Until 31 December 2020, the Principal VAT Directive 2006/112 (PVD) was the ultimate source of UK VAT law, via European Communities Act 1972 (ECA 1972) s 2. From 1 January 2021, there is a different legal gateway for EU rules to come into UK law. The new gateway in the European Union (Withdrawal) Act 2018 (EUWA 2018) is narrower than the old gateway in ECA 1972. As a result, some – but only some – effects of the PVD remain supreme over UK VAT legislation, and only over some UK VAT legislation.
The starting point is now EUWA 2018 s 4(1). This provides that any right, obligation, etc. (which we call here a ‘point’, ‘effect’ or ‘proposition’) in EU law as it stood on 31 December 2020 continued in UK law on 1 January 2021. Generally speaking, all EU law propositions not preserved by another provision of EUWA 2018 passed through that gateway. The wording of s 4(1) is detailed but does not appear to restrict this general approach: as acknowledged by Dominic Raab MP in Hansard debates, it merely tracks the wording of ECA 1972 s 2. Under EUWA 2018 s 5, those EU law propositions override UK legislation in force on 31 December 2020, and also any subsequent amendments to UK legislation where that would be consistent with the intention of the amendment.
But s 4(2) narrows the gateway where the proposition arises from a Directive, such as the PVD. The only propositions which pass through are ones ‘of a kind’ recognised by the CJEU by the end of 2020, or (by virtue of EUWA 2018 Sch 8 para 38) by a UK court or tribunal in litigation started before then. The ‘of a kind’ gateway is therefore the key to identifying which EU law propositions from the PVD survive. Until clarified in case law, the width of the gateway is uncertain.
The explanatory notes to EUWA 2018 propose (at para 98) a narrow meaning for ‘of a kind’ – effectively that a specific point recognised in litigation involving one taxpayer can also be applied to another taxpayer. Statements by government ministers in the Hansard debates also take this approach. But there is a problem with this. It would make the words ‘of a kind’ redundant. As Lord Pannick QC said in Hansard debates, they must mean something.
Neither the explanatory notes nor government Hansard statements appear central to determining that meaning. The relevant Oxford English Dictionary definition of a ‘kind’ appears to be ‘a class or category of things distinguished by common characteristics and attributes possessed by its members; ... a particular variety or type; a sort’. On this basis, an EU law proposition passes through the gateway if it is properly seen as part of a relevant ‘class’ with one or more propositions recognised in relevant case law. Lord Pannick QC in Hansard debates also said that he saw the question as whether the relevant ‘principles’ had been ‘stated by the Court of Justice, or a domestic court’. The Law Society’s 2021 book Retained EU Law: a practical guide, co-authored by a key former government lawyer, sees similar possibilities.
This appears to mean that effects of the PVD recognised after Brexit by the CJEU (or UK courts after the relevant cut-off) pass through the gateway if they are ‘of a kind’, but even quite straightforward effects will not if they are not, i.e. they are sufficiently different from ones litigated before Brexit.
Another part of the Brexit legislation makes clear that broad case law propositions which go well beyond the specific words of the PVD can pass through the gateway. The Taxation (Cross-border Trade) Act 2018 (TCTA 2018) s 42 explains this both for case law tied to a specific PVD provision, citing Kittel (Case C-439/04) and for case law which is not, citing Halifax (Case C-255/02). Other case law– such as BLP (Case C-4/94) on input tax deduction, and CPP (Case C349/96) and Levob (Case C-41/04) on complex supplies – should equally pass through. But, to take an example, where there is no CPP/Levob caselaw directly on point for a given situation, and in the past a taxpayer has filed based on propositions sensibly inferred from the caselaw that does exist, EUWA 2018 only supports that ongoing filing position if those propositions are ‘of a kind’ in terms of s 4(2).
We see five ways in which a specific legal point could be ‘of a kind’ with a recognised point or points, or not. These are broad categories, intended as a tool to help analysis. The approach they represent will be refined as more practical examples are considered in practice and by the courts.
1. This specific point has been recognised. This is the narrowest possible reading. It is consistent with an approach HMRC has in the past taken following CJEU rulings against them. It is arguably the government’s ‘intention’ as set out in the EUWA 2018 explanatory notes. As above, we consider it to be too narrow. It is also not easily reconciled with the drafting of TCTA 2018 s 42(5) which preserves certain propositions of the VAT implementing regulation.
2. One or more broad points have been recognised but, as at the cut-off, whether the PVD covers the specific point required analysis. In our view, propositions within the scope of that analysis clearly passed through. For example, the analysis of the scope of ‘management’ of a special investment fund continues (see, for example, the current reference to the CJEU in K v Austria (Case C-58/20)). Whatever propositions that analysis might support are very likely to be ‘of a kind’ with pre-2021 decisions.
3. One or more broad points have been recognised, but to cover the specific point they require extension. This is discussed below in relation to Danske Bank (Case C-812/19). In our view, whether the specific point passed through will be highly context-sensitive.
4. A similar point has been recognised. This is where a different point or points have been recognised, but an analogy can be made with the specific point in question (say, an analogy between the exemptions for insurance intermediation and for the negotiation of credit). Again, this will be fact-sensitive. Retained EU Law even suggests that this approach might encompass provisions drafted in the same style, or with the same broad purpose.
5. There is doubt because the recognised point is so generic. There is a credible argument that all propositions which do not go ‘against the grain’ of the UK legislation and so can be implemented via the Marleasing (Case C-106/89) principle of interpretation, however novel, pass through the gateway. However, this relies on all Marleasing interpretation itself being ‘of a kind’. This analysis may not be consistent with the broader purposes of EUWA 2018.
There are also two possible dimensions to ‘of a kind’: is the issue technically ‘of a kind’, and is it ‘of a kind’ in effect? For instance, although merely a technical consequence of the doctrines of equivalence and effectiveness, if the availability of compound interest were to be considered now for the first time, it is realistic to think that an effect of that sort would not be ‘of a kind’ with previously recognised consequences.
Much of the VAT-specific discussion of EUWA 2018 has focused on ss 6(1) and 6(2) under which post-2020 CJEU decisions are not binding in the tax tribunals and UK courts, but they may be taken into account. If the discussion suggests a broad discretion for the UK courts, it may miss the point. CJEU decisions on PVD provisions in force by the end of 2020 simply recognise the effect of those provisions. Those effects, whenever recognised, appear to have passed through the gateway if they are ‘of a kind’ already recognised. The question then becomes simply whether the UK courts consider that the CJEU has got EU law wrong.
Section 5 has also been discussed. It provides for EU law to remain supreme. However, this only describes the status of EU law which EUWA 2018 retains. EU VAT law is only retained in so far as ss 2 to 4 – and 6(3) – permit.
Section 6(3) provides that any question as to the validity, meaning or effect of any retained EU law is to be decided, so far as that law is unmodified after the end of 2020 and so far as they are relevant to it, in accordance with any retained case law. However, unless the relevant effect of the PVD passed through the s 4(2) gateway it is hard to see how EU case law interpreting the PVD can be ‘relevant’.
Section 2(1) needs more serious consideration. It provides that ‘EU-derived domestic legislation, as it has effect in domestic law immediately before [the end of 2020], continues to have effect in domestic law’ afterwards. On one view, the words ‘as it has effect’ preserves the full effect of any Directive which has been implemented, like the PVD, whatever that might be, as it was on 31 December 2020.
This argument is credible. Retained EU Law also supports it. However, like the Marleasing argument it should for now be treated with caution. That is because, on its face, this argument directly conflicts with the Directive-specific gateway in s 4(2). It would be necessary to argue that the restriction in s 4(2) focuses on situations where, unlike in VAT, there is no UK implementing legislation.
There has already been significant discussion of the statement in para 29 of Danske Bank that article 11 of the PVD requires VAT groups to be strictly territorial. This would be a novel consequence in the UK, going against HMRC’s response to the decision in Skandia (Case C‑7/13). That is a reason to think that this effect of the PVD is not ‘of a kind’ with ones previously recognised.
However, (1) the CJEU has previously recognised legal effects of article 11, and its predecessor in the Sixth VAT Directive; (2) those include limits on the approaches member states can apply to the constitution of a VAT group (for example, Larentia & Minerva (Cases C-108/14 and C-109/14)); (3) when assessing those limits, the CJEU has taken into account concepts from outside the words of the article, where those relate to matters of overriding importance in relation to taxation (again, see Larentia & Minerva); and (4) the ‘fiscal principle of territoriality’ has been previously recognised as such a matter, at least in relation to direct tax.
As such, this statement in Danske Bank is in category 3 above: it represents some extension of previously recognised propositions. Is it ‘of a kind’ with those propositions? If HMRC or a taxpayer wanted to argue that it was, it would be relevant that the Danske Bank decision itself has a somewhat similar effect to the decision in Skandia, and that the CJEU considered its decision in Danske Bank to be sufficiently straightforward in light of previous caselaw that it needed no advocate general’s opinion. If this statement is indeed ‘of a kind’, then, as above, the only remaining question for a UK court or tribunal would be whether it disagreed with the CJEU on the EU law meaning of the PVD.
HMRC’s Compliance Handbook says (at CH81120) in the context of avoiding penalties, that if a taxpayer is confronted with something ‘with which they are not familiar’, they should ‘take care to find out about the correct tax treatment or to seek appropriate advice’. The current consultation on disclosing uncertain tax positions envisages that (for calendar quarter groups, from Q1 2022) large taxpayers will disclose to HMRC any filing position worth £5m or more if it is ‘in some way novel such that it cannot reasonably be regarded as certain’. It is reasonable to say that the interaction of EUWA 2018 and the PVD is unfamiliar, novel and uncertain.
Until 31 December 2020, the Principal VAT Directive 2006/112 (PVD) was the ultimate source of UK VAT law, via European Communities Act 1972 (ECA 1972) s 2. From 1 January 2021, there is a different legal gateway for EU rules to come into UK law. The new gateway in the European Union (Withdrawal) Act 2018 (EUWA 2018) is narrower than the old gateway in ECA 1972. As a result, some – but only some – effects of the PVD remain supreme over UK VAT legislation, and only over some UK VAT legislation.
The starting point is now EUWA 2018 s 4(1). This provides that any right, obligation, etc. (which we call here a ‘point’, ‘effect’ or ‘proposition’) in EU law as it stood on 31 December 2020 continued in UK law on 1 January 2021. Generally speaking, all EU law propositions not preserved by another provision of EUWA 2018 passed through that gateway. The wording of s 4(1) is detailed but does not appear to restrict this general approach: as acknowledged by Dominic Raab MP in Hansard debates, it merely tracks the wording of ECA 1972 s 2. Under EUWA 2018 s 5, those EU law propositions override UK legislation in force on 31 December 2020, and also any subsequent amendments to UK legislation where that would be consistent with the intention of the amendment.
But s 4(2) narrows the gateway where the proposition arises from a Directive, such as the PVD. The only propositions which pass through are ones ‘of a kind’ recognised by the CJEU by the end of 2020, or (by virtue of EUWA 2018 Sch 8 para 38) by a UK court or tribunal in litigation started before then. The ‘of a kind’ gateway is therefore the key to identifying which EU law propositions from the PVD survive. Until clarified in case law, the width of the gateway is uncertain.
The explanatory notes to EUWA 2018 propose (at para 98) a narrow meaning for ‘of a kind’ – effectively that a specific point recognised in litigation involving one taxpayer can also be applied to another taxpayer. Statements by government ministers in the Hansard debates also take this approach. But there is a problem with this. It would make the words ‘of a kind’ redundant. As Lord Pannick QC said in Hansard debates, they must mean something.
Neither the explanatory notes nor government Hansard statements appear central to determining that meaning. The relevant Oxford English Dictionary definition of a ‘kind’ appears to be ‘a class or category of things distinguished by common characteristics and attributes possessed by its members; ... a particular variety or type; a sort’. On this basis, an EU law proposition passes through the gateway if it is properly seen as part of a relevant ‘class’ with one or more propositions recognised in relevant case law. Lord Pannick QC in Hansard debates also said that he saw the question as whether the relevant ‘principles’ had been ‘stated by the Court of Justice, or a domestic court’. The Law Society’s 2021 book Retained EU Law: a practical guide, co-authored by a key former government lawyer, sees similar possibilities.
This appears to mean that effects of the PVD recognised after Brexit by the CJEU (or UK courts after the relevant cut-off) pass through the gateway if they are ‘of a kind’, but even quite straightforward effects will not if they are not, i.e. they are sufficiently different from ones litigated before Brexit.
Another part of the Brexit legislation makes clear that broad case law propositions which go well beyond the specific words of the PVD can pass through the gateway. The Taxation (Cross-border Trade) Act 2018 (TCTA 2018) s 42 explains this both for case law tied to a specific PVD provision, citing Kittel (Case C-439/04) and for case law which is not, citing Halifax (Case C-255/02). Other case law– such as BLP (Case C-4/94) on input tax deduction, and CPP (Case C349/96) and Levob (Case C-41/04) on complex supplies – should equally pass through. But, to take an example, where there is no CPP/Levob caselaw directly on point for a given situation, and in the past a taxpayer has filed based on propositions sensibly inferred from the caselaw that does exist, EUWA 2018 only supports that ongoing filing position if those propositions are ‘of a kind’ in terms of s 4(2).
We see five ways in which a specific legal point could be ‘of a kind’ with a recognised point or points, or not. These are broad categories, intended as a tool to help analysis. The approach they represent will be refined as more practical examples are considered in practice and by the courts.
1. This specific point has been recognised. This is the narrowest possible reading. It is consistent with an approach HMRC has in the past taken following CJEU rulings against them. It is arguably the government’s ‘intention’ as set out in the EUWA 2018 explanatory notes. As above, we consider it to be too narrow. It is also not easily reconciled with the drafting of TCTA 2018 s 42(5) which preserves certain propositions of the VAT implementing regulation.
2. One or more broad points have been recognised but, as at the cut-off, whether the PVD covers the specific point required analysis. In our view, propositions within the scope of that analysis clearly passed through. For example, the analysis of the scope of ‘management’ of a special investment fund continues (see, for example, the current reference to the CJEU in K v Austria (Case C-58/20)). Whatever propositions that analysis might support are very likely to be ‘of a kind’ with pre-2021 decisions.
3. One or more broad points have been recognised, but to cover the specific point they require extension. This is discussed below in relation to Danske Bank (Case C-812/19). In our view, whether the specific point passed through will be highly context-sensitive.
4. A similar point has been recognised. This is where a different point or points have been recognised, but an analogy can be made with the specific point in question (say, an analogy between the exemptions for insurance intermediation and for the negotiation of credit). Again, this will be fact-sensitive. Retained EU Law even suggests that this approach might encompass provisions drafted in the same style, or with the same broad purpose.
5. There is doubt because the recognised point is so generic. There is a credible argument that all propositions which do not go ‘against the grain’ of the UK legislation and so can be implemented via the Marleasing (Case C-106/89) principle of interpretation, however novel, pass through the gateway. However, this relies on all Marleasing interpretation itself being ‘of a kind’. This analysis may not be consistent with the broader purposes of EUWA 2018.
There are also two possible dimensions to ‘of a kind’: is the issue technically ‘of a kind’, and is it ‘of a kind’ in effect? For instance, although merely a technical consequence of the doctrines of equivalence and effectiveness, if the availability of compound interest were to be considered now for the first time, it is realistic to think that an effect of that sort would not be ‘of a kind’ with previously recognised consequences.
Much of the VAT-specific discussion of EUWA 2018 has focused on ss 6(1) and 6(2) under which post-2020 CJEU decisions are not binding in the tax tribunals and UK courts, but they may be taken into account. If the discussion suggests a broad discretion for the UK courts, it may miss the point. CJEU decisions on PVD provisions in force by the end of 2020 simply recognise the effect of those provisions. Those effects, whenever recognised, appear to have passed through the gateway if they are ‘of a kind’ already recognised. The question then becomes simply whether the UK courts consider that the CJEU has got EU law wrong.
Section 5 has also been discussed. It provides for EU law to remain supreme. However, this only describes the status of EU law which EUWA 2018 retains. EU VAT law is only retained in so far as ss 2 to 4 – and 6(3) – permit.
Section 6(3) provides that any question as to the validity, meaning or effect of any retained EU law is to be decided, so far as that law is unmodified after the end of 2020 and so far as they are relevant to it, in accordance with any retained case law. However, unless the relevant effect of the PVD passed through the s 4(2) gateway it is hard to see how EU case law interpreting the PVD can be ‘relevant’.
Section 2(1) needs more serious consideration. It provides that ‘EU-derived domestic legislation, as it has effect in domestic law immediately before [the end of 2020], continues to have effect in domestic law’ afterwards. On one view, the words ‘as it has effect’ preserves the full effect of any Directive which has been implemented, like the PVD, whatever that might be, as it was on 31 December 2020.
This argument is credible. Retained EU Law also supports it. However, like the Marleasing argument it should for now be treated with caution. That is because, on its face, this argument directly conflicts with the Directive-specific gateway in s 4(2). It would be necessary to argue that the restriction in s 4(2) focuses on situations where, unlike in VAT, there is no UK implementing legislation.
There has already been significant discussion of the statement in para 29 of Danske Bank that article 11 of the PVD requires VAT groups to be strictly territorial. This would be a novel consequence in the UK, going against HMRC’s response to the decision in Skandia (Case C‑7/13). That is a reason to think that this effect of the PVD is not ‘of a kind’ with ones previously recognised.
However, (1) the CJEU has previously recognised legal effects of article 11, and its predecessor in the Sixth VAT Directive; (2) those include limits on the approaches member states can apply to the constitution of a VAT group (for example, Larentia & Minerva (Cases C-108/14 and C-109/14)); (3) when assessing those limits, the CJEU has taken into account concepts from outside the words of the article, where those relate to matters of overriding importance in relation to taxation (again, see Larentia & Minerva); and (4) the ‘fiscal principle of territoriality’ has been previously recognised as such a matter, at least in relation to direct tax.
As such, this statement in Danske Bank is in category 3 above: it represents some extension of previously recognised propositions. Is it ‘of a kind’ with those propositions? If HMRC or a taxpayer wanted to argue that it was, it would be relevant that the Danske Bank decision itself has a somewhat similar effect to the decision in Skandia, and that the CJEU considered its decision in Danske Bank to be sufficiently straightforward in light of previous caselaw that it needed no advocate general’s opinion. If this statement is indeed ‘of a kind’, then, as above, the only remaining question for a UK court or tribunal would be whether it disagreed with the CJEU on the EU law meaning of the PVD.
HMRC’s Compliance Handbook says (at CH81120) in the context of avoiding penalties, that if a taxpayer is confronted with something ‘with which they are not familiar’, they should ‘take care to find out about the correct tax treatment or to seek appropriate advice’. The current consultation on disclosing uncertain tax positions envisages that (for calendar quarter groups, from Q1 2022) large taxpayers will disclose to HMRC any filing position worth £5m or more if it is ‘in some way novel such that it cannot reasonably be regarded as certain’. It is reasonable to say that the interaction of EUWA 2018 and the PVD is unfamiliar, novel and uncertain.
1c3280ef-80d4-447d-8006-e2abc6265712.jpg?sfvrsn=333a2b58_1)
