Revenue and Customs v Isle of Wight Council & Ors

In paragraph 72 of its judgment in Halifax plc and Others v. Commissioners of Customs and Excise, Case C-255/02, 21 February 2006, the Court of Justice of the European Communities (“the ECJ”) observed as follows:

The wisdom of that observation is perhaps obvious, and so it is unfortunate that the ECJ has found it necessary to keep re-affirming it. The present appeal, against a decision dated 23 January 2006 of the VAT and Duties Tribunal (Stephen Oliver QC, Chairman, and Kenneth Goddard MBE), turns on certain provisions of a Community tax directive whose formulation appears to have eschewed any attempt to achieve foreseeable certainty. It is not surprising that it has led to the four cases under appeal.

The relevant provisions are the first two sub-paragraphs of article 4.5 of the Sixth Council Directive of 17 May 1977 (77/388/EEC), a directive aimed at harmonising the system of value added tax (“VAT”) and achieving a uniform basis of assessment. Article 4.5 is in a provision dealing with the identification of “Taxable Persons” for VAT purposes. The first sub-paragraph exempts certain public authorities (including local authorities) from the status of taxable persons in respect of certain activities in which they engage as such; and the second sub-paragraph cancels that exemption if its conferring “would lead to significant distortions of competition.” Absent agreement, it is inherent in the directive that a public authority’s liability to account for VAT on its revenue from relevant activities will depend upon the outcome of an inquiry into the economic consequences of the conferring upon it of an exemption.

The appeal arises from the provision by four local authorities of off-street car-parking facilities. The private sector also provides like facilities. There is, therefore, the potential for competition between the public and private sectors; and, if the public sector is exempt from VAT on its car-parking revenue, it has the opportunity to undercut the private sector. Of course a local authority’s provision of any such facilities will be within its own area, so that it is obvious that the market affected (if at all) by its exemption from VAT will be a local one falling within that area or at least not ranging materially beyond it. Consistently with that, the tribunal held that the economic investigation that the second sub-paragraph of article 4.5 intends is a local one. It conducted such an investigation in relation to each of the four local authorities and concluded in each case that their exemption from taxable status would not lead to significant distortions of competition. The result was that all were exempted from having to account for VAT on their off-street car-parking revenue.

By this appeal the Commissioners challenge the correctness of those decisions. Their main point is that the tribunal was in error in conducting an investigation of the impact of any tax exemption on the particular market in which each local authority was respectively operating. Their submission is that article 4.5 does not intend any “real world” investigation of that sort. They say that what it intends is a nationwide investigation extending across the United Kingdom directed at identifying an overall picture as to the impact upon the private sector generally of the exemption of local authorities generally from taxation in respect of off-street car-parking. If the outcome is that such exemption would lead to significant distortions of competition generally, then it is said to follow that every local authority providing off-street car-parking will be a taxable person even if it is plain in the case of any particular authority that its exemption from VAT would not significantly distort competition in the local market in which it is providing such facilities.

If the Commissioners are right on that, the appeal must be allowed since it would follow that the tribunal had misdirected itself as to the nature of the required investigation. The Commissioners also submit that the tribunal misdirected itself in its approach to the question of whether the non-taxable status of the local authorities “would lead to” significant distortion of competition. Their submission is that the Community learning shows that the quoted words mean “could lead to”, a phrase whose true sense is that “there is a real risk that etc”. That is a question of importance to the application of article 4.5 since, if the Commissioners are right, it probably means that a public authority’s scope for avoiding classification as a taxable person is more limited than if the words of the directive are interpreted as meaning what they might appear to be saying; and the Commissioners say that in the present case the tribunal did not apply the right test. They further say that the tribunal misdirected itself in its interpretation of “significant distortions of competition”. Finally, if they are wrong in their criticisms of the tribunal’s directions on the law, they complain that its fact-finding was in material respects either perverse or at least insufficiently reasoned for them to understand why they had lost.

Having considered the parties’ submissions, I have decided against attempting to rule at this stage on the outcome of the Commissioners’ appeal. As to the point on the nature of the economic investigation, Mr Christopher Vajda QC, for the Commissioners, satisfied me that it raises a question of general importance that only the ECJ can answer; and I am satisfied that the other two points raise like questions. Without answers to these questions, I regard it as impossible to embark on a reasoned assessment of whether the tribunal’s decision was in any respect wanting. I propose to refer these questions to the ECJ. I should explain how I have arrived at that decision.

Article 2 (in Title II, headed “Scope”) provides:

Article 4 (in Title IV, headed “Taxable Persons”) provides, so far as material:

The directive does not in fact number the four sub-paragraphs of article 4.5 but I have done so for ease of subsequent reference. As regards the “would” in article 4.5(2), the material before me suggests that the French, Danish, Italian, Dutch, German and Swedish versions may or do also support a “would” interpretation, whereas the Portuguese and Spanish versions may or do instead support a “could” interpretation. Annex D, referred to in article 4.5(3), lists 12 categories of activities, which do not include off-street parking.

Historically, local authorities in the United Kingdom have accepted that they were taxable persons and so accountable for VAT on their revenue from off-street car-parking activities. The inference is that, until this century, none focused on the potential for asserting an exemption under article 4.5(1) and all meekly accounted to the Commissioners for VAT. The ECJ’s decision on 14 December 2000 in Fazenda Publica v. Camara Municipal do Porto [2001] STC 560 (“Porto”) appears to have provoked them into relevant thought, following which some 127 local authorities made claims for repayment of the VAT they had previously paid but which they now say was not due. Their total claim is for some £129m. I am told that, ignoring parish councils and the like, there are some 468 local authorities in the United Kingdom, so that there is a considerable potential for yet greater claims. In the four cases under appeal the claims total some £1.6m. The cases are in the nature of test cases and the four respondent local authorities have been identified as covering a representative cross-section of the national local authority community. The Isle of Wight Council operates (dare I say) on an island; the Mid-Suffolk District Council covers a rural area; the West Berkshire District Council covers a provincial area; and the South Tyneside Metropolitan Borough Council covers an urban area.

There is no doubt that the issues raised by the appeals are of general importance and potentially involve huge sums of money. The Commissioners emphasise that, if they are wrong on their assertion that the economic question posed by article 4.5(2) requires a nationwide inquiry on what they refer to as an “activities” (i.e. off-street car-parking) basis, they may be faced with hundreds of local inquiries directed at answering the article 4.5(2) question on an “authority by authority” basis. From their perspective, the appeals are therefore of great practical importance. That may be so, but they are not the only players; and the local authorities’ response is that the Commissioners’ interpretation of the relevant provisions makes neither legal nor economic sense, and that its correct interpretation does indeed carry with it the potential for an endless multitude of inquiries and repeat inquiries. Whatever the correct answer to that issue, it is an odd feature of the legislation that the taxability or otherwise of any particular authority in respect of its off-street car-parking activities should turn on an inquiry as to the economic effects upon the private sector, which is not itself a party to the inquiry.

Article 4.5 has not been the subject of direct implementation by any national law in the United Kingdom, but whether it has direct effect was considered by the ECJ in Ufficio Distrettuale delle Imposte Dirette di Fiorenzuola d’Arde v. Comune di Carpaneto Piacentino (Case 129/88) and a conjoined case, reported at [1991] STC 205 (“Carpaneto No.1”). The reference arose out of the rectification of tax returns submitted by the local authority of Carpaneto Piacentino (“Carpaneto”) on the ground that they did not take account of fees received in respect of certain public activities, including concessions in respect of graves and vaults. The Customs and Excise Enforcement Service claimed that they should have done, since article 4 of the applicable Italian Presidential decree was said to have identified the relevant activities as being within the scope of VAT. Carpaneto claimed that, under article 4.5, it was not a taxable person and so was entitled not to charge VAT on the transactions. One of the questions referred to the ECJ was as to the direct effect or otherwise of article 4.5. The ECJ decided that it has direct effect

As to the correct interpretation of article 4.5, Mr Vajda emphasised that this depends not just on an interpretation of its text but on its consideration against the relevant principles of Community law. He said the most relevant principle is that of fiscal neutrality, one which “precludes, inter alia, economic operators carrying on the same activities from being treated differently as far as the levying of VAT is concerned” (paragraph 20 of the ECJ’s judgment in Gregg and another v .Commissioners of Customs and Excise [1990] STC 934).

The importance of that principle in VAT law is not in dispute although it is unclear to me how it can be relevant to the interpretation of article 4.5. The scheme of article 4.5 is that article 4.5(1) confers an exemption from taxability on public bodies in relation to their activities as such; but that article 4.5(2) cancels it in relation to any activity where its conferring would significantly distort competition. It does not, however, cancel it in cases in which an exemption would only distort competition insignificantly, a situation in which those carrying on the same activities are intended to be treated differently as regards the levying of VAT. If fiscal neutrality is the sacred watchword, there would be no scope for the article 4.5(1) exemption in such a case; and the only case in which any exemption might harmonise with the principle would or might be one in which the local authority is performing an activity which is not carried on in competition with others at all.

The ECJ touched on this in Carpaneto No. 1. One question referred was as to the scope of the expression “such activities” in article 4.5(2). The ECJ offered the following views (in setting them out - as in subsequent citations – I have modified them so as to reduce the verbiage resulting from the lack of sub-numbering in article 4.5):

Paragraphs 22 and 23 reflect the manner in which the ECJ regarded the scheme of articles 4.5(1) and (2) as respecting the principle of fiscal neutrality. With respect, and for reasons given, it appears to me that the ECJ may perhaps there have slightly overstated the case, since article 4.5(2) will not in every case achieve the respecting of that principle. In paragraph 38 of his opinion in Waterschap Zeeuws Vlaanderen v. Staatssecretaris van Financien [2005] STC 1298 (“Waterschap”), Advocate General Jacobs observed that “[i]t is inherent in the existence of exceptions to the VAT system that they will interfere to some extent with the application of the principles of neutrality and equality.” The ECJ endorsed that in paragraph 43 of its judgment. Putting it at its lowest, it is inherent in articles 4.5(1) and (2) that local authorities will or may carry out activities in competition with the private sector and – provided that it will not significantly distort competition – they will enjoy an exemption from VAT on their revenue which the private sector will not. That involves a deliberate departure from the principle of fiscal neutrality. It is no doubt the case that, in other circumstances, article 4.5(2) is intended to limit the potential for a departure from that principle; and the appeal raises the question of what those circumstances are. The answer cannot, however, lie in an appeal to the principle of fiscal neutrality.

I add that in the above citation I have highlighted the uses of “would”, “may” and “could” in paragraphs 22, 23 and 24. I will return to these paragraphs when considering Mr Vajda’s submissions as to the correct interpretation of “would lead to” in article 4.5(2).

The tribunal decided that the directive pointed to a “taxable person by taxable person” approach: the economic question raised by article 4.5(2) had to be answered in relation to each of the four local authorities separately. Taking each separately, the question was whether the conferring of an exemption would lead to significant distortions of competition in its local market. If not, it was entitled to exemption. Different answers could in theory be arrived at for each authority. The tribunal rejected the submission that the relevant inquiry was a nationwide “activities” one. They regarded that as at odds with the text of article 4.5, with the decisions of the ECJ and with the economic evidence. It is unclear to me how any “economic evidence” could assist in identifying the correct approach to article 4.5.

Mr Vajda’s submission was that whilst the tribunal’s approach might be right, it is not unequivocally clear that it is. He said there is a solid argument for the view that what the directive in fact intends is a general investigation requiring the determination of whether the provision nationwide of local authority off-street parking would distort competition significantly. If the result of that investigation is a yes, then all such local authority activities are taxable even if (for example) any particular authority might be able show that its own activities neither would nor could distort competition in its own market. He said that the logic of the local authorities’ submission was that each authority has a community law right not to be taxed unless it can be established, by a local inquiry peculiar to its own circumstances, that the conferring of a tax exemption upon it would lead to article 4.5(2) economic distortions. He said there is no support for that in the European jurisprudence.

Mr Vajda set the scene for his submissions with the ECJ’s decision in Finanzamt Gladbeck v. Linneweber (C-453/02) and a conjoined case, decided on 17 February 2005. Article 13B(f) of the directive provided for member states to exempt from VAT “betting, lotteries and other forms of gambling, subject to conditions and limitations laid down by each Member State.” The relevant German law exempted from VAT the turnover of licensed public casinos. The Linnewebers owned amusement arcades and provided gaming machines in restaurants and amusement arcades. They claimed exemption from VAT on the ground that the income from like machines in licensed casinos was exempt. Their claim was rejected since their income was not derived from a licensed public casino. The matter went to the ECJ, which recognised that, whilst the directive required member states to exempt gambling from VAT, the member state retained the power to lay down the conditions and limitations of such exemption. The court continued:

Mr Vajda derived two principles from Linneweber, which he said apply in the present cases. First, that the principle of fiscal neutrality precludes the different fiscal treatment of similar goods or services which are in competition with each other; and for that purpose it is not necessary to carry out an investigation into the effect of the differential VAT treatment once it is established that the relevant services are in such competition. Second, that proof of its breach in Linneweber did not depend on showing that there was in fact competition between the Linnewebers’ machines and those in any local licensed casino (if any). The ECJ instead approached the case on the basis that it was necessary to look at it on a global member state approach, being an approach which had established two tax regimes for like activities. It did not consider the position at the local level: it looked at it on an activities basis. Whilst accepting that Linneweber was not an article 4.5 case, Mr Vajda said that it supports an “activities” approach to article 4.5. He said the only reason why article 4.5 imposes a more complicated exercise is because of the inclusion of the word “significant” (“significant distortions of competition”) in article 4.5(2), an exercise involving an economic analysis. He emphasised that “distortion” of competition does not mean its elimination. There was distortion of competition in Linneweber: but it was not suggested that the differential tax treatment would cause the Linnewebers to go out of the market, it simply involved the removal of a level playing field.

With respect to Mr Vajda’s careful submission, I derive no assistance from Linneweber in the present context. It was not an article 4.5 case and appears to have turned upon the principle of fiscal neutrality. I have indicated my views on the relevance of that to the interpretation of article 4.5.

Turning to the text of article 4.5, Mr Vajda drew attention to the various plurals in article 4.5(1), which identifies public bodies in general and refers to general classes of activities or transactions in which they engage as such. It is not looking at markets. Article 4.5(2) qualifies article 4.5(1), but again looks at the position by reference to the public sector and the activities in which it engages. Article 4.5(3) is, he said, obviously linked to article 4.5(2), and again looks at the position on a class of activity basis. He compared this with the ECJ’s approach in Linneweber. Mr Vajda’s submission was, therefore, that article 4.5 should be interpreted as concerned with classes of activity, not with the taxability of individual public bodies. That approach made, he said, solid sense in the context of a taxing statute which should be certain in its interpretation and application and capable of easy administration. It involves looking at the United Kingdom as the relevant market. Mr Vajda accepted that the consequence of the “activities” approach is that in principle a local authority may find itself accountable for VAT on its off-street car parking revenue even if, as a matter of economic analysis, its own car-parking activities neither do, nor are likely to, cause any significant distortion of competition at the local level. But such a consequence is, he submitted, supported by the Advocate General’s opinion in Carpaneto No.1.

The first question the Advocate General considered was whether article 4.5 had direct effect. He said it would have such effect if it was unconditional and sufficiently precise. He regarded article 4.5(1) as sufficiently precise. But in paragraph 14 he discussed whether it was also unconditional having regard to the fact that the subsequent provisions of article 4.5 gave member states:

Having so said, the Advocate General opined in paragraph 15 that article 4.5(1) was “unconditional and precise in so far as a given activity can in no case come within the scope of the exceptions.” He continued:

I should refer also to certain further observations of the Advocate General upon which Mr Vajda relied, in the part of his opinion headed “Obligations of member states concerning the method of transposition of the directive.” He there said:

Mr Vajda drew on those various observations in support of the submission that articles 4.5(1) and (2) intend an “activities” rather than a “public body by public body” approach. He relied in particular on those in paragraphs 21, 22 and 92. In relation to paragraph 92, he said it had in mind the creation at national level of a list of activities “deemed” to create significant distortions of competition for article 4.5(2) purposes. He recognised that the Commissioners do not have the benefit of such a list, because the directive has not been implemented in the United Kingdom. But they had sought to make a like “evidence based” case before the tribunal. He said these observations by the Advocate General gave strong support to the “activities” approach and that the tribunal had not referred to them in its decision.

Mr Julian Ghosh QC, for the local authorities, submitted that Mr Vajda’s reliance upon the Advocate General’s opinion was misplaced. First, he said the opinion was confined to a consideration of article 4.5(1) in the context of mixed regimes, namely situations governed in part by public law and in part by private law. Whatever weight that point may have, it appears to be clear that in paragraphs 89 to 92 the Advocate General was dealing with the implementation at national level of article 4.5(2), where he appears to have been favouring an “activities” approach; and my reading of what he was saying in paragraphs 22 to 24 is that it was similarly generally based. Secondly, Mr Ghosh submitted that the Advocate General’s opinion had proceeded from his premise that articles 4.5(1) and (2) had no direct effect in relation to cases such as the present, whereas the ECJ held otherwise, paragraph 24 being the critical paragraph in their judgment. It followed that the opinion proceeded from a false start, it was not endorsed by the ECJ and so provided no help to the Commissioners. Mr Vajda accepted that the ECJ took a different view from the Advocate General on direct effect, but submitted that his views in the cited passages (which the ECJ did not criticise) still provided valuable guidance on the manner in which article 4.5 is intended to operate. In my judgment, the Advocate General’s observations in Carpaneto No.1 cannot simply be dismissed from the present discussion in the summary way that Mr Ghosh would have, and I consider they can be said to provide at least some support for the Commissioners’ approach, although whether that approach is the correct one is another matter. The tribunal did not refer to them.

Carpaneto No.1 was decided on 17 October 1989. The same Advocate General’s opinion in Carpaneto No.2 (Comune di Carpaneto Piacentino and Others v. Ufficio provinciale imposta sul valore aggiunto di Piacenza Case C-4/89) was delivered on 28 March 1990. The ECJ delivered its judgment on 15 May 1990. Carpaneto No.2 appears to have raised for consideration much the same questions as those raised and answered in Carpaneto No.1, although in paragraph 4 of his opinion the Advocate General described a new question that arose, which Mr Ghosh submitted was of no help to the questions raised by the appeal, being one confined to the scope of 4.5(1) (see paragraphs 3, 10 and 14). Mr Vajda placed reliance on what the Advocate General said in paragraph 15, which he said was in line with what he had said before. I do not propose to discuss Carpaneto No.2 further in this context. Assuming in Mr Vajda’s favour it is not irrelevant, it appears to me neither to add to nor subtract from the Commissioners’ argument. The tribunal did not refer to it.

The ECJ decided the Porto case on 14 December 2000. It was in part about the accountability of a local authority for VAT on the revenue from its off-street car-parking activities. The Porto local authority was assessed to VAT on its receipts. Its challenge to this led to the ECJ. One question referred was whether the Finance Minister of a member state might be authorised by a national law to define on a case by case basis the activities liable to bring about significant distortions of competition. This was the approach that had been adopted in Portugal, the inference from the report being that car-parking had been an activity so defined. Article 2 of the Portuguese VAT Code had provided that “… the Minister for Financial Affairs and the Plan shall define, case by case, those activities which are liable to give rise to distortions of competition and those which are exercised on a negligible scale.” After referring to paragraph 13 of the ECJ’s judgment in Carpaneto No.2 (which simply referred back to paragraph 24 of its judgment in Carpaneto No.1), the Advocate General said:

Mr Ghosh submitted that those paragraphs, in particular paragraphs 50 and 58, show that the Advocate General’s view was that the national implementation of article 4.5 required a focus not on any “activities” approach, but on a case by case approach by reference to particular public bodies. Mr Vajda recognised that paragraph 58 lent some support to that view, but he said that a consideration of the Community jurisprudence as a whole (essentially a reference to the Carpaneto cases) showed that it was not the correct one. I do not myself read what the Advocate General was saying as providing unequivocal support for Mr Ghosh’s interpretation of the operation of article 4.5. In particular, in paragraph 50 the favoured approach appears to have been to the effect that it was open to the member state to identify particular activities, perhaps also by reference to quantitative limits, as being liable (if carried out by tax exempt public bodies) to lead to significant distortions of competition, subject always to the proviso that any such identification must be open to judicial review. The ECJ judgment, at paragraph 32, was to the same effect:

I consider that Mr Vajda can therefore derive some assistance from Porto for the “activities” approach, an approach which, once adopted, can then be applied to individual public bodies. The ECJ has made it clear, however, that any such implementation must be open to review by the national courts, a point which perhaps leaves open the nature of the inquiry on such a challenge: is it local or national? Mr Vajda’s answer is that it is the latter, Carpaneto No.1 supporting the view that a particular authority cannot complain that the national implementation had not taken account of special local situations.

The tribunal considered the Porto case but declined to accept that it endorsed the “activities” approach, saying in paragraph 30 that neither the Advocate General nor the ECJ had been considering a wider market than the actual markets in which Porto had been providing its car-parking services. They held that the “activities” approach had not been in issue. Mr Vajda said that, even if the tribunal were right about that, it did not go to the lengths of saying that Porto was authority against the activities approach, and nor was there any ECJ authority to the effect that it is wrong. He also cited the reference to the ECJ by the German Bundesfinanzhof in Gotz (Preliminary reference C-408/6), which – albeit in a field remote from car-parking – raised the question of the width of the market that had to be looked at in assessing “significant” distortions of competition within the meaning of article 4.5(2).

It is convenient to take these issues together. The phrase in article 4.5(2) is “would lead to significant distortions of competition”. I have cited paragraphs 20 to 24 of the ECJ’s judgment in Carpaneto No.1, where I highlighted their uses of “may”, “would”, “therefore” and “could”. Mr Vajda said the ECJ’s use of the first “may” (in paragraph 22) was deliberate and was intended to encompass the consideration not just of actual competition but also of potential competition that might arise in the future. Paragraph 23, in using the word “would”, was simply referring to the language of article 4.5(2). The key to the ECJ’s decision was in paragraph 24, which (a) repeats the “may” (so re-embracing potential as well as actual competition) and (b) interprets “would” as meaning “could”. That is a potentially critical interpretation, since the evidential difference between “would” and “could” is obvious. When England fields its cricketers against the Australians in 2009, there are those who will take the view that they could win the Ashes. It is likely, however, that fewer people will take the view that they would do so. “Could”, in the present context, is primarily the language of possibility, whereas “would” is the language of (at least) probability. It is unfortunate that the ECJ did not find it necessary to explain their chosen interpretation of article 4.5(2). The “therefore” in paragraph 23 suggests that they regarded paragraph 24 as summarising what had gone before whereas it was not: the “could” appears to have been a leap into new territory. Mr Vajda submitted that this decision shows that “could” is the true sense of “would”; and that the “may” includes actual and potential competition from the private sector.

The “could” approach is said to have been endorsed by the Advocate General in Carpaneto No.2. He referred to paragraph 24 of the ECJ’s judgment in Carpaneto No.1, and in paragraph 22 re-described the concept as follows:

If I may respectfully say so, that appears to reflect two inconsistent thought processes. The Advocate General opines first that there is an obligation to tax if exemption would lead to significant distortions of competition, a view which does not appear to recognise the ECJ’s preferred “could”. He then opines that there is a duty to exempt public bodies from taxation even if the “likely” consequence of such exemption is distortions of competition, provided that such distortions are not “significant”. The Advocate General was there simply looking at the other side of the same coin, but now appears to consider that the inquiry reposes in the field of what is “likely”, which is again in line with a “would” concept. But that interpretation is obscured by the fact that he then appears to equate “is likely to” with “possible”, whereas the latter is closer to a “could” concept. I therefore find myself uncertain as to what principle he was intending to identify. When the matter went to the ECJ, the court confined itself in to referring to its holding in Carpaneto No.1 in paragraph 24.

The tribunal grappled in paragraph 44 with the sense of “would lead to significant distortions of competition ….” They recorded that the local authorities’ case was that, in discharging the evidential burden, the Commissioners had to prove their factual case to a high degree of probability. The Commissioners contended that it was enough to engage article 4.5(2) if there was a possible distortion of future competition. The Commissioners relied on paragraph 22 of the opinion of the Advocate General in Carpaneto No. 2. Having cited that paragraph (adding a mistaken “not” before the “possible”), the tribunal concluded that “That passage and particularly the Advocate General’s statement that one can only exclude a ‘possible’ distortion of competition if that distortion is not significant, shows that a possible distortion of competition in the future is sufficient.” For reasons given, I am not sure I would myself have been content to regard the Advocate General’s opinion as providing reliable guidance on this point, although it could be said that the tribunal’s approach was justified by the ECJ’s views in the Carpaneto cases. But in paragraph 45 the tribunal then promptly directed itself that it had “to decide on the evidence whether the disapplication of the VAT on local authority parking activities would lead to significant distortions of competition” (my italics). So what test was it applying?

As I will explain, the matter is said by Mr Vajda anyway to have been overtaken by the recent opinion given on 7 September 2006 (post-dating the tribunal’s decision) of Advocate General Kokott in Hutchison 3G UK Ltd and Others v. Commissioners of Customs & Excise (Case C-369/04) (“3G”). Mr Vajda said that opinion identifies the correct approach as to the meanings of both “would lead to” and “significant distortions of competition” in article 4.5(2). But before coming to it, I must revisit Waterschap..

In Waterschap a water authority had built a sewage plant in 1990 in its capacity as a public authority. It had incurred input tax in the process, in respect of which some four years later it sought a deduction when it engaged in a taxable sale and leaseback transaction. The claim resulted in a reference to the ECJ, which rejected the argument. I need not take time in discussing its essence, which appears to have been optimistic. Waterschap’s importance in the present context turns on the expression of view by the Advocate General in paragraph 41 of his opinion that:

The ECJ did not comment on those propositions in their judgment. Mr Vajda submitted that they must not be interpreted as expounding a general principle as to the true sense and purpose of article 4.5, but as being confined to the particular facts of the Waterschap case. I am unable to accept that submission, which appears to me to be wrong, and fairly obviously so. It is apparent from his chosen language that the Advocate General was expounding his perception of how article 4.5(1) and (2) work and that he was expounding it as a general principle. He would not have said what he did in the way he did if this had not been his intention. It is no surprise that, in paragraphs 36 and 40, the tribunal adopted that guidance as identifying the principles that it should itself apply, its approach being that:

Given what the Advocate General had said in Waterschap, that approach appears to have been fully justified; and it is also no surprise that the like approach had been adopted in paragraph 6 of the expert report of Helen Jenkins, a witness called by the local authorities. What, however, Mr Vajda is entitled to do, as in effect he did, is to submit that the statement of principle outlined by the Advocate General was wrong and that, in consequence the tribunal was misdirected by it into applying a like wrong approach. And Mr Vajda’s submission was indeed that, contrary to what the Advocate General had said, it is unnecessary in the course of an article 4.5(2) inquiry to look for exceptional circumstances. He said that all that has to be looked at are the normal or natural economic consequences that the conferring of any tax exemption may have on the private sector; and, if they are “significant”, article 4.5(2) is engaged. If he is right on that, the tribunal applied the wrong test.

Mr Vajda said that his suggested approach is expressly supported by the recent opinion of the Advocate General in 3G. One of the questions raised was:

In answering that question, the Advocate General referred first to the ECJ’s “could” decision in Carpaneto No.1. The report of her opinion records, in paragraph 127, that she had also considered Waterschap, although she referred to it only in support of the proposition that “capacity as a taxable person or non-taxable person has to be determined at the time of the transaction.” Her opinion does not reflect any consideration of the statement of principle the Advocate General had uttered in paragraph 41. That is unfortunate, because after reciting various other considerations, she concluded that the answer to the fourth question required an approach apparently – or at least arguably - different from that which he had expressed, namely that:

Relating that to the present case, that is an opinion that the words “…treatment as non-taxable persons would lead to significant distortions of competition” in article 4.5(2) mean “…treatment as non-taxable persons would result in a real risk of a materially adverse effect on the competitive position of present or potential suppliers of off-street car parking facilities.” The time in relation to which any such economic inquiry has to be held is said to be from 2000 onwards, the local authorities’ claims being continuing claims. The inquiry requires the consideration not just of actual competitors but of potential competitors. “Would lead to” means “would result in a real risk that”, which is said to be essentially in line with the ECJ’s simpler “could” but in fact appears to be out of line with the Advocate General’s own statement in paragraph 135 that the relevant verb is “would”. “Significant” is equated to “material adverse effect”. That approach is said by Mr Vajda to be inconsistent with the Waterschap approach, as interpreted and applied by the tribunal, and he may be right. Mr Vajda’s submission was that the 3G approach is the one that should be regarded as governing the day, whereas that is not the approach the tribunal adopted. That of course involves no criticism of the tribunal, since 3G post-dated its decision. Mr Vajda pointed to various paragraphs of the tribunal’s decision which he said reflected that they were applying a “would” rather than either a “could” or a 3G approach.

Having outlined his criticism of the tribunal’s approach to the legal bases on which it answered the questions before it, Mr Vajda challenged the tribunal’s overall conclusions on the basis that he was wrong in those submissions. If, however, he was correct in any of them, it may follow that there will be little alternative but to remit the cases for a re-hearing.

I received convincing arguments from Mr Ghosh in support of the view that the only intelligible interpretation of articles 4.5(1) and (2) is that they reflect a “taxable person by taxable person” approach. Article 4.5 is, he said, looking at particular public bodies, with the relevant question being whether or not that particular person’s exemption from VAT would affect its competitors. He submitted that the Commissioners’ nationwide “activities” approach is not only inconsistent with any sensible interpretation of the language of the article; it invites the undertaking of an unintelligible and unworkable economic inquiry in relation to a global market which, in the real world, does not exist: what conceivable effect can the exemption of the Isle of Wight Council from VAT on its off-street car-parking revenue have on supply and demand in Manchester? And, given that the directive is addressed to all member states, why pick the United Kingdom as the relevant market? He said that the Carpaneto and Porto cases provide no support for the contrary view.

As regards the “would/could” point Mr Ghosh’s approach was perhaps a little fluid. At one point he submitted that the substance of the matter is that “would” in article 4.5(2) means “would”, a concept embracing that which is foreseeable with certainty, and he drew on paragraph 135 of the Advocate General’s opinion in 3G. The difficulty with that is that, by the time the Advocate General arrived at her conclusion in paragraph 139, she expressed it in terms of “real risk” which appears to be involve a wider inquiry – assuming, that is, that the Advocate General can there fairly be interpreted as intending to equate “would” with meaning a “real risk that”. Mr Ghosh accepted that “would” does not mean inevitable but also asserted that it does not mean “could” – even though that was the verb apparently favoured by the ECJ in the Carpaneto cases. With an admirable sense of compromise, Mr Ghosh suggested that its true sense lay somewhere between “would” and “could”. Ultimately he favoured the view that it embraces a present, realistic prospect of distortion – a definition apparently closer to “would” than “could”. To the question how he would define the concept of “would” in synonymous language, Mr Ghosh had no immediate response. He was, however, content to adopt the view that the relevant question was whether tax exemption would on the balance of probability lead to a significant distortion of competition. That appeared to be a reversion to the proposition that article 4.5(2) means what it appears to say. But in Community law things are not always so simple, and the ECJ can be said to have favoured a different interpretation.

As regards the “significant distortions of competition” in article 4.5(2), Mr Ghosh supported the correctness of the Advocate General’s general statements in paragraph 41 of his opinion in Waterschap and submitted, wrongly in my view, that they had been endorsed by the ECJ. But the difficulty with those statements, as the discussion during argument revealed, is to identify what they mean. What, in particular, did the Advocate General mean by the “normal consequences”? As regards the meaning of “significant distortions”, I also understood Mr Ghosh to be content to adopt the expression of opinion by the Advocate General in 3G that it means “material adverse effect”. But is there in this respect a material difference between the approaches respectively favoured by the Advocates General in Waterschap and 3G?

I intend no discourtesy to Mr Ghosh in not explaining his careful submissions as fully as they deserved, or as fully as I have referred to Mr Vajda’s. I hope it will suffice if I say that the outcome of counsel’s arguments was that I was satisfied that I cannot safely embark upon an assessment of the correctness or otherwise of the tribunal’s decision until I first have the benefit of clear answers from the ECJ on the three legal issues that Mr Vajda has identified, namely (i) the “activities” approach as against the “taxable person by taxable person” approach; (ii) the true sense of the words “would lead to” in article 4.5(2); and (iii) the true sense of the “significant distortions of competition” in the same subparagraph.

As regards the first question, Mr Vajda did not persuade me that the Community jurisprudence positively supports his submission as to the “activities” approach, but I accept that it at least enables him to mount the argument. My instinct is to prefer the tribunal’s approach, which appears to me to reflect the more natural interpretation of article 4.5. On the other hand, if that is the right approach, it produces the perhaps surprising result that (for example) each of the 468 United Kingdom local authorities is entitled to say to the Commissioners that it is only accountable for VAT on its car-parking activities if the Commissioners can first prove a factual case for that by reference to the impact of a tax exemption of that authority on its local market; and that if any such local inquiry should go against the authority, it would be open to it to say in the following year that it had changed its charging policy and that the Commissioners must now look afresh at its local circumstances before it could hope to recover any more VAT from it. If that is the true effect of the direct effect of article 4.5, so be it. But the consequences of such an interpretation appear to me to be sufficiently far-reaching to justify a reference to the ECJ for clarification as to exactly how in practice article 4.5 is supposed to work in the context of the type of difference that has arisen between the Commissioners and the four respondent authorities. As regards the two other questions, I also take the view that clear and unequivocal guidance upon them is required before the correctness of the tribunal’s decisions under appeal can be considered on a measured basis. I find the Community guidance on them thus far to be insufficiently clear. Those considerations satisfy me that I should refer these questions to the ECJ (see Regina v. International Stock Exchange of the United Kingdom and the Republic of Ireland Ltd., Ex parte Else (1982) Ltd. and Another [1993] QB 535, at 545, per Sir Thomas Bingham MR).

May I invite counsel to endeavour to agree the forms of the questions to be referred.