Anglian Water Services Ltd v Revenue and Customs (VAT - ZERO-RATING : Sewerage services and water) [2017] UKFTT 386 (TC)

FTT-Tax

Anglian Water Services Ltd v Revenue and Customs (VAT - ZERO-RATING : Sewerage services and water)

[2017] UKFTT 386 (TC) · 2017-03-24

[1]The appellant (‘AWSL’) appealed against a refusal by HMRC to repay its voluntary disclosure of overpaid output tax of over £12 million charged on ‘infrastructure charges’ which were treated as standard rated by the appellant. HMRC accept that the infrastructure charges should not have been subject to VAT and accept that VAT was overpaid by AWSL so that is not in dispute.

[2]The exact amount the subject of the appeal has not been agreed: nevertheless, the parties expect to be able to reach agreement on the sum concerned if the appellant succeeds in principle in its claim, so this Tribunal is also not called upon at present to decide quantum.

[3]The claim relates solely to VAT charged in the period 1 April 1990 to 4 December 1996 because the infrastructure charges were first introduced on 1 April 1990 (when water undertakings were privatised) and were treated as zero rated after 4 December 1996. Nevertheless, time limits are also not in issue: this is a so-called Fleming claim made on 30 March 2009: HMRC do not suggest the claim was made out of time.

[4]What is in dispute is whether HMRC is entitled to refuse to repay the overpaid VAT on the basis of the unjust enrichment defence contained in s 80(3) Value Added Tax Act 1994 (‘VATA’). That was the sole subject of the hearing. The law The law on unjust enrichment

[5]S 80(3) Value Added Tax Act (‘VATA’) provides: It shall be a defence, in relation to a claim under this section by virtue of subsection (1) or (1A) above that the crediting of an amount would unjustly enrich the claimant.

[6]To a large extent the parties were agreed how this section should be interpreted. It should be interpreted in accordance with the case law of the CJEU. Both parties referred to Lady & Kid

[2012] STC 854

at §§18-21. In that case the CJEU said that unjust enrichment of the taxpayer was the exception to the right of the taxpayer to repayment of overpaid taxes (§18) and that that exception was justified because otherwise the taxpayer would be paid twice over, once by its customer and once by HMRC (§19). The CJEU went on to say that the exception had to be interpreted narrowly (§20) and that even where the VAT was passed on to the customer, there might be no unjust enrichment in repaying the taxpayer because the taxpayer may have suffered as a result of ‘a fall in the volume of his sales.’(§21).

[7]I accept that that the reference to avoiding a narrow interpretation in §20 was explained in §21 as meaning that the court had to consider the taxpayer’s loss in the round and not just the blinkered view of whether the VAT charge was passed on.

[8]Much the same was said in the earlier case of Weber’s Wine World Handels-GmbH C-147/01 at §95-102. Whether VAT has been passed on is a question of fact to be determined by the court, and even if it is shown that the VAT charge was wholly passed on, the taxable person may have suffered from a fall in volume of sales (§99). In §100 the CJEU said that the existence and extent of unjust enrichment could only be established following an economic analysis. Certainly both parties in this appeal relied on expert economic evidence.

[9]I accept, as Mr Mantle said, and Mr Rivett did not suggest otherwise, that there are no presumptions or assumptions in favour of either party in resolving the issue in this appeal other than that the burden of proof lies on HMRC. Burden of proof

[10]The parties were agreed that the burden lay on HMRC to prove that the VAT was passed on. And to the extent that HMRC could not prove that, the appellant would win the appeal. I agree; there is binding authority on the Tribunal to that effect such as Baines & Ernst

[2006] EWCA Civ 1040

“[12]....the burden of proof lies on the Member State, and no presumptions are to be applied, including any assumption that because the tax has been included in the price, it has been borne by the customer.”

Per Lloyd LJ

[11]Mr Rivett relied on Le Fils de Jules Bianco

[1989] 3 CMLR 36

for the proposition HMRC would have the burden of proof in a case where taxpayer was regulated as much as in any other case. HMRC did not dispute this and I accept that that is right.

[12]The appellant relied on Baines & Ernst for the proposition that unless HMRC could prove a particular amount of VAT was passed on, the appellant was entitled to a full repayment of the VAT overpaid. In other words, HMRC was liable to repay the full amount overpaid save to the extent that they could prove that a quantified part of that amount had been passed on. It seems to me that the appellant must be right on this: not only is that clear from Baines & Ernst , it is a matter of logic. If HMRC have the burden of proving that VAT was passed on, they can only win to the extent that they can prove it was not passed on, as stated in Baines & Ernst: [13] ....[HMRC] has to prove that the burden of the tax was passed on to customers in whole or in part and, if the latter, to what extent.

[13]So, as I understand it, the gist of the appellant’s case is that the incidence of VAT would have affected the level of infrastructure charges, even though I do not understand them to suggest it would necessarily have been a £ for £ reduction, but because there is no evidence to indicate the extent by which VAT would have affected infrastructure charges, then the entirety of the appellant’s claim must succeed. I agree with that proposition in principle, but it relies on HMRC being unable to prove that VAT had no effect at all on infrastructure charges, and that I deal with below.

[14]HMRC considered that Baines & Ernst was not directly relevant because in that case, unlike this case, the overpayment of VAT carried with it the necessary implication that the taxpayer had recovered output tax to which it was not entitled: HMRC suggested in that case that had the taxpayer known its supplies were exempt, it would have charged the net fee plus a figure equal to what was its irrecoverable VAT, whatever that was divided amongst all its customers. The Court of Appeal did not consider that in putting that case HMRC had even sought to establish a figure that the taxpayer would have charged had the true VAT liability been known: and certainly that it had no evidence to support it. But I consider that although the facts were different, the underlying point is the same: HMRC must repay the overpaid VAT save to the extent that they can prove the repayment unjustly enriches the appellant and that would require them to prove in this case the amount the infrastructure charges would have been had they been known not to be subject to VAT.

[15]I note in passing that it appeared to be HMRC’s position, relying on a passage in Marks & Spencer

[1999] STC 205

at 239 that in a case where economic loss was in issue, the burden of proof, or at least an evidential burden of proof, would pass to the taxpayer. I consider this no further because, as I explain below at §24, this is not a case where I have to consider potential economic loss.

[16]I also note that the standard of proof is the balance of probabilities. The issue for the Tribunal

[17]The appellant is entitled to recover from HMRC the VAT which it overpaid on infrastructure charges unless HMRC can show that the repayment would lead to AWSL’s unjust enrichment.

[18]AWSL was a water and sewerage undertaker. In 1989 it was, as part of the privatisation of the provision of water and sewerage services, appointed by the Water Act 1989 as the Water Undertaker for the Anglian region.

[19]AWSL had the power to and did levy charges on its customers. However, its permitted charging levels were regulated first by the Secretary of State for the Environment and then by Ofwat.

[20]It charged for the supply of water and the use of mains sewers. Those charges are not in issue in this appeal. AWSL also levied various charges for the first-time connection of premises to a water supply and/or sewer. These included requisition charges and connection charges: they are not in issue in this appeal.

[21]This appeal concerns the third type of charge made for first-time connection, the infrastructure charge. Under s 79(2) of the Water Act 1989 water undertakings were given the right to levy infrastructure charges for the first-time connection to the water supply of premises for ‘domestic purposes’ and a sewerage infrastructure charge had to be paid for the first-time connection to a mains sewer of premises for ‘domestic purposes’. ‘Domestic purposes’ included residential properties, but also included commercial and industrial premises where the purpose of the connection to the mains was for domestic use, such as the purpose of providing cooking, washing and toilet facilities to workers. AWSL’s input tax

[22]AWSL fully recovered its input tax and HMRC accept that that was correct. Therefore, so far as unjust enrichment is concerned, this Tribunal is only concerned with the output tax charged by AWSL to its customers and accounted for to HMRC. AWSL had no blocked input tax, and should have had no blocked input tax, attributable to the supplies reflected in the infrastructure charges. AWSL’s input tax is not relevant to this appeal. AWSL’s output tax

[23]AWSL charged VAT on its infrastructure charges. But as I have said above at §§6-8, it is well understood that the mere fact that a charge is ‘plus VAT’ does not prove that the VAT has been recovered from the taxpayer’s customers. The taxpayer may have reduced its prices because the prices were (wrongly) understood to be subject to VAT, or, if prices were not reduced, the taxpayer may have suffered economic loss through a reduction in sales. In other words, some customers may have been put off buying the product because of the gross VAT inclusive price, but would have made the purchase had the price been the equivalent of the net of VAT price.

[24]However, AWSL accept that it did not suffer economic loss from the addition of VAT to infrastructure charges. AWSL had a regional monopoly. Its customers had no other possible supplier of connection to mains water and sewers. And while a customer, faced with no alternative supplier, might choose not to make the purchase at all, in this case the level of charges for connection compared to the overall costs of new buildings were so minimal that in practice it was unlikely that the addition of VAT would be result in a falling off in demand. In other words, prices would have had to be considerably higher, well in excess of the gross price in issue, to affect demand for connection. The experts agreed on this and Mr Rivett did not suggest the contrary.

[25]So the parties were agreed that in this appeal the question was whether (bearing in mind HMRC have the burden of proof) it can be proved that AWSL’s net infrastructure charges were not reduced because of the incidence of VAT. Taking into account that AWSL always charged infrastructure charges at the maximum level permitted by the regulator (see §§47-50), both parties were therefore agreed that the only question of fact for the Tribunal was whether that maximum infrastructure charge set by the regulator was less than it would have been but for the imposition of standard rated VAT. In other words, did the regulator reduce the amount of the maximum infrastructure charge because of the incidence of VAT?

[26]I note that HMRC expressly disclaimed any argument that, even if the regulator did reduce the amount of the maximum charge because of the imposition of VAT, nevertheless the VAT was fully passed on because the regulator would have set other charges higher in compensation to the water services suppliers: on the contrary, HMRC accept that one of the regulator’s objectives was to avoid cross-subsidisation so the regulator would have been unlikely to act in this manner. In other words, the question for the Tribunal was simply whether the regulator reduced the amount of the maximum infrastructure charges AWSL was able to charge because of the incidence of VAT. More accurately, because HMRC have the burden of proof, the question is whether HMRC can prove on the balance of probability that the regulator did not reduce the amount of the maximum charge for infrastructure charges because of the incidence of VAT.

[27]HMRC’s case is that it is unlikely that the regulator set the maximum infrastructure charges at a lower level than he would otherwise have done but for the incidence of VAT; the appellant’s case is that the regulator is likely to have set the maximum infrastructure charges at a lower level than otherwise due to the incidence of VAT. HMRC did not have a ‘fall-back’ or secondary case that, even if they cannot show it was unlikely that the maximum infrastructure charge levels were entirely unaffected by the incidence of VAT, nevertheless it was unlikely that the charge was affected by more than a certain amount. Both parties treated the case as an all or nothing scenario.

[28]That brings me to the relevant statutory provisions on regulation. The law on regulation

[29]The Water Act 1990 set out the duties of the regulator. So far as relevant it provided: 7 General duties with respect to water supply and sewerage services (1).... (2) The Secretary of State or, as the case may be, the Director shall exercise and perform the powers and duties mentioned in subsection (1) above in the manner that he considers is best calculated – (a).... (b)...to secure that companies holding appointments ....as water undertakers or sewerage undertakers are able (in particular, by securing reasonable returns on their capital) to finance the proper carrying out of the functions of such undertakers. (3) Subject to subsection (2) above, the Secretary of State or, as the case may be, the Director shall exercise and perform the powers and duties mentioned in subsection (1) above in the manner that he considers best calculated – (a) to ensure that the interests of every person who is a customer or potential customer of a company which has been or may be appointed....to be a water undertaker or sewerage undertaker are protected as respect the fixing and recovery by that company of – (i) charges in respect of any services provided in the course of the carrying out of the functions of a water undertaker or sewerage undertaker; and (ii)... and, in particular, that the interests of customers and potential customers in rural areas are so protected and that no undue preference is shown, and that there is no undue discrimination, in the fixing of those charges.... ..... (d) to promote economy and efficiency on the part of any such company in the carrying out of the functions of a water undertaker or sewerage undertaker; and .....

[30]It was accepted, and I find, that the above legislation was in force at the time relevant to this appeal. The Water Industry Act 1991, as amended by the Water Act 2003 with effect from 1 April 2005, recast the duties so that the obligation on the regulator to consider the interests of customers (s 7(3) above) was no longer stated to be ‘without prejudice’ to consideration of the ability of the water undertakers to finance their undertakings (s 7(2) above): it was to be given equal consideration from 1 April 2005. This Tribunal is only concerned with the position 1990-1996.

[31]Having dealt with the applicable law, I move on to make findings of fact. The facts The evidence

[32]To a large extent the factual evidence was not in dispute. The expert opinion evidence, on the other hand, was very much in dispute, and I deal with that below.

[33]The appellant called two witnesses of fact. The evidence of Mr Iain Amis was accepted by HMRC and he was not called to give evidence. He joined AWSL in 1996 but did not deal with infrastructure charges until 2003 and since 2012 has been head of developer services responsible for collecting infrastructure charges. I accept his evidence, although I do not summarise it as it was not really relevant to the issue for this Tribunal.

[34]Mr Richard Allen was an accountant specialising in tax and employed by AWSL since 1997. His evidence was largely accepted by HMRC but he was cross examined on elements of it. I accept his evidence and deal with it below at §66-69. Findings of Fact

[35]From the witness and documentary evidence I find as follows: The purpose of infrastructure charges

[36]Connection charges were intended to be charges for work necessitated by actually making the connection from the new premises to the mains; requisition charges were intended to reflect the cost of works needed to remote infrastructure necessitated by new connections; infrastructure charges were intended to cover the cost of increased capacity in the system as a whole each time new premises were joined to the mains. There was the possibility of requisition and infrastructure charges overlapping but that was not relevant to the appeal.

[37]Both experts were agreed that the rationale for specific charges for infrastructure charged for new connections was to avoid ‘cross-subsidisation’ between different groups of customers. In particular, infrastructure charges were intended to reverse the historic position whereby the cost of increasing capacity fell on all users, and to ensure instead that, under the privatised system, the cost of increased capacity within the system fell on those newly connecting to the system who were thus causing the need for increased capacity.

[38]The experts were also agreed that the substantial reduction in infrastructure charges which took effect on 1 July 1995 (following the 1994 Ofwat report referred to at §§58-61 below and set out at §41-43) was because of concern that infrastructure charges were set at a level which meant new users were paying for all increased capacity need, including that generated by existing users. Regulation of AWSL

[39]As I have said, the Water Act 1989 (and then the Water Industry Act 1991) gave AWSL a regional monopoly over water supplies and sewerage in the Anglian Region and at the same time provided for its regulation. In particular, infrastructure charges were subject to price limits set by the regulator.

[40]The first regulator was the Secretary of State for the Environment (‘SOSE’). The SOSE at the time was the Right Hon. Nicholas Ridley MP and he set the maximum rates for infrastructure charges for all water undertakings. Responsibility for regulation was then passed on to the Director General of Water Services (‘Ofwat’), Sir Ian Byatt, with effect from 1 August 1989, which was before privatisation took effect. The maximum rates set by the SOSE came into effect on 1 April 1990 and they increased by the RPI each year. The experts inferred that the SOSE set the infrastructure charge rates on or before 1 August 1989 because that is when regulation was handed over to Ofwat: neither party questioned this inference. I accept that it was more likely than not that the infrastructure charges in force from 1 April 1990 were set before 1 August 1989 and so find as a fact.

[41]The maximum rates of infrastructure charges for AWSL, set on or before 1 August 1989, which came into effect on 1 April 1990, were:(a) £479 for water;

(b) £597 for sewerage. These rates increased by the RPI on 1 April each year.

[42]For the first five years, each water undertaking had different maximum rates. Six water companies had what would have been infrastructure charge levels in excess of £1000 for water but which were capped at £1000. The lowest water charge was £111. For sewerage, the lowest was £240 and the highest £983.

[43]Infrastructure charges, as well as other charges, were then reviewed five years later in line with the legislation. The new maximum rates (referred to in §38) came into effect on 1 April 1995 for all water undertakings, including AWSL and were:(a) £200 for water;

(b) £200 for sewerage.

[44]They increased by the RPI on 1 April 1996. Infrastructure charge rates exclusive of VAT

[45]Both parties accepted that the maximum charge rates set by the regulators were exclusive of VAT. I was not pointed to any document where this was stated to be the case: nevertheless, it was accepted that all the water companies (including AWSL) and the regulator had proceeded on the basis that the water companies were entitled to charge the maximum charge rates plus VAT, and, moreover, that the £1,000 overall maximum referred to in §42 was £1000 plus VAT.

[46]Both parties accept that AWSL levied VAT on all its infrastructure charges in the period 1 April 1990 to 31 March 1994 when they were considered to be in all cases subject to VAT. However, after a review in 1993, HMRC decided that infrastructure charges should be zero rated if charged in respect of new connections to ‘qualifying’ buildings. This followed the Tribunal decision in Rannoch School Ltd

[1993] STC 389

. So from 1 April 1994 when this new view came into effect, as I understand it, it is accepted that AWSL zero rated infrastructure charges on new dwellings and other qualifying buildings. This fact is no doubt relevant to quantum. It has some further relevance which I discuss below at §§172-177. AWSL’s charge rates

[47]During the entire period at issue in this appeal, AWSL charged the maximum rate of infrastructure charges with two exceptions. Those exceptions were: a) AWSL did not apply the RPI increase on 1 April 1993 or 1 April 1994. b) AWSL was legally entitled to charge the infrastructure charges at the rate applying at the date of connection but as a matter of practice only charged the infrastructure charges at the rate applying at the date they were paid.

[48]Dealing with (a), as I have said, by 1 August 1989, the maximum infrastructure charge was set by the regulator with effect from 1 April 1990; each anniversary it increased by the RPI. But AWSL, like all the other water undertakers, and most likely because of pressure from Ofwat who it appears from the evidence had by this time reached the view that the SOSE had set infrastructure charge levels too high, did not apply the permitted RPI increase on 1 April 1993 nor on 1 April 1994.

[49]Despite not applying the RPI increase on these two occasions, both parties accept that AWSL always charged the de facto or practical maximum infrastructure charges bar point (b).

[50]Dealing with (b), the object of this policy of AWSL’s was to encourage early payment of infrastructure charges so that AWSL would have early access to the money. AWSL accept that to the extent that it was not charging the maximum infrastructure charge, the reason was nothing to do with the incidence of VAT. It does not affect the unjust enrichment analysis and the case must be considered on the basis that AWSL in practice charged the maximum level of infrastructure charges. The setting of the infrastructure charges in 1989

[51]House of Lords Debate: The provisions in the Water Act relating to infrastructure charges was debated in the House of Lords on 22 May 1989. Lord Caithness, in supporting the Bill, explained the Government’s intention that persons making new connections to the mains water and sewers should meet the cost of the resulting need for greater capacity in the system as a whole; and that by levying an upfront charge for connections, water undertakers would not be able to delay making new connections for lack of money to invest in the infrastructure.

[52]The prospectus: The prospectus (November 1989) for the public offering of the water companies contained a statement by the first Director of Ofwat, Sir Ian Byatt. He stated that he recognised the need for water undertakers to have a reasonable return on capital while at same time the regulator needed to protect the interests of consumers. He stated that he saw these duties as complimentary as it would not be in the interests of consumers if undertakers were unable to carry out their functions. The regulator went on to say that, as the water undertakers were in a monopoly position, he would fulfil these duties by using his regulatory powers to obtain the same balance as would otherwise be achieved by a competitive market.

[53]1991 consultation paper: Ofwat published a consultation paper in 1991 which set out its view of how the SOSE had set the first infrastructure charge rates. It explained that the infrastructure charges were intended to cover the cost to the water undertaker concerned of works to remote infrastructure to cater for new connections (ie new connections creating the need for larger capacity in the system generally). It was calculated by estimating the cost of works in the next 20 years to provide increased capacity discounted back to a present day figure.

[54]The SOSE also capped the infrastructure charge at £1,000 per connection (ie £1000 for a water infrastructure charge charge and £1000 for a sewerage infrastructure charge) and I have referred to this above at §42. To the extent that the water undertaking’s uncapped infrastructure charge exceeded £1000, the difference was recovered from general water and sewerage charges. The cap did not apply to AWSL (as its infrastructure charges were £479 and £597) so this cross-subsidization did not occur in respect of AWSL.

[55]House of Commons Debate: Infrastructure charges were also debated in the House of Commons on 17 December 1991. Mr Baldry, the Under-Secretary of State for the Environment at the time, also reflected what Lord Caithness had said which was that infrastructure charges to new customers had been intended to remedy the unsatisfactory nature of the situation before privatisation. In other words, the new system was meant to be simple (ie a fixed charge known in advance) and at the same time to ensure new customers paid for the full cost of new connection.

[56]NAO report: There was a National Audit Office report in 1992 examining the role of the SOSE in privatisation of the water undertakings. It gave the NAO’s view of the regulator’s duties which were primarily to ensure water undertakings were properly funded but secondarily to protect consumers from the monopoly position of water undertakings both as to quality provided and price charged. It discussed the setting of charges and had a section on ‘taxation’ in which only direct taxes were mentioned (in other words, privatised water undertakings would become subject to corporation tax and ACT). There was no consideration of VAT.

[57]Parker’s History: I was also referred to Parker’s The Official History of Privatisation, in particular Vol II Chap 8 Regulating the Water Industry. Parker, talking about charges generally, referred to the fact that calculating charges for a privatised water industry was not simply a matter of taking costs, projecting them forwards and building in efficiency gains, as there was a need for new investment to update the water systems and comply with environmental standards. This was likely to result in increased prices to consumers rather than a price reduction which they might have expected on privatisation. The setting of the 1994 infrastructure charges

[58]Ofwat’s 1994 report: Ofwat set revised infrastructure charge rates with effect from 1 April 1995 (see §§38 and 43). The new rates were explained in the 1994 Ofwat paper: Future charges for water and sewerage: the outcome of the periodic review. The report followed three years of consultation with the water undertakings and their customers (see §53 for a reference to one of Ofwat’s consulation papers).

[59]The summary of this report referred to the rationale for the 1989 level of infrastructure charges being to enable the water undertakings to ‘recover all the costs associated with servicing additional customers’ but that the Ofwat Director had decided that in future infrastructure charges should be restricted to the costs of ‘developing the local network’. This led, as I have said, to a significant reduction in charges for most water undertakers, including AWSL (§43).

[60]The section of the report relating to infrastructure charges was not particularly long. It elaborated on the summary: the concern was that by paying infrastructure charges at existing rates as well as the normal charges for water usage, new customers were subsidizing existing customers. So Ofwat opted for an across the board single charge designed to cover local distribution costs only. This provided simplicity and was intended to avoid cross-subsidization between new and existing customers.

[61]It was an extremely long report and made no mention of VAT at all: it did not even state whether the charges it set were inclusive or exclusive. Corporation tax, advance corporation tax, and income tax were mentioned but nothing about VAT.

[62]MMC reports: Water undertakings could challenge the rates set by the regulator and two did so in respect of the rates set by Ofwat in 1994 to take effect from 1 April 1995. Dr Rubin (HMRC’s expert) referred to both reports and included extracts from them with his expert report. It was his evidence and HMRC’s case that the submissions recorded in the reports as made by both Ofwat and the two companies concerned did not make any reference to VAT.

[63]Nevertheless, the full documents were not before the Tribunal and Mr Rivett’s position was that HMRC could not therefore prove the proposition on which they relied. However, while it is not a major point nor one on which this appeal turns, I accept below that Dr Rubin gave an independent view to the Tribunal and understood his duty to bring to the Tribunal relevant information, irrespective of the interests of the party calling him as a witness. If he had found anything in either of the MMC reports which referred to VAT, he would have mentioned it. Therefore, I conclude that there was nothing in the MMC reports about VAT. Evidence of other regulators considering VAT

[64]Ofgem report 2002: I was referred to an Ofgem ‘decision’ document from 2002 which mentioned VAT in setting maximum retail prices (MRP). It had asked for comments on whether the MRP should be VAT inclusive or exclusive and went on to decide that the MRP should be VAT exclusive so that changes in the rate of VAT were borne by the consumer. The premise underlying the decision seemed to be that the suppliers ought to have the ability to pass on costs which they could not control. Could AWSL’s customers recover the VAT on infrastructure charges?

[65]There was a dispute between the parties as to what extent AWSL’s customers would have been able to recover the VAT charged to them on the infrastructure charges as input tax. Both parties seemed to accept that the majority of customers by value would have been able to recover the VAT as they would have been businesses, such as developers, which were fully taxable. Both parties accepted that a significant minority by value of customers would not have been able to recover VAT, such as DIY builders and housing associations.

[66]Mr Allen analysed AWSL’s customer base for infrastructure charges in the year 2015/16. Everyone was prepared to operate on the assumption that the breakdown in the 1990s would have been similar and I accept that. His analysis shows 59% by value of customers were house builders, 13% paid one infrastructure charge only and the rest were ‘others’ (in other words the ‘other’ category were persons who paid more than one infrastructure charge in the year but were not classed by Mr Allen as being obviously developers).

[67]I find that the ‘one connection’ group was not limited to DIY builders: it included a high percentage of companies, many with terms like ‘builder’ in their name which suggested many were likely to be VAT registered and fully taxable. While there were some obvious charities and parish councils in the list, they were very much a small minority. So far as the ‘others’ were concerned, a substantial majority appeared to be businesses, so it is likely significant numbers were VAT registered and fully taxable.

[68]This does suggest that a very high proportion of the infrastructure charge payers were developers and other taxable businesses. However, as HMRC accepted, some developers in the 1990s were likely to have been unable to fully recover all their VAT as (as per Briararch

[1992] STC 732

) the 1990’s property recession meant some new builds were let rather than sold. While that situation was not really predictable in 1989, it would have been predictable even in 1989 that some developers/housing associations would have built with the intention of letting rather than selling, and therefore it was predictable that some developers would not have been able to recover VAT on infrastructure charges.

[69]On the evidence I have, it is not possible to conclude that a particular proportion of AWSL’s customers paying infrastructure charges were able to recover the VAT charged to them: but it is reasonable to conclude that it would have been obvious in 1989 and 1994 that a significant proportion were able to do so. History of HMRC’s ruling on VAT liability on water/sewerage supply charges

[70]In 1989, the water undertakers, acting via the Water Services Association (‘WSA’) were in discussion with HMRC on the liability of the new charges to VAT. This is apparent from various letters in evidence before the Tribunal, although no one suggested that the letters and notes discovered by HMRC were the complete picture. They were, after 30 years, all that HMRC and the appellant could find.

[71]I find that as at 29 August 1989, when there was a meeting between HMRC and the WSA relating to industrial water supplies, HMRC were of the view that infrastructure charges would follow the same liability as a supply of water or sewerage services.

[72]At the time of that meeting all supplies of water and sewerage services were zero rated, but HMRC clearly knew, and it would have been public knowledge, that certain supplies of water were to cease to be zero rated with effect from 1 July 1990. This was because the Value Added Tax Act 1983, as amended by the Finance Act 1989 (enacted 27 July 1989) provided that, for supplies of water on or after 1 July 1990, only those other than in connection with the carrying on of a business of a relevant industrial activity (defined as divisions 1-5 of SIC) would be zero rated; mains sewerage services, however, all remained zero rated whatever the classification of the customer.

[73]Therefore, HMRC’s view expressed in the notes of the meeting was that if the customer was in standard industrial classification (‘SIC’) division 1-5 the infrastructure charge would be standard rated but, otherwise, the charge would be zero rated. Sewerage infrastructure charges would always be zero rated.

[74]Nevertheless, within 5 days of this meeting, I find officers within HMRC had doubts whether the premise that infrastructure charge charges would have the same VAT treatment as water and sewerage charges was correct. On 4 September 1989, an HMRC officer wrote to the WSA to put them on notice that HMRC was in the process of reconsidering its view and in particular considering whether, instead, infrastructure charges should be treated for VAT purposes as if they were charges for civil engineering works (normally standard rated).

[75]There followed correspondence between HMRC and Price Waterhouse on behalf of the WSA. HMRC notified the WSA on 20 September 1989 that they had completed their review and that they did consider that certain supplies, including infrastructure charges, were supplies of civil engineering works. HMRC clarified their position in a letter of 16 October 1989, stating that all infrastructure charges to developers would be standard rated but supplies to consumers would be zero rated if for a new dwelling or for a building for relevant charitable or residential purpose but, the letter seemed to imply, only in respect of work adjacent to the new building.

[76]In January 1990, WSA requested a revision of the ruling on the basis of administrative simplicity so that zero rating would be limited to work wholly within the development site of a dwelling and not works which were ‘adjacent’. Then in April 1990, Price Waterhouse on behalf of the WSA wrote to HMRC entirely rejecting the civil engineering work analysis and requesting a ruling that all infrastructure charges be standard rated, again because of practical difficulties with implementing HMRC’s ruling.

[77]The correspondence must be incomplete: in July 1990 HMRC replied to the January letter agreeing that only works to a DIY builder within the development site would be zero rated. Nevertheless, HMRC must have accepted the position in the April 1990 letter, as (both parties were agreed) at the time infrastructure charges were introduced (1 April 1990) they were all treated as standard rated.

[78]In summary, HMRC moved from a position as on 29 August 1989 that all water and sewerage infrastructure charges were zero rated unless they related to water and were supplied to businesses within SIC 1-5 (which included developers) to a position in mid- 1990 that all water and sewerage infrastructure charges were standard rated.

[79]As I have mentioned at §46, HMRC changed their minds again in 1993. Certain infrastructure charges were zero rated from 1 April 1994 and all infrastructure charges were zero rated from 4 December 1996 (§3). Infrastructure charges are now treated as outside the scope. That is why HMRC accept that the AWSL overpaid VAT from 1 April 1990 until it ceased to charge VAT on them on 4 December 1996. The Expert Evidence Admissibility of expert evidence - independence

[80]Mr Rivett challenged the independence of the expert witness called by HMRC, Dr Rubin, who was an HMRC employee. My concern with this challenge was whether it was made at the proper time. The appellant had known for some time that Dr Rubin was to give evidence: a challenge to his independence could have been made earlier. It is clear from Armchair Passenger Transport Ltd v Helical Bar Plc

[2003] EWHC 367 (QB)

per Nelson J that it is best practice to challenge independence at an early stage in the litigation because a late but successful challenge deprives the other party of the chance to obtain expert evidence from different expert: i) It is always desirable that an expert should have no actual or apparent interest in the outcome of the proceedings; ii) The existence of such an interest, … does not automatically render the evidence of the proposed expert inadmissible. …. iv) The decision as to whether an expert should be permitted to give evidence in such circumstances is a matter of fact and degree. The test of apparent bias is not relevant to the question of whether or not an expert witness should be permitted to give evidence. v) the questions which have to be determined are whether(i) the person has relevant expertise and

(ii) he or she is aware of their primary duty to the Court if they give expert evidence, and willing and able, despite the interest or connection with the litigation or a party thereto, to carry out that duty. vi) ... vii) If the expert has an interest which is not sufficient to preclude him from giving evidence the interest may nevertheless affect the weight of his evidence.

[81]Mr Rivett appeared to accept this as it was his position that he was not asking for Dr Rubin’s evidence to be excluded on the grounds of lack of independence but simply that less weight be put on it. Indeed, Nelson J said something similar at (vii) in the above cited case.

[82]HMRC accepted that I ought to put less weight on Dr Rubin’s evidence if I considered he was not giving fully independent evidence, but did not accept that that Dr Rubin failed to give fully independent evidence.

[83]It is clear that ‘apparent bias’ is not relevant (see above citation at (iv)). Being an HMRC employee does not bar Dr Rubin from being an expert witness on behalf of HMRC. The question is whether he understood his duty to the Tribunal and actually gave an opinion uninfluenced by partisan considerations (see above citation at (v)). Duty to Tribunal

[84]As all parties recognised, the CPR, and in particular CPR 35, are not directly relevant in this Tribunal. Nevertheless, as a matter of natural justice I had to consider the extent to which, if any, Dr Rubin’s evidence was independent and that included consideration of whether his statement that his report complied with CPR 35 was correct and honestly given.

[85]So far as his duty to the Tribunal was concerned, Mr Rivett took Dr Rubin through the requirements of CPR35 in some detail and asked him whether he still maintained that his report complied. Dr Rubin clearly gave thought to the requirements and said his only concern with whether he had complied was that perhaps in retrospect he ought to have suggested calling Sir Ian Byatt (first Director of Ofwat) as a witness (who, unlike Lord Ridley, was still alive).

[86]I consider that the extent to which this lack of this recommendation made his report non-compliant with CPR 35, the same criticism could be levelled at Dr Koboldt. In any event, Dr Rubin’s willingness to accept that he might be in error indicated a very genuine intention to give unbiased evidence.

[87]Dr Rubin was criticised for listing in his report only the documents he considered relevant and not all the documents he had consulted in connection with drawing up his report. Yet CPR 35 PD 3.2(2) requires details only of material relied upon.

[88]He was also criticised for not including the full text of his instructions from HMRC, but CPR PD 3.2 (3) only requires the expert to set out the substance of his instructions, and I find that Dr Rubin did that.

[89]In conclusion, I find Dr Rubin did understand the requirements of CPR 35 and his statement that he complied with them was, in essentials, true. Report uninfluenced by partisan considerations?

[90]Mr Rivett pointed out, as Dr Rubin’s report made clear, that Dr Rubin was asked by HMRC to advise on whether or not this case should be defended, long before the litigation commenced.

[91]However, the point in time at which he was asked to give expert evidence does not seem relevant to me save if the circumstances were such that it meant he was not independent. I accept that Dr Rubin, as he says, was not at any point tasked with defending the appeal. From the first, he was asked for his expert opinion on whether or not repayment would result in the unjust enrichment of the appellant. He was not asked to provide an opinion to support HMRC’s position: indeed, his instructions indicate that if Dr Rubin had advised that the counterfactual price was likely to have been higher than the actual infrastructure charges, then HMRC was unlikely to defend the appeal.

[92]Mr Rivett said that various robust statements by Dr Rubin in his report, an example being his statement that the regulator’s decision on infrastructure charges in 1994/5 was well-documented, indicated a lack of balance and therefore a lack of independence. I do not agree with the criticism: the regulator did publish a long report about his 1994 decision on charges, including with respect to infrastructure charges, so the decisions he made at that time could justifiably be described as ‘well-documented’.

[93]In conclusion, I accepted Dr Rubin as an independent expert and did not put any less weight on his evidence because of the appellant’s concerns over his independence. There was no challenge to Dr Koboldt’s independence and I accept he was an independent expert witness. Subject of expert evidence

[94]Mr Rivett was also concerned that Dr Rubin’s report strayed beyond the boundaries of subjects on which expert evidence was admissible. He referred me to two English High Court decisions ( Clarke (executor of the will of Francis Bacon) v Marlborough Fine Art [2003] CP Rep 30 (Patten J) and Barings plc v Coopers & Lybrand [2001] Lloyds Rep PN 370 at 386 (Evans-Lombe J)) which had both approved and relied on the statement of the Australian Chief

Justice King in the case of R v Bonython [1984] SASR 45, as follows:

“Before admitting the opinion of a witness into evidence as expert testimony, the Judge must consider and decide two questions. The first is whether the subject matter of the opinion falls within the class of subjects upon which expert testimony is permissible. This first question may be divided into two parts: (a) whether the subject matter of the opinion is such that a person without instruction or experience in the area of knowledge or human experience would be able to form a sound judgement on the matter without the assistance of witnesses possessing special knowledge or experience in the area and (b) whether the subject matter of the opinion forms part of a body of knowledge or experience which is sufficiently organised or recognised to be accepted as a reliable body of knowledge or experience, a special acquaintance with which of the witness would render his opinion of assistance to the Court. The second question is whether the witness has acquired by study or experience sufficient knowledge of the subject to render his opinion of value in resolving the issue before the Court.”

I accept that what was said in Bonython was a correct statement of the law.

[95]Mr Rivett was happy to accept that expert evidence on two areas in this hearing fulfilled the above criteria in Bonython in that they were areas in which the Tribunal would be assisted by expert evidence and areas in which there was a sufficiently reliable body of knowledge and experience. Those two areas were:(1) Economic theory in so far as relevant eg on the law of supply and demand;

(2) The ‘economics’ of regulatory decision making and in particular the question of what factors a regulator would take into account in setting prices.

[96]Mr Rivett considered that large chunks of Dr Rubin’s report went beyond either of these two areas and in particular gave evidence about what Dr Rubin thought SOSE and Ofwat had actually done in 1989 and 1994 respectively.

[97]However, I do not consider this criticism justified. Both experts gave their opinion on what Mr Rivett described as the economics of regulatory decision making and both gave their opinion on what the SOSE and Ofwat was likely or not likely to have taken into account in their decisions 30 years ago. Dr Rubin did not suggest that he knew as a fact what matters the SOSE and Ofwat took into account: he gave his opinion on what he thought they would have taken into account. That falls within (2). Each expert’s opinions were different on (2) but both concerned the same subject matter and were equally admissible (bar the point below on expertise).

[98]Mr Rivett referred me to the very different case of Altus Group UK Ltd v Baker Tilly Tax and Advisory Services Ltd

[2015] STC 788

at [89] where the judge ruled inadmissible an expert’s evidence because he was giving an opinion on what someone else would have done and because the Judge thought it biased. The Judge in that case did appear to accept, however, that it was possible for an expert witness to give evidence about the typical behaviour of a third person. In my view, there is a very narrow line between giving evidence how about how a person could be expected to act in given circumstances, and making inferences as to how they did act, and if either expert slipped from one to the other, the Tribunal is not misled.

[99]As I have said, both experts gave their opinion on factors which the regulator would, or would not, have taken into account when setting maximum infrastructure charges. Expert evidence on that subject was in my view admissible as both(a) on a subject matter that a person without instruction or experience in the area of knowledge or human experience would be unable to form a sound judgement on the matter without the assistance of witnesses possessing special knowledge or experience in the area and

(b) of a subject matter which forms part of a body of knowledge or experience which is sufficiently organised or recognised to be accepted as a reliable body of knowledge or experience.

[100]I note in passing that HMRC agreed large sections of Dr Rubin’s report were not relied on because they concerned matters that were no longer in issue and I do not consider them here. Expertise

[101]HMRC and the appellant accepted that Dr Rubin was an expert economist; HMRC did not put him forward as, nor did the appellant consider him to be, an expert on the economics of regulated industries .

[102]While Dr Rubin had written a paper on the regulation of the Post Office, he did not specialise in regulatory economics. He was an economist and his employment history had been mostly in positions involving economics, research and analysis. His normal daily work at HMRC was giving advice on transfer pricing, which, by its nature, was unlikely to involve regulated monopolies. He had previously given advice to HMRC on economics in a number of unjust enrichment cases but none involving regulated industries.

[103]Dr Koboldt, on the other hand, was an economist with a specialism in, amongst other things, regulatory economics. He had given advice on many occasions to clients operating in regulated industries and to regulators themselves. I accept, as HMRC did, that Dr Koboldt was an expert in regulatory economics.

[104]So while it appears that the appellant considered Dr Rubin could give an expert view on(1) (Economic theory), they did not accept he could give an expert opinion on

(2) (the economics of regulatory decision making).

[105]I consider that Dr Rubin could evidence on economic theory and, because the regulator would be seeking to replicate an outcome that a free market would generate (both experts were agreed on this and also see §52), economic theory was relevant to what regulator would have done. So, to some extent, I accept that Dr Rubin could give an expert opinion on (2) because he was an expert economist, however, I accept Mr Rivett’s point that only Dr Koboldt was an expert on regulatory economics, and therefore that Dr Koboldt should be in a position, unlike Dr Rubin, to give evidence on what a regulator would do in practice.

[106]I accept, therefore, that Dr Koboldt’s opinion on what a regulator would be likely to consider in practice was a more expert view than that of Dr Rubin’s and I refer to this again at §125.

[107]I move on to consider Mr Rivett’s case that the entirety of the evidence before the Tribunal, factual and opinion, was insufficient to conclude the case in favour of HMRC. Does the Tribunal have sufficient evidence to reach a conclusion?

[108]The burden of proof is on HMRC: if there is insufficient evidence to reach a conclusion either way then the appellant must win the appeal. Another way of putting this is that HMRC must raise a ‘prima facie’ case that infrastructure charges levels were not affected by the incidence of VAT, by which I mean that HMRC must have sufficient evidence to convince me (in the absence of rebuttal evidence) that it was more likely than not that infrastructure charge levels were not affected by the incidence of VAT. Of course, even if HMRC can raise a prima facie case, the appellant may be able to rebut it. But a part of the appellant’s case was that there was insufficient evidence for HMRC even to establish a prima facie case that infrastructure charge levels were unaffected by VAT.

[109]Mr Rivett referred me to Alliance & Leicester BS v P Robinson (2000) unrep where the Court of Appeal ruled that the trial judge had been entitled to reject as unreliable the evidence of two witnesses about what they would have done had the circumstances been other than what they were. However, the Court of Appeal were not saying that this evidence, being speculative, was inadmissible, but that it was reasonable to conclude it was unreliable in circumstances where it was not supported by other evidence such as to guidelines that would have applicable.

[110]He also referred me to Laker Vent Engineering Ltd v Templeton Insurance co Ltd (2008) unrep where the High Court found that the party with the burden of proving that it would have acted in a certain way in different circumstances had not discharged it because it had not produced convincing evidence of what was alleged.

[111]Lastly I was referred to E Surv Ltd v Goldsmith Williams Sols

[2016] 4 WLR 44

where the Court of Appeal overturned the trial judge because the Court of Appeal considered that the trial judge did not have the evidence to conclude that the party with the burden of proving that a particular person would have acted in a particular way in a given set of circumstances had discharged it.

[112]However, these cases are fact specific. The Tribunal is not prohibited from reaching a conclusion on what a person was likely to have done in a given set of circumstances that did not in fact occur, if there is sufficient reliable evidence to reach a conclusion on that matter.

[113]There is rarely direct evidence in litigation of the reasons why a person did whatever it is that is in issue: if there were direct evidence, the matter would not normally be in dispute. So it is normal for a tribunal to rely on circumstantial evidence, and in that sense to speculate. What the above cases show is that is that an opinion on what a decision would have been, even by the person who would have made the decision, may not be enough. But where there are established guidelines or framework for the decision, there may be enough evidence for a Tribunal to decide what the decision was likely to be.

[114]So I move on to consider the evidence and whether HMRC have made out their case and, if they have, whether the appellant has rebutted it. As can be seen from the above summary of the factual and opinion evidence, the evidence before this Tribunal fell into two types: evidence of what actually happened, and other evidence, including opinion evidence, on what factors a regulator would take into account in setting infrastructure charge levels. There is some overlap, so that what actually happened in 1994 might (or might not) be indicative of what happened in 1989.

[115]I consider first the circumstances in which the decisions were made and start with what the experts said about how regulator would approach his task of determining infrastructure charge levels. I move on to consider the evidence of what actually happened. It is convenient to consider the evidence by reference to the propositions HMRC put forward in support of their case: Would SOSE and/or Ofwat take the incidence of VAT into account when setting infrastructure charge levels?(1) Incompatible with the regulator’s statutory objectives to reduce infrastructure charge to reflect the incidence of VAT?

(2) The risk of cross-subsidization indicated the regulator would not choose to reduce infrastructure charges to reflect the incidence of VAT?

(3) the regulator would not choose to reduce infrastructure charges to reflect the incidence of VAT because that would result in undue preference to those customers able to recover VAT?

(4) Incompatible with economic theory for a regulator to take into account the incidence of VAT when setting prices?

(5) infrastructure charges were thought to be zero rated?

(6) VAT was not a cost to many payers of infrastructure charges?

(7) VAT is a tax on final consumer and incompatible with Parliament’s intent to reduce infrastructure charges for the incidence of VAT?

(8) The regulator would not choose to reduce infrastructure charges to reflect the incidence of VAT because AWSL’s infrastructure charge maximum was well within the overall £1,000 maximum? The SOSE and/or Ofwat did not in fact take VAT into account because:

(9) The regulator did not choose to reduce infrastructure charges to reflect the incidence of VAT because the infrastructure charges were net;

(10) The regulator did not choose to reduce infrastructure charges to reflect the incidence of VAT because there is no mention of VAT being considered.

[116]I have to consider each of the above points in relation to the SOSE in 1989 and Ofwat in 1994, in respect of both water and sewerage infrastructure charges. Would SOSE and/or Ofwat take the incidence of VAT into account when setting infrastructure charge levels? Statutory objectives of regulator:

[117]Primary and secondary objective: HMRC’s point was that the legislation gave the regulator the primary obligation to set infrastructure charges to cover the water undertaker’s costs. S7(2)(b) of the Water Act provided: (2) The Secretary of State or, as the case may be, the Director [of Ofwat] shall exercise and perform the powers and duties mentioned in subsection (1) above in the manner that he considers is best calculated – (b)...to secure that companies holding appointments ....as water undertakers or sewerage undertakers are able (in particular, by securing reasonable returns on their capital) to finance the proper carrying out of the functions of such undertakers.

[118]The obligation to consider the interests of consumers was only secondary as it was expressly stated to be ‘subject to subsection (2)’: (3) Subject to subsection (2) above , the Secretary of State or, as the case may be, the Director [of Ofwat] shall exercise and perform the powers and duties mentioned in subsection (1) above in the manner that he considers best calculated – (a) to ensure that the interests of every person who is a customer or potential customer of a company which has been or may be appointed....to be a water undertaker or sewerage undertaker are protected as respect the fixing and recovery by that company of – (i) charges in respect of any services provided in the course of the carrying out of the functions of a water undertaker or sewerage undertaker; and...

[119]Both experts and the parties seemed agreed, and I accept, that the regulator would have considered the costs to the water undertakings of providing the necessary increases in infrastructure capacity: those costs would have been very hard to estimate, including many uncertainties and variables. Future costs had to be estimated and discounted backwards to a present day cost (§53). In order to incentivise efficiency, the regulator may well only have allowed as costs what he considered to be the costs of a reasonably efficient water undertaking (see s 7(3)(d) Water Act 1990 at §29). Costs had to include a reasonable return on capital but also include an amount to incentivise necessary investment (s 7(2)(b)).

[120]Where the experts and the parties diverge, is on the question of whether the regulator would have considered price to the consumer when setting infrastructure charges. HMRC’s and Dr Rubin’s view was that the regulator would not have considered price as it was irrelevant to the question of costs to the water undertaker. Having estimated the water undertaking’s costs and arrived at an appropriate infrastructure charge, the regulator could not then reduce the infrastructure charge in order to reduce gross price as that would mean the water undertakers were not receiving the full amount of their costs, so that the regulator would not meet the primary objective of the legislation in setting the infrastructure charges.

[121]The appellant’s and Dr Koboldt’s view was that was that protecting the interests of consumers required a regulator not only to look at the undertakers’ costs but at the actual price paid by the consumers. Even if the regulator’s duty to consumers was only secondary to its duty to the undertaking, a regulator had to think of the gross price. The reality was that an estimate of costs included many variables and things impossible to predict with accuracy: the regulator would look at final price as a check on whether his costs estimate was too generous to the water undertakings. If he considered the gross price too high, he would be more conservative in what he considered to be allowable costs.

[122]Both experts and parties were agreed that the uncertainties over the water undertaking’s present and future costs in relation to infrastructure charges would be very difficult to resolve. Dr Rubin’s view was that would not make the regulator any more likely to take gross price into account; Dr Koboldt’s view was that VAT would be just one more factor to be considered when the regulator had to pick a price level, and the incidence of VAT on top of net price would tend to depress his estimate of costs.

[123]It seemed to me that Dr Rubin’s view might fairly be described as that of a purist economist. There is logic in what he says, even if I take what he says as a submission rather than putting any weight on it as the opinion of an expert in regulatory economics: The law required the regulator to set price by reference to the undertaker’s costs. The interests of the consumer were secondary and in any event, as stated by the regulator (§52), it was in the interests of the consumer that the water undertaking charged sufficient to cover costs (costs being understood to include future investment). It was also in the interests of the consumer that the water undertakers did not charge prices which allowed monopolistic profits (ie prices which exploited their monopoly position) but, because prices were set by reference to costs, by definition the price should not include monopolistic profits. In other words, by limiting price to costs (defined as above) the consumer was protected as much as possible: the water undertaking remained viable but unable to exploit their monopolistic position. The regulator could and should arrive at this position without any consideration of gross or net price to the consumer and what was ‘affordable’. HMRC accept that the regulator would have been aware of his secondary duty to consider the interests of consumers, but considered that consumers were protected by(a) infrastructure charges being set by reference to an accurate as possible estimate of the undertaking’s costs and no more; and

(b) having a viable water undertaking, which would not be the case if infrastructure charges were reduced to less than the undertaking’s costs.

[124]However, Dr Koboldt was the expert in regulatory economics and, as I have said, his view on what regulators would actually do in practice carries more weight than Dr Rubin’s. Nevertheless, it seems to me that, even accepting, as I do, Dr Koboldt’s evidence that the SOSE and Ofwat would look at price to consumers when setting infrastructure charge rates, that does not necessarily mean that infrastructure charge rates would be lower than they would otherwise have been. All the circumstances should be considered, at least if it is likely they would have been known to the SOSE and Ofwat at the relevant time. Cross-subsidization

[125]I accept, as the experts said, that one of the objectives of the regulator in setting infrastructure charges was to avoid cross-subsidization, by which was meant ensuring that the water undertakings’ existing customers did not pay towards the cost of connection of new customers. This was apparent from what was said in Parliament (see §§51 and 55) if not expressly stated in the Water Act.

[126]HMRC’s submission was that reducing infrastructure charges because of the incidence of VAT would mean that other prices would be raised to compensate and that would result in cross-subsidisation which the regulator must be supposed to wish to avoid.

[127]As a matter of fact, it was true that where a water undertaking’s infrastructure charge level was greater than £1000, the infrastructure charge was capped but the water undertaking was then allowed to compensate by increasing other charges, thus resulting in cross-subsidization. This was an undisputed finding of fact at §54. But there was nothing to indicate that the regulator would have permitted cross-subsidization for water undertakings, like AWSL, whose infrastructure charge were less than £1000.

[128]Dr Rubin’s view was that this meant that infrastructure charges would not have been less then otherwise because of the incidence of VAT, because otherwise the undertaking would not be fully compensated for its costs. Dr Koboldt considered that that view was too simplistic. His view was, as already stated, that the regulator would have looked at gross price as a check on costs; the regulator’s estimate of costs may have been more conservative than otherwise because of incidence of VAT but he would not have permitted cross-subsidization because of his more conservative costs estimate.

[129]I do not think there is anything in this point: in other words, I accept Dr Koboldt’s point that if the regulator took gross price into account and took a more conservative view of AWSL’s costs than he would have done in the absence of VAT, I do not think the regulator would have seen any cross-subsidization as necessary to compensate the undertaking as he would still have regarded the undertaking’s costs as covered by the infrastructure charge. So the desire to avoid cross-subsidization would not logically have deterred the regulator from considering the VAT inclusive price. Undue preference?

[130]HMRC also put the view that the regulator was required not to show undue preference to one set of customers over another and the appellant agreed with this as I do (see S7(3)(a) at §29 above). But I do not see it as helping HMRC’s case on this: while looking at the gross price might be thought to favour those customers who were charged irrecoverable VAT over those who were not, the infrastructure charges levels applied equally to all. So if infrastructure charges were less than they would have been but for the incidence of VAT, it favoured all customers (at the expense of AWSL) equally. The point on ‘undue preference’ is neutral so far as these proceedings are concerned. Economic theory

[131]Both experts gave their opinion on economic theory and to what extent that would influence the regulator, and what they said was much the same as what they said in relation to the regulator’s duties. So far as I understood it, pure economic theory was on Dr Rubin’s side. The regulator would aim for a ‘perfect market’ in which no monopolistic prices could be charged but instead only the costs of a reasonably efficient operator (costs to be understood as including a reasonable return on capital and an incentive to invest) would be covered. In a perfect market, economic theory was that VAT would be 100% passed on. So a regulator aiming for a perfect market would ignore the incidence of VAT on the customers. The addition of VAT to a price by the water undertaker was not an abuse of monopolistic power and it was not overcharging.

[132]And while the regulator ought to take into account the need to incentivise the regulated water undertakers to become efficient and reduce costs, and therefore might only allow the water undertakings the costs of a reasonably efficient operation, VAT was not a cost which any water services company controlled and they could not be incentivised to reduce this cost. The joint experts agreed on this.

[133]While Dr Koboldt accepted all the above as correct economic theory, his view was that a regulator would know he could not achieve the price that a perfect market could set: he could only aim for an approximation. And his approximation might well be influenced by an awareness of the final price that consumers would have to pay. The knowledge of the gross price might influence him to take a more conservative view of the water undertakings’ costs with a view to keeping down the gross price; if there was no VAT on the price, or if the customers were able to recover VAT, the regulator’s view may have been less conservative.

[134]What Dr Koboldt actually said was that it was ‘likely’ that regulators would consider VAT and ‘not unlikely’ that counterfactual net prices (ie the price that the regulator would have set had there been no VAT) would have been higher. He also said that he could not say with certainty that the counterfactual prices would not have been lower. In oral evidence, he said VAT was ‘conceivably’ something which the regulator would take into account. He was cross examined on what he meant: he explained that English was not his native language and he did not mean it was possible but not probable that the infrastructure charge levels were affected by VAT. He appeared to indicate he thought it more likely than not the regulator considered VAT, but at the same time he indicated that that did not necessarily mean that the counterfactual price would have been higher.

[135]While Dr Rubin’s more ‘purist’ view of how a regulator ought to set prices taking into account both his statutory obligations and economic theory is logical and attractive, I do put weight on Dr Koboldt’s view that as a matter of practical reality any regulator would look at gross price as a check on whether or not their estimate of the regulated industry’s costs was correct. But, as I think Dr Koboldt accepted, the fact that VAT was added to the gross price would not necessarily mean that the regulator would set a price at a lower level than if VAT was not added to the price.

[136]Dr Koboldt pointed out that Ofgem had taken VAT into account in a decision dating to 2002 (§64). The summary of that decision given by Ofgem indicates that VAT was considered to the extent there was an issue whether the price set should be net or gross: the question was whether the regulated industry or its customers should bear the risk of VAT rate changes. Ofgem decided that that was a risk for the customers and so set a net VAT price. However, the brief reference to VAT in that decision leaves it entirely unclear whether VAT actually affected the level of the net price in that instance and so on this point it is of little use.

[137]The consideration of statutory duties and economic theory has not resolved the dispute: even though, because he is the regulatory expert, I prefer Dr Koboldt’s evidence that the regulators would consider gross price when setting infrastructure levels, I do not accept that that necessarily means that the infrastructure charge levels were lower than they would otherwise have been. However, HMRC have not, by relying on these two points, satisfied me of their case on the balance of probability and so if I had nothing else to consider the appellant’s appeal would succeed.

[138]But I do have other factors which must be considered so I go on to consider the other factors, specific to the decisions at issue in this appeal, which the regulator may have considered, and also what evidence there is of what was actually considered. However, as the position in respect of the following issues was different in 1989 to 1994, I consider them separately: Did the SOSE know that infrastructure charges would be subject to VAT?

[139]In practice, from 1 April 1990 standard rated VAT was charged on all infrastructure charges until 1 April 1994. HMRC discovered evidence, I am told somewhat late in the litigation process, that that VAT liability of infrastructure charge was in some doubt before 1 April 1990 (§§70-78). But in their reports both experts assumed that the regulator knew infrastructure charges would be subject to VAT. Their reports must be considered with that in mind because this assumption is in my opinion incorrect .

[140]It seems to me that if the regulator adopted Dr Rubin’s view that the incidence of VAT was strictly irrelevant in economic theory and by statute to the job he was required to do, then he would have considered VAT no further and it would not have affected the price AWSL was permitted to charge.

[141]But if the regulator adopted a view similar to Dr Koboldt’s, and looked to gross price as a check on whether his estimate of the water undertaker’s costs was right, then it seems to me considerably more likely than not that the regulator would have taken steps to inform himself of the gross price. He would not have assumed it. Mr Rivett says this is speculation. I do not agree. The regulator, as the experts accepted, had to deal with many variables but there is no suggestion that he would not have taken appropriate steps to inform himself on matters he considered relevant. Bearing in mind he was subject to a statutory duty and exercising power conferred on him by Parliament I find that, if he considered VAT relevant at all, he would not have made assumptions about it but would have taken reasonable steps to inform himself whether VAT was chargeable on the net price and whether VAT was necessarily a cost to those who paid it.

[142]Further, I accept HMRC’s case that it seems more likely than not that the SOSE, as head of one government department, if he wished to be informed on the correct VAT position of infrastructure charge charges, would have asked the government department with responsibility for such matters, HMCE. The SOSE may also have asked the water undertakings or WSA but it seems to me more likely that he would have relied on the view of HMCE, as the government department in charge of collecting the VAT due. It does not really matter as there is no suggestion that the views of HMRC and WSA diverged before 1 September 1989.

[143]The parties were agreed that the infrastructure charges set by the SOSE were set no later than 1 August 1989 when regulation was handed over to Ofwat (§40). Do I have sufficient evidence to decide what the SOSE would have been told at that point in time if he had enquired of HMCE about the incidence of VAT on infrastructure charges?

[144]There is clear evidence (§73) as at 29 August 1989, in other words, nearly a month after the rates were set, that HMRC and, it seems, the WSA, thought that the VAT liability of infrastructure charges would follow the VAT liability of supplies of water and sewerage services. As at that date, the VAT supply of all water and sewerage services were zero rated; it was to become standard rated for most industries after 1 July 1990 and as the charges were not due to come into force until 1 April 1990, I find it was, as would have been expected, the post- 1 July 1990 position that was discussed at the meeting on 29 August 1989 (see §73).

[145]But what would HMRC have considered the VAT treatment of infrastructure charge charges to have been if they had been asked a month earlier?

[146]Although HMCE/HMRC over the years have vacillated between seeing infrastructure charge charges as zero rated, standard rated or outside the scope, the only suggested reasons for considering them to be standard rated, other than as a supply of water to SIC 1-5 industries, was by seeing them as charges for civil engineering works, and that appears to have first occurred to HMCE a few days after 29 August 1989 and was not formally adopted until about 20 September 1989 (see §74-75). So it seems more likely than not prior to 29 August HMCE would have seen infrastructure charges as zero rated as part of the charge for the supply of water/sewerage services, acknowledging that from 1 July 1990 they were to become standard rated in so far as supplied to taxpayers in SIC 1-5 industries. The only other alternative is that they considered them to be outside the scope. Either view amounts to the same so far as this appeal is concerned. The point is that the evidence is clear that the only reason given for considering infrastructure charges to be subject to VAT first occurred to HMCE over a month after infrastructure charge levels were set.

[147]In short, I am satisfied that if SOSE had considered gross price to consumer relevant to his ascertainment of the infrastructure charges, he would have asked HMRC sometime on or before 1 August 1989 and most likely would have been informed that all sewerage charges were zero rated and further that water charges were zero rated unless supplied to SIC 1-5 industries.

[148]Mr Rivett considers such a conclusion too speculative. His case is that I ought to conclude that HMRC have not proved their case that the regulator would have made enquiries of HMCE around July 1989, and have not proved that the regulator would have been told by HMCE that the VAT liability of infrastructure charges would be the same as the VAT liability of charges for the supply of water and sewerage services. But I do not agree for the reasons given above.

[149]In any event, even if Mr Rivett was right and HMRC were muddled and kept changing their mind so that they would not have been in a position to give a categoric answer to the question of the VAT liability of infrastructure charges in July/August 1989, it would not really help their case. Dr Koboldt took the view, logically, that if the incidence of VAT was uncertain, the regulator was less likely than otherwise to take it into account when setting infrastructure charge rates. I think that if the regulator was unable to get a reliable answer to the question of infrastructure charge VAT liability he was unlikely to factor VAT into his calculation of infrastructure charges.

[150]However, my finding is that it is more likely than not that, the regulator would have asked HMRC at around the time infrastructure charges were set and HMRC would have communicated to the SOSE that sewerage charges were expected to be zero rated. I note that Dr Koboldt agreed that if the regulator had known that VAT was not imposed on infrastructure charges, then he would not have reduced the infrastructure charges by the incidence of VAT. I agree. Therefore, so far as sewerage charges set in 1989 are concerned, I determine the appeal in HMRC’s favour. Either the SOSE would not have considered the incidence of VAT at all, or he would have been aware that no VAT was expected to be chargeable on sewerage infrastructure charges. Therefore, VAT would not have had any impact on the level of the sewerage infrastructure charge set by the SOSE at around 1 August 1989.

[151]It is the case that in the event HMRC determined that VAT was payable on all infrastructure charges (see §77) but that was neither known nor predictable as at the time SOSE set the infrastructure charges: the evidence shows that it was an unanticipated and sudden change of mind by HMRC in September 1989 which was after infrastructure charge charges were set. (Indeed, it is evidence in favour of HMRC’s view that, because the change in HMRC’s views on the liability of infrastructure charges to VAT did not cause a revision in infrastructure charge rates, the SOSE did not consider VAT when setting the infrastructure charges in the first place and I refer to this at §189).

[152]Not all water infrastructure charges were expected to be zero rated as at 1 August 1989 because, as was public knowledge, standard rating was being introduced for ‘industrial’ users of water. I find that as at the date the infrastructure charge were set, if the regulator thought VAT relevant to setting infrastructure charges, more likely than not, he would have made enquiries of HMCE and, more likely than not, would been informed that some payers, in particular developers, of the water infrastructure charge would have to pay VAT.

[153]As I have said the SOSE, and later Ofwat, were undertaking a statutory duty and would have taken steps to inform themselves on matters that were relevant. So I also consider that if the SOSE had decided it was appropriate to take VAT into account, he would also have considered the question of whether VAT was actually a cost to those who had to pay it. VAT not a cost to many users?

[154]Mr Allen (§§66-69) produced a breakdown of AWSL’s current customers and gave evidence, which I accept, that there was no reason to suppose the breakdown was any different in 1990-1996. There were some submissions about the percentage of those customers who were able to recover VAT. The appellant’s point was that HMRC had to prove it but the evidence was too vague. I have dealt with this at §69 and my conclusion is that HMRC could not prove a percentage of customers who were able to recover VAT in 1989 although on the evidence I do accept that a significant proportion were likely to be able to do so. In any event, the uncertainty, as I have said at §149, does not operate in the appellant’s favour: it seems to me that in circumstances were the percentage of users able to recover VAT was uncertain but clearly significant, the regulator was less likely than not to take VAT into account.

[155]But, as I have said, the SOSE could only have operated on information available to him in July/August 1989. He would not have known that infrastructure charge charges would be subject to VAT for DIY builders as that was not clear until September 1989. If he considered VAT relevant at all, he most likely would have made enquiries of HMRC and have would have known that VAT was only expected to be charged on infrastructure charges paid by businesses in SIC 1-5. He must have understood many businesses, including many developers, are VAT registered and taxable, while others are not registered, or not fully taxable.

[156]Therefore, HMRC’s case was that the regulator would have understood that reducing infrastructure charge rates to take into account the incidence of VAT would fly in the face of his primary duty to ensure that the water undertaking received reimbursement of its full costs, while at the same time without the reduction being necessary to protect the interests of many of the undertakings’ customers. On the contrary, it would hand many a windfall.

[157]As I understood it, the appellant accepted that in fact a significant number of customers would be able to recover the VAT on infrastructure charges and the regulator would have known this. As I understood Dr Koboldt’s view, adopted by the appellant, it was that it was unlikely that the regulator would in these circumstances reduce the regulated price by the full amount of VAT, but he might reduce it by some part of the VAT to take into account the gross price when setting infrastructure charge rates.

[158]My decision is that while I accept it was possible that the SOSE might have considered the gross price, it seems to me less likely than otherwise: it is more likely that that, because VAT was either not going to be paid, or if it was paid, not going to be a cost to what was a significant percentage of the water undertakings’ customers, it was more likely the regulator, if he considered price at all, looked at net rather than gross price. VAT is a tax on final consumption

[159]It was also HMRC’s case that the regulator would not have wished to protect consumers from the incidence of VAT. While VAT was a cost (in the sense wholly or partly irrecoverable) to some of the water undertaking’s customers, VAT was intended to be a cost to them because they were, in a VAT sense, final consumers, and VAT was meant to fall as a cost on final consumption. HMRC’s case was that this was expressly stated in the Sixth Vat Directive and reiterated by the CJEU in Elida Gibbs (1996) C-317/94.

[160]I discount what was said in Elida Gibbs as it was said in 1996 and would not have been in the regulators’ minds when the infrastructure charge charges at issue in this appeal were set.

[161]Nevertheless, it seems to me that a regulator, choosing to take gross price into account, must have understood that VAT was a tax. Where VAT liability was imposed, Parliament intended the purchasers to pay it. Where the purchasers were unable to recover it, that was also an outcome intended by Parliament. Why should the regulator attempt to set infrastructure charge levels to ‘correct’ a result which Parliament intended?

[162]Dr Koboldt’s view was that while that was probably true, the regulator would not be trying to ‘correct’ the incidence of VAT, but simply be taking the gross price into account when setting infrastructure charge levels. But it seems to me the fact that VAT would affect different customers differently is a reason why a regulator, choosing to take price into account in setting infrastructure charge levels, would consider net price more appropriate to take into account than gross price. It might be different if all purchasers had to pay VAT and none could recover it, but that was not the case here.

[163]In other words so far as water infrastructure charge setting was concerned, I am satisfied that the regulator in 1989 more likely than not would not have reduced infrastructure charges because of the incidence of VAT. While I accept it was more likely than not that he would look at the price to consumer as well as cost to the water undertaking, if he did consider price to the final consumer, the varying impact of VAT on consumers and in particular(a) the fact that a significant section, albeit probably a minority, of customers were not expected to pay VAT on water infrastructure charges,

(b) VAT would reasonably be expected to be recoverable by a significant section of customers who would have to pay VAT on infrastructure charges and

(c) that VAT was intended to be a tax on final consumption and intended to be irrecoverable by those making exempt supplies, I consider the regulator was more likely to look at net than gross price. The appellant’s case that the regulator, in these circumstances, would simply make a smaller reduction in infrastructure charges than he would have made if VAT was a cost to all payers of infrastructure charges doesn’t persuade me because it does not seem logical as in such circumstances net price would appear a more appropriate measure of whether the estimate of the undertaking’s costs was reasonable. The £1000 limit

[164]HMRC’s case was that the regulator set a maximum level for infrastructure charges of £1000 (plus VAT). In practice, this affected 6 water undertakings which were unable to charge the infrastructure charges which they would otherwise have been able to charge calculated on the same basis as other water undertakings. In other words, their cost of investment discounted over the 20 years per new customer was higher than £1000 plus VAT (see §§42 and 54). AWSL was not one of the six.

[165]HMRC say that the natural inference is that the regulator chose the £1000 limit taking into account affordability to customers. It was clear, say HMRC, that the £1000 limit was not set with the undertakings’ costs in mind: that was because it was provided (see §54) that those affected undertakings could recoup their extra costs in excess of the £1000 through other charges (in other words, cross-subsidisation was permitted to that extent). So the £1,000 limit must have been set with affordability to customers in mind. I agree with that view: the £1,000 bore no reference to costs, as can be seen from the fact that the water infrastructure charges themselves varied from §111 to over £1,000 and sewerage infrastructure charges varied from §240 to §983.

[166]HMRC, and Dr Rubin, drew from that the proposition that the majority of water undertakings whose infrastructure charge was below £1000 per VAT, such as AWSL, had their infrastructure charge levels set without considerations of price to customer.

[167]As I understand it, Dr Koboldt’s view was that, while £1000 may have been a view on affordability, the regulator was nevertheless required to look at each water undertaking individually. Looking solely at the costs of an undertaker might lead to an infrastructure charge level of somewhat less than £1000 (as in the case of AWSL) but the regulator might still consider gross price as an indicator of whether he had been too generous in his cost estimate.

[168]Nevertheless, I consider HMRC’s view valid. £1,000 did seem more likely than not to reflect the SOSE’s view on affordability. So even if the regulator considered gross price should be taken into account when deciding whether an assessment of the undertaking’s costs was too generous, the regulator might well think that an assessment of costs which was less than £1,000 before gross price was even considered, was within the acceptable range. Moreover, the facts show a wide variety in infrastructure charges between the water undertakers, which is inconsistent with the regulator having a fixed view on affordability, other than a view that all charges below £1,000 were affordable.

[169]In conclusion, that a maximum charge of £1000 was set and AWSL’s infrastructure charges were well below indicates to me that regulator would be less likely than not to reconsider his costs estimate on the basis that the gross price was too high.

[170]That conclusion only reinforces the conclusion which I reached in §163. In summary, my conclusion so far is that that the SOSE in 1989 when setting infrastructure charge rates(a) more likely than not did not reduce sewerage infrastructure charge rates because of the incidence of VAT. I take this view because on the basis that even though he may well have considered price relevant, the information likely to have been available to him at around the time the rates were set was that net price was to be the same as gross price (in other words, no VAT would be charged). In the event, this information would have been wrong but he could not have known that as at the time the rates were set.

(b) more likely than not did not reduce water infrastructure charge rates because of the incidence of VAT. I take this view because even though he may well have considered gross price relevant, if he had thought it relevant, on the information likely to have been known to him at the time, he would have been aware that for many customers net price was effectively the same as gross price and that AWSL’s infrastructure charge rates would be well below the maximum level of £1000 in any event. So that taking all these variables into account, it is more likely than not that the infrastructure charge rates were set without any reduction because of the incidence of VAT. Ofwat’s 1994 decision on maximum infrastructure charge rates

[171]Both parties were agreed that the vast majority in value of the fees at issue in this appeal were paid under the infrastructure charges in force from 1 April 1990 to 1 April 1995. Not only were infrastructure charges much reduced with effect from 1 April 1995, some had become zero rated. Nevertheless, I need to determine what factors, on the balance of probability, Ofwat would have considered in 1994 when setting the infrastructure charge levels with effect from 1 April 1995. I do not consider water and sewerage separately as the factors which related to each were the same.

[172]My conclusion is the same in respect of the general factors relating to statutory duty and economic theory considered at §§117-138. There was no objective reason why Ofwat’s reasoning would be any different to the SOSE’s on the question of statutory powers and economic theory.

[173]But the VAT position was different in 1994. Moreover, Ofwat set no overall maximum so considerations of the SOSE’s £1,000 maximum was irrelevant.

[174]For the same reasons as before, while I accept Dr Koboldt’s expert opinion that a regulator would be likely to consider gross price when estimating a regulated supplier’s costs, nevertheless if he did consider gross price relevant, I find he would have taken reasonable steps to inform himself of what the gross price actually was. In other words, he would, more likely than not, have consulted HMCE on the proper VAT liability of infrastructure charges.

[175]As I have said, while, unlike the SOSE, he would have known infrastructure charges had been standard rated since 1 April 1990, he would also have been informed both water and sewerage infrastructure charges were zero rated for new connections to domestic and qualifying buildings as from 1 April 1994 (§46). The appellant’s case as I understand it is that Ofwat would not have known what percentage of AWSL’s customers to whom VAT was not a cost, and I accept that that could not be known with precision. Yet Mr Allen’s evidence shows that it was possible to identify that at least 59% were developers and some 13% were likely to be small builders/DIY builders/parish councils and charities. (Neither party suggests the information available would have been different in 1994 to 2015/16). While it seems unlikely the SOSE could have known what percentage of developers could actually recover VAT, that question would have been irrelevant in 1994 when it was apparent charges to developers, DIY builders, parish councils and so on would be zero rated. Moreover, while businesses other than developers would pay VAT on infrastructure charges, Ofwat would have to assume that a significant proportion would be VAT registered and able to recover VAT.

[176]In conclusion, if the regulator in 1994/5 considered gross price, more likely than not he would have known that a clear majority of AWSL’s customers would not pay VAT on the infrastructure charges which he was setting and of those who did pay VAT, some of them would be VAT registered businesses and able to recover it. It is inconceivable in these circumstances that the regulator would have reduced the infrastructure charge pound for pound for the incidence of VAT in the circumstances when VAT was not a cost to the majority of payers of the charge. The appellant’s case is that there would have been a lesser (albeit not de minimis) reduction. But I do not agree for the reasons given at §157/8 and 163. The rationale given by Dr Koboldt for a regulator, bound to set prices by reference to cost, considering price was affordability to consumers. In my view, in circumstances where a majority of payers would only in practice pay a net price, it was more rationale to consider net price than gross price, and therefore more likely than not the regulator would have considered it inappropriate to reduce infrastructure charges to reflect the incidence of VAT.

[177]I move on to consider the evidence of what the regulators actually did in 1989 and 1994. Both sides were agreed that actual documentary evidence of what the regulators did in 1989 and 1994 would be for more persuasive than opinion evidence of how regulators would act. The Dog that didn’t Bark?

[178]Referring to the dog in Conan Doyle’s Silver Blaze which only barked at strangers, HMRC’s case was that I could infer that VAT had not been taken into account by the regulators in setting the infrastructure charge levels as it was not mentioned in any of the documentary evidence of what factors were actually taken into account.

[179]The appellant disagreed; it considered there was virtually no direct evidence of what the SOSE took into account in 1989, and while there was more evidence of what Ofwat took into account in 1994, it was not comprehensive. Either way it would be wrong, it said, for me to conclude from the lack of direct mention of VAT that VAT was in fact not considered.

[180]Both experts were agreed that regulators in general rarely refer to VAT but that cuts both ways: it does not tell me anything as it is not clear whether VAT is rarely mentioned because it is rarely taken into account by regulators or because it is taken into account but regulators choose not to say so.

[181]I accept that there is more evidence from 1994, and I find it makes more sense to work backwards and consider the evidence from 1994 first. I also accept that the 1994 Ofwat report (§§58-61) did not comprehensively list all factors considered by Ofwat in setting the infrastructure charge rates. Nevertheless, I do find it indicates that VAT was not considered at all: it suggests that Ofwat, contrary to Dr Koboldt’s opinon of what a regulator was likely to do, actually followed the line that Dr Rubin considered a regulator ought to of considering costs without reference to gross price, because there is simply no mention of affordability (other than a brief mention of affordability in reference to water charges generally and even then there was no suggestion that charge levels were reduced to make them affordable). VAT in particular was not mentioned anywhere.

[182]Indeed, I have already referred to the fact that while the parties were agreed the infrastructure charges (and the maximum of £1,000) were all VAT exclusive, they appeared agreed, and I find, that this was not expressly stated anywhere. In my opinion, that is a further support for the case that VAT was simply not considered. If there had been consciousness of VAT, it seems more likely than not that the regulator would state, for the avoidance of doubt, whether the charges were inclusive or exclusive.

[183]Dr Koboldt’s view was that VAT would have been one of many factors considered but it may have been a more minor factor and for that reason it may not have been referred to. I don’t accept that. Firstly, for the reasons given at §123 above, it is not obvious that VAT is even relevant to determining infrastructure charge levels so if it was to be taken into account, this was likely to have been explained; secondly, VAT is a one of a kind cost. It is a tax over which the water undertakings have no control and which applied differently to different customers. If it was thought relevant, the basis of this was likely to be explained. Thirdly, other taxes (eg corporation tax) were mentioned in the report making the absence of VAT suggestive of the fact it was not considered.

[184]So the fact that VAT was not mentioned in the 1994 Ofwat report means I find that the direct evidence is that it was more likely than not that VAT (or gross price) was not considered relevant to the setting of infrastructure charge levels.

[185]The Tribunal did not have the full MMC reports in which two water undertakings challenged the regulator’s 1994 decision. But as I have said, I accept that they corroborate the conclusion I reached in relation to the Ofwat report. They indicate that the regulator was not making out a case that VAT was a concern for customers, nor were the water undertakings complaining that the maximum limit of infrastructure charges was unfairly lowered due to perceived incidence of VAT. While not conclusive by itself, it supports the conclusion in §184 that more likely than not in fact the Ofwat did not consider VAT when setting infrastructure charge levels. Factors actually taken into account in 1989?

[186]There is very little direct evidence of what was taken into account in 1989. In particular, the debates in Parliament and prospectus to which I was referred (§§51, 52 and 55) were far too general to be of assistance. The 1991 Ofwat report (§53), the NAO report in 1992 (§56) and Parkers (§57) all suggest that it was only costs which was taken into account by the SOSE when setting infrastructure charge levels and that the only taxes considered were direct taxes, but by themselves are far too general to establish that the SOSE most likely did not consider price when estimating costs.

[187]But it is relevant that the 1994 Ofwat report (§§58-61) did not mention VAT as, in my view, the report would have been likely to mention VAT if Ofwat was taking a different line to the SOSE on the relevance of VAT (I have already concluded that Ofwat did not take VAT into account in 1994: §184-185). I also consider it relevant that in 1990 the WSA persuaded HMCE to rule more infrastructure charges were standard rated than HMCE had originally suggested: this is also suggestive, without being conclusive by itself, that VAT had not featured in the SOSE’s original determination of infrastructure charges. Had the WSA thought the incidence of VAT led to reduced infrastructure charges they could have been expected to challenge HMRC’s revised ruling which increased the incidence of VAT, not request that HMRC go even further.

[188]It is also suggestive if inconclusive by itself that there was no reconsideration of infrastructure charge levels when it became clear in early 1990 that all infrastructure charges would be standard rated, contrary to what I have found would have been understood to be the position at the time the infrastructure charges were set. Similarly, the fact that net rates were set without any statement that they were rates net of VAT is indicative of the rates being set without regard to VAT as it makes VAT look like an afterthought.

[189]My conclusion is that although the evidence of what actually happened in respect of 1989 is weaker than that in respect of 1994, the evidence summarised in the previous two paragraphs does amount to a prima facie case that the SOSE did not in fact take VAT into account when setting infrastructure charge rates. And that has not been rebutted by the appellant: there is no evidence that the SOSE actually did take VAT into account. That may be because he took a purist view, in line with his statutory duties and economic theory, as described by Dr Rubin, and considered only the water undertakings’ costs, or it may be that he did as Dr Koboldt says most regulators do, and his costs estimate was influenced by his view of the affordability of the price to consumers, but if he did the latter, on the basis of the evidence at §§187-188, I find it is more likely than not that he only considered the net price to consumers. And that would have been a quite reasonable view where, so far as he would have known at the time, the net price would be the real price to a very significant number of the consumers. Conclusions

[190]The appellant’s case was that there was insufficient evidence to reach a conclusion on the speculative question of what the regulators would have done in 1989 and 1994. But it was not entirely speculation: the regulators had a statutory framework and would have taken into account economic theory; I had the benefit of an expert in economic regulation and evidence of what was known about VAT at the time and some direct evidence of whether they actually took VAT into account. I consider that there was sufficient evidence for HMRC to make out their prima facie case that the maximum infrastructure charges were not affected by the incidence of VAT, and the appellant failed to rebut it. Sewerage

[191]1989 : I consider it much more likely than not that the level of infrastructure charges in 1989 would have been set at £597 (their actual level) irrespective of the incidence of VAT because as explained above it is more likely than not that, if the regulator considered gross price relevant at all, he would have known that at the time HMCE’s view was that all sewerage infrastructure charges would be zero rated.

[192]1994 : I consider it more likely than not that the level of infrastructure charges in 1994 would have been set at £200 irrespective of the incidence of VAT because(a) if the regulator considered VAT relevant at all, he would have known that most payers of sewerage infrastructure charges would not pay VAT and a significant number of those which did would be able to recover it and

(b) in any event, the evidence is that in practice Ofwat did not consider VAT relevant to setting infrastructure charge levels as it was not mentioned in the report or otherwise. Water

[193]1989 : I consider it more likely than not that the level of infrastructure charges in 1989 would have been set at £479 (their actual level) irrespective of the incidence of VAT because as explained above it is more likely than not that, if the regulator considered price relevant at all, he would more likely have considered net price rather than gross price as he would have known that at the time HMCE’s view was that only water infrastructure charges to industry would not be zero rated and so he ought to have considered that a significant percentage of payers would either not pay VAT at all or would be able to recover it: this is a more finely balanced question than in respect of sewerage but I am satisfied that HMRC have made out the burden of proof on this because in practice the indications are that VAT was not considered relevant (§§187-188).

[194]1994 : I consider it more likely than not that the level of water infrastructure charges in 1994 would have been set at £200 irrespective of the incidence of VAT because(a) if the regulator considered VAT relevant at all, he would have known that most payers of water infrastructure charges would not pay VAT and a significant number of those which did would be able to recover it and

(b) in any event, the evidence is that in practice Ofwat did not consider VAT relevant to setting infrastructure charge levels as it was not mentioned in the report or otherwise.

[195]As I said at §14, HMRC must repay the overpaid VAT save to the extent that they can prove the repayment unjustly enriches the appellant and that would require them to prove the level of infrastructure charge charges would have been had they been known not to be subject to VAT. As I have found at §§190-193 that the infrastructure charges more likely than not would have been set at the level at which they were set irrespective of the incidence of VAT, it follows that even if it had been known that they were zero rated it is more likely than not that they would have been set at the same level. HMRC have therefore proved their case as the standard of proof is only balance of probability.

[196]The appeal is dismissed.

[197]This document contains full findings of fact and reasons for the decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009. The application must be received by this Tribunal not later than 56 days after this decision is sent to that party. The parties are referred to “Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)” which accompanies and forms part of this decision notice. BARBARA MOSEDALE TRIBUNAL JUDGE RELEASE DATE: 4 MAY 2017

Anglian Water Services Ltd v Revenue and Customs (VAT - ZERO-RATING : Sewerage services and water) [2017] UKFTT 386 (TC)

Justice King in the case of R v Bonython [1984] SASR 45, as follows:

Cited in 2 later judgments