WHA Ltd & Anor v Revenue and Customs [2007] EWCA Civ 728

[2007] EWCA Civ 728

Case No C3/2003/1412

IN THE SUPREME COURT OF JUDICATURE

COURT OF APPEAL (CIVIL DIVISION)

ON APPEAL FROM The Chancery Division of the High Court

Mr Justice Lloyd

CH2002AP0400

Venue Royal Courts of Justice, Strand, London, WC2A 2LL

Date 17/07/2007

Before

LORD JUSTICE WALLER

Vice-President of the Court of Appeal, Civil Division

LORD JUSTICE LATHAM

LORD NEUBERGER OF ABBOTSBURY

Between

WHA Ltd & AnrRespondent

HM Revenue and CustomsAppellant

Roderick Cordara QC (instructed by Messrs Deloitte & Touche) for Respondent

Jonathan Peacock QC and Aidan Robertson (instructed by HM Revenue and Customs) for Appellant

Hearing Hearing dates : 20th, 21st June 2007

Judgment

Lord Neuberger of Abbotsbury:

[1]On 14 May 2004, we handed down our judgments in relation to a scheme (“the Scheme”) which was claimed to have the effect of minimising overall liability to VAT in the context of the supply of repairs and parts provided pursuant to contracts of motor breakdown insurance (“MBI”). The matter came before us as the first part of an appeal from a decision of Lloyd J. He had allowed an appeal by the taxpayers against the decision of the VAT and Duties Tribunal (“the Tribunal”), in favour of the Commissioners for Customs and Excise, now Her Majesty’s Revenue and Customs (“HMRC”). In effect, we held, in part for different reasons from Lloyd J, that the Scheme had the effect for which the taxpayers contended. That decision is reported at

[2004] STC 1081

[2]As I have mentioned, that decision concerned only the first part of the appeal. As I explained in paragraph [4] of my earlier judgment (with which Waller and Latham LJJ agreed), the arguments before us, and our decision, were subject to certain further arguments raised by HMRC “based on the alleged artificiality of the Scheme, including a contention based on abuse of rights (or abus de droit)”. Those arguments give rise to the issues raised on the second part of the appeal, which now have been considered at a further hearing and are the subject of this judgment.

[3]In paragraphs [2] and [3] of my earlier judgment the Scheme was described in these terms:

“MBI policies are issued to members of the public by an English company, [NIG]… . NIG reinsures its liabilities under these policies with a Gibraltar based company called [Crystal] which in turn retrocedes 85% of the reinsurance to another Gibraltar based company, Viscount. Viscount contracts with an English company, WHA, to instruct garages to carry out any works required to be effected under the policies, and to pay for those works. . . . On each occasion that such work is carried out by a garage . . . on WHA’s instructions, the garage renders an invoice to WHA. It is common ground that VAT is payable on this invoice. The effectiveness of the Scheme primarily depends on WHA being able to treat this VAT as input tax. WHA renders an invoice to Viscount which WHA contends is exempt from VAT. If that contention is correct, WHA is able to claim repayment from [HMRC] of the input tax. Alternatively, if WHA is wrong and VAT is chargeable on its invoice to Viscount, then Viscount contends that it is entitled to recover the VAT it has to pay in respect of the invoice from WHA.”

[4]The Scheme was explained in more detail, together with some further relevant facts found by the Tribunal, in paragraphs [5] to [23] of my earlier judgment. I do not propose to set out the contents of those paragraphs again; they may be treated as incorporated into this judgment. However, it is appropriate for present purposes to identify or emphasise certain aspects of the Scheme (some of which are in the passage quoted from my earlier judgment), as they are highly relevant to the issues which we now have to determine.

[5]First, WHA, Viscount and Crystal are and always have been part of the same group of companies. WHA is a wholly owned English subsidiary of an entity called Oriel, itself an English company. Oriel also owns all the shares in another English company called Warranty, which in turn owns all the shares in a Gibraltar company called Practical, of which both Crystal and Viscount are wholly owned Gibraltar subsidiaries. Secondly, under the Scheme, NIG’s insurance liabilities to motorists with MBI policies are 100% reinsured by Crystal, which in turn retrocedes 85% of its reinsurance liability to Viscount. Thirdly, under the Scheme, claims handling and contracts with garages for repair works and parts (“claims handling”) are, as it were, subcontracted by NIG to Crystal, and by Crystal to Viscount, and, finally, by Viscount to WHA.

[6]Fourthly, the Scheme was set up in 1998, and replaced an earlier arrangement, under which NIG’s liability under the MBI policies was 100% reinsured by Practical, which in turn retroceded 100% of its liability to Warranty, and the claims handling was subcontracted by NIG directly to Warranty. Thus, the two main differences between the Scheme and the predecessor arrangement were(a) the involvement of a second Gibraltar company (Viscount) as an 85% retrocedent rather than the English claims handler (Warranty/WHA) as a 100% retrocedent, and

(b) a claims handling contractual “chain” (through what the Tribunal called the “Gibraltar loop”), which included Viscount, from NIG to the ultimate claims handler (WHA), as opposed to a direct contract between NIG and the claims handler (Warranty).

[7]As already explained, the first part of the appeal was argued in 2004 on the basis that the Scheme should be accepted at what one might call face value, that is on the assumption that the documentation embodying the Scheme was valid and effective for all purposes. Accordingly, the questions we had to consider in 2004 involved determining the VAT consequences on that basis. In accordance with well established principles, it was therefore necessary to consider each step in the Scheme individually – see in this connection the observations of Lord Hoffmann in Customs and Excise Commissioners v Robert Gordon’s College

[1995] STC 1093

at 1099, and paragraph 29 of the opinion of the Advocate-General in Optigen Ltd v Customs and Excise Commissioners (Joined cases C 354/03, C355/03 and C 484/03)

[2006] STC 419

[8]The effect of our decision, as summarised in paragraph [157] of my earlier judgment, was as follows.(i) WHA and Viscount (“the respondents”) succeeded in their argument that there was a supply of services to WHA by the garages, and that the VAT payable by WHA was therefore input tax capable, at least in principle, of being recoverable;

(ii) WHA failed in its argument that it made no taxable supply of services to Viscount;

(iii) Viscount succeeded in its argument that it was entitled to recover the input tax it had to pay WHA. Accordingly (subject to the issues which have now been argued before us) we concluded that, although WHA was not entitled to recover the input tax it paid to the garages, as it had failed on point (ii), the Scheme succeeded in minimising overall liability to VAT, as Viscount was entitled to recover the input tax it paid to WHA, in the light of its success on point (iii). Thus, we held that (subject to the issues which now fall to be determined), to put it slightly over-simply, the VAT paid by one member of the Oriel group, WHA, on the repair works and parts provided by the garages, could be recovered by Viscount, another member of the group.

[9]The issue which now has to be decided is whether the Scheme should be struck down, or, to put it more accurately, “redefined”, on the basis that it, or one or more of the steps, or transactions, involved in it, amounts to an “abusive practice”, as recently and authoritatively discussed and explained by the Grand Chamber of the European Court of Justice in Halifax plc v Customs and Excise Commissioners (Case C 255/02)

[2006] 2 WLR 905

. In that case, the European Court made two important rulings in relation to schemes whose purpose was to avoid or mitigate liability to VAT. First, that a transaction entered into solely for that purpose “and without any other economic objective” can constitute a supply of goods or services, and an economic activity (see paragraph [60]). Secondly, that the concept of abusive practice is applicable in the field of VAT (see paragraph [85]).

[10]Both rulings of the European Court in Halifax are relevant to the present dispute. The first has caused HMRC to abandon certain of the arguments it had raised (unsuccessfully) before the Tribunal, and it is therefore unnecessary to consider that aspect of Halifax further. The second ruling is central to the issue we have to decide. The kernel of the European Court’s decision in relation to its second ruling is to be found in the following conclusions:

“[85] …[T]he Sixth Directive must be interpreted as precluding any right of a taxable person to deduct input VAT where the transactions from which that right derives constitute an abusive practice. [86] For it to be found that an abusive practice exists, it is necessary, first, that the transaction concerned, notwithstanding formal application of the conditions laid down by the relevant provisions of the Sixth Directive and of national legislation transposing it, result in the accrual of a tax advantage the grant of which would be contrary to the purpose of those provisions. Secondly, it must also be apparent from a number of objective factors that the essential aim of the transactions concerned is to obtain a tax advantage. ….. [94] … [T]ransactions involved in an abusive practice must be redefined so as to re-establish the situation that would have prevailed in the absence of the transactions constituting that abusive practice.”

[11]As just mentioned, the decision in Halifax has resulted in many of HMRC’s arguments before the Tribunal attacking the Scheme being abandoned, and the remainder of those arguments being refined. The Tribunal had proceeded on a mistaken basis as to the law on abuse (through no fault on its part, as its decision of course preceded Halifax), and Lloyd J did not consider the abuse issue (as he had been invited not to do so, at least for the time being). In these circumstances, following our decision in 2004 and the decision of the European Court in Halifax in 2006, the parties agreed that, as all the relevant facts had been found by the Tribunal in its full and clear decision, the most sensible course in terms of time and cost was for us to consider the abuse issue now. Particularly in the light of the delays which had occurred in relation to these proceedings, we acceded to that course.

[12]The abuse issue can usefully be considered by answering four questions, which appear to emerge from the passages I have quoted from the judgment in Halifax. First, does the Scheme, or an aspect of the Scheme, result in the accrual of a tax advantage which, as HMRC assert, is “contrary to the purpose of” the provisions of the Sixth Directive? Secondly, if so, was it, as HMRC contend, the “essential aim” of the Scheme, or of the relevant aspect, that a tax advantage be obtained? Thirdly, if so, are there any special features of the Scheme itself, or of the law relating to it, which should nonetheless prevent the abuse argument succeeding? Fourthly, if not, can (and must) the Scheme, or the relevant part, be “redefined”?

[13]While one can analyse the issue in this case by breaking it down into these four questions, it is right to acknowledge that the answers may overlap to some extent, and that it may be a matter of opinion as to which question a particular argument or point goes. Nonetheless, I propose to consider the four questions in turn, as that makes it less difficult to achieve a structured and tolerably clear approach to what is, to my mind at least, a potentially confusing problem.

Is the Scheme or part of it contrary to the purpose of the Sixth Directive?

[14]In paragraph [69] of its judgment in Halifax, the European Court drew a distinction between transactions entered into “in the context of normal commercial operations” and those entered into “solely for the purpose of wrongfully obtaining advantages provided for by Community law”. The latter type of transaction is capable of constituting an abuse, provided it satisfies the two tests identified in paragraph [86]. Such a transaction or scheme will not satisfy the first test unless it is “contrary to the purpose” of the principles governing the payment of VAT, which include the “provisions of the Sixth Directive” (see paragraph [74]), as well as “the principle of fiscal neutrality” (see paragraph [80]).

[15]The purposes of the VAT provisions with which we are concerned in the present case is, as in Halifax, to be found primarily by reference to the provisions of the Sixth Directive, Council Directive 77/388 (although the law has subsequently been consolidated into the VAT Directive, Council Directive 2006/112/EC – “the 2006 VAT Directive”). Those purposes have been discussed by the European Court in a number of earlier decisions including Elida Gibbs Ltd v Commissioners of Customs and Excise (Case C 317/94)

[1996] STC 1387

. In paragraph [19] of its judgment in that case, the court said that the “basic principle of the VAT system is that it is intended to tax only the final consumer”. In the following paragraph, the court confirmed that the system was “based on neutrality”, which means that “within each country similar goods should bear the same tax burden whatever the length of the production and distribution chain”. In paragraph [22], the court stated that “taxable persons [other than the final consumer do not] bear the burden of VAT”, but “collect the tax on behalf of the tax authorities and account for it to them”. Thus, as explained in paragraph [23] of Elida Gibbs, VAT is, in general, charged on the consumers in each transaction in the chain, and is recoverable by taxable consumers in the chain, provided they make taxable supplies, so that the tax is ultimately borne by the final consumer.

[16]For HMRC, Mr Peacock QC contended that such fiscal neutrality requires the conclusion that an insurer, who provides, in the EU, insurance services which are exempt for VAT purposes, cannot recover input tax attributable to those services. Thus, in what one can fairly characterise as transactions in the context of “normal commercial operations” of an insurer and a claims handler, such as that embodied in the arrangements which were replaced by the Scheme, there would be no question of the input tax attributable to the cost of repairs and parts being recoverable. Given that the effect of the Scheme, according to our 2004 decision, is that input tax, incurred in the provision of exempt insurance services, is recoverable, HMRC accordingly argue that the Scheme is, at least to the extent that it has such an effect, contrary to the purposes of the legislative purposes of the VAT legislation, as embodied in the Sixth Directive (and now in the 2006 Directive).

[17]It seems to me that, subject to any arguments to the contrary by reference to the background and details of the Scheme or the legislation, this argument is correct. Although Gibraltar companies, namely Viscount and Crystal, are involved in the chain, the truth is that the provision of the services comprising the repairs and parts are provided in the EU to WHA, and what WHA provides, albeit through two Gibraltar companies in the same group, is the provision of claims handling, again in the EU, to a supplier of exempt services in the EU, namely NIG. On the face of it, at any rate, the VAT regime would plainly require that arrangement to result in an overall liability to VAT equal to the tax chargeable on the services, rather than, as results from the Scheme, no net liability whatever to VAT (as Viscount recovers the input tax paid by WHA).

[18]Mr Cordara QC, who appeared for the respondents, contended that this analysis was impermissible because the present scheme is governed by, or includes steps which are governed by, Article 17(3) of the Sixth Directive, whereas Halifax, involving as it did no transactions outside the EU, was only governed by Article 17(2), which pursues a different policy objective. Although it is true that, unlike in Halifax, Article 17(3) is engaged here, I do not accept that that represents a satisfactory answer to HMRC’s contention on the first question, at least on the facts of this case. In paragraph [80] of its decision in Halifax, the European court made what appears to me to be a perfectly general point (which was echoed in their conclusion in paragraph [86] quoted above):

“To allow taxable persons to deduct all input VAT even though, in the context of their normal commercial operations, no transactions conforming with the deduction rules of the Sixth Directive or of the national legislation transposing it would have enabled them to deduct such VAT …. would be contrary to the principles of fiscal neutrality and therefore contrary to the purpose of those rules”

. Similarly, in paragraph [68], the court, when explaining why the abuse argument applied in the VAT field, invoked the general principle that “according to settled case law, Community law cannot be relied on for abusive or fraudulent ends”.

[19]No sensible reason has been advanced as to why there should be a different approach on the issue of abuse in relation to a scheme which involves transactions to which Article 17(3) applies as opposed to a scheme all of whose transactions fall within the ambit of Article 17(2). In those circumstances, I am unpersuaded that the fact that, unlike in Halifax, some of the steps in the present Scheme involve transactions which fall within Article 17(3), as opposed to Article 17(2) of the Sixth Directive, affects the application of the abuse principle, as explained in Halifax, to the Scheme.

[20]Turning to the facts of this case, what enables the Scheme to achieve its end is not merely that one of the links in the chain is in Gibraltar: that is well illustrated by the fact that the earlier scheme did not have the allegedly abusive consequences of the Scheme, but still involved a link in Gibraltar (namely, Practical). As already mentioned, the most essential differences between the Scheme and its predecessor arrangement are what Mr Cordara, with characteristic candour, accepted was its most vulnerable aspect for present purposes, namely the arrangements between Crystal and Viscount, as well as the claims handling chain from NIG to WHA, via Viscount. I shall discuss these aspects further in the next section of this judgment, but the essential point for present purposes is that, in essence, it is the arrangements between Crystal and Viscount, together with the claims handling chain including Viscount, which enable HMRC to make out their case on the second question, namely that the purpose of the Scheme is to obtain a tax advantage. In those circumstances, it would be surprising if the mere fact that Article 17(3) applies to those arrangements could prevent the abuse principle applying, particularly given that it is a principle of general application.

[21]The services under the Scheme (with the exception of the reinsurance) were ultimately all provided in the EU. The imposition in the claims handling chain of a second Gibraltar company which also retrocedes 85% of the reinsurance liabilities of the other Gibraltar company in the chain, when both companies are wholly owned subsidiaries of a third company, emphasises that the only commercial service provided outside the EU is the reinsurance. As I have mentioned, these was not enough to achieve the tax-minimising aim of the Scheme, as is demonstrated by the fact that, as is common ground, the preceding arrangements did not achieve (or, indeed, seek to achieve) that aim.

[22]It was also contended by Mr Cordara that it would be wrong to characterise the Scheme as an abuse because it involves inappropriately considering the Scheme as a whole, whereas questions relating to VAT are to be determined by reference to individual transactions – i.e. by treating each step in the Scheme separately. Thus, he argued in his written submissions that the “abuse test is not satisfied by WHA viewed alone”. While I accept the soundness of the approach in classic VAT cases (indeed, we adopted it when considering whether the Scheme worked when considered at face value), I do not consider that it can possibly be appropriate when considering whether a scheme infringes the purpose of the Sixth Directive. Otherwise, a scheme would never be liable to attack on the basis of the principle established in Halifax. Effectively by definition, each step of such a scheme would be unassailable (as it would otherwise be unnecessary to invoke the abuse principle). Accordingly, on this argument, the scheme itself would be unassailable. Indeed, if this argument were correct, the European Court would have decided Halifax differently. The whole point of the principle is that, although each step of the scheme in question works, the overall effect of the scheme is unacceptable.

Was the essential aim of the Scheme to obtain a tax advantage?

[23]In Halifax, the European court said in paragraph [82] that, in order to establish abuse, tax saving had to be “the sole purpose of the transactions at issue”, which could be said to be a little stricter than the test laid down in paragraph [86] (“the essential aim of the transactions concerned”). Mr Peacock accepted that the arguably looser test in paragraph [86] should not to be treated as diluting the stricter test suggested in the earlier paragraph, a concession which can be said to be supported by what the court said in paragraph [75]. As Mr Peacock himself observed, it is unnecessary to decide whether he was right to make that concession on the facts of this case, so I am content to proceed on the assumption that the “sole purpose” test is correct.

[24]The European court also indicated on more than one occasion in its judgment that the question of purpose was to be judged objectively and not subjectively, i.e. by reference to the terms of the scheme concerned and the commercial realities, not by reference to what the parties concerned say their intention was (or what their subjective intention is found to have been). Thus in paragraphs [75] and [86], the court made reference to the necessity of basing one’s conclusion as to the intention on “objective factors”. The point was more fully made in paragraph [87] of the Advocate General’s opinion.

[25]In this case, I agree with Mr Peacock that the Tribunal made findings which appear to be conclusive against the respondents as to the purpose for which the Scheme was entered into. As the Tribunal pointed out in paragraph [13] of its Decision, to recover the input VAT paid by WHA, it was “necessary to install another non-EU ‘trader’ [Viscount] interposed between the Gibraltarian reinsurer [Crystal] and the UK ‘users’ of the supplies originating from WHA”. In paragraph [38], the Tribunal concluded that “the purpose of those responsible for implementing [the Scheme] was to obtain that tax advantage”, that is the recovery of the input tax paid by WHA. In so far as it is relevant and admissible, this conclusion was supported by the minutes of all the relevant meetings, as the Tribunal mentioned in paragraph [53].

[26]In paragraph [61], the Tribunal described Crystal and Viscount as “creatures” of the Scheme. In paragraph [71], the Tribunal said that “the documentation and the arrangements” comprising the Scheme were “designed to divert the supplies of labour and parts from their normal direct route from garage to insured by routing them via the Gibraltar loop” in order “to facilitate” the Scheme. Mr Cordara rightly said that this observation was directed to a finding (unsustainable in the light of the first point decided in Halifax) that all or some of the transactions comprised in the Scheme were not to be treated as involving a supply or an economic activity. However, the fact that this finding was invoked by the Tribunal to justify a conclusion which was wrong in law does not mean that it was not a valid finding which can properly be invoked by this court for justifying another conclusion whose legal basis is valid.

[27]It is also of note that the Tribunal observed in paragraph [74] that “the claims handling supplies ha[d] been put through the Gibraltar loop to facilitate” the Scheme, and that it was “unusual to say the least for a retrocedent insurer such as Viscount to take on responsibility for claims handling”. In the same paragraph, the Tribunal stated that there was “of course no ‘commercial’ need to have a second Gibraltarian captive” which they described as “a requirement purely of” the Scheme.

[28]In the light of these findings, it seems to me that only one answer is possible in relation to the second question. That answer is that it was the sole purpose of the Scheme, and in particular the retrocession arrangement between Crystal and Viscount and the creation of the claims handling chain so as to include Viscount, to avoid, or at any rate to minimise, any net liability to VAT, by enabling the input tax paid by WHA to be reclaimed by Viscount (or alternatively to be reclaimed by WHA, but that was not achieved in the light of our conclusion on point (ii) in our earlier conclusion as summarised in paragraph [8] above).

[29]Of course, in one sense at any rate, the purpose of the Scheme was to enable NIG’s liabilities under the MBIs to be performed and to be reinsured. So, it may be contended, tax avoidance cannot be said to be the sole, even arguably the main, purpose of the Scheme, viewed as a whole. However, as I see it, when considering the purpose of the Scheme for present purposes, one must primarily address the aspects of the Scheme which are artificial. Otherwise, many schemes, however abusive, would succeed: indeed, on the basis of this contention, the decision in Halifax might very well have gone the other way. It seems to me that I am supported in this opinion by the reference in paragraph [80] of the judgment in Halifax to “normal commercial operations”, and the requirement in the following paragraph that, where abuse is established, the national court must “determine the real substance and significance of the transactions concerned”. This plainly seems to envisage that a scheme may be abusive while having a genuine underlying commercial purpose.

[30]Another way of making the same point, which I think was in the mind of the Tribunal in the light of what it said in the passage quoted above from paragraph [13] of the Decision, is that the Scheme only differed from the previous arrangement by involving(a) the interposition of a second Gibraltar company between the ultimate claims handler and the reinsurer, and

(b) the creation of a claims handling chain which includes that company. It is true that the identity of the reinsurer changed from Practical to Crystal, and that the ultimate claims handler changed from Warranty to WHA. However, all those four companies, and the new interposed company, Viscount, were ultimately 100% owned by Oriel and were 100% members of the same group of companies.

[31]So, at the risk of a little over-simplification, the real substantive change was Viscount’s involvement, which had two functions. The first was to retrocede 85% of the liabilities of the reinsurer, Crystal; given that both Viscount and Crystal were 100% owned by Practical, this seems, as the Tribunal found, to have been commercially pointless, by which I mean of no value except to enable WHA’s input tax to be reclaimed. Viscount’s second function was to be another link in the claims handling chain; that function was similarly commercially pointless, and was either contingent on the first function or came into existence to enable the input tax to be reclaimed. The claims handling chain was part of the “Gibraltar loop” as described by the Tribunal: there was no apparent commercial reason for NIG not sub-contracting the claims handling directly to WHA, as it had done to Warranty.

[32]Mr Cordara submitted that the Tribunal nonetheless were not entitled to find, as they effectively did in the passages I have quoted from their Decision, that the sole purpose of the Scheme, and in particular the involvement of Viscount as a retrocedent and as a link in the claims handling chain, was to avoid or minimise overall liability to VAT. This submission was based on the evidence which the Tribunal received of other benefits of the Scheme. In this connection, he concentrated on the centrally artificial aspect of the Scheme, namely the existence and involvement of the two Gibraltar companies, rather than one such company. Mr Ross-Roberts, the finance director of Warranty at the time, suggested that there were capitalisation and cash-flow advantages for the Oriel group as a whole in having two Gibraltar companies rather than one, and that it was good practice to separate the claims handling from MBI policy sales. As the Tribunal expressly accepted Mr Ross-Roberts’s evidence on this topic, and as Mr Peacock accepted the “sole purpose” test, ran the submission, HMRC’s case on abuse falls at the second of the two hurdles identified in Halifax.

[33]I would reject that submission for two separate reasons. First, in paragraph [42] of their Decision, immediately after saying that they accepted Mr Ross-Roberts’s evidence on the cash flow benefits, the Tribunal immediately went on specifically to reject the notion that it “need[ed] two Gibraltar captives to achieve the cash-flow benefits”. In the following paragraph, they dealt in similar terms with his evidence as to the division of functions. It is true that the Tribunal did not expressly deal with the capitalisation benefits, but it seems clear from the passages of the Decision from which I have already quoted that the Tribunal considered that they had no practical significance. Indeed, I find them impossible to understand. Mr Cordara had no answer to the apparently logically inevitable point put by Waller LJ in argument that, if by retroceding most of its reinsurance liability to Viscount, Crystal’s minimum capitalisation under Gibraltar law was reduced, Viscount’s must have been correspondingly increased. Accordingly, there can have been no overall benefit to the Oriel group on this ground.

[34]Further, as the Tribunal pointed out, in none of the contemporaneous documentation concerning the setting up of the Scheme was there reference to any purpose other than the minimising of liability to VAT. At best from the respondents’ point of view, the Tribunal found that, to the extent that any of the advantages identified by Mr Ross-Roberts existed, they were collateral beneficial results of a scheme, which was conceived and implemented purely for tax avoidance purposes. Particularly when one takes into account the very high hurdle which the respondents face in challenging a decision of fact on the part of the Tribunal (see e.g. Edwards v Bairstow

[1956] AC 14

at 29 and 36), I would therefore reject the contention that the Tribunal was not entitled to find that the sole reason for the Scheme (particularly the involvement of Viscount and the claims handling chain) was to minimise the Oriel group’s net liability to VAT.

[35]There is a second reason for reaching this conclusion. It appears to me that evidence from witnesses such as Mr Ross-Roberts, as to the subjective intention or purpose of the Oriel group companies in setting up the Scheme, and, in particular, in including Viscount as a retrocedent and in the claims handling chain, was probably not admissible in the light of what was said by the European Court in paragraphs [75] and [86] of its judgment, and, more fully, by the Advocate General in paragraph [87] of his opinion, in Halifax. On that basis, I consider it plainly more likely than not, indeed inevitable, that the Tribunal would have rejected the reasons advanced by Mr Ross-Roberts as representing any part of the purpose in developing or implementing the Scheme.

[36]In these circumstances, I would leave open the further point that, even if any of the reasons advanced by Mr Ross-Roberts could be relied on as representing part of the purpose of the Scheme, they were so minor and unimportant when compared with the purpose of avoiding overall liability to VAT, that it would not be a misuse of language to say that tax avoidance was the “sole purpose” or “essential aim” of the Scheme, or, to be more precise, of the involvement of Viscount (and, possibly, the claims handling chain), as those expressions were used by the court in Halifax.

Arguments as to why HMRC’s abuse argument should nonetheless fail

[37]On behalf of the respondents, a number of reasons were raised to support the contention that, despite the conclusions reached so far, namely that the two requirements for the establishment of abuse laid down in Halifax have been satisfied, HMRC’s case on abuse should fail.

[38]The point that it is convenient to take first (although by no means the most strongly pressed) is that the court in Halifax accepted at paragraph [73] that a taxpayer who has alternative courses open to him is entitled to choose that which minimises his liability to VAT. I do not consider that there is anything in that point. There may be cases where it is difficult to decide whether a particular arrangement is one which includes a step or steps which amount to an abuse or whether it is a course which is properly open to the taxpayer as a way of minimising his liability to VAT. However, this is not such a case. In this connection, one comes back yet again to the insertion of Viscount and the claims handling chain: especially in the light of the findings of the Tribunal, the Scheme was plainly not simply one of two (or more) courses for the Oriel group reinsuring NIG’s MBI liabilities and arranging the claims handling. It was an ingenious, but wholly artificial, step included purely to enable the input tax paid by WHA to be reclaimed.

[39]A second argument, raised in a somewhat inchoate way, was based on the European Court’s emphasis in paragraph [72] of its judgment in Halifax on the need for legal certainty. I do not consider that this point can assist the respondents here. While there may, I suppose, be exceptional cases, once it is established that HMRC can clear the two hurdles which were identified by the court in Halifax, then that, as I see it, must be the end of the matter so far as legal certainty is concerned. The point is that the two hurdles were fashioned by the court partly in order to satisfy the requirement for legal certainty, so if they are cleared, it is hard to see how the need for legal certainty can present a difficulty.

[40]Mr Cordara’s third argument I turn to consider is that HMRC, having succeeded on point (ii) on the first part of the appeal, namely having established that WHA is obliged to charge output tax for the supplies it makes to Viscount, are not entitled on this part of the appeal to make submissions which are inconsistent with that conclusion. In my view, that argument misunderstands the nature of the abuse argument. As I have explained, the conclusions reached at the previous hearing were on the express, clear and common assumption that the Scheme was not an abuse. With, I suppose, one exception, any conclusion reached on that basis, whether in favour or against HMRC, whether contested or agreed, is not binding when one turns to this stage, where the enquiry is whether the Scheme is abusive. (The exception is, of course, whether the Scheme succeeds in its aim of minimising liability to VAT). To put the point a slightly different way, the first part of the appeal was concerned with the VAT analysis of each step of the Scheme, whereas this part is directed to the effect of the Scheme as a whole.

[41]The fourth argument on behalf of the respondents was that HMRC’s abuse argument offends against the principle of freedom of establishment enshrined in Articles 43 and 48 of the EC Treaty, as recently considered and applied by the Grand Chamber of the European Court in Cadbury Schweppes plc v Inland Revenue Commissioners (Case C 196/04)

[2006] STC 1908

. This argument rests on the proposition that the Scheme succeeds in minimising the Oriel group’s net VAT liability because Viscount is established in Gibraltar, and the conclusion that the Scheme is an abuse therefore impermissibly penalises the group because it has exercised its freedom to set up Viscount in Gibraltar. (In this context, I note in passing the apparent paradox that, while the effectiveness of the Scheme depends on Gibraltar being outside the EC for VAT purposes, this argument relies on the fact that it is inside the EU so far as freedom of establishment is concerned. This is because although the EC Treaty applies to Gibraltar generally pursuant to Article 227(4), the effect of Article 28 of the UK Act of Accession is to exclude Gibraltar from the territorial scope of VAT.)

[42]The short answer to this argument is, in my judgment, that advanced by Mr Peacock. He submitted that Cadbury-Schweppes was concerned with the right of a business to establish itself in another EU member state to take advantage of that state’s beneficial tax regime, whereas the principle relied on here by HMRC, as in Halifax, is that, once it is established, in exercise of its right of establishment, in an EU member state, a business is not permitted to abuse the VAT system within the EU. The correctness of this submission in its application to the facts of this case is borne out by considering the essential nature of the allegedly abusive and artificial contrivance of the Scheme. It was not the establishing of Viscount in Gibraltar: after all, the previous arrangement was not abusive, and the reinsurer, Principal, was established in Gibraltar. It was the involvement of Viscount in the Scheme which was the artificial contrivance, because there was no need or purpose (other than the avoidance of tax) in dividing Principal’s former functions between two companies, namely Crystal and Viscount, in creating another link in the claims handling chain, or in retroceding 85% of Crystal’s reinsurance liabilities to Viscount.

[43]Fifthly, the respondents relied on the fact that the reason the Scheme succeeded in avoiding VAT was not because of European legislation, but because of domestic law. In our decision on the first part of this appeal, we concluded that Viscount was able to recover the input tax paid by WHA, not because of the relevant provisions of the Sixth Directive (namely Articles 9(2)(e) and 17(3)) or the Thirteenth Directive, but because of the effect of the domestic legislation enacted with the intention of implementing the relevant provisions of those Directives (namely regulations 186 and 190 of the Value Added Tax Regulations 1995, SI 1999/3121, “the 1995 order”) - see paragraphs [113] to [155] of my earlier judgment. In those circumstances, Mr Cordara contended that the abuse principle cannot be invoked, because it is only in relation to EU legislation that it can apply.

[44]In my opinion, that contention should be rejected. It is clear that the provisions of the 1995 order (and in particular, for present purposes, regulations 186 and 190) were enacted to give effect to the provisions of the Sixth and Thirteenth Directives. Mr Peacock argued that, provided that the domestic legislation in question has been enacted with such an intention, then the fact that it imperfectly transposes the Sixth Directive should not justify non-application of the abuse principle. I agree. It appears to me that it would be illogical and unsatisfactory if the abuse principle could not apply to a scheme, merely because one of its steps relied for its efficacy on domestic VAT legislation. As already mentioned, whether a particular scheme is abusive must be determined by considering that scheme as a whole, first in terms of its effect and secondly in terms of its purpose. The reason a particular step in the scheme achieves its end is scarcely in point.

[45]Further, it would seem odd if the mere fact that a national rule relating to the recovery of input tax was more generous than what was provided in the Sixth Directive should enable a scheme which included a step which relied on that national rule to be immune from being defeated on abuse grounds. Apart from being wrong in principle in my view, if that were the case, it appears to me that it could often give rise to difficult questions to which the answers might often be somewhat subjective, such as whether the step in question was the (or an) artificial step and whether that step was an essential ingredient of the scheme.

[46]I draw some support for my view on this point from the observations in the European Court’s judgment in Halifax. Not only is there the referenceto “the deduction rules” (paragraph [80]) or “the relevant provisions” (paragraph [86]) “of the Sixth Directive [and/or] the national legislation transposing it” (emphasis added). There is also the observation in paragraph [67] that the problems raised in the appeal “at least in part, stem from national rules …. which allow a taxable person [in certain circumstances] to deduct the total input VAT …” (emphasis again added). The words I have emphasised could, as a matter of language, be read as limited to the national legislation or rules only in so far as they reflect the requirements of the Directive, but I think that their more natural meaning in their context is that they also extend to legislation and rules intended to implement, or to reflect the terms of, the Directive. That is not only for the reasons already mentioned, but also because those words, at least in paragraphs [80] and [86], otherwise appear to add nothing in practice to what precedes them, as the Sixth Directive has direct effect.

[47]I do not see any reason in principle or logic why the applicability of the abuse principle as considered in Halifax should be limited to schemes which depend for their efficacy solely on Community law, whether transposed into domestic legislation or of direct effect. Because abuse is a relatively new concept in the field of VAT and I am uneasy about the possible ramifications of basing a conclusion on a wider ground than is strictly necessary, I prefer to decide this point on the limited basis, which Mr Peacock advanced as his primary contention on this point, namely that the domestic legislation relied on was intended to implement the relevant provisions of the Sixth Directive.

[48](There is a potential irony if this fifth argument had succeeded, in the light of developments subsequent to our earlier decision, according to what we were told by counsel. Following our judgment, the UK government has changed the domestic legislation so that it follows what we held to be the effect of the Sixth Directive. The European Commission has now given formal notice to the UK government that, as a result of this change, the domestic law no longer complies with the provisions of the Sixth Directive. In other words, contrary to the view we formed, the Commission considers that the former provisions of the domestic legislation, as we construed them, did comply with the provisions of the Sixth Directive, contrary to the way we construed them. If the Commission’s opinion is correct, then Mr Cordara’s fifth argument as discussed in the immediately preceding paragraphs would not be open to him).

[49]The sixth argument pursued by the respondents was that it would have been only too easy for the UK legislature to change the national law, pursuant to the provisions of Article 9(3) of the Sixth Directive (now Article 58(b) of the 2006 VAT Directive) so that the Scheme could not achieve its end. This would be achieved, Mr Cordara argued, by redrafting the provisions of Article 19 (or paragraph 18) of the Value Added Tax (Place of Supply of Services) Order 1992, SI 1992/3121, so that reinsurance effected outside the EU of a liability in the UK should be treated as the supply of a service in the EU. The most attractive way he put the argument, at least from my perspective, was when he said that the doctrine of abuse should not be invoked to put right an oversight in the drafting of national legislature.

[50]However, it appears to me that this argument is really the same as the fifth argument I have just considered and rejected. It strikes me as illogical that a Community-based rule against abuse of the VAT regime should not apply merely because the otherwise abusive step happens to rely on a national rule. The observations in paragraphs [67], [80] and [86] in Halifax are again in point. Further, the court in that case impliedly rejected a similar (if much broader) argument raised by the taxpayer, namely that, as the UK could have exercised its right under Article 27 of the Sixth Directive to defeat the scheme in that case, the scheme could not be characterised as abusive. (The point is expressly and fully dealt with in the opinion of the Advocate General at paragraphs [76] to [79] of his opinion).

[51]Seventhly, Mr Cordara pointed out that, in a number of places in its Decision, the Tribunal had said that each of the steps constituting the Scheme was a genuine transaction. That seems to me to take matters no further. It is entirely consistent with the first point decided in Halifax, where the European Court emphasised that the fact that a transaction was entered into for purely tax avoidance or tax mitigation purposes did not prevent it from being a supply of goods or services for VAT purposes. However, as is plainly demonstrated by the decision of the court on the second point in that very case, that does not prevent a scheme, which includes or consists of such transactions, from being an abuse. The fact that a transaction is genuine does not by any means demonstrate that it is a “normal commercial operation”.

[52]Eighthly, it was said on behalf of the respondents that the principle that subjective motive was irrelevant assisted them here. As I understood it, this was on the basis that, as the question of intention should be judged objectively, there was no abuse here, because a person, who formulated an arrangement in connection with MBI reinsurance and claims handling, for a purpose other than minimising VAT liability, might have come up with the Scheme. There are two answers to that point, both of which involve focussing on the inclusion of Viscount as a retrocedent coupled with the claims handling chain. First, it is clear that this contention falls foul of the Tribunal’s findings in connection with what they called “the Gibraltar loop”, as discussed above, Secondly, for the same reasons as the Tribunal, it seems to me fanciful to think that there was any reason other than minimising liability to tax for including Viscount in the Scheme either as a retrocedent insurer or as a link in the claims handling chain.

[53]Finally, Mr Cordara sought to justify the Scheme on the ground that the introduction of Insurance Premium Tax (“IPT”) in 1998 meant that, unless an arrangement such as the Scheme was introduced, reinsurers such as the Oriel group could not compete with competitors established outside the EU. Although I have some sympathy with the predicament in which the Oriel group found itself by 1999, apparently at least partly as a result of the introduction of IPT, that extraneous factor cannot assist the respondents here. There is always a good commercial reason to minimise liability to VAT, so that plainly cannot be enough to defeat an abuse claim. It appears equally plain to me that it cannot matter whether the commercial reason is to make or increase profits or to avoid or reduce a loss.

[54]In these circumstances, I am of the view that, given that the two requirements identified in paragraph 86 of Halifax are satisfied, no convincing reason has been advanced as to why the Scheme should not be redefined as discussed by the European Court in paragraph [94] of that case.

How is the Scheme to be redefined?

[55]This final question concerns whether, and if so how, the Scheme must or can be redefined, given my conclusion that it represents an abuse. The question raises an issue which is, at least as I see it, academic, although a substantial amount of the written submissions, particularly on the part of the respondents, was devoted to it. The reason the issue is academic in my view is that, in the light of the events that have happened, there is no need to redefine the Scheme in any specific way because, on the assumption that the Scheme is an abuse and therefore does not achieve it aim, there is no money claim either by HMRC or by any of the respondents. WHA has paid the input tax, but, based on their contention that the Scheme does not work, HMRC have not paid either Crystal (as they should have done on Lloyd J’s analysis of the Scheme at face value) or Viscount (as they should have done on our analysis of the Scheme at face value). As that contention has turned out, as a result of this judgment, to be justified, nothing remains to be done.

[56]Mr Cordara argued that it was necessary to seek to redefine the Scheme (or identify a “counter factual” as he usefully called it) essentially for two reasons, as I understood it. First, he contended that there was, on analysis, no satisfactory way in which the Scheme could be redefined, with the result that HMRC’s abuse case should ultimately fail. He also contended (although not, I think, with great force) that, even if it was only of academic interest, we should identify the way in which the Scheme was to be redefined, as that is what the European Court said was required of a national court in paragraph [94] of its judgment in Halifax. I am not impressed with either point.

[57]As to the first point, it is clear to my mind from what the court said in the very paragraph to which I have just referred that an abusive scheme “must be redefined”. In other words, once the two hurdles identified in paragraph [86] of the European Court’s judgment have been crossed, there is no third hurdle to be crossed before an abusive scheme is, as it were, neutralised. The only reason for a formal redefinition as mentioned in paragraph [94], at least in my judgment, is to work out the VAT consequences, as seems apparent from the following three paragraphs of the judgment in Halifax. These deal with the rights of the tax authorities to demand tax, the rights of taxpayers not to be over-taxed, and the rights of third parties, as a result of a scheme being abusive, and needing to be redefined. Where, as here, the consequences of the scheme in question being abusive can involve no such rights, because the tax due has been fully paid, but not overpaid, and no third party rights need protection, redefinition has no purpose.

[58]As to the second point, it may be of academic interest to see how the Scheme could and should be redefined, but that is scarcely a sufficient reason for carrying out such an exercise. Having said that, it is right to mention that two possible redefinitions were suggested. The first (preferred by the respondents, if we were to redefine) simply involved the excision of Viscount; the second (favoured, I think, by HMRC) was rather more radical, and involved the garage’s supply being treated as being to the insured alone, and in effect involved the whole claims handling chain being struck down. Each of the two possible redefinitions has its attractions. The former option is relatively uncontroversial; it fastens on what is perhaps the most artificial aspect of the Scheme, namely the involvement of Viscount; and it notionally restores the position to what it was before the Scheme was devised. (The former option also leaves WHA’s position effectively intact, and would therefore represent a second reason for rejecting Mr Cordara’s third point considered in paragraph [40] of this judgment). On the other hand, the latter option reflects the Tribunal's opinion of commercial reality, and it gets rid of the claims handling chain, which is an artificial feature. However, it is a more radical option, and it goes further than merely notionally restoring the position to what it was before the Scheme was introduced. There are, I suspect, other possible redefinitions, but it would be pointless, in the light of the wholly academic nature of the redefinition exercise, to examine this aspect further. It suffices to say that either of the two proffered redefinitions appears to me to be acceptable.

[59]Arguments were developed as to why each of the two proffered redefinitions would be impermissible, but I did not find those arguments convincing, essentially because they appeared to overlook the fact that the redefinition is not for any purpose other than achieving an appropriate outcome for VAT purposes in the light of the principles which inform the Sixth Directive. I mean no disrespect to those arguments by not dealing with them specifically or more fully. However, as explained, I consider that redefinition is not necessary on the facts, and that, even if the facts required it, then it would not be permissible to reject redefinition. Accordingly, it would be otiose to take up further time and paper dealing with Mr Cordara’s objections to the two redefinitions which have been proposed, including the redefinition which he himself would have preferred us to adopt if we identified any specific redefinition.

Conclusion

[60]In the event, then, having held in 2004 that the Scheme, if taken at face value, achieved its tax-saving aim and that Viscount was entitled to claim payment of the input tax paid by WHA, I would now hold, having heard arguments based on the reasoning and principles laid down by the European Court in Halifax, that the Scheme is abusive and that Viscount is not so entitled. I believe that it follows that HMRC’s appeal succeeds, and that the financial consequences of the decision of the Tribunal (who, like Lloyd J, did not have the benefit of the European Court’s guidance) are reinstated, albeit for somewhat different reasons.

Lord Justice Latham :

[61]I agree.

Lord Justice Waller :

[62]I also agree.

Cited in 11 later judgments

and 2 more of the most senior