDDs, the KBA was bound to investigate that question before granting Type Approval . It follows from this, in turn, that the reasonable recipient would understand that the grant of Type Approval meant the KBA had investigated whether the vehicle type contained any PDDs and had concluded that it had not. This should therefore be part of the binding tenor.

Whilst Ms Davies KC was clearly right that, under German Law, it is necessary to consider what the reasonable recipient would understand objectively, it is my view that this test – applied to the present facts – would lead to the conclusion adopted by the Schleswig court, as stated above, and as articulated by Professor Hofmann in the importance of what might be regarded as the need, objectively, for a ‘positive decision’. Ms Davis KC conceded that the effect of Dr Neumann’s contention was that in circumstances where (a) the information required to be provided by the KBA for Type Approval was objectively insufficient to have disclosed the existence of a prohibited defeat device and (b) prior to it becoming “known that various manufacturers in the USA and the European Union were violating the ban on the use of defeat devices” (adopting the words from the quotation above), such that the reasonable Type Approval authority would not have had cause to ask for different or additional material, it was nevertheless the objective tenor of a Type Approval that the relevant authority was making a declaration – binding inter omnes under German Law- that no prohibited defeat device existed. I consider that, on the basis of the reasoning of the BGH and the Schleswig Administrative Court, this somewhat unattractive conclusion is not demanded by German Law. I do not consider that the making of a ‘positive decision’ as part of the tenor can be implied simply by the existence of the prohibition against defeat devices within the Emissions Regulation coupled with the duty to investigate upon the KBA alone. Both the BGH and the Schleswig decisions explicitly or impliedly identify a third factor: the disclosure by the manufacturer of all defeat devices (albeit ones the manufacturer considers to be permissible) to the regulating authority so they are, in effect, on clear notice as to what to investigate. It is the fact of this disclosure within the information provided by the manufacturer that, on my understanding of the two decisions and as supported by the reasoning of Professor Hofmann, provides the objective backdrop against which it may then be appropriate to conclude that, having had a particular defeat device brought to the attention of the authority, the authority, in line with its duties, has duly investigated. It is only on this specific basis, therefore, that, in granting Type Approval, the authority has determined positively (rightly or wrongly) that no defeat device exists.

There is some tension between this third factor and the assumption upon which this preliminary issue has been ordered to be tried, namely that the KBA was “properly advised as to the nature and existence of any defeat devices which were the subject of any KBA decision”. The assumption has greater immediate resonance once, in the wake of the Dieselgate revelations, the information that was required to be provided began to focus specifically on the existence or absence of defeat devices (and the development of this I shall consider further below specifically in the context of TAD3 and TAD4). However, for TAD1 and TAD2, dated before Dieselgate erupted, I take the assumption as meaning no more than the manufacturers acted properly in providing, without concealment, the information that they were required by law to provide (and no more). It makes no sense to assume that, in 2012, the manufacturers were either required to or did in fact draw the nature and existence of defeat devices to the attention of the KBA in order for them to be investigated as to their permissibility when there is no suggestion, as far as I am aware, from either the Claimants or Defendants that this assumption reflects either what they were required by law to do, or the reality of what they did.

In this respect, Professor Hofmann was right to accept, as he did in cross-examination, that the determination that the vehicle type meets the technical requirements of the relevant regulatory acts is a conditio sine qua non; and that the conditio sine qua non can form part of the tenor of an act. This was, at least in part, the reasoning he provided before Waksman J in Crossley 1 as to why the Recall Decisions were binding on VW as a matter of law as to the existence of PDDs. But Professor Hofmann was also right to point out, at the same time, that the BGH did not, at least as at today, agree with him in the specific context of KBA Type Approval. Indeed, it seems to me that the Schleswig Administrative Court in its 17^th January 2024 decision also disagreed with him. Whilst he accepted that the pertinent statements in both cases were obiter, it is of some considerable relevance when determining what German Law is that there is not a single final decision in the German Courts in which the existence of Type Approval has, by virtue of a tenor which extends to a declaration of the non-existence of a PDD, in fact precluded a civil law claim for damages against a manufacturer. I regard the better view of what German Law is to be that set out, albeit obiter, in the reasoning of the BGH and the Schleswig Administrative Court in the decisions referred to above. The fact that the vehicle type meets the technical requirements of the relevant regulatory acts may be a conditio sine qua non, but this alone is insufficient to introduce into the binding tenor a declaration about the existence or non-existence of PDDs. Thus, there needs to be, objectively, material upon which to conclude that a positive decision was made by the regulator as part of the conditio sine qua non of the act in circumstances where the existence of defeat devices (albeit said by the manufacturer to be permissible) was required to be, and was, properly disclosed and considered.

Applying this analysis to the sample decisions, TADs 1 and 2 were granted at a time when there is no suggestion that any specific information or disclosure relating to the potential existence of defeat devices was either legally required or actually provided. In accordance with the decisions of the BGH and the Schleswig Court referred to above, and the evidence of Professor Hofmann, I conclude that under German Law the tenor of these administrative acts does not extend to any declaration touching upon the existence or non-existence of PDDs. This conclusion is supported by the evidence of Professor Hofmann as to the present state of German Law, at least as it would apply to civil actions for damages, on the basis of the decisions referred to above, and irrespective of his own views from a normative perspective.

The position in relation to TADs 3 and 4 is more complicated.

The two TADs are dated 2^nd September 2016 (TAD3) and 24^th February 2017 (TAD4). The Mercedes Defendants rely upon the amendment to the Implementing Regulation (Commission Regulation (EC) No 692/2008 implementing and amending the Emissions Regulation) (‘2016 IR’). There is no dispute that this required the manufacturer to provide information on the base emissions strategies (‘BES’) and the auxiliary emissions strategies (‘AES’). Article 5(11)(a) of the 2016 IR, introduced by amendment, required manufacturers to provide “information on the operation of all AES and BES, including a description of the parameters that are modified by any AES and the boundary conditions under which the AES operate, and indication of the AES or BES which are likely to be active under the conditions of the test procedures set out in this Regulation”.

The Mercedes Defendants contend that this information would identify (i) which AES and/or BES would be likely to be active in test conditions, (ii) by necessary implication which AES and/or BES would be unlikely to be active in such conditions, and (iii) information as to the parameters modified by any AES. Taken together, it is said that this would in substance amount to “describing the engine and emission control strategies and any condition(s) under which they would not operate as they do during testing”.

In his expert report, Professor Hofmann explained that the 2016 IR (1) did not require the manufacturers to provide a declaration that the vehicle did not contain any PDDs, and (2) did not expressly require the manufacturer to identify conditions under which the Emissions Strategies would not operate as they did during testing. It was not suggested that this construction of the 2016 IR was, of itself, wrong (although Professor Hofmann accepted in cross-examination that, depending on the transparency of the information in fact provided, it may be sufficient for the authority to work out for itself what happens outside the test procedure). On the basis of (1) and (2), respects in which the 2016 IR was less rigorous than that required by appendix 3a of EC Regulation 2017/1151 (the New Implementing Regulation (‘New IR’)), Professor Hofmann therefore concluded that if the KBA was supplied only with the information strictly required by EU type approval law, the assumption on which the preliminary issue proceeds does not lead to the inevitable conclusion that the Mercedes Defendants communicated to the KBA exactly what defeat devices were present in the vehicle types for which they applied for approval. On that basis, and in accordance with the BGH and Schleswig decisions, Professor Hofmann concluded that the absence of PDDs would not be part of the tenor.

Professor Hofmann was cross-examined on the basis of a document dated 24 May 2016, produced by the KBA (document ‘01-16’). Ms Davies KC put that the document was an order from the KBA with which manufacturers in Germany had to comply subsequently when making type approval applications. Professor Hofmann, entirely fairly, queried the legal status of the document, and the cross-examination proceeded on the assumption that Ms Davies KC’s characterisation of the document was correct. On the basis of this assumption, Professor Hofmann fairly accepted that the conclusion he had reached as to the position following the New IR, would apply equally to the period from the assumed implementation of document 01-16. This conclusion was that the words ‘Type Approval Granted’ could be understood as including a (binding) finding that there were no PDDs present for the vehicle types (not, of course, applicable to individual vehicles in the context of a civil law damages claim).

In his closing submissions, Mr Campbell KC on behalf of the Claimants advanced a number of serious issues with this. The first, and in my view justifiable, criticism is that there is no evidence as to what legal status the document has. It is not a document that was considered by Dr Neumann in his report. There is no evidence about whether such a promulgation from the KBA can, as a matter of law, have binding effect, noting as I do that the other instruments by which the information to be provided by manufacturers are set down are in the formal Implementing Regulations. There is no factual evidence about, for example, its dissemination or de facto application. The document is not referred to on the face of the relevant sample Type Approvals, nor are the documents (potentially) submitted in accordance with the document appended to the Type Approval, whether as part of the Information Package or otherwise. Its status cannot be assumed from the content of the document alone: as pointed out by Mr Campbell, the document states in respect of the information to be provided, ‘The KBA will disclose details on the extent of documentation. Annex 2 illustrates a first draft, preliminary and not conclusive.’ This is not resonant of a document with immediate binding legal force.

Ms Davies KC’s assertion of the document’s status, no doubt properly on instructions, is not evidenced. As a result, in the circumstances I have described above, the cross-examination by Ms Davies KC proceeded upon an assumption which has not, as matter of either fact or German Law, been established.

In these circumstances, I consider that the reservations made by Professor Hofmann about the lacuna in information required to be provided to the KBA, and which was not resolved until the New IR from July 2017, were fair ones, and are not satisfactorily met by document 01-16. Based upon the BGH and Schleswig decisions, which I consider to reflect the correct approach to the determination of the tenor of a Type Approval , at least in the context of a civil claim for damages, Professor Hofmann’s reservations remain justified. I accept his evidence that the regulatory regime in existence prior to the New IR precludes a finding in the tenor of TADs 3 and 4 that no PDDs were present.

Thereafter, I consider as a matter of generality (and not in the context of any particular sample) that after 1^st September 2017 EC Type Approval decisions under the New IR were capable of being understood as including a finding that there were no PDDs present in those vehicle types, for the reasons explained by Professor Hofmann. It remains the case that whether this is so for any particular Type Approval would remain to be considered on a case-by-case basis (or at least category by category). It obviously remains the case that even if such an administrative act was binding in respect of the particular vehicle type, it would not be binding under German Law in respect of any individual vehicle in the context of a civil claim for damages for the reasons already addressed.

In the circumstances, it is not necessary for me to decide three other points advanced by the Claimants. I nevertheless address them briefly below.

The first is that, even if further information was required and provided pursuant to document 01-16 or otherwise, it appears from the face of the collateral clause in sample TAD3 that the information was, by definition, provided at some date after the Type Approval itself was given. TAD 3 was issued on 2^nd September 2016, and certain further information was required to be provided by 30^th November 2016. As stated within the decision, there is a one-month period in which to appeal the administrative decision. Without having to decide the matter, I go no further than expressing the view that there appears to be some force in the Claimants’ submission that, as matter of logic:

The second point is that the whole vehicle Type Approval cannot contain a binding tenor relating to compliance with the prohibition on defeat devices within the Emissions Regulation because, at the point of approving whole car Type Approval, the KBA did not put its mind specifically to the substance of compliance with that Regulation. Instead, it was a task of collating the different previous approval decisions relating to all of the relevant components, likened in submissions by Mr de la Mare KC, on behalf of the Claimants, to collecting stickers in a Panini book. It may have been a number of years earlier, when type-approving the emissions control unit as a system, that the KBA actually put its mind to system’s compliance with the Emissions Regulation. When then granting Type Approval, it merely checks that the system within the vehicle has previously been the subject of its own Type Approval (i.e. the sticker is collected in the book). It may well be that there is no ‘one size fits all’ answer to the step-by-step analysis advanced, if looked at through the lens of the appeal process to test when decisions of substance are made. If a complete unit is or contains without more a prohibited defeat device, the point at which a decision is made about that unit’s compliance with the technical requirements is when it, as a unit, is given the relevant Type Approval . That decision needs to be appealed (and – on the Claimants’ broader analysis – this would include appeal by a consumer or consumer body with standing) within a month from the unit’s Type Approval decision. It would seem odd if the time for appealing the approval of that unit ran afresh each time it formed part of a wider whole vehicle Type Approval. Conversely, if the non-compliance with the Emissions Regulation stems from the interaction between two different units, or by reference to particular parameters introduced to control the unit, which happens only as part of the assembly of the whole vehicle type, then that is the point at which any decision of substance in respect of the absence, or otherwise, of a PDD is made (if it is made at all).

It is not therefore possible to apply with any certainty the ‘step by step’ objection advanced by the Claimants to any of the sample decisions: there is insufficient factual basis to do so. In any event, the Mercedes Defendants take a pleading point: nowhere in the Reply to the Defence do the Claimants make the point that if there is any KBA decision with binding effect in respect of the emissions control unit, it would reside as a matter of principle not in the whole vehicle Type Approval but in the underlying unit approval. I reject the suggestion that this positive case, if it is to be run, did not have to be pleaded. It is, for the reasons I have explained, not a pure question of legal analysis but would, or at least could, turn on the precise facts. It is not a point dealt with by the experts, or in respect of which any factual evidence has been called by either party (unsurprisingly, in light of the absence of any pleading). It is not a point open to the Claimants to take at this stage.

Finally, the Claimants submit that a Type Approval cannot be a binding decision as to the absence of a PDD in circumstances where this would be inconsistent with Article 30 of the Framework Directive. This is the Article pursuant to which steps are to be taken in circumstances where a vehicle accompanied by a certificate of conformity or bearing an approval mark do not conform to the type approved. The steps are required to being the vehicle back into conformity. The Claimants contend that the Recall Decisions are predicated on the basis that changes are required to bring the vehicle back into conformity. If the Type Approval was a binding decision that there are no PDDs, the authority could not mandate changes to bring it back into conformity: the only option would be to revoke or amend the Type Approval. Whilst this submission has a superficial logical attraction, it does not in fact add analytically to the position under German Law in circumstances where both experts agreed that the tenor of an act cannot be construed by matters which post-date the decision. It is not necessary to consider this argument further.

The answer to the first preliminary issue in respect of the sample Type Approval Decisions, on the basis of my findings of fact as to relevant German Law, is ‘No’.

There are two sample Recall Decisions, dated 23^rd May 2018, and 26^th October 2022. The binding effect of Mandatory Recall Decisions is not in dispute: they are binding as to the presence of PDDs.

The Mercedes Defendants rely upon the evidence of Professor Hofmann in respect of the binding effect of the Recall Decisions (given both in Crossley 1 and in the present case), to support their arguments on the binding effect of other types of decisions. Whilst there is some forensic merit to this submission, it does not in fact deal with the substance of the German Law decisions which underlie Professor Hofmann’s conclusions and, indeed, my conclusions above (and below).

The answer to the first preliminary issue in respect of Recall Decisions is, by agreement, ‘Yes’.

There are two sample Mandatory Update Decisions dated 12^th August 2019 (which relates to the sample Recall Decision dated 23^rd May 2018) and 20^th September 2019, following a Recall Decision dated 21^st June 2019 (which does not relate to a sample otherwise considered).

The Claimants rely upon a number of BGH decisions that make clear that, in the context of civil law claims for damages, the tenor of a Mandatory Update Decision is limited to the release of the update (effectively granting permission to make the update), and that statements in the decisions that no PDDs were found are elements of the reasoning and not part of the tenor:

In the relevant section of his report, Professor Hofmann noted the first two of the foregoing decisions, summarising them, correctly in my view, as concluding that the part of the Mandatory Update Decision containing the statement “the existing defeat devices were classified as permissible” was an element of the statement of reasons and not part of the (binding) tenor. He also identified a case of the Schleswig Administrative Court decision of 17^th January 2024 which, in respect of this point, disagreed with the BGH. At paragraph 371 the Court stated:

“…the court does not share the view of the Federal Court of Justice that the decisive regulatory object of the so-called release notices is exclusively the release of the software update and that further statements according to which existing defeat devices are permissible are merely justificatory elements that are not covered by the regulatory content and thus the factual effect of the administrative act itself (see, for example, BGH, judgment of 8 December 2021 - VIII ZR 190/19 -, para. 81 with further references).”

At paragraphs 144 and 145 of his report, Professor Hofmann sought to distinguish the Schleswig decision, contending that the Mandatory Update Decisions under consideration in this case appeared to resemble those considered by the BGH rather than the decision by the Schleswig Court. This attempt did not, in my judgment, stand up to scrutiny, and Ms Davies KC effectively extracted an acceptance in cross-examination that the distinction was untenable.

However, at paragraph 147 of his report he expressed the view that another reason for the BGH’s approach (in addition to the attempt to distinguish the underlying material) is the different procedural perspective: while the BGH considered whether there is something binding in the KBA decisions that prevented the Civil Courts from granting damages to the claimants if the courts’ analysis revealed the presence of PDDs, the VG Schleswig reviewed the KBA’s Release Decision with regard to possible errors on the side of the KBA, for example the erroneous certification of comprehensive compliance with EU Type Approval law. He concluded that, for the case at hand, the BGH’s jurisprudence applied particularly with regard to the question of a binding effect of an administrative decision for a Civil Court of law.

I accept this evidence. Indeed, ultimately, there appears to be no disagreement between the experts on this question as regards applicable German Lawin the context of civil law claims for damages. Dr Neumann gave the following evidence (Day 4, 101, 106):

Dr Neumann also confirmed that there had been no appeal against the BGH decisions referred to above, and so, at least in respect of the specific difference highlighted by the Schleswig Administrative Court, there was no immediate route to a higher court resolving the approaches (to the extent that they are truly irreconcilable, as opposed to co-existing as different approaches by different branches of the German legal system tasked with different types of case).

Dr Neumann was also asked about a more recent case in the Higher Regional Court of Celle, dated 20^th December 2023, relating to a Mercedes vehicle. The court concluded:

“Thus, according to the convincing interpretation of the Federal Court of Justice, the approval notice, understood as a (modifying) EC Type Approval (cf. VG Schleswig, judgement of 20 February 2023 - 3 A 113/18, juris margin number 210), only the legality of the approved vehicle type, not the specific vehicle (cf. BGH, judgement of 26 June 2023 - VIa ZR 335/21, juris margin number 14). The statements of the Federal Motor Transport Authority, according to which existing defeat devices are permissible, are elements of justification that are not covered by the regulatory content and thus the factual effect of the administrative act itself (cf. BGH, judgement of 8 December 2021 - VIII ZR 190/19, BGHZ 232, 94 margin number 81).

…In accordance with Article 3(2) of Regulation (EU) 2018/858, "EU type-approval" means the procedure whereby an approval authority certifies that a type of vehicle, system, component or separate technical unit satisfies the relevant administrative provisions and technical requirements of this Regulation. Except for editorial adjustments, this corresponds to the wording of the definition of EC type-approval in Art. 3 No. 5 of Directive 2007/46/EC. Even then, it is not a specific vehicle that is authorised, but the vehicle type.”

Dr Neumann fairly accepted that the Court was making the point that (1) statements in Mandatory Update Decisions that defeat devices were permissible was part of the reasoning and not the tenor; and (2) that type approval decisions, including any modifications to Type Approval, only relate to the vehicle type and not to the specific vehicle. There then was the following short exchange:

I am therefore left in no doubt that, as regards claims for damages in the Civil Courts in Germany, it is ‘settled jurisprudence’ that statements within Mandatory Update Decisions as to the presence or absence of PDDs are not part of the binding tenor of those decisions. I have already decided in the context of Type Approval Decisions that where the approach in German Law in the Civil Courts in the context of a damages claim might conflict with the German Administrative Court’s approach, it is the former which should properly guide me as to the applicable German Law in the present proceedings. It would be absurd if a claimant in Germany was not precluded from a damages claim by reference to the potential divergent approach of the German Administrative Courts to the identification of the ‘tenor’ of an act, but a claimant in England and Wales was so precluded (because of a finding on my part that the highest Civil Court in Germany had got it wrong). If authority were needed for such a proposition, support is found in the judgment of Lord Reid in Carl Zeiss Stiftung v Rayner & Keeler Ltd (No 2) [1967] AC 853 in which he said:

“it seems to me to verge on absurdity that we should regard as conclusive something in a German judgment which the German courts themselves would not regard as conclusive. It is quite true that estoppel is a matter for the lex fori but the lex fori ought to be developed in a manner consistent with good sense.”

Although this case was not dealing with the proper approach in the face of different strands of German Law, it is directly relevant to the effect of Ms Davies KC’s submission, namely that this Court should preclude a claim for damages on the basis of the conclusive effect of the KBA’s decisions when the same decision would not be regarded as conclusive in the context of precisely the same claim, if brought in the German courts.

This view is also supported by the expert evidence of Professor Hofmann who, as I have indicated, considers that the BGH jurisprudence applies particularly on the question of the binding effect of an administrative decision for a Civil Court of law.

Whilst no more than a forensic point, it is of some note that in the Schleswig decision, which concerned the standing of the DUH and upon which the Mercedes Defendants place considerable weight as stating the correct view of German Law as regards the tenor of a Mandatory Update Decision, it was (1) the position of the KBA that the tenor of the Mandatory Update Decision was limited solely to release of the update; and (2) the position of VW that the tenor did not include a finding that the vehicle contained no PDDs. Mr Kennelly KC fairly accepted, when asked directly, that VW advanced, and are maintaining on appeal from the Schleswig court, the diametrically opposite argument to the one being maintained by the Defendants (including VW) in this Court as to the effect of the Mandatory Update Decision.

Finally, I cannot accept the Mercedes Defendants’ submission, in addition, that it follows from the binding effect of Recall Decisions on individual vehicles that the resultant Mandatory Update Decisions must also so bind individual vehicles. Whilst I agree that a certain logic underpins this submission:

I therefore conclude that the answer to the first preliminary issue as regards Mandatory Update Decisions, on the basis of my findings of fact as to relevant German Law, is ‘No’.

There are two sample Voluntary Update Decisions, dated 14^th September 2018 (VUD1) and 16^th July 2020 (VUD2).

The Voluntary Update Decisions are in materially identical form. They state on the first page that they are a “National Type Approval”, and “issued by [the KBA] according to §20 and 22 Straßenverkehrs-Zulassungs-Ordnung (StVZO) for a type of the following approval object”, namely, “modification of engine management system”. The test report underlying the approval is identified, and the certificate state that “The offer for sale is only allowed on the listed vehicles under the specified conditions”. Section 10 of the certificate states: “The requirements of Article 31, paragraphs 5, 6, 8, 9 and 12 of directive 2007/46/EC - Sale and entry into service of parts or equipment which are capable of posing a significant risk to the correct functioning of essential systems - are met.” Section 12 states “Approval granted”. An annex to the certificate includes a table which identifies by category, with Vehicle Identification Numbers “from” and “to” the specific vehicles to which the approval applies (i.e. the vehicles to which the software update which is the subject of the Voluntary Update Decision is to be applied).

There is not, on the face of the Voluntary Update Decision (and in contrast to the Recall Decisions or Mandatory Update Decisions), any express reference to the existence, non-existence or detection of PDDs.

It is the Claimants’ case that Voluntary Update Decisions are made without legal basis and are a nullity under German Law. However, their primary position, irrespective of whether a Voluntary Update Decision are legally flawed, is that the tenor of the Voluntary Update Decision cannot be any wider than applicable to a Mandatory Update Decision.

This short point was fairly accepted by Dr Neumann, who agreed in terms that assuming the BGH’s view as to the tenor of a Mandatory Update Decision is correct, then it would follow from that reasoning that the tenor of the Voluntary Update Decision is limited simply to modification and nothing else. This was also the thrust of Professor Hofmann’s evidence, who specifically pointed out the distinction between the approaches in the Administrative Courts and the Civil Courts in the context of answering questions about the Voluntary Update Decisions. I have, as set out above, determined that the settled jurisprudence in German Civil Courts restricts the tenor of Mandatory Update Decisions so that Type Approval and Mandatory Update Decisions are not construed so as to preclude damages claims in the Civil Court system. In line with the evidence of both Dr Neumann and Professor Hofmann – and indeed, in line with what seems fairly obvious – it is plain that if presented with the argument that the tenor of a Voluntary Update Decision includes a binding determination on the absence of a PDD (so as to preclude a civil claim for damages), the German Civil Courts would almost inevitably reject it.

This is sufficient to answer Preliminary Issue 1 in the negative, insofar as it relates to Voluntary Update Decisions.

Given my conclusion above, it is not strictly necessary to consider the individual applicability of the Voluntary Update Decisions to specific vehicles rather than simply vehicle types or the question of nullity. However, I do so (at least in part) in case I am wrong in respect of the foregoing determination.

As to the application to specific vehicles, I have already determined that the relevant German Law as reflected in the decisions of the BGH adopts the approach that Type Approval or Mandatory Update Decisions apply merely to ‘type’ and not individual vehicles. The Mercedes Defendants rely, in the context of Voluntary Update Decisions, on the fact that – in contrast to the sample Type Approval and Mandatory Update Decisions – the sample Voluntary Update Decisions specifically identify by VIN numbers the specific vehicles to which the decision applies.

I do not accept, as submitted by Ms Davies KC, that it is the inexorable logic of the fact of the list of VINs, and the necessary limitation on the application of the determination to specific vehicles within the type rather than the type as a whole, that the binding effect of the tenor under German Law says anything about the particular status of an individual vehicle. The BGH makes clear that the specific status of a particular vehicle remains always for the Civil Courts to determine on the evidence relating to that particular vehicle. In other words, a specific vehicle’s condition, in respect specifically to the presence or absence of PDDs in the context of a civil claim for damages, is not pre-determined by the KBA’s administrative approval. The fact that the KBA approval is limited to an identifiable sub-set of a type (by reference to groups of VINs) is not likely, in my view, to dislodge this approach were the issue to come before the German Civil Courts.

Considerable time was spent by the parties exploring the Claimants’ contention that the Voluntary Update Decisions were void under German Law. Whilst not strictly necessary to decide in light of my conclusions above, I consider the matter below.

The starting point is that the Voluntary Update Decisions were issued as National Type Approval Decisions rather than as decisions made, or purported to be made, as an amendment to the EU Type Approval through the provisions within the Framework Directive that permit amendment. It is to be noted that the definition of a National Type Approval under the Framework Direction is “a type-approval procedure laid down by the national law of a Member State, the validity of such approval being restricted to the territory of that Member State”.

The Mercedes Defendants’ pleaded case was, initially, that the Voluntary Update Decisions were decisions granting a certificate authorising the sale and supply of Voluntary Updates “corresponding to” the requirements of Article 31(5) of the Framework Directive and/or Article 56(2) of the Framework Regulation (see paragraph 148.4) of the Generic Defence. When further information was sought as to whether the decisions were “made…pursuant to” the identified provisions, the Mercedes Defendants clarified that they were not; and that the decisions were made pursuant to §20 and 22 StVZO, certifying that (as quoted above) Article 31 paragraphs 5, 6, 8, 9 and 12 of the Framework Directive had been met. Although the pleaded case remains linked to Article 31(5), the legal analysis now advanced by the Mercedes Defendants accepts (as did Dr Neumann in evidence) that in fact Article 31(5) is not relevant to such power as may have been exercised by the KBA. The argument now hinges on Article 31(12) of the Framework Directive.

Article 31 of the Framework Directive reads as follows:

“Sale and entry into service of parts or equipment which are capable of posing a significant risk to the correct functioning of essential systems’

1.

Member States shall permit the sale, the offer for sale or entry into service of parts or equipment which are capable of posing a significant risk to the correct functioning of systems that are essential for the safety of the vehicle or for its environmental performance, only if those parts or equipment have been authorised by an approval authority in accordance with paragraphs 5 to 10.

2.

Parts or equipment subject to authorisation in accordance with paragraph 1 shall be inserted in the list to be established in Annex XIII. […].

3.

Paragraph 1 shall not apply to original parts or equipment which are covered by a system type-approval with respect to a vehicle and to parts or equipment which are type-approved in accordance with the provisions of one of the regulatory acts listed in Annex IV, except where those approvals relate to aspects other than those covered in paragraph 1. Paragraph 1 shall not apply to parts or equipment exclusively produced for racing vehicles not intended for use on public roads. If parts or equipment included in Annex XIII have a dual use for racing and on the road, these parts or equipment may not be sold or offered for sale to the general public for use in on-road vehicles unless they comply with the requirements of this Article.

Where appropriate, the Commission shall adopt provisions for identifying the parts or equipment referred to in this paragraph. […]

12.

As long as a decision as to whether or not a part or piece of equipment is to be included in the list referred to in paragraph 1 has not been taken, Member States may maintain national provisions dealing with parts or equipment which are capable of posing a significant risk to the correct functioning of systems that are essential for the safety of the vehicle or its environmental performance.”

There is no dispute that the list referred to in paragraph 3 of Article 31 does not exist. Paragraph 2 makes clear that parts or equipment subject to authorisation pursuant to the paragraph 1 approval process (as then described in more detail in paragraphs 5-10 of the Article, which have not been reproduced above) would be inserted into such a list.

Paragraph 3 makes clear, moreover, that paragraph 1 does not apply to “original parts or equipment” which are covered by a system type approval with respect to a vehicle and to parts or equipment which are type-approved in accordance with the provisions of one of the regulatory acts listed in Annex IV. Annex IV includes the Emissions Regulation.

The Mercedes Defendants argue that this Article must be read in conjunction with Recital 14 and Article 1(2). Recital 14 states:

“The main objective of the legislation on the approval of vehicles is to ensure that new vehicles, components and separate technical units put on the market provide a high level of safety and environmental protection. This aim should not be impaired by the fitting of certain parts or equipment after vehicles have been placed on the market or have entered service. Thus, appropriate measures should be taken in order to make sure that parts or equipment which can be fitted to vehicles and which are capable of significantly impairing the functioning of systems that are essential in terms of safety or environmental protection, are subject to a prior control by an approval authority before they are offered for sale. These measures should consist of technical provisions concerning the requirements that those parts or equipment have to comply with.”

Article 1 of the Framework Directive states:

“This Directive establishes a harmonised framework containing the administrative provisions and general technical requirements for approval of all new vehicles within its scope and of the systems, components and separate technical units intended for those vehicles, with a view to facilitating their registration, sale and entry into service within the Community.

This Directive also establishes the provisions for the sale and entry into service of parts and equipment intended for vehicles approved in accordance with this Directive.”

It is said that these provisions demonstrate a clear distinction between new vehicles and systems, components and separate technical units intended for type approved vehicles on the one hand, and certain essential ‘parts or equipment’ to be fitted to vehicles which are not covered by the provisions on the other hand. It is said that the former require type approve pursuant to Articles 6 to 11 of the Framework Directive. The latter are subject to Article 31. Ms Davies KC argues that the distinction is carried through to the definitions section in Article 3, in which mutually exclusive definitions are provided for “system”, “component”, “separate technical unit” and “original parts or equipment”. Whilst it was submitted that none of these phrases are found in Article 31, this is not correct: the phrase, “original parts or equipment” is found in the first line of Article 31(3).

It is clear, as accepted by Ms Davies KC in oral submissions, that the (original) emissions control system for any particular variant included its software, and was subject to a system Type Approval which included approval of that (original) software. However, the Mercedes Defendants’ argument is that instead of being able to amend the system Type Approval, or the whole vehicle type approval, in light of a modification of the original software, the replacement software is to be considered, instead, a separate “part” or “equipment” for the purposes of approval under a national scheme pursuant to Article 31. Ms Davies KC, importantly, accepted in closing submissions that if there had been an ability to amend a Type Approval in respect of the software update then Article 31(12) of the Framework Directive would not have been available. She contended that in circumstances where production of the variant had ceased (which is the case for the two sample decisions) and/or where it was not necessary to amend the information package in respect of the introduction of revised software for the emissions control system, it was not possible to amend the emissions system Type Approval, or whole vehicle Type Approval.

It is sensible, therefore, to consider first whether Ms Davies KC is right in her contention (advanced, as the Claimants correctly point out, in closing for the first time) that it is not possible to amend Type Approval insofar as it applies to variants of a type in respect of which production has ceased. The first point is that, even on Ms Davies KC’s argument, this difficulty does not arise in the context of any Voluntary Update Decisions in respect of vehicle variants that were still in production (although it does arise in respect of the two samples).

Ms Davies KC relies upon Article 17(1) and (2) of the Framework Directive. This Article (‘Termination of Validity’) states:

“1.

An EC type-approval of a vehicle shall cease to be valid in any of the following cases:

(a)

new requirements in any regulatory act applicable to the approved vehicle become mandatory for the registration, sale or entry into service of new vehicles, and it is not possible to update the approval accordingly;

(b)

production of the approved vehicle is definitively discontinued voluntarily;

[…]

2.

Where only one variant within a type or one version within a variant becomes invalid, the EC type-approval of the vehicle in question shall lose validity only in so far as the particular variant or version is concerned.”

Mr de la Mare KC argues that, first, termination of ‘validity’ is limited to the ability to rely upon the type approval certificate for the purposes of putting a new car on the market. It does not mean that the determination has been revoked or repealed. To support this submission, reliance was placed on the judgment of Waksman J in Crossley & Others v Volkswagen AG & Others [2021] EWHC 3444 (‘Crossley 2’), in which he determined in the context of the ‘validity’ of a Certificate of Conformity (not a Type Approval) that the existence of a PDD -or some other non-conformity with the underlying requirements - did not, of itself, render such a certificate ‘invalid’. This is clearly not a direct analogy.

Similarly, the Claimants referred to the case of Case C-639/11 European Commission v Republic of Poland (ECLI:EU:C:2014:173), from which it is clear that the (re-)registration of a vehicle may take place within a Member State following its purchase and registration some time earlier in another Member State. That case concerned people returning to Poland having bought (and registered) a passenger vehicle in the United Kingdom. However, the requirement to register a vehicle – which may well happen in the case of re-registration in a different Member State from that where the vehicle was originally purchased - is dependent not necessarily upon the continuing validity of the Type Approval, but upon the existence of a valid Certificate of Conformity, which is issued at the point of manufacture. The lapsing of the validity of a Type Approval does not, of itself, seemingly impact upon the continuing validity of a Certificate of Conformity which was issued at the time when the underlying Type Approval was, by definition, valid.

Neither of these authorities, therefore, quite hit the mark. Nevertheless, it is clear from Article 9 of Commission Regulation (EC) (No 692/2008) (the ‘Implementing Regulation’), implementing and amending the Emissions Regulation, that the overall regulatory framework includes in-service conformity measures to confirm the functionality of any pollution control devices during normal use after manufacture. The vehicles to be selected for conformity testing are required to have been in service for at least 15,000km or 6 months, and for no more than 100,000km or 5 years. It is obviously readily conceivable that during this time the variant itself will have ceased production and, under Article 17 of the Framework Directive, the Type Approval become ‘invalid’. Paragraph 4.4 of ‘In-Service Testing’ forming part of Appendix 1 to Annex 2, which is referenced in Article 9, provides:

“If a component or system operates in a manner not covered by the particulars in the type-approval certificate and/or information package for such vehicle types and such deviation has not been authorized under Article 13(1) or (2) of Directive 2007/46/EC, with no malfunction indication by the OBD, the component or system shall not be replaced prior to emission testing, unless it is determined that the component or system has been tampered with or abused in such a manner that the OBD does not detect the resulting malfunction.”

This provision comprehends the situation where the in-service testing requires consideration of the potential for deviations from the type approval certificate which have not (and, by definition, those which have) been authorised under Article 13 of the Framework Directive (the Article permitting amendments). For this to be workable in practice, the Type Approval certificate must have continuing ‘validity’ in the context of in-service testing, and amendments to that certification may need to be made: for example, where some fault is detected in-service, and a recall is put in place to replace a particular part (or update the software) for the safety of consumers, at some point after production of the variant has ceased but prior to the 5-year period envisaged by this system. This provision expressly envisages that deviations from the original Type Approval may well have been authorised under Article 13 of the Framework Directive (which permits amendments). There is no reason why deviations, if introduced by a replacement part (including software), should not be the subject of Framework Directive compliant authorisation during the 5 years of in-service conformity testing whether or not the variant has ceased production.

This analysis provides clear support for the Claimants’ submission that the word “validity” in Article 17 is limited to the context of the production and sale of new vehicles pursuant to type. Article 17 does not mean that Type Approval is somehow “void” for all purposes. The IR demonstrates that it remains the (valid) benchmark against which in-service testing conformity is measured, and it must in my judgment, in order to allow this sensibly to work, be capable of amendment to reflect any necessary in-service changes which have been introduced notwithstanding the fact that (new) production of the variant has permanently ceased.

The Claimants’ further point is that Article 17 deals only with whole-vehicle Type Approval as is made plain by the first sentence of the first sub-paragraph. It does not refer at all to a system type approval. As conceded by Ms Davies KC, the original software was approved as part of the emission control unit system approval. There is no equivalent provision to Article 17 relating to a point at which the validity of a system type approval becomes invalid. This is unsurprising: whilst, once a variant of a vehicle has been permanently discontinued, there is no need for the whole vehicle type approval to continue to be valid (at least for the purposes of registering new such vehicles), this does not apply to systems forming part of a vehicle variant. The systems may be used in many variants over many years, and potentially by a variety of manufacturers.

There is, therefore, no basis upon which Article 17 of itself can be said to have prevented an amendment to the emissions system type approval, as the Voluntary Update Decision purported to do, insofar as this was otherwise possible in accordance with the Framework Directive.

The second reason advanced by the Mercedes Defendants specific to the inability to amend the system type approval is that the ability to amend depends upon the need to change the information package. Article 13 of the Framework Directive states:

“The manufacturer shall inform without delay the Member State that granted the EC type-approval of any change in the particulars recorded in the information package.”

The Mercedes Defendants argue that the required content of the information package is defined in the Implementing Regulation, in Article 3 and Appendix 3 to Annex I, which specified the required information document for an ETA. The Mercedes Defendants rely upon the Claimants’ own submission that this document mentioned software, and even hardware, of the emission control system only in outline.

The Claimants take the On-Board Diagnostics system (‘OBD’) as an example of an area where the information which was required to be provided as part of the Type Approval ought to have been affected by the changes which the information provided by the Voluntary Update Decision describes.

So, for example, at 3.2.12.2.7.5, Appendix 3 to Annex I requires a list of all OBD output codes and formats used is required. It is clear from the “comprehensive document describing all sensed components with the strategy for fault detection” described at 3.2.12.2.7.6.3, and the example table at 3.2.12.2.7.6.4 that the information package ought to include full details of all the fault codes generated by the OBD for the relevant components monitored.

It is clear from documentation supplied in respect of VUD2 that there were changes to the OBD as part of the software updates which affected, amongst other things, two NOx sensors, and the fault codes generated. For example, NOx Sensor 2 removal detection monitoring included a new diagnostic function, and the relevant fault code changed from P229F to P22FD. On the basis of this example, I do not accept the Mercedes Defendants’ general submission that, approached properly, the software changes did not introduce changes which, providing the information package produced with the original type approval complied with the requirements of the Implementing Regulation, would have led to the information package being changed. Given the significance of the changes to the way the system operated, introduced by the software updates (reducing NOx emissions by 20-30%), it might be thought odd if this were not the case.

Similarly, it is clear that further testing was required to be undertaken, and was in fact undertaken, in respect of the software updates. The results of the tests are appended to both VUD1 and VUD2. The Mercedes Defendants deny, however, that there would nevertheless be a need to amend the particulars by the provision of a new test report to support the grant of type approval unless the regulatory requirements in relation to the continued production of vehicles of the relevant type had changed, which they contend was not the case in respect of the sample Voluntary Update Decisions.

I accept the Claimants’ submission that this argument is based upon a misreading of the relevant Articles from the Framework Directive. Once a Member State has been informed of any change to the particulars of the information package, that Member State may inform the manufacturer that fresh tests are “necessary”. Article 14(2) then identifies three circumstances in which the amendment to Type Approval shall be designated an ‘extension’ rather than a “revision”. The third of these relates to new requirements under regulatory acts; the first relates to where further inspections or fresh test are required. This, which should plainly be read disjunctively, relates to where (as here) the Member State (through the KBA) in fact required further testing.

I am therefore unable to accept the Mercedes Defendants’ second reason advanced as to why the KBA was unable to require the proposed software amendments to be channelled through the EU regulatory framework by an amendment to the vehicle type approval and/or the underlying emissions type approval. In circumstances where amendment to type approval was permissible (and necessary) in the circumstances of the Mercedes Defendants’ proposed software changes there was, as Ms Davies KC accepted, no room for the application of Article 31(12) by way of “national” type approval.

Standing back from the detail of the arguments, this is an unsurprising conclusion. It was no doubt readily foreseeable that software updates would be introduced to vehicles to prevent obsolescence or introduce new features even if the variant of the vehicle itself had ceased production. The inability to amend Type Approval in these circumstances would seem to be an improbable lacuna in the EU approval scheme, particularly in circumstances where it would potentially lead to precisely the fragmentation of a harmonised scheme which, in the Mercedes Defendants’ broader submissions on EU law, is so essential to the efficacy of the system. That such a lacuna would not have been intended gains support, as submitted by Mr de la Mare KC, from the Opinion of Advocate General Rantos in the conjoined cases of C-145/20 DS v Porsche Inter Auto GmbH & Co. KG- Volkswagen AG, Case C-128/20 GSMB Invest GmbH & Co. KG v Auto Krainer GesmbH and Case C-134/20 IR v Volkswagen AG, in particular at [138] (a passage expressly approved in the subsequent judgment of the CJEU at [89] of their judgment):

“Lastly, as has already been stated, the objective of Regulation No 715/2007 is to guarantee a high level of protection of the environment. The assumption that the sole standpoint to be adopted is that of the date of manufacture of the vehicle would mean that vehicle manufacturers would not be required, once a vehicle has been put into service, to install a defeat device that complies with the provisions of that regulation. Such an assumption would be contrary to the objective of the regulation. Indeed, in order to circumvent the obligations laid down in that regulation, manufacturers would need only to replace the original pollution control device, which would be compliant with [the Emissions Regulation], with a less effective defeat device that does not ensure compliance with the limit values laid down for NOx.”

It is next necessary to consider the fact that the KBA did not, as already explained, even purport to make the Voluntary Update Decisions expressly pursuant to Article 31(12). The KBA purported to make its decisions pursuant to §20 and 22 of the StVZO. Dr Neumann sought to contend that this approach was legitimate in his report, where he said, “According to Sections 20 and 22 of the StVZO ‘general operating permits’ can be issued for parts of vehicles of vehicle class M, which also includes the relevant vehicles.” The authority he cited for the applicability of Sections 20 and 22 was a textbook entitled ‘Road Traffic Law’ by Freymann and Wellner (2^nd Edition). As Dr Neumann accepted in terms, however, this text says precisely the opposite, namely that Sections 19 and following of the StVZO (including Sections 20 and 22) have no application to the type M vehicles which are of concern in this litigation. As such, Dr Neumann was unable to provide any independent support for his opinion.

I also consider that that there is a fundamental inconsistency in the Mercedes Defendants’ case in this regard. Section 22 (the only section upon which they relied in Closing) relates only to “parts of vehicles if the part forms a technical unit that can be dealt with independently in the authorisation procedure”. Yet in the context of their arguments about Article 31(12), it is necessary for the Mercedes Defendants to insist that the software update is not a “separate technical unit” (see paragraph 84 of their Closing Note). If that is right, there was no power under Section 22 (irrespective of its general inapplicability as described by Freymann and Wellner).

Even if, therefore, Article 31(12) created a pathway by which a national authority could make a decision which had the effect, as Dr Neumann contended, of amending or supplementing a type approval granted under the EU harmonised framework, such a decision could not have been effected through Sections 20 and 22 of the StVZO, as the KBA purported to do. I therefore consider that the approach of the KBA was unlawful.

In light of this conclusion, the question arises as to whether, as a matter of German Law, the Voluntary Update Decisions are a nullity (albeit the question is moot in this case where, as I have determined above, the tenor of the Voluntary Update Decisions does not, in any event, contain a binding determination as to the presence or absence of PDDs).

The test for a nullity is prescribed by Section 44(1) of the Administrative Procedure Act as follows:

“an administrative act is null and void if it is vitiated by a particularly serious error and this is obvious from a reasonable assessment of all the relevant circumstances”.

It is implicit in this test, as was accepted by Professor Hofmann in cross-examination, that under German Law there is a distinction between a decision which is a nullity and a decision which is unlawful but nonetheless effective.

Whilst the Claimants are right that the question of pure law is the test I have to apply, and that it is for me to determine whether the test is satisfied, this does not mean, as the Claimants’ submissions imply, that the evidence of the German Law experts as to how this test might be approached in practice in the German Courts is irrelevant. It is not.

Both experts agreed that the test represents ‘a very high bar’. In answers to questions from Ms Davies KC, Professor Hofmann accepted that the 1998 decision of the BVerWG relating to tax issues expressed generally applicable principles of German Law to the question of the nullity of an administrative act. The relevant passage stated as follows:

“According to the consistent case law of the Federal Administrative Court and the Federal Fiscal Court, the consequence of the nullity of an administrative act resulting from legal defects has always been regarded as a special exception to the principle that an act of state authority carries the presumption of its validity … Particularly serious within the meaning of § 125 I AO 1977, which literally corresponds to the regulation of § 44 I VwVfG, is therefore only an error which makes the administrative act affected by it appear absolutely intolerable, i.e. incompatible with fundamental constitutional principles or essential values inherent in the legal system…. On the other hand, the nullity of an administrative act cannot be assumed simply because it lacks a legal basis (so-called “lawless” administrative act - …) or the legal provisions in question have been incorrectly applied. The serious error of the administrative act must be obvious to a reasonable citizen…. An administrative act can therefore only be considered invalid if the requirements of proper administration are violated to such an extent that no one can be expected to recognise the administrative act as binding….”

Professor Hofmann also accepted, and I find, that if it were to be concluded that there are arguments that can reasonably be made to the effect that the KBA was entitled to act as it did in relation to the Voluntary Update Decisions, in accordance with the test of German Law, the decisions would not be void, even if those arguments were ultimately shown to be wrong.

Lacking a degree of definitiveness, the language used by Professor Hofmann, both in his report and in answer to questions was that the Voluntary Update Decisions were ‘candidates’ for nullities, and that he accepted that it is at least ‘possible’ that the decisions are not void.

I readily accept, as contended by the Claimants, that the consequences of seeking to act by way of national type approval had potentially important consequences in the context of transparency and engagement with other Member States which would have been inherent if the route to approval had been by way of amendment to the existing EU Type Approval. On the other hand, I have also concluded that the Voluntary Update Decisions have a limited binding tenor, and in no way do they affect the rights of private citizens to allege and prove that, notwithstanding the KBA’s views of the effect of the update, PDDs remain in the relevant vehicles. This limited tenor, in my view, significantly dilutes the ability of the decision ‘to cut across fundamental constitutional principles’. Perhaps it is in light of that that, as a matter of fact, no party with standing (which would cover anyone with a legitimate interest in doing so) has in fact sought to challenge the Voluntary Update Decisions in the German Administrative Courts, notwithstanding the significant passage of time since they were made.

Taking account of all the circumstances, and notwithstanding my determination that on my analysis the KBA erred in law in making national Type Approval Decisions with regard to the software updates, I accept the evidence of Dr Neumann on this point: that the error was not so obvious or serious that it made the administrative act affected by it completely incompatible with fundamental constitutional principles or essential values inherent in the legal system. As such, I conclude that the decisions were not lawful, but nevertheless, are effective under German Law.

In circumstances where I have determined under German Law that the tenor, or regulatory content, of the KBA Decisions (Recall Decisions aside) includes no determination about the presence or absence of PDDs, the existence of additional potential criteria for the binding effect of an administrative act posited by Professor Hofmann, in addition to the three agreed between the experts and set out at paragraph ‎23 above, does not arise.

The fourth condition was that the act should not be suspended. This is not relevant to the present case.

Professor Hofmann’s fifth condition was that an affected party must have had the opportunity to challenge the administrative act.

Insofar as it did arise, I would have concluded that no such fifth condition universally exists in German administrative law. Professor Hofmann himself identified that “the overwhelming majority of courts and legal literature in Germany seem to differ”, taking the view that no such additional condition exists. I would also have formed the view that individuals did have standing to challenge the KBA decisions, even though that pre-existing right was only more recently recognised in the CJEU judgment in case C-100/21 QB v MB. That case did not create a fresh right; it stated the law as it always had been, as Professor Hofmann fairly accepted. Whether that right amounts to effective judicial protection in the context of EU law is, of course, a different matter.

This does not strictly arise in circumstances where I have concluded that the Type Approval Decisions, Recall Decisions and Voluntary Update Decisions do not include determinations as to the presence or absence of PDDs which are binding as a matter of German Law in civil claims for damages. However, if I am wrong about this, it is necessary to consider whether, if binding under German Law, these decisions would be binding on this Court under English law (through retained EU Law). The preliminary issue does not, of course, touch upon an underlying question of whether a breach of the Emissions Regulation is, in itself, actionable in private law (which, as Mr Peretz KC representing Ford was keen to emphasise) remains in dispute between the parties, notwithstanding the decision of the CJEU in QB v MB.

I consider first the position under the Framework Directive.

There is no dispute between the parties that the Type Approval system under the Framework Directive constitutes total harmonisation. See, for example, Recital 2, which states:

“For the purposes of the establishment and operation of the internal market of the Community, it is appropriate to replace the Member States’ approval systems with a Community approval procedure based on the principle of total harmonisation.”

See also Article 1:

“This Directive establishes a harmonised framework containing the administrative provisions and general technical requirements for approval of all new vehicles within its scope and of the systems, components and separate technical units intended for those vehicles, with a view to facilitating their registration, sale and entry into service within the Community. This Directive also establishes the provisions for the sale and entry into service of parts and equipment intended for vehicles approved in accordance with this Directive. Specific technical requirements concerning the construction and functioning of vehicles shall be laid down in application of this Directive in regulatory acts, the exhaustive list of which is set out in Annex IV.”

As noted by Waksman J at [294] of Crossley 1, under the Framework Directive, only the relevant approval authority (here, the KBA) can grant the type approval for the relevant vehicles, and it is not open to authorities (including courts) in other Member States to go behind the grant of type approval by the KBA which is valid across the EU. Article 30 then provides that it is only the relevant approval authority which is to impose the relevant measures in the event of non-conformity. Article 3(3) then allows for an approval authority in any other Member State to request the granting relevant authority to take appropriate measures where the former can show a lack of conformity. It is then for the granting relevant authority to undertake the measures required. If there is a dispute between the two authorities as to whether, and if so what, action is required, there is a settlement mechanism which would involve the Commission effectively as a mediator.

The Claimants do not dispute that this is the effect of the totally harmonised regime at least in the context, as per Article 2, that harmonisation is “with a view to facilitating their registration, sale and entry into service within the Community”. The purpose of the regulatory regime is to set the conditions in a harmonised framework within which products are allowed to be placed on the market in all Member States. It is not, argue the Claimants, to determine all questions of private law liability.

The nature of total harmonisation was considered expressly in Crossley 1 in the context of Recall Decisions. Ms Davies KC was right to point out that, in his decision, Waksman J was addressing two limbs of a preliminary issue, as set out in paragraph [287] of the judgment:

“Is the High Court of England and Wales bound (having regard to the terms and operation of the EC Type-approval legislation and pursuant to its duty of sincere cooperation) by the finding of the competent EU type-approval authority (the... KBA, or by the [UK's] Vehicle Certification Agency (VCA) in this case) that a vehicle contains a defeat device in circumstances where that finding could have been, but has not been, appealed by the manufacturer; and/or is it an abuse of process for the Defendants to seek collaterally to attack the KBA's and VCA's reasoning or conclusions by denying that the affected vehicles contain defeat devices?”

The first limb required an analysis of the type approval legislation to determine whether the KBA decision was binding on VW; and the second limb considered the same question through the lens of abuse of process.

Given the centrality of Waksman J’s decision to many of the arguments canvassed before me, I set out the relevant passages below in full:

“The binding nature of the KBA Letters as a matter of EU law

It can be seen that the judge answered ‘yes’ to both limbs. Ms Davies KC was, therefore, also correct in her submission that, as the extract further below demonstrates, the decision in respect of bindingness in answer to the first limb was not an analysis of ‘res judicata’/abuse of process: it was the result of an analysis of the harmonised regime on the ability of authorities (which would include courts) to second-guess the decision of the approving authority, supported by the principle of sincere co-operation. In short, the central thrust of Waksman J’s reasoning was that, as urged upon me by the Defendant Manufacturers, if it was open to the courts of other Member States to challenge the findings made by the relevant approval authority that there was a defeat device this would run contrary to the principle of “full effectiveness” of the approvals regime: it would simply play havoc with the whole harmonised scheme.

However, it is also fair to point out that Waksman J considered the principle applied to “at least” the addressee, and noted that no point had been taken in respect of non-addressees. Given that embedded within the preliminary issue was whether the addressee was bound (with the specific assumption that the addressee could have appealed the decision of the KBA but chose not to), Waksman J’s decision cannot be taken to have decided the question of whether or in what circumstances a decision by a Type Approval authority binds a non-addressee in a private action for damages.

It is, of course, not without irony and in the best tradition of the Bar that it was, in Crossley 1, the defendant manufacturer VW (ably represented then, as now, by Mr Kennelly KC amongst others) seeking to persuade the Court that the Recall Decision was not binding; and Mr de la Mare KC, then as now, equally ably representing the Claimants, arguing that the decision was binding as a matter of EU law. Before me, Ms Davies KC and Mr Kennelly KC rely upon the decision of Waksman J and contend that his analysis is equally applicable to each variety of KBA Decision I am considering. Although accepting that paragraphs 394 to 396 were obiter, they contend that this passage nevertheless represents the correct position in law as to the binding effect of a KBA decision on the Claimants. Mr de la Mare KC argues that Waksman J’s conclusion that the KBA decision was, as a matter of EU law, binding on the addressee was entirely correct; but this does not mean such decisions would be binding on the Claimants. He says, squarely, that the obiter remarks in paragraphs 394 to 396 were wrong. Such a conclusion flows from the principle of effective judicial protection, and in particular consumer protection principles at the EU level.

There are a number of reasons why, in my judgment, the Claimants’ argument is to be preferred.

First, the Defendant Manufacturers, notwithstanding no doubt exhaustive research, were unable to point to any previous authority in which a regulator’s approval of a product within a totally harmonised regime has been successfully used to preclude a private law action for damages in respect of a defect in that product. Of course, the absence of authority by definition cannot be determinative of the question; but it is striking, particularly given the large number of categories of products, from medical devices to fertilising products to toys, which are subject to EU harmonised regulation, that there is no example of a defence against a claim by reference to the existence of Type Approval.

Second, there are a number of CJEU decisions which are inconsistent with the proposition that the existence of valid Type Approval, or a mandatory update decision approving the installation of software, has the effect of precluding a national court from determining whether, as a matter of fact, a particular vehicle contains a prohibited defeat device and, in doing so, form a different view from the KBA.

Three cases, C-145/20 DS v Porsche Inter Auto GmbH & Co. KG- Volkswagen AG, Case C-128/20 GSMB Invest GmbH & Co. KG v Auto Krainer GesmbH and Case C-134/20 IR v Volkswagen AG were dealt with jointly in the Opinion of Advocate General Rantos (although the cases were not, themselves, joined), already cited at paragraph ‎128 above. As set out in paragraph 4 of the Opinion, these three cases concerned vehicles equipped with software installed in the electronic engine controller which, on the basis of certain outside temperature and driving altitude conditions, limited the reduction of NOx emissions, which resulted in the limit values laid down in the Emissions Regulation being exceeded.

In GSMB, the claimants brought an action before the Landesgericht Klagenfurt (Regional Court, Klagenfurt, Austria), the referring court, seeking the cancellation of that sale, on the basis of Paragraph 879(1) and Paragraph 932(4) of the Allgemeines bürgerliches Gesetzbuch, the Austrian General Civil Code, (‘the ABGB’). This provided for certain consumer rights in respect of defective products. GSMB claimed that, on the date of the sale, it had believed that it was purchasing a new, environmentally friendly vehicle, the exhaust gas emissions of which complied with the statutory requirements. However, following an update of the software installed in the electronic engine controller fitted, an update carried out by VW, the purification of exhaust gas was deactivated at an outside temperature of below 15 °C and above 33 °C, and at driving altitude above 1000m. It was claimed that that window was an unlawful system since none of the derogations from the prohibition of a defeat device, as provided for in Article 5 of the Emissions Regulation could justify it. The defendant dealer contended that a temperature window was used by all diesel vehicle manufacturers in the Euro 5 category, that the KBA had always regarded that window as a lawful measure for the purposes of the Emissions Regulation, and that, in addition, when the software at issue was examined, the KBA found, after conducting an in-depth review, that the update had had no negative impact on the durability of the air pollution control devices. In other words, the vehicle had a valid Type Approval following the issue of a mandatory or voluntary update decision, with the KBA having been satisfied that no prohibited defeat device remained. Nevertheless, according to the referring court, the temperature window was a prohibited defeat device within the meaning of that regulation. Whilst the questions referred related to the circumstances in which a defeat device may be regarded as a prohibited one, there was no suggestion that the fact of the Type Approval or the sanctioning of the software update by the KBA precluded the very investigation that the Austrian Court (and, indeed, then the CJEU in terms of the proper construction of the regulations) embarked upon.

In IR, a consumer purchased a VW Touran Comfortline BMT with a Euro 5 generation EA189 type diesel engine, with an EGR valve. The vehicle originally contained software installed in the electronic engine controller which had a ‘mode 0’ and a ‘mode 1’ (‘the switch system’). Mode 1 was used for the approval test for pollutant emissions, called the ‘New European Driving Cycle’ (NEDC), which is conducted in a laboratory. If the characteristic conditions of that approval test did not exist, mode 0 was applied and, in that case, the exhaust gas recirculation rate decreased and the injection timing and duration changed. In real-world operation, the vehicle was almost exclusively in mode 0, meaning that it did not comply with the NOx limit values laid down in the Emissions Regulation. The reference was made on the basis that the switch system was therefore a prohibited defeat device within the meaning of Article 5(1) and (2) of the Emissions Regulation. By letter of 8^th October 2015, the general importer of VW vehicles in Austria informed IR that changes had to be made to the vehicle and that the manufacturer would bear all costs associated with the repair work required in that connection.IR was subsequently asked to install the software at issue in his vehicle, which he did. The purpose of that update was to establish a temperature window. The KBA approved the software at issue and did not therefore withdraw the EC type approval. In that regard, it found, amongst other things, that there was no prohibited defeat device within the meaning of the Emissions Regulation. The KBA did not have any information about that software at its disposal because it did not request that such information be sent to it.IR brought an action before the Landesgericht Eisenstadt (Regional Court, Eisenstadt, Austria), the referring court, seeking the cancellation of the sale of the vehicle pursuant to the ABGB.VW contended that the installation of the software at issue in that vehicle gave satisfaction to IR such that he had no interest in bringing proceedings. As set out at paragraph 62 of the Opinion, the referring court considered that for judgment to be given, it was necessary to determine whether the software at issue was a technical solution that complied with the requirements of the Emissions Regulation and the Implementing Regulation. This, of course, is the very question that the KBA explicitly or implicitly considered when issuing the Mandatory Update Decision. Assuming the Defendant Manufacturers were correct as to the tenor of such a decision (which assumption this analysis necessarily proceeds upon), the KBA would have determined this question in the positive: the vehicle, following the mandatory update, complied with the regulations. Yet the referral proceeded on the uncontroversial basis that this fact alone did not preclude the national court from considering the question for itself in the context of a civil law claim. Indeed, in the opinion of the Advocate General (in the context of answering the fourth question in the referral), Mr Rantos stated at [132]:

“Here, it is apparent from the order for reference in Case C-134/20 that the purpose of the installation of the defeat device in the form of the software at issue was to remedy the prohibited switch system and to comply with the provisions of the [Emissions Regulation] by means of a repair. This question is based on the assumption that that software enabled the vehicle manufacturer concerned to achieve that objective. It is for the referring court to determine, having regard to the answers given to the questions previously examined, whether that is the case. 72 If not, the defeat device would, in any event, be prohibited on the basis of Article 5(1) and (2) of that regulation.”

If the effect of the Mandatory Update Decision was, as a matter of EU law, to impose upon the referring court the KBA’s view about whether the software had in fact remedied the existence of a PDD, it is perhaps surprising that this was not identified by the Advocate General or the CJEU as a direct preliminary answer to the question raised by the referring court (i.e. the answer should have been: your question is not relevant in circumstances where you have no jurisdiction to question the decision of the KBA).

In DS, a consumer purchased a VW with a Euro 5 generation type EA 189 which had an EGR valve, and software which operated the exhaust gas recirculation system on the basis of a switch system. The KBA ordered VW to ensure the compliance of Euro 5 generation EA189 type engines with the national legislation and EU law in force (by way of what has been described in this litigation as a Recall Decision). Some months later, the KBA informed VW that it confirmed that the software at issue was suitable for re-establishing the conformity of the vehicles concerned (effectively a Mandatory Update Decision). As a result, EC Type Approval for the vehicle was not withdrawn or revoked by the KBA. DS duly installed the software at issue. The update replaced the switch system with the temperature window.DS brought an action before the Landesgericht Linz (Regional Court, Linz, Austria) seeking the reimbursement of the purchase price of the vehicle against return of that vehicle or, in the alternative, a reduction in the price of the vehicle or, in the further alternative, a declaration that Porsche Inter Auto (the dealer) and Volkswagen were liable for damages as a result of the presence of a PDD within the meaning of Article 5(2) of the Emissions Regulation. By a judgment of 12 December 2018, the Landesgericht Linz (Regional Court, Linz) dismissed DS’s action. Following an appeal lodged by DS, the Oberlandesgericht Linz (Higher Regional Court, Linz, Austria) upheld that judgment by a judgment of 4 April 2019. That court took the view, that, even assuming that the vehicle was initially defective, the software at issue remedied that defect. In addition, it found that the system by which exhaust gas recirculation was reduced when the outside temperature was below 15 °C or above 33 °C was permissible under Article 5(2) of the Emissions Regulation because it was necessary to protect the engine against damage. Neither court took the view that investigating such a question in the context of the civil law claims before it was itself precluded by the existence of the Type Approval or the mandatory update decision.

DS lodged an appeal on a point of law before the Oberster Gerichtshof (Supreme Court, Austria), the referring court before the CJEU. Before that court, DS claimed that the vehicle was defective because the switch system was a prohibited defeat device and that the software at issue did not remedy that defect. The defendants contended that the temperature window was a defeat device within the meaning of the Emissions Regulation, but that the device was lawful, and that that view was shared by the KBA. The referring court stated that it was called upon to rule on whether the vehicle was defective at the time of delivery, whether that defect has been remedied and whether DS suffered damage caused by the vehicle manufacturer concerned. Since the software at issue was approved by the KBA, the referring court asked, first of all, whether that approval was sufficient, on its own, to achieve the improvement of the item purchased, within the meaning of Paragraph 932(1) of the ABGB. This led to the following question being referred.

Directive 1999/44 is a consumer protection directive. Article 2(2)(d) is part of a rebuttable presumption of conformity. At paragraph [141], the Opinion stated:

“As a preliminary point, I would observe that that question is based on the premise that the vehicle concerned is equipped with a prohibited defeat device within the meaning of Article 3(10) of Regulation No 715/2007, read in conjunction with Article 5(2) of that regulation. As I have stated, it is for the referring court to determine whether that is the case.”

In the latter sentence, Mr Rantos was referring (…”As I have stated”…) to paragraph [126] in the same Opinion, in which he said, in a part of his Opinion dealing with questions arising in each of the three cases (I have included paragraph [125] for context):

“125.

In accordance, once more, with the judgment in X, only immediate risks of damage that trigger a specific danger whilst the vehicle is being driven are capable of justifying the use of a defeat device such as a temperature window. In my view, that situation could arise if the malfunctioning of the EGR has a sudden effect on the functioning of the engine itself, and the regular and appropriate maintenance of the vehicle could not prevent such an effect.

126.

Only in such a situation could the defeat device at issue be authorised on the basis of Article 5(2)(a) of Regulation No 715/2007. Since this involves conducting an analysis of a factual nature, it falls to the referring courts, who alone have jurisdiction to find and assess the facts in the cases before them, to determine whether the potential malfunctioning of the EGR could give rise to sudden, immediate risks of damage to the engine itself, thus triggering a specific danger whilst the vehicle is being driven, even if that vehicle undergoes regular and appropriate maintenance.”

The opinion that it falls to the referring courts to determine on the facts of the cases before them whether there existed a prohibited defeat device is in direct contradiction to the suggestion before me that the matter, in the context of the civil claims which were before the referring courts, had been determined conclusively by the KBA as part of the totally harmonised scheme.

At paragraph [146], Mr Rantos expressed the view that since an average consumer, who is reasonably well informed and reasonably observant and circumspect, can expect the regulatory requirements for EC Type Approval to be satisfied, even in the absence of specific contractual terms, the vehicle concerned is not in conformity with the contract of sale within the meaning of Directive 1999/44 if not all those requirements are met. This opinion formed the basis of the judgment of the CJEU at [55]. At paragraph [56], the CJEU went on:

“As the Advocate General observed in point 149 of his Opinion, that interpretation is not called into question by the fact that the vehicle concerned is EC type-approved, allowing that vehicle to be driven on the road.”

The CJEU then reasoned that it was inherent in the harmonised system by which Type Approval can be withdrawn or amended if non-conformities exist that a vehicle may in fact be out of conformity with the relevant regulations yet be type approved. It flows from this reasoning, in my judgment, that the fact of Type Approval of itself cannot, as a matter of EU law, answer the question of whether the vehicle in fact contains a defeat device and/or is therefore of unsatisfactory quality in the context of a civil law claim. That does not impinge upon the system by which it is determined that (by reference to the Type Approval), that same car may be lawfully registered and sold in accordance with Type Approval unless and until that Type Approval is revoked or amended in accordance with the regime.

A fourth CJEU case, this time involving Mercedes, is in essence consistent with this approach. The case involved QB who bought a used Mercedes model C220 CDI equipped with a Euro 5 generation diesel engine. The vehicle included what has been called a temperature window. QB brought an action before the Landgericht Ravensburg (Regional Court) seeking compensation for the damage allegedly caused to him by the Mercedes-Benz Group by equipping the vehicle in question with a defeat device prohibited by the Emissions Regulation.

The referring court considered that the temperature window was a prohibited defeat device. It also considered that, on the face of it, the exception laid down in Article 5(2)(a) of the Emissions Regulation did not apply. It therefore considered that QB may have had a right to compensation under paragraph 823(2) of the BGB, which required proof of negligence. However, that provision presupposed the infringement of a law intended to protect others, which, under German Law, meant that that law was intended to protect an individual or a group of persons against a failure to have regard to a specific legal interest. The referring court was uncertain whether Article 18(1), Article 26(1) and Article 46 of the Framework Directive and Article 5(2) of the Emissions Regulation were intended to protect, in addition to public interests, the interests of an individual who was the purchaser of a vehicle which does not comply with EU law where that vehicle is fitted with a defeat device prohibited under the latter provision.

At paragraph [59], following GSMB, the CJEU reiterated that, “It is therefore for the referring court to decide, where appropriate, whether, in the light of the clarifications provided in the case-law cited in paragraph 58 above, the software referred to in paragraph 24 above constitutes a ‘defeat device’, within the meaning of Article 3(10) of [the Emissions Regulation]”. After setting out various principles relating to the proper construction of the regulations, the CJEU then repeated at [67]: “It is for the referring court to carry out the factual assessments necessary for the purposes of applying the conditions referred to in paragraphs 60 to 66 above.”. There is no sign from the judgment, in a case involving the Mercedes at the highest level in Europe, that this was not an entirely uncontroversial statement of the law. It is fundamentally inconsistent with the suggestion, now, that as a matter of EU law the question of the existence or absence of a defeat device is determined for all purposes by the existence of Type Approval, or any Mandatory or Voluntary Update Decision by the KBA.

At paragraphs 83 to 85, the CJEU then held:

“83.

It cannot be ruled out that a vehicle type covered by an EC type-approval allowing that vehicle to be driven on the road may, initially, be approved by the approval authority without the presence of the software referred to in paragraph 24 above having been disclosed to it. In that respect, the Framework Directive envisages the situation in which the unlawfulness of an element of design of a vehicle, for example in the light of the requirements of Article 5 of Regulation No 715/2007, is discovered only after that approval has been granted. Thus, Article 8(6) of that framework directive provides that that authority may withdraw the approval of a vehicle. Furthermore, it follows from the first and third sentences of Article 13(1) of that framework directive that, where a manufacturer informs a Member State which has granted EC type-approval of a change in the information package, that Member State may, where necessary, decide, in consultation with the manufacturer, that a new EC type-approval is to be granted (see, to that effect, judgment of 14 July 2022, Porsche Inter Auto and Volkswagen, C-145/20, EU:C:2022:572, paragraph 56). Lastly, Article 30(1) of the Framework Directive provided that, if a Member State which had granted an EC type-approval found a lack of conformity to the vehicle type it had approved, it was to take the necessary measures, including, where necessary, the withdrawal of that type-approval, to ensure that the produced vehicles were brought into conformity with that type.

84.

Consequently, the unlawfulness of a defeat device equipped in a motor vehicle, discovered after the grant of EC type-approval for that vehicle, is capable of calling into question the validity of that type-approval and, by extension, the validity of the certificate of conformity intended to certify that that vehicle, belonging to the series of the type approved, complied with all regulatory acts at the time of its production. In the light of the rule laid down in Article 26(1) of the Framework Directive, that unlawfulness is thus liable, inter alia, to create uncertainty as to the possibility of registering, selling or entering into service that vehicle and, ultimately, to harm the purchaser of a vehicle equipped with an unlawful defeat device.

85.

In the light of the foregoing considerations, the answer to the first and second questions is that Article 18(1), Article 26(1) and Article 46 of the Framework Directive, read in conjunction with Article 5(2) of Regulation No 715/2007, must be interpreted as protecting, in addition to public interests, the specific interests of the individual purchaser of a motor vehicle vis-à-vis the manufacturer of that vehicle where that vehicle is equipped with a prohibited defeat device, within the meaning of the latter provision.”

In these paragraphs, and in the context of the specific questions asked in the context of the German civil law right, the CJEU clearly considered that the situation may arise where a vehicle which is being legally driven on the road pursuant to EU Type Approval may in fact contain an unlawful defeat device. In QB, that question of ‘unlawfulness’ had been determined by the national court in the context of a damages claim, notwithstanding the existence of Type Approval or update decisions. The CJEU then identified the provisions of the Framework Directive by which that unlawfulness may be dealt with, for the purposes of substantiating its conclusion that the uncertainty caused by the operation of the Framework Directive in the context of registering or selling the car can cause harm to the purchaser of a vehicle equipped with an unlawful defeat device. Of course, the CJEU did not deal head on with the argument now before me: that the national court was precluded from making any finding of unlawfulness in the first place whilst Type Approval or other KBA decisions remain unchallenged. It is right to acknowledge that there has (recently) been a request for a preliminary ruling on this precise point from the Landgericht Duisburg in Germany lodged on 7^th December 2023 in a case involving Mercedes (C-751/23). But, unless and until the CJEU rules otherwise, it seems to me that the conclusion of the CJEU in the foregoing cases is wholly inconsistent with the contention that a national court is so precluded. Put another way, if I were to succumb to the Defendant Manufacturers’ arguments, I consider that I would be cutting across a foundation fundamental to these CJEU decisions, namely the national court’s ability (and, indeed, obligation) to conclude on the facts of each case before it the potential existence of a prohibited defeat device notwithstanding extant KBA Type Approval and/or Mandatory Update Decisions pursuant to which the vehicles were being driven legally in accordance with the harmonised regime. Given that it may therefore fairly (if unattractively, given the identity of the defendants in the CJEU cases) be said that the point before me was not directly argued in the CJEU or, therefore, expressly decided, I accept that these cases are not determinative of the issue. I also accept, as urged upon me by Ms Davies KC, that they have to be seen in the context of the specific questions being asked. However, even bearing this in mind, in my judgment they constitute, at the very least, a powerful backdrop against which to consider the merits of the Defendant Manufacturers’ arguments on this preliminary issue. They certainly cause me considerable caution before accepting the contention that a determination by national courts of the existence of PDDs notwithstanding decisions by the KBA in the context of civil law claims strikes fundamentally at the heart of total harmonisation or would cause complete havoc. Were that so, it is surprising that such a dramatic conclusion did not occur to the Advocate General, or to any of the enormously experienced EU judges sitting in the Grand Chamber of the CJEU on 4 occasions in high profile cases (nor, it might be said at least in the context of these decisions, to VW or Mercedes, until the point was taken recently in C-751/23).

This conclusion is not inconsistent, in my view, with the havoc of which Waksman J spoke. I regard this as a reference to the havoc that would be caused if the validity of the administrative acts by the KBA was open to challenge by the national courts in the context, as Article 2 of the Framework Directive makes clear, of facilitating the registration, sale and entry into service within the Community of vehicle and components, etc; or if the procedural exclusivity envisaged by the KBA by which addressees can challenge decisions was also disrupted by the national courts. Neither of these factors, however, are relevant to the proper role a national court has in determining for itself the facts and the law in respect of a claim for civil damages brought against a manufacturer on the basis of the alleged existence of a PDD. That such cases can be brought and determined on the facts by national courts irrespective of Type Approval and Mandatory Update Decisions is demonstrated by the very fact of the cases underlying the CJEU jurisprudence referred to above. Those cases are consistent with the conclusion that, in determining such a case, a national Court is not required to question the validity of the grant of approval, or a decision to require measures, by a relevant approval authority in another Member State referred to by Waksman J at paragraph [379], which is not permitted in a totally harmonised scheme. To the extent that Waksman J’s comment should be construed more broadly, I respectfully disagree.

Third, there is in my view no contradiction between this conclusion and the approaches of the CJEU in Astellas, followed by Lewis J in Orion Corp v Secretary of State for Health and Social Care [2019] EWHC 689 (Admin), or the CJEU in Synthon BV v MHRA (C-452-06); or that of Mr Justice Supperstone in R (ex parte Generics (UK) Limited t/a Mylan v SSH [2018] EWHC 228 (Admin); or of Mr Justice Keene in R v The Licencing Authority & Ors (ex parte Monsanto plc [1997] CMLR 402; or of Mr Justice Jay in R (Teva BV) v SSH acting as the Licensing Agency & Biogen IDEC Ltd [2018] EWHC 1051 (Admin), the latter three cases being relied upon in particular by Ms Howard KC representing Nissan. That is because these decisions were all made in the context of challenges to the legality and/or effect of an administrative act by which permission was given to place a product on the market within a harmonised (albeit decentralised) scheme. The cases make clear the inability of the courts to second-guess that regulatory act in the context of challenges by parties whose commercial interests were impacted by the authorisation or refusal of authorisation.

So, in Orion, the claimant was seeking judicial review of the Secretary of State’s decision to grant marketing authorisation for the interested party’s generic product. It was held, amongst other things, that while the holder of a marketing authorisation could bring legal proceedings to protect the exclusivity of its data and, in particular, to challenge an error in the calculation of the period of protection conferred by article 10 of Parliament and Council Directive 2001/83/EC, that did not extend to challenging in the courts of one member state the compatibility with EU law of marketing authorisations granted by the competent authorities in another member state; and that any such challenge had to be brought in the courts of that other member state in accordance with the relevant rules of national procedure, including any applicable time limits for bringing such a challenge.

In Mylan, the challenge was, by way of judicial review, to a decision of the Medicines and Healthcare products Regulatory Agency refusing to validate and/or revisit the claimant’s application for a generic marketing authorisation. The claimant in that case described its own judicial review proceedings as “an avowed and necessary collateral attack on the reasoning contained in and the validity of” the Commission Decision. Unsurprisingly, having concluded that the decision was a regulatory act which was of direct concern to the claimant, Mr Justice Supperstone concluded – just as Mr Justice Jay had in the similar case of Teva, which he cited – that the claimant had alternative remedies sufficient to defeat any claim for judicial review. The court also found that the substance of the complaint was an impermissible challenge to the factual and scientific conclusions and regulatory judgment of the relevant expert competent authorities. Mr Justice Keene’s approach in the judicial review claim in Monsanto, again related to pharmaceutical licencing, was no different in principle.

The observations and outcomes in these cases quite properly reflect the orthodox position that a national court cannot second guess the validity of a regulatory authorisation in the context of a harmonised framework. They do not purport to touch upon the right of someone to sue, under consumer protection laws, a pharmaceutical manufacturer for damages said to have been caused by a harmful drug (harmful in fact notwithstanding its authorisation for sale pursuant to the relevant directive), which would be the analogous position to the case before me. As Mr de la Mare KC correctly submitted, there is no authority which suggests that, as a precursor to a claim against a pharmaceutical manufacturer under consumer protection laws, a consumer must bring a claim in the relevant member state in order to set aside the decision of the licencing authority on grounds that the product is in fact unsafe and should not have received authorisation for sale in the first place.

Fourth, accepting the Defendant Manufacturers’ submissions would lead to a significant curtailment of private law rights relating to the quality of products. Mr de la Mare KC argued that, if the Defendant Manufacturers’ argument was correct in respect of the existence of PDDs, it would apply equally to product liability claims in all regulated areas. By way of example, submissions focussed on the Safety Glazing Directive. Ms Davies KC, by contrast, sought to single out the Emissions Regulation by the manner in which, at Article 5(2), it contained an express prohibition on the use of defeat devices that reduce the effectiveness of emission control systems, save in particular circumstances. The existence of this prohibition meant, Ms Davies KC argued, that the question the Court needed to ask to determine liability in the present case is identical to the question that the KBA asked itself, and that (by reason of the harmonised scheme) one national authority (i.e. this Court) cannot second guess the very same question determined by another (the KBA). The existence of the express prohibition, it was said, could be contrasted to virtually all the other directives and, as such, the practical impact of the decision to be bound by the KBA decision in respect of PDDs was extremely limited. Accepting her argument in respect of the KBA’s decision as to PDDs would not, Ms Davies KC contended, impact on general product liability cases. This was because in most other areas of regulation (including Safety Glazing), the question of whether a product was safe in all the circumstances was not precisely the same question asked by the regulator in approving a product’s compliance with a particular requirement. Thus, in such cases, the binding effect of the regulatory approval would not impact on a national court’s ability to determine the question of product liability.

I cannot accept that the very specific express prohibition within the Emissions Regulation considerably narrows the impact of the Defendant Manufacturers’ contended for analysis on consumer rights. Taking the Safety Glazing Directive as an example, this includes compliance with the UNECE Regulation 43, and this in turn requires all glazing material to be “such that, in the event of shattering, the danger of bodily injury is reduced as far as possible”. If the tenor of a regulatory act was to include a binding declaration that particular glass had been designed to comply with this requirement, and the Defendant Manufacturers’ argument as to the impact of such a decision on private law rights was correct, a claimant would not be able bring a private law action to the extent that it depended upon demonstrating that, in the event of shattering, the glass did not in fact reduce the danger of bodily injury as far as possible. The application of the Defendant Manufacturers’ argument would not be as limited as Ms Davies KC said. It is inevitable that considerable incursions into private law rights would exist throughout regulatory frameworks dealing with all sorts of products if the argument was right.

The Defendant Manufacturers’ contention would also run counter to the orthodox proposition under English law that compliance with an industry standard does not, of itself, amount to a defence to a claim for losses caused by some attribute of the product in question: see, for example, as summarised in Clerk & Lindsell on Torts 24^th Edn at 10-60. At 10-67, the editors state “It would be odd, to say the least, if the effect of safety regulations was actually to reduce the exposure of the manufacturer to civil liability.” The same irony would exist in the present case if, as Ms Davies KC contended, the regulators’ express prohibition of defeat devices has had the effect of removing from consumers any ability to sue in respect of their existence (i.e. the exposure to civil liability will have been reduced, not enhanced, by the strength of the regulation).

Fifth, the requirement of effective judicial protection, particularly in the context of consumer rights, is inconsistent with the conclusion that unless the regulatory act is challenged in an administrative court, a consumer claim against the manufacturer in respect of a prohibited defeat device is precluded. The starting point for a consumer is that they have rights conferred upon them by EU law. It must follow that there must be effective judicial protection in respect of those rights.

Ms Davies KC fairly accepts that effective judicial protection is a fundamental principle which must form part of the overall regulatory framework. As summarised at paragraphs 45-46 of the decision of the CJEU in Impact (C-268/06):

“In that regard, it is important to note that the principle of effective judicial protection is a general principle of Community law (see, to that effect, Case C‑432/05 Unibet [2007] ECR I‑2271, paragraph 37 and the case‑law cited). …

On that basis, as is apparent from well‑established case‑law, the detailed procedural rules governing actions for safeguarding an individual’s rights under Community law must be no less favourable than those governing similar domestic actions (principle of equivalence) and must not render practically impossible or excessively difficult the exercise of rights conferred by Community law (principle of effectiveness) (see, in particular, Rewe-Zentralfinanz and Rewe-Zentral, paragraph 5; Comet, para‑ graphs 13 to 16; Peterbroeck, paragraph 12; Unibet, paragraph 43; and van der Weerd and Others, paragraph 28).”

It is also clear from Impact that procedures that render the protection of rights derived from a directive excessively difficult in practice would be contrary to the principle of effectiveness (see [51]).

However, together with Ms Howard who provided specific submissions in this regard, Ms Davies KC contends that the ability for a consumer to bring an action in the German courts challenging the validity of a KBA decision provides sufficient effective judicial protection within the harmonised framework so as not to contradict the principle of effectiveness.

I cannot agree. The judicial protection the Defendant Manufacturers identify involves the consumer fighting the validity of the Type Approval or update decision in the relevant, foreign Administrative Courts through to the conclusion of the appeals process. In respect of a private law consumer claim against Mercedes (or, indeed, in respect of a defence and counterclaim if sued under a contract of sale by Mercedes or a dealer), the civil proceedings would need to be stayed. Whether a stay would be granted would of course be a matter of discretion. The manufacturer would no doubt argue that the claim should not be stayed but dealt with on its merits subject to the extant regulatory act binding on the national court (as the Defendant Manufacturers seek to do before this Court). Even if stayed - rather than determined on the basis of the binding KBA decision - resolution of the German administrative proceedings would take many years and come at significant cost. The administrative claim would, moreover, relate to a decision to which the claimant was neither a party nor an addressee. In practical terms, they would almost certainly not be capable of appealing within the one-month period permitted (to addressees) for appeals from the decisions from the KBA and, as such, the decision will have become prima facie incontestable. As Professor Hofmann pointed out under German Law, the only option would be to file for restitutio in integrum under Section 60(1) VwGO, which permits the Court to grant reinstatement if a person did not manage to observe a legal deadline because of an obstacle without fault of his or her own. Professor Hofmann’s view was that once an administrative act has become final and binding for others to rely upon (both manufacturers and other consumers), German procedural law considerably limits the options to third parties who were not addressees to challenge and have the act revoked. This conflict between legal certainty and the substantial accuracy of an administrative action made it difficult to predict what the outcome of the administrative process and appeals would be. The same cannot, of course, be said for the manufacturer who is the addressee of the decision, made submissions leading to the regulatory act, and can obviously appeal it to the Administrative Courts within the required timeline to the extent they consider that appropriate. There is clear and effective judicial protection for the manufacturer built into the framework. However, if the Defendant Manufacturers are right, there is no such effective judicial protection for the consumer: rather, the regulator’s binding decision would in all practical senses determine the matter against the consumer by reason of the cost and duration of administrative proceedings required as a pre-requisite to bringing any claims under consumer protection rules. The complete asymmetry between consumer and manufacturer in the context of effective judicial protection explains readily why Recall Decisions are binding on the addressee manufacturer (who chooses not to appeal), but Mandatory Update Decisions or Voluntary Update Decisions – insofar as they purport to determine the absence of PDDs and in the context of a civil claim for damages which does not itself seek to impeach the effectiveness of the administrative act – are not binding on the consumer.

Finally I return, then, to Waksman J’s obiter remarks at [394]-[395] of Crossley 1 about the binding effect of a regulatory decision that a PDD did not exist on the Claimants. I note first that the context at paragraph 394 is stated to be “an appeal from the KBA letters” (i.e. a Recall Decision), in which there is an ‘adverse’ decision for the Claimants. Presumably, in such an appeal, that would be a determination in favour of the manufacturers that there was no prohibited defeat device and, as such, that no measures were required to be carried out under threat of withdrawal of Type Approval. The German Law experts were both agreed, and I find, that the effect of a successful appeal against the Recall Decision under German administrative law would be that the Recall Decision would be regarded as a nullity. Waksman J’s remarks in paragraphs 394 and 395 (made without the benefit of the agreed position in German Law from the experts on this point) are not therefore strictly correct: as Ms Davies KC accepted, there would in fact be no remaining regulatory act to bind either a claimant consumer or a defendant manufacturer in respect of the existence or non-existence of a defeat device following a successful appeal against the Recall Decision.

Insofar as Waksman J dealt with an ‘adverse’ decision to the claimants ‘from the KBA’ (which he refers to at paragraph 395), I consider that it is clearly correct that the claimants would be ‘bound’ by such a decision in the sense that they could not then challenge the legal validity of Type Approval. So, a consumer interest organisation as posited by Waksman J in paragraph 395 would not be able to bring a claim in this Court for a declaration that the UK’s Vehicle Certification Agency’s decision to permit cars the subject of such decisions to be registered and sold in the UK was unlawful. However, I respectfully disagree to the extent that Waksman J was suggesting that a determination by the KBA as to the absence of a PDD in the context of a Type Approval-validity related question involving the manufacturers would, without more, bind this Court in a private law action for the five interrelated reasons I have given above.

I note in conclusion, as will be apparent from the foregoing, that I did not consider that either side’s submissions relating to principles that could – or could not – be drawn from competition law particularly advanced their respective arguments.

Therefore, I find that if (contrary to my determination above) Type Approval Decisions, Mandatory Updates Decisions and Voluntary Update Decisions are binding in German Law as to the presence or absence of defeat devices, I would take the view that:

In circumstances where I consider that the KBA decisions would not bind this Court in respect of those made pursuant to the Framework Directive, I conclude that the Defendant Manufacturers are in no better position under the Framework Regulation. However tempting, it is not therefore necessary for me to consider the extent to which the Framework Regulation fundamentally changed the landscape, whether before or after IP Completion Day.

I therefore conclude that the answer to the first preliminary issue as regards Type Approvals, Mandatory Update Decisions and Voluntary Update Decisions is ‘No’.

The answer in respect of Recall Decisions is (as agreed by the parties, and in any event) is ‘Yes’.

The answer to the second preliminary issue is (as agreed by the Claimants and the Mercedes Defendants, and in any event) ‘No’.