BT Plc v The Commissioners for HMRC

It was accepted both before the judge and this court that for present purposes there was no practical difference between the tests for striking out and reverse summary judgment. There was also no issue as to the correct legal test to apply, namely that summarised by Lewison J in Easyair Ltd v Opal Telecom Ltd [2009] EWHC 339 (Ch) (“Easyair”) at [15] in the context of summary judgment. I am content to proceed on that basis, although it is worth bearing in mind that there are some differences between the summary judgment procedure under CPR Part 24 and applications for striking out for lack of reasonable grounds to bring or defend a claim under rule 3.4(2)(a). In particular, questions of striking out are generally determined by reference to the pleaded case and are not apt for the determination of any factual dispute, whereas applications under Part 24 are more likely to involve the scrutiny of evidence.

As Lewison J explained in Easyair, the court must consider whether the claimant has a realistic as opposed to a fanciful prospect of success. Realistic means that the claim must carry some degree of conviction. A mini-trial must not be conducted and the court must also take into account evidence that can reasonably be expected to be available at trial. However:

“…it is not uncommon for an application under Part 24 to give rise to a short point of law or construction and, if the court is satisfied that it has before it all the evidence necessary for the proper determination of the question and that the parties have had an adequate opportunity to address it in argument, it should grasp the nettle and decide it. The reason is quite simple: if the respondent’s case is bad in law, he will in truth have no real prospect of succeeding on his claim or successfully defending the claim against him, as the case may be. Similarly, if the applicant’s case is bad in law, the sooner that is determined, the better.”

In contrast, if further evidence is likely to be available at trial that would affect the position then summary judgment should not be given, albeit that:

“…it is not enough simply to argue that the case should be allowed to go to trial because something may turn up…”

A number of authorities have considered whether a legislative scheme is intended to exist alongside common law remedies or is intended to be exhaustive and exclude them. Some of these are tax cases and some concern other areas. As Miles J explained, they include Johnson v Unisys Ltd [2001] UKHL 13, [2003] 1 AC 518, Deutsche Morgan Grenfell Group plc v Inland Revenue Commissioners [2006] UKHL 49, [2007] 1 AC 558 (“DMG”: discussed further below in the context of unjust enrichment), Revenue and Customs Commissioners v Total Network SL [2008] UKHL 19, [2008] 1 AC 1174 (“Total Network”), Monro v Revenue and Customs Commissioners [2008] EWCA Civ 306, [2009] Ch 69, R (on the application of the Child Poverty Action Group) v Secretary of State for Work and Pensions [2010] UKSC 54, [2011] 2 AC 15 (“CPAG”), Investment Trust Companies v Revenue and Customs Commissioners [2017] UKSC 29, [2018] AC 275 (“Investment Trust Companies”) and Littlewoods Limited v Revenue and Customs Commissioners [2017] UKSC 70, [2018] AC 869 (“Littlewoods”).

Miles J relied on the following summary by Hickinbottom LJ of the principles to be derived from these cases in Southern Gas Networks Plc v Thames Water Utilities Ltd [2018] EWCA Civ 33, [2018] 1 WLR 5977 (“Southern Gas”) at [37]:

“It is unnecessary to quote at length from the cases to which I have already referred. The following propositions can be drawn from them.

(i)

Where Parliament has legislated for a statutory remedy to apply in certain circumstances, whether that remedy ousts any common law remedy which would or might have arisen on the same facts depends upon whether, on the true construction of the particular statutory provisions, Parliament intended that provision to oust, or co-exist with, the common law remedy. The courts will not maintain a common law remedy in the case of an evident intention of Parliament to displace it (see, eg, Johnson v Unisys Ltd [2003] 1 AC 518 at para 58, per Lord Hoffmann, and para 80, per Lord Millett, Deutsche Morgan Grenfell Group Plc v Inland Revenue Commissioners [2007] 1 AC 558, para 19 per Lord Hoffmann, and CPAG [2011] 2 AC 15, para 27, per Dyson JSC).

(ii)

Where that intention is not express, the threshold for inferring ouster of common law rights is high; but it is not helpful to approach the question on the basis that there is a presumption against ouster. Nor, before common law rights are displaced, does ouster have to be a necessary implication, in the sense that the common law remedy is only displaced if, as a matter of logic, it cannot co-exist with the statutory regime (although, of course, common law remedies can be ousted by such necessary implication) (CPAG, para 31, per Dyson JSC).

(iii)

Whether common law remedies are ousted is dependent upon the true construction of the particular statutory provisions. However, where the statutory remedy covers precisely the same ground as the common law remedy, the latter will almost certainly have been excluded by necessary implication (above, at para 33). Furthermore, where the statutory regime provides a special or qualified remedy, it may (although not necessarily will) be inferred that Parliament intended to exclude any common law remedy that would or might arise on the same facts (see, e g, Deutsche Morgan Grenfell at para 19 per Lord Hoffmann, and at para 135 per Lord Walker of Gestingthorpe).

(iv)

The identification of some differences between the statutory scheme and the common law remedy will not necessarily lead to an inference that Parliament intended the former to oust the latter. As Dyson JSC put it in CPAG [2011] 2 AC 15, para 34:

“The question is not whether there are any differences between the common law remedy and the statutory scheme. There may well be differences. The question is whether the differences are so substantial that they demonstrate that Parliament could not have intended the common law remedy to survive the introduction of the statutory scheme. The court should not be too ready to find that a common law remedy has been displaced by a statutory one, not least because it is always open to Parliament to make the position clear by stating explicitly whether the statute is intended to be exhaustive. The mere fact that there are some differences between the common law and the statutory positions is unlikely to be sufficient unless they are substantial … The question is whether, looked at as a whole, a common law remedy would be incompatible with the statutory scheme and therefore could not have been intended by [sic] coexist with it.” (Emphasis in the original.)

Rather than “incompatible”, in Revenue and Customs v Total Network SL [2008] AC 1174, para 130, Lord Mance used the phrase “positively inconsistent”.”

As Lord Dyson explained in CPAG at [32], the test is “whether in all the circumstances Parliament must have intended a common law remedy to coexist with the statutory remedy”. He did not consider that there was a material difference between the principle to apply in tax cases and the social security legislation under review in that case, noting that Lord Mance had referred to non-tax cases in Total Network. This is reflected in this court’s application of the same principles in Southern Gas to the question whether a statutory compensation scheme for damage to gas pipes caused by a burst water main precluded a claim in negligence.

In Investment Trust Companies the Supreme Court had to consider whether s. 80 VATA 1994 excluded a common law claim to recover VAT from HMRC where the claim was made not by the supplier (to whom s.80(7) would clearly apply) but by a customer of the supplier. In concluding that it did, Lord Reed applied a purposive construction which took account of the fact that the legislation and regulations created a scheme for refunds to suppliers to be passed on to customers. He also relied on the time limit in s.80 (at the relevant time three years), commenting at [83] that:

“The evident aim is to protect public finances against the risk of a liability to repay tax emerging more than three years after the tax was received. It would be inconsistent with that intention for there to be a potentially far longer period within which non-statutory claims might be brought against the commissioners by the supplier’s customers.”

The combination of these features meant that:

“Parliament cannot sensibly be taken to have intended, when it created this scheme for the reimbursement of suppliers (with provision for them in turn to reimburse their customers), subject to strict time limits, that it should exist concurrently with non-statutory liabilities towards suppliers and their customers which were potentially wider in scope and were subject to a longer and less certain limitation period. Such an intention would be inconsistent with the rationale of the statutory scheme.” ([84])

Lord Reed was also unpersuaded by an argument that a right to recover money paid under a mistake of law was not intended to be excluded by legislation enacted before that remedy was established by the decision in Kleinwort Benson Ltd v Lincoln City Council [1999] 2 AC 349 (paragraphs [85]-[86] of Lord Reed’s judgment). He concluded at [87] that the statutory scheme was “inconsistent” with the existence of a non-statutory liability to consumers, and commented further that:

“In the absence of section 80(7), one would therefore conclude that section 80 impliedly excluded such liability (assuming that it might otherwise exist).”

Littlewoods raised the question whether the right to interest on overpaid VAT conferred by s.78 VATA 1994 precluded a common law claim in restitution for compound interest. It was held that such a claim was impliedly excluded, despite the fact that s.78(1) expressly applied only “if and to the extent” that interest would not be due apart from s.78. Further, an award of interest at common law was not required to comply with the EU law principle of effectiveness.

Lord Reed and Lord Hodge relied on the limitations in s.78, including a three year time limit, and at [34] said that:

“[Those limitations] would therefore be defeated if it were possible for the taxpayer to bring a common law claim. Parliament cannot have intended the special regime in section 78 to be capable of circumvention in that way.”

Lord Reed and Lord Hodge went on to acknowledge that the wording giving priority to other rights to interest would apply to common law rights on a literal interpretation, but applied a purposive approach, in that case taking account of the fact that when the provision was enacted s.78 was more generous than the common law as then understood, prior to Sempra Metals Ltd (formerly Metallgesellschaft Ltd) v Inland Revenue Commissioners [2007] UKHL 34, [2008] 1 AC 561 (“Sempra Metals”). (As explained below, Sempra Metals has since been departed from.)

Drawing some threads together:

A useful starting point in determining whether unjust enrichment has occurred is to ask the questions posed by Lord Steyn in Banque Financière de la Cité v Parc (Battersea) Ltd [1999] 1 AC 221 at 227, namely: (a) Has the defendant benefited, in the sense of being enriched? (b) Was the enrichment at the claimant’s expense? (c) Was the enrichment unjust? (d) Are there any defences? However, as Lord Reed explained in Investment Trust Companies at [41]-[43], these four questions are no more than “broad headings for ease of exposition”, intended to ensure a structured approach to the analysis. That analysis must be carried out having regard to the purpose of the law of unjust enrichment, which is to correct “normatively defective” transfers of value, meaning defective in a way recognised by the law of unjust enrichment, for example because of a failure of basis. (Failure of basis occurs when a benefit is conferred on the joint understanding that the right to retain it is conditional and the condition is not fulfilled: see Goff & Jones, The Law of Unjust Enrichment (10^th ed.) at 12-01.)

Investment Trust Companies contains a detailed analysis of the second of the questions posed by Lord Steyn, namely whether the defendant’s enrichment was at the claimant’s expense. As Lord Reed explained at [52] the receipt of a benefit is not enough. The claimant must incur a loss through the provision of the benefit, and a “but for” causal connection between the claimant’s loss and the defendant being better off is not sufficient. Moreover, as he went on to say at [59] and [60] the requirement that the defendant is enriched at the claimant’s expense is not satisfied by a connection that exists between the benefit and loss merely as a matter of economic or commercial reality. Apart from potential difficulties in determining where the economic burden of an unjust enrichment has fallen, a fundamental difficulty with that approach is that it conflicts with the purpose of restitution, which is “not to compensate for loss, but to reverse the defective transfer”.

Lord Reed’s analysis in Investment Trust Companies was applied and developed the following year in Prudential Assurance Co Ltd v Revenue and Customs Commissioners [2018] UKSC 39,[2019] AC 929 (“Prudential”). Lord Reed, Lord Hodge and Lord Mance explained at [69]-[77] that where money is paid by mistake a benefit is thereby conferred by the payment, and the transaction is normatively defective because the benefit is provided as a result of the mistake. In those circumstances an obligation arises immediately to reverse the enrichment. However, contrary to the conclusion in Sempra Metals, the defendant’s retention of the money thereafter does not result in any separate or additional transfer of value which must be reversed by an award of compound interest. Rather, the original mistaken payment creates a debt on which simple interest can be awarded under s.35A Senior Courts Act 1981. A failure to pay an amount which is legally due is not itself a transfer of value. The benefit of the use of the money is obtained from a failure to pay the debt.

As is well known, s.32(1)(c) can defer the commencement of a limitation period where the action is for “relief from the consequences of a mistake”. In such a case time will not start to run until the claimant has either discovered the mistake or could with reasonable diligence have discovered it.

It was confirmed in Test Claimants in the FII Group Litigation v Revenue and Customers Commissioners [2012] UKSC 19, [2012] 2 AC 337 (“FII 2012”) that in order to fall within s.32(1)(c), the mistake must constitute an essential element of the cause of action (see also FII 2020 at [14]). Further, FII 2020 departed from earlier authority in its analysis of what it means to “discover” a mistake. In summary, the Supreme Court departed from the House of Lords’ decision in DMG that a mistake of law was discovered only when the position was determined by a final judicial decision. The test established by FII 2020 was conveniently summarised by Sir Geoffrey Vos MR in Gemalto Holding BV v Infineon Technologies AG [2022] EWCA Civ 782, [2023] Ch 169 at [43] as follows:

“… time runs from the point in time when the claimant has discovered, or could with reasonable diligence have discovered (a) that it had been mistaken with sufficient confidence to justify embarking on the preliminaries to the issue of a writ, such as submitting a claim to the proposed defendant, taking advice and collecting evidence, or (b) its mistake in the sense of recognising that a worthwhile claim arises.”

After summarising the legislative background, Miles J considered the decisions of this court in BT CA and GMAC CA. He outlined BT’s claim for restitution, the principles to apply in deciding whether to grant summary judgment and the basis on which HMRC made their application, namely that (in respect of the main period) the Old Scheme excluded common law claims from 1 October 1978 and there was in any event no unjust enrichment, and in respect of the 9-month period that the claim was time-barred. Miles J noted at [63] that HMRC did not seek to contend that the claim for the 9-month period was excluded by the later enacted Old Scheme, and that they did not rely on the Limitation Act in respect of the main period.

Miles J decided that the question whether the Old Scheme ousted common law remedies was a short question of law which was suitable for summary determination. He applied Hickinbottom LJ’s summary in Southern Gas and listed seven features of the Old Scheme that led to the conclusion that Parliament intended it to be exhaustive (paragraphs [74]-[88]). In summary these were as follows:

Miles J concluded that BT’s submission that a common law remedy could coexist with the statutory scheme would make the various qualifications in that scheme optional for the taxpayer, which did not seem to be a sensible view of what Parliament intended (paragraph [91]). It was not material that claims for mistake of law had not been recognised when the Old Scheme was enacted (relying on Lord Reed’s rejection of that argument in Investment Trust Companies) (paragraph [92]).

The judge then addressed BT’s reliance on FII 2012, which is discussed further below. He preferred the submission of HMRC that that decision did not assist BT (paragraphs [93]-[102]). He also accepted that the absence of an express ouster was important, but it had to be considered with the other circumstances (paragraph [103]).

Miles J then went on to consider HMRC’s argument that there was no unjust enrichment on the basis that BT had not overpaid output tax but had simply failed to make a claim. He concluded that BT had a realistically arguable case on that point, and it was not suitable for summary determination (paragraph [119]). Similarly, the restitutionary claim for the 9-month period, in respect of which HMRC relied only on limitation, should not be determined summarily. BT had a realistically arguable case that any enrichment of HMRC was the consequence of a mistake, and as regards discoverability there were factual issues which should be resolved with the benefit of further evidence (paragraph [134]).

There are two preliminary points on which we sought clarification during the hearing.

The first point, relevant to issues 2 and 3, was what BT maintained was its mistake. BT’s pleaded case refers to a mistaken overpayment of VAT, and a Part 18 response refers to BT having “mistakenly [taken] the UK legislative position to be correct”. In oral submissions Mr Cordara identified the mistake as being BT’s belief that the UK domestic regime was compliant, and specifically that the insolvency condition was valid. However, some elements of his submissions also focused on BT mistakenly “missing the train” by failing to bring a claim under the Old Scheme before the enactment of s.39(5) FA 1997.

In relation to the 9-month period, Mr Cordara submitted that it could not be viewed in isolation from the main period. The mistaken belief that the insolvency condition was valid meant that BT was also mistaken in relation to the 9-month period, on the basis that any directly effective entitlement in respect of that period would have been thought to be no better than BT’s rights under the Old Scheme in respect of the main period. Put another way, the (assumed to be valid) Old Scheme with its insolvency condition provided the yardstick for the 9-month period.

The second point was exactly what was covered by the 9-month period. Given that quantum has been agreed, the lack of clarity about this is somewhat surprising. However that may be, it was unclear to us whether the amount attributable to the 9-month period (calculated as around £900,000, according to Miles J’s costs judgment and BT’s skeleton argument for this appeal) related to (a) supplies made during that period which gave rise to bad debts (either before or after the end of that period) or (b) bad debts arising during that period. A helpful explanation given by Ms Frawley, junior Counsel for BT, indicated that the calculations had been done on the basis of bad debts estimated as having arisen during the 9-month period, and that for those purposes debts had been treated as going bad after six months.

That suggests that the figure in dispute for the 9-month period would include bad debts arising in respect of supplies during the last six months of 1977, before Article 11C(1) came into effect on 1 January 1978, as well as supplies made during the first three months of 1978. Although Ms Mitrophanous KC, for HMRC, suggested that Article 11C(1) could have no effect in relation to supplies made before it took effect, it is not necessary to resolve that point and I will proceed on the basis of Ms Frawley’s explanation.

Mr Cordara’s submissions had some central themes, of which the most prominent were the following:

I cannot accept Mr Cordara’s submission that issue 1, which involves an exercise of statutory construction, needs to be conducted in this case in a manner affected by any EU law requirement to provide an effective remedy. That submission does not take proper account of the decisions reached and the reasoning in the statutory appeal, and in particular BT CA.

BT CA did not proceed on the basis that there was (or even might be) an alternative common law remedy in respect of the main period. Indeed, the only references to common law restitution are at [86] and [118], which relate to the 9-month period. Instead, what BT CA determined was that BT’s directly effective EU law rights were addressed by “moulding” the Old Scheme to exclude the insolvency condition. What was fatal to BT’s claim for relief under the Old Scheme was the court’s conclusion that s.39(5) FA 1997 did not fall to be disapplied, with the result that its claim under the Old Scheme in respect of the main period was out of time. That issue turned precisely on whether disapplication was required to provide an effective remedy under EU law. The Upper Tribunal had held that it was required and the Court of Appeal disagreed, with the result that BT’s claim under the Old Scheme failed.

It is obviously correct that BT’s statutory appeal did not determine that BT had no common law remedy. Indeed, there would have been no jurisdiction to make that determination in the context of a statutory appeal. But what BT CA did determine was, as Christopher Clarke LJ put it in BT CA at [138]:

“… the question …whether s 39(5) of the Finance Act 1997 must be disapplied because it amounted to an unlawful curtailment of the exercise by BT of its directly effective EU rights under the Old Scheme.”

The conclusion that s. 39(5) did not require to be disapplied was reached, and could only have been reached, on the basis that BT’s EU law rights were regarded as sufficiently protected in respect of the main period by the existence of the Old Scheme until s.39(5) was enacted. The Court of Appeal clearly did not have in mind the existence of a common law remedy in respect of the main period as a means of enforcing BT’s directly effective rights. Rather, Rimer LJ (with whom Kitchin and Christopher Clarke LJJ both agreed) reached his conclusions on the basis that the Old Scheme had provided a suitable mechanism to enforce BT’s directly effective rights, and that BT could not properly complain about the fact that they could no longer be exercised after s.39(5) was enacted. This is clear from the critical paragraphs of Rimer LJ’s judgment on this issue:

“119.

As for BT’s claims in respect of supplies during the period 1 October 1978 to 31 March 1989, I respectfully disagree with the Upper Tribunal that section 39(5) did anything amounting to an unlawful curtailment of the exercise by BT of its directly effective EU rights under the Old Scheme.

120.

I admit to having derived little assistance in the resolution of this issue from the European authorities to which we were referred. I also regard the case as having little about it that is akin to what happened in Fleming. I would, however, agree with Mr Lasok [for HMRC] that, if at the time of the introduction of the New Scheme, BT had any expectations as to the future of the Old Scheme, it would, as a prudent and circumspect operator, be likely to have foreseen its eventual repeal. I also accept his submission that a claim by a litigant to a national court inviting a disapplication of a provision of its national law as infringing the litigant’s EU law rights is one that must be decided on the basis of the facts affecting the particular litigant, in this case BT.

121.

The essential question for us is, I consider, whether BT is right that the enactment of section 39(5) of the FA 1997, which barred the making of any Old Scheme bad debt relief claims after 19 March 1997, infringed its directly enforceable rights under the Directive to claim VAT bad debt relief in respect of the bad debts the subject of the claim that BT eventually made in March 2009. Those debts all arose from supplies made prior to 31 March 1989, 20 years earlier. I explained in [48] above how, having regard to the nature of BT’s business, the badness of the latest bad debts to accrue would have been apparent to BT within (at most) months of the end of 1989 – and within several years before 19 March 1997 (and see also the last part of [105] above).

122.

It had, therefore, been open to BT from the dawn of the Old Scheme in 1978 down to 1990 to make bad debt relief claims under the machinery of the Old Scheme in respect of each bad debt now relied upon as it arose; and it continued to be open to it to make belated such claims during the remaining years of the 1990s in which the Old Scheme machinery remained on the statute book. Of course, in making good such claims BT would have had to show that the insolvency condition was incompatible with its EU law rights. I consider, however, that we must approach the case on the basis that it could and would have done so. The only reason BT did not make such claims is, I presume, because it was unaware that it was open to it to do so.

123.

I do not understand how such unawareness can be a relevant consideration. EU law has been flowing up our estuaries since 1972 and BT had every opportunity to obtain the most expert advice as to its rights. I therefore fail to understand how BT can now say that the eventual demise by the Finance Act 1997 of a bad debt scheme that had included provisions that, so it claims and I would hold, infringed its directly enforceable EU rights was a change in the law that also infringed its directly enforceable EU rights. It did not. BT had literally had almost decades in which to enforce its rights, but did nothing towards doing so. The suggestion that the four-month warning of the impending change in the law was too short a warning for BT, or those in a like position, is one with which I also disagree. BT could in fact have sought to enforce its directly enforceable EU rights during that period, although in the event it still did nothing towards doing so for a further 12 years. It is in my view counter-intuitive that BT should now be entitled to bring such a stale claim. The Upper Tribunal’s further suggestion that the four-month warning was insufficient to enable traders to seek the opportunity, should they wish to, of satisfying the domestic insolvency condition may be correct as a matter of fact. But the more relevant question is, I consider, whether the enactment of section 39(5) is one that infringed BT’s directly enforceable EU rights to claim bad debt VAT relief in respect of its supplies made prior to 31 March 1989. In my judgment, it did not.”

Section 39(5) FA 1997 did not establish a limitation period, but it had an analogous effect and similar principles must apply. As Lewison LJ explained in Leeds City Council v Revenue and Customs Commissioners [2015] EWCA Civ 1293, [2016] STC 2256 (“Leeds CC”) at [41], citing Fantask A/S v Industriministeriet (Erhvervsministeriet) (Case C-188/95) [1998] All ER (EC) 1, [1997] ECR I-6783 at [52], the fact that a Directive has not been properly transposed into UK law does not preclude taxing authorities from relying on a limitation period against a person who seeks to assert a directly effective EU law right. Further, and as Lewison LJ went on to say at [43] with reference to Lord Sumption’s judgment in FII 2012 at [151] and BT CA, there is no EU law requirement for a limitation period to be deferred, and ignorance of legal rights is not a basis to disapply a limitation period.

The conclusions in BT CA were echoed in Floyd LJ’s discussion of s.39(5) FA 1997 in GMAC CA. The points made included that:

“127.

…it is common ground that the right to claim a reduction in the case of non-payment is not absolute. It is open to member states to impose formal conditions on the exercise of the right, and to subject its exercise to reasonable time limits. It is also open to member states to make changes to the scheme, whether by changing the conditions, or by exercising, in a proportionate way, its power to derogate from the right. To that extent, therefore, the debate about whether the regime is a mandatory or discretionary one is not susceptible of a binary answer. It is mandatory in the sense that effect must be given to the EU law right, but the manner in which that is done, in terms of conditions and time limits, affords the member state a margin of discretion. Taxpayers cannot rely on the conditions for relief remaining the same.”

Floyd LJ also pointed out at [130] that the effect of s.39(5) was the same as an alteration in time limits, which was the kind of procedural hurdle that was within the discretion of the member state (subject to the principles of effectiveness and protection of legitimate expectations), and noted at [131] the “prolonged crossover arrangement” between the Old Scheme and the New Scheme. He then said this:

“133.

Against that background, was the exercise of GMAC’s EU law rights rendered excessively difficult or virtually impossible by section 39(5)? I do not consider that it was. GMAC had more than adequate time to exercise their EU law rights and were given adequate notice of the withdrawal of the scheme. It is no answer to say that GMAC’s claim would have been rejected by the Commissioners in the same terms as their actual claim was rejected when it was eventually made in 2006. As I have said, this part of the case must be approached on the basis of EU law for which GMAC contend and which I have held to be correct.

134.

I do not therefore consider that it is necessary for the court to find some other route to give effect to GMAC’s EU law rights, so as to avoid collision with section 39(5)...”

It follows from BT CA and GMAC CA that BT and other taxpayers in its position had until the repeal of the Old Scheme to bring claims under it based on the direct effect of Article 11C(1). That was found to be compatible with their EU law rights, and there is no need to identify any other remedy for the pursuit of those rights.

It also follows that Mr Cordara’s reliance on FII 2012 to support the argument that the exercise of determining whether the Old Scheme provided an exclusive and exhaustive remedy should be affected by EU law considerations is misplaced.

The relevant statutory provision considered in FII 2012 was s.33 Taxes Management Act 1970 (“TMA”), which at the relevant time provided a limited statutory right to repayment of tax paid by mistake. Among other things it precluded recovery if the tax return in question had been made in accordance with generally prevailing practice. The question which the Supreme Court was required to consider was whether s.33 provided an obstacle to recovery by the test claimants as an exclusive remedy and, if so, whether it could be given a conforming interpretation under the Marleasing principle (Marleasing SA v La Comercial Internacional de Alimentación SA (Case C-106/89) [1990] ECR I-4135). That issue, which formed a small part of the overall appeal, was addressed by Lord Walker and Lord Sumption, with whom their other Lordships agreed on this point.

The Court of Appeal had applied a conforming interpretation of s.33 by reading the proviso relating to prevailing practice as subject to a limitation that it applied only to the extent that it was consistent with Treaty obligations. Lord Walker doubted that s.33 could be given a conforming interpretation in that form on the basis that the prevailing practice requirement appeared to be a cardinal feature of the legislation. He preferred a simpler approach under which s.33 was read as non-exclusive, saying at [119]:

“In my view the Marleasing principle can be applied in a simpler and more natural way by not construing section 33 as impliedly setting itself up as an exclusive provision (which it did not do expressly, unlike section 80 of the Value Added Tax Act 1994). The test claimants submit that the application of Marleasing cannot rework section 33 in a way that serves any relevant purpose. But to read it as non-exclusive does not go against its grain. It would merely exclude an implication which is itself no more than a process of statutory construction.”

Lord Sumption also considered that s.33 should be treated as non-exclusive, saying this:

“[204] … No one suggests on this appeal that such a limited remedy could possibly be enough in itself to satisfy the virtually unqualified obligation of the United Kingdom to provide an effective means of recovering tax overcharged contrary to EU law. This does not of course matter if it is an additional remedy as opposed to an exclusive one. There is certainly nothing in the provision which expressly excludes the availability of other causes of action at common law. If that is its effect, it must be by implication. In the ordinary way, such an exclusion might be implied, on the ground that where Parliament confers a restricted right of recovery, that must impliedly displace a corresponding right at common law which would be unrestricted. However, it is axiomatic that the courts cannot imply an exclusion of unrestricted rights of action at common law where that would be inconsistent with an overriding rule of EU law that an unrestricted right must be available. Section 33 cannot therefore be an exclusive right to recover tax overcharged contrary to EU law. Whether it is an exclusive right in other circumstances, is not a point which needs to be considered on this appeal.

[205] The Court of Appeal held that section 33 did impliedly exclude a right of action at common law, even in relation to claims for tax overcharged contrary to EU law. They then dealt with the resulting inconsistency with EU law by reinterpreting the section so as remove the offending restrictions and the element of discretion. I think that this was wrong in principle. I very much doubt whether such radical surgery can be justified even under the extended principles of construction authorised in Marleasing. Its effect would be fundamentally to alter the scheme of the provision. But, however that may be, it seems, with respect, eccentric to imply an ambit for section 33 which is inconsistent with EU law and then to torture the express provisions so as to deal with anomalies that but for the implication would never have arisen.”

As can be seen from these passages, both Lord Walker and Lord Sumption based their approach on what was required to provide an effective EU law remedy. If s.33 appeared to stand in the way of that then it had to be interpreted so as not to do so, and that could be done by treating it as non-exclusive so as to permit the remedy that EU law required.

That is not this case. BT CA determined that the “moulded” Old Scheme provided an effective remedy. To use Lord Sumption’s analogy, surgery has already been performed. No further surgery is required to treat the Old Scheme as non-exclusive.

Mr Cordara also sought to rely on Autologic Holdings plc v Revenue and Customs Commissioners [2005] UKHL 54, [2006] 1 AC 118 as a case where a statutory scheme did not exclude a common law remedy in an EU context. That case does not assist BT. It primarily concerned a procedural question. The House of Lords decided by a majority that claimants were first required to seek to vindicate their EU law rights by exhausting the statutory appeals procedure, because the alternative of seeking relief through the High Court would circumvent that and amount to an abuse of process. Only where claims could not be dealt with by that route could other relief be sought. That included cases where the claimant was too late to use the statutory procedure (see at [39], per Lord Nicholls), but the focus in that case was not on whether the statutory scheme precluded such other claims altogether and, apart from a brief reference in the dissenting judgment of Lord Walker at [126], the point appears not to have been considered.

Mr Cordara’s second overarching point was that regard had to be paid to the nature of the remedy sought to be excluded, and specifically to the mistake made in this case, which related to the words of the statute in question and only became apparent long after BT’s ability to access the Old Scheme had been removed.

I cannot accept this submission. Quite apart from the notion that Parliament might be more inclined to intend to permit claims based on a mistaken failure to use the scheme rather than other types of claim being one that is somewhat difficult to grasp, as I have sought to explain (see the summary at [35] above), the question is whether Parliament intended the Old Scheme to be exclusive. That is not a question that can or should be answered by reference to a specific alternative remedy, and still less one that had not even been identified at the time that the Old Scheme was enacted.

BT’s real complaint is about the enactment of s.39(5) FA 1997 before it appreciated that it had a claim. I would give two answers to that. First, s.39(5) has been held not to conflict with BT’s directly effective rights: see above. Secondly, to the extent that what is being said is that the Old Scheme would have provided an appropriate remedy while it existed, but once it has gone then other remedies should be treated as available, then that cannot assist BT. The question whether Parliament intended the Old Scheme to be exclusive must necessarily be answered on the hypothesis that the scheme exists.

Having addressed Mr Cordara’s overarching submissions, I now turn to the interpretation of s.12 FA 1978.

In determining that the domestic bad debt relief scheme could be “moulded” to comply with EU law, the UT found in GMAC/BT UT at [165] that the “Old Scheme was intended to reflect the obligations of the UK under Article 11C”. That is obviously correct. The Old Scheme was enacted nine months after the provisions of the Sixth Directive were required to be implemented. The predecessor directive included no equivalent to Article 11C(1) (see Grattan v Revenue and Customs Commissioners (Case C-310/11) [2013] STC 502 at [35]-[36]).

As Ms Mitrophanous submitted, this is material. Parliament clearly gave careful consideration to the terms on which it was prepared to permit the new relief to be granted. They included important limitations in the form of the insolvency condition and the property condition. They also included the following:

The original regulations were the Value Added Tax (Bad Debt Relief) Regulations 1978, SI 1978/1129. They illustrate the significance of some of these points, for example by requiring a claim to be made only for the period in which proof of insolvency was received (albeit that, as the judge found, the Commissioners did not insist on that until 1991 when a later set of regulations were introduced), by provisions which determined which supplies should be treated as having been paid for in the event of part payment by a debtor, for a deemed set off of amounts owed by the claimant to the debtor and for a requirement to repay to the Commissioners any refund obtained in the event of a failure to comply with the duties imposed by the regulations to preserve and produce records.

It is true that the most material conditions were the insolvency condition and the property condition, and that a number of the procedural requirements related to the insolvency condition. But I think the judge was right to conclude that it was not correct to disregard those elements because they have subsequently been found not to comply with EU law. As he said at [82], to ignore those parts of the Old Scheme would be to “exclude significant indicators of what Parliament meant to achieve”. It does not seem to me to be realistic to attempt to enquire, for example, into what Parliament might have intended to happen in the event that any of the conditions it was imposing proved unlawful. The intention of Parliament must be ascertained by reference to the words actually used, and not on the basis of the text as it might read following a “moulding” process to comply with EU law.

It is the case that the Old Scheme covers the same ground as the common law claim that BT seeks to advance. Fundamentally, BT seeks bad debt relief pursuant to Article 11C(1). The Old Scheme was enacted to give effect to the bad debt relief required to be conferred by that provision.

Further, Parliament chose to do so on qualified terms, in particular the insolvency condition but also the property condition, the detailed requirements that were to be governed by regulations and the route of appeal. Those qualifications would have little purpose if a taxpayer could simply circumvent them by bringing its claim in another form. They are clear indications that the Old Scheme was intended to be exclusive. An ability to circumvent them by bringing another form of claim would be incompatible or inconsistent with the statutory scheme.

As the authorities show, although the position is obviously more straightforward where there is an express provision for exclusivity and its absence is a relevant factor, such a provision is by no means essential. Littlewoods provides a powerful illustration of that in relation to s.78 VATA 1994, despite its proximity to the express exclusivity provision in s.80(7).

Mr Cordara sought to rely on the fact that other provisions of Article 11C(1) were not the subject of domestic legislation (apart from a regulation introduced years later to deal with price reductions: see [85] below) as indicating that Parliament must be taken to have permitted the existence of other remedies. I prefer Ms Mitrophanous’s submission that this point assists HMRC more than BT. Parliament chose to enact a specific statutory scheme to set out the conditions under which bad debt relief could be claimed. Given the practical difficulties in determining whether a debt is bad and when that occurs, that is unsurprising. In contrast, it is not generally difficult to determine whether a price has been reduced or a contract has been cancelled.

The judge also referred to the Old Scheme being expressed to be separate from the procedure for recovering overpaid VAT pursuant to s.80 VATA 1994 and its predecessor, and to the enactment of s.39(5) FA 1997 (at [87] and [88]: see [42] (vi) and (vii) above). I would not myself rely on either of those points because they materially post-date the enactment of the Old Scheme. As far as s.39(5) is concerned the most that can be said is that its existence illustrates that any scheme enacted by Parliament may be altered, or indeed removed, by it, and that Parliament must be taken to have been aware of that.

In conclusion on this issue, I consider that the judge was correct to conclude that Parliament cannot be taken to have intended that other means of seeking bad debt relief should coexist with the Old Scheme.

The conclusion that the Old Scheme provided an exclusive remedy means that it is not strictly necessary to address the question of whether HMRC were unjustly enriched in respect of the main period, which was the context in which this issue was raised by HMRC’s Respondent’s notice. However, it is necessary to consider the question of whether HMRC were unjustly enriched in respect of the 9-month period. What follows reflects the fact that the submissions we heard on this issue were directed primarily at the main period, but both parties also sought to rely on them in HMRC’s cross-appeal in relation to the 9-month period. I will deal in more detail with the latter.

In outline, HMRC’s case is that BT’s complaint is one of failure to make a claim in time, or “mistaken failure to claim”. BT cannot (and do not seek to) argue that the VAT originally accounted for on the supplies was not due for the period in which the supplies were made. Although BT could have claimed bad debt relief in a later period by setting it off against VAT due on later supplies, it did not do so. HMRC maintain that that involved no transfer of value beyond the initial payment.

Ms Mitrophanous used an example of BT accounting for £100 of VAT at time T1. At T2 BT reaches the view that £10 of this relates to debts that have become bad. BT makes further supplies and at T3 accounts for a further £150 of VAT in respect of those supplies without setting off the £10 from T1 that relates to bad debts. Ms Mitrophanous submitted that in doing so BT had simply failed to make a claim, and that involved no transfer of value. It is true that HMRC were better off as a result of BT’s failure, but as Investment Trust Companies explains the test is not merely one of economic or commercial reality.

Mr Cordora maintained that in this example HMRC were enriched by £10 and that this occurred at, or at least no later than, T3. Mr Cordara relied on Iveco Ltd v Revenue and Customs Commissioners [2017] EWCA Civ 1982, [2018] STC 364 (“Iveco”) and DMG to support BT’s argument that HMRC were unjustly enriched on the basis that, absent its mistake, BT would have paid £140 and not £150 at T3. Alternatively, Mr Cordara relied on Gibb v Maidstone & Tunbridge Wells NHS Trust [2010] EWCA Civ 678 as authority for the proposition that forbearance to bring a valid claim can amount to enrichment. He also submitted that HMRC’s case relied on the “machinery” of the VAT system, wrongly overlooking the distinction between the procedures for assessing and collecting tax on the one hand and the underlying liability on the other. I will deal with these points in turn.

Iveco was a decision of this court concerning the part of Article 11C(1) that dealt with price reductions. The appellant, Iveco, brought a statutory appeal against HMRC’s refusal to refund VAT related to price rebates granted by Iveco between 1 January 1978 and 31 December 1989. During that period no domestic legislation had been in place implementing Article 11C(1) in relation to price reductions. A provision was first included in the VAT regulations with effect from 1 January 1990 (see now regulation 38 of the 1995 Regulations).

Iveco’s claim was first made to HMRC in November 2011, and the issue which was appealed to this court was whether its claim was time-barred. Reversing the FTT, the UT had held that it was, on the basis that Article 11C(1) had direct effect at the point that rebates were paid, that s.80 VATA 1994 provided the means to enforce Iveco’s directly effective rights, and that Iveco’s claims were time-barred by reference to that provision.

Iveco appealed, contending that there was no automatic reduction in the taxable amount for the purposes of Article 11C(1) at the point that the price was reduced, but instead that it had a directly effective right to bring about a reduction in the taxable amount at a time of its choosing. The appeal was dismissed. Newey LJ considered both BT CA and GMAC CA, and at [38] rejected Iveco’s attempt to rely on a passage in the judgment of Floyd LJ in GMAC CA on the basis that Floyd LJ was considering bad debt relief, which “has always been something that a creditor must claim”. He concluded that there was nothing in Article 11C(1) to suggest that a reduction in the taxable amount could legitimately be deferred to any great extent, and on the contrary it required a refund to be available within a reasonable period. The United Kingdom was therefore required to provide for a reduction in the taxable amount at or soon after the time when a rebate was paid, and suppliers were therefore entitled to enforce their directly effective rights by making a negative entry in their VAT account when the rebate was paid. On that footing there was an overpayment of VAT (paragraphs [46] and [47]). However, that did not translate into a right to reduce the taxable amount whenever the taxable person chose (paragraph [48]).

In determining how Iveco’s Article 11C(1) right should be given effect, Newey LJ concluded at [54] that its remedy was to claim to recover overpayments under s.24 FA 1989 and (later) s. 80 VATA 1994. Iveco was out of time to do so. In the light of that conclusion it was unnecessary to consider whether some other form of remedy would have been required if Iveco had sought to enforce its rights before s.24 FA 1989 came into force (paragraphs [56] and [57]).

Mr Cordara submitted that Iveco demonstrates that, in the absence of effective domestic legislation, Article 11C(1) applied automatically and was not neutered by a requirement to make a claim as HMRC maintained. As the CJEU had said in Minister Finansów v Kraft Foods Polska SA (Case C-588/10) EU:C:2012:40, [2012] STC 787:

“26.

… it must be noted that art 90(1) of the VAT Directive [i.e. Article 90 of the Principal VAT Directive, previously Article 11C(1)] requires the member states to reduce the taxable amount and, consequently, the amount of VAT payable by the taxable person whenever, after a transaction has been concluded, part or all of the consideration has not been received by the taxable person (see Goldsmiths (Jewellers) Ltd v Customs and Excise Comrs (Case C-330/95) [1997] STC 1073, [1997] ECR I-3801, para 16).

27.

That provision embodies one of the fundamental principles of the VAT Directive, according to which the basis of assessment is the consideration actually received and the corollary of which is that the tax authorities may not in any circumstances charge an amount of VAT exceeding the tax paid by the taxable person (see, to that effect, Goldsmiths (para 15)).”

The difficulty with BT’s reliance on Iveco in respect of the main period is twofold:

The analysis in relation to the 9-month period is somewhat different. Although HMRC’s position in the proceedings as a whole is that the moulded Old Scheme can be “read back” to apply to the 9-month period, for the purposes of their strike out application they place no reliance on the Old Scheme. On that footing there was therefore no domestic implementing legislation in place for the 9-month period that provided for bad debt relief, and in that respect there is a direct parallel with Iveco. If the reasoning in Iveco were applied, therefore, BT would have been able to rely on its directly effective rights to obtain bad debt relief, including by setting off the £10 at T3 in the example above. In that respect Ms Mitrophanous correctly accepted that there is a distinction between the 9-month period and the main period.

For the main period there was a domestic scheme in place, and the lawfulness of the feature that it required a claim to be made has never been put in issue. That feature meant that s.80 was not engaged in respect of bad debt relief in respect of the main period, as BT CA determined. In contrast, on the hypothesis that the Old Scheme is not read back to the 9-month period there would have been no requirement for a claim, and it could then be argued that in the example only £140 was owed at T3, because the £10 was to be treated as automatically set off against the VAT of £150 that would otherwise have been due.

However, it does not follow from this that BT can successfully rely on Iveco in respect of the 9-month period, because point (b) at [90] above still applies. Iveco’s claim under s.80 failed because it was out of time. BT made an in time claim under s.80, but the effect of BT UT 2023 is that its claim under s.80 has been finally determined against it by the statutory appeal being struck out.

The reasoning relating to s.80 is put shortly in Rimer LJ’s judgment in BT CA. In summary, he agreed with the UT and with the submissions of Mr Lasok for HMRC. He said this:

“[126] The Upper Tribunal’s view, succinctly expressed at [181], was that s 80 applies to cases where the taxpayer has brought into account as output tax an amount that was not output tax due. When GMAC (and likewise BT) made its supplies, it accounted for tax which was then due. The subsequent failure of the customer to pay for the supply gave rise to a bad debt, and a possible claim for bad debt relief, which would be for the repayment of all or part of the output tax originally paid by BT. The arising of such bad debt did not, however, mean that the output tax earlier paid was not output tax due within the meaning of s 80. It was and remained so, and the arising of the bad debt did not retrospectively change that.

[127] Mr Lasok supported the Upper Tribunal’s reasoning and I too agree with it. Not even BT seems to have believed that s 80 was relevant. Its somewhat ill-drawn claim letter of 30 March 2009 was in respect of bad debt relief and made no suggestion that it had made any payment of tax which was not output tax due. This, I take it, is the sense of the second sentence of the Upper Tribunal’s answer to preliminary issue 3.”

(Preliminary issue 3 was whether s.80 VATA 1994 and s.121 FA 2008 applied to BT’s claims.)

Earlier in his judgment Rimer LJ had recorded Mr Lasok’s submissions in the following terms:

“[30] By way of anticipation of the arguments on the cross-appeal, Mr Lasok’s responsive submission was that s 24 and its successors have nothing to do with claims for bad debt relief. They were and are concerned only with cases in which an overpayment of tax had been made to the Commissioners and provide a restitutionary basis for the recovery of the excess. In a bad debt case, there was no payment of VAT that was not due, or therefore any overpayment of VAT. A failure to make a bad debt relief claim in a subsequent return still did not mean that there was any overpayment in the tax actually paid. If no relief claim was made, the tax paid was the tax due.” (Emphasis supplied.)

(The cross-appeal was BT’s challenge to the UT’s decision on preliminary issue 3.)

Ms Mitrophanous submitted, and Mr Cordara did not dispute, that Rimer LJ accepted Mr Lasok’s submissions in these terms, and that BT had not disputed that that was the case in submissions to the UT in BT UT 2023. I agree. It explains why the second half of the passage above is cited in BT UT 2023 at [52] on the basis that it records Rimer LJ’s rejection of the argument that, although VAT was properly due at T1 (in the example above), there was an automatic set-off at T3 such that there was an overpayment for the purposes of s.80 at that point.

The point was an important one. Mr Cordara’s submissions in BT UT 2023 were to the effect that, even though BT CA had determined that there was no overpayment at T1 pursuant to s.80(1), the (differently worded) s.80(1B) was not addressed, and that provision, together with the later guidance in Iveco, enabled BT to claim that there was an overpayment at least by T3.

The UT determined at [47] that Rimer LJ’s judgment extended to s.80 as a whole, including s.80(1B). In considering Iveco, it also recorded at [53] its acceptance of HMRC’s submission that there are good reasons to treat bad debts separately from other reductions in consideration, including a lack of clarity as to when a debt becomes bad. The UT’s conclusion as to what BT CA decided in relation to s.80 was expressed in the following terms at [65(c)]:

“Neither s80(1) nor s80(1B) can be utilised for the purposes of claiming relief for bad debts.”

Importantly, the UT’s conclusions draw no distinction between the 9-month period and the main period as far as s.80 is concerned. Indeed, the UT expressly confirmed at [57] that BT CA had determined the claim for both periods. BT’s statutory appeal was accordingly struck out in full. The reason that no distinction was drawn is clear from an extract from the transcript of Mr Cordara’s submissions to the UT that Ms Mitrophanous showed us. Those submissions focused solely on the main period and indicated that BT was “not concerned” in the statutory appeal with the 9-month period, which had “very little money in it”.

It seems to me that this places an insuperable obstacle in the way of BT’s argument based on Iveco, in the form of s.80(7), set out at [21] above. As Lewison LJ said in Leeds CC at [13], s.80 is intended to be a complete statutory code for the repayment of overpaid VAT. Section 80(7) excludes other remedies for overpaid VAT. Iveco cannot therefore assist BT in relation to the 9-month period. This is because, if BT was right in its submission that Iveco demonstrates that BT’s directly effective rights for the 9-month period meant that it had overpaid VAT, the only route that it would have to vindicate those rights would be under s.80. However, its claim under s.80 has been finally determined, and in circumstances where BT chose not to press its claim for the 9-month period. Having made that choice it cannot assert a claim that is expressly excluded by s.80.

I should also refer to Floyd LJ’s comments in GMAC CA about s.80 and what at that time was the UT decision in Iveco. Floyd LJ confirmed the conclusion in BT CA that s.80 is not engaged in respect of bad debt relief. He said this at [134] (immediately after the passage cited at [58] above):

“…I would simply record my view, which is in conformity with the view which GMAC expressed to the Commissioners in their original claim, that section 80 is not the appropriate domestic provision for giving effect to bad debt relief. When GMAC accounted for VAT on the whole value of the supply it did not account for VAT which was not due. That did not change at the point when GMAC considered the debt to be bad. To that extent, to the extent they are different, I prefer the views of the UT expressed in the present case to those expressed in Iveco.”

Although we did not hear detailed argument on the point and I do not need to rely on it, it seems to me that there is also a broader objection to BT’s reliance on Iveco based on res judicata and, potentially, abuse of process. The statutory appeal determined that BT did not have a s.80 claim on the basis that there was no overpayment of VAT. The conclusion that there was no overpayment was a necessary ingredient of the outcome of the appeal on that issue. That seems to me to fall within the classic definition of issue estoppel (see Virgin Atlantic Airways Ltd v Zodiac Seats UK Ltd [2013] UKSC 46, [2014] AC 160 at [20] per Lord Sumption, citing the speech of Lord Keith in Arnold v National Westminster Bank plc [1991] 2 AC 93). Although there are very significant limitations on issue estoppel in a tax context because of the way that tax is charged on a periodic basis (see Caffoor v Income Tax Commissioner [1961] AC 584, discussed in detail in a VAT context by Henderson J in the High Court in Littlewoods [2014] EWHC 868 (Ch), [2014] STC 1761), in this case the common law claim relates to precisely the same tax as was the subject of the statutory appeal. For that reason it seems to me that it is strongly arguable that it is simply not open to BT to raise arguments now based on an alleged overpayment of VAT. And if that were not correct, then the abuse of process principle derived from Henderson v Henderson (1843) 3 Hare 100 at 115 would in any event appear to be relevant. In respect of the 9-month period, BT chose not to pursue the argument that it overpaid VAT, even though it must have been aware that there was a difference in the analysis between that period and the main period.

DMG is an important case that established that English law recognises a restitutionary claim for tax paid under a mistake of law. It remains authority for that proposition, although it was departed from in FII 2020 as regards its conclusions about when the relevant mistake was discoverable.

The context of DMG was the ECJ decision in Metallgesellschaft Ltd v Inland Revenue Commissioners and Hoechst AG v Inland Revenue Commissioners (Joined Cases C-397 and 410/98) [2001] Ch 620 (“Metallgesellschaft”). In Metallgesellschaft the ECJ had determined that the regime under which UK subsidiaries could pay dividends to their UK parents without accounting for advance corporation tax (“ACT”) by making a group income election (“GIE”), but could not do so where the parent company was based in Germany because a GIE was unavailable, was contrary to European law.

The key parts of the decision in Metallgesellschaft are set out in Lord Walker’s speech in DMG. In principle, member states are required to repay “improperly levied” charges (ECJ decision at [84] and [85]). However, because the ACT in that case had subsequently been set off against corporation tax, the breach arose from the tax being levied prematurely rather than from the payment of tax itself. The ECJ ruled at [87] that in those circumstances an award of interest appeared to be “essential” to restore equal treatment. Further, the UK’s argument that it could rely on the failure of the taxpayers to attempt to make GIEs was rejected as contrary to the principle of effectiveness (at [106]):

“The exercise of rights conferred on private persons by directly applicable provisions of Community law would, however, be rendered impossible or excessively difficult if their claims for restitution or compensation based on Community law were rejected or reduced solely because the persons concerned had not applied for a tax advantage which national law denied them, with a view to challenging the refusal of the tax authorities by means of the legal remedies provided for that purpose, invoking the primacy and direct effect of Community law.”

In DMG there was no statutory mechanism that provided the remedy that the ECJ required domestic law to provide (the House of Lords having rejected an argument based on s.33 TMA). In those circumstances it is not surprising that the Commissioners conceded that some form of restitutionary remedy was required, albeit only by analogy with the Woolwich principle (Woolwich Equitable Building Society v Inland Revenue Commissioners [1993] AC 70), which would have been subject to a six-year time limit (see [7], [107], [135] and [160]). DMG had pleaded that it was entitled to restitution based either on an unlawful demand (i.e., Woolwich) or money paid under a mistake of law. The decision of the majority was that DMG was effectively entitled to choose the latter and benefit from a longer limitation period.

As to the existence and nature of the mistake, Park J had accepted the evidence of Mr Thomason, DMG’s head of tax, that he believed that he was bound to act in accordance with the UK statute. Park J had also concluded that the mistake was not about whether ACT was payable but whether DMG should have been permitted to make a GIE.

Lord Hoffmann observed at [20]-[21] that the effect of Metallgesellschaft was that “the Inland Revenue had not been entitled to the money”, and English law provided a remedy in unjust enrichment for money paid without a legal basis. Instead of Park J’s “rather sophisticated” view of the nature of the mistake, the mistake was “about whether DMG was liable for ACT”, the election provisions being “purely machinery” to enforce DMG’s right to exemption ([32]).

Lord Walker reached a similar conclusion, commenting at [143] that the judge’s approach was “rather over-analytical” and that the procedural requirement for a GIE did not alter the substance of the mistake.

In contrast, at [62] Lord Hope accepted Park J’s analysis as to the nature of the mistake but concluded as Park J had that the ACT was nonetheless paid by mistake because the mistake about the availability of a GIE led inevitably to the liability to pay ACT. Lord Brown’s judgment does not cast much further light on the point because he agreed with the speeches of Lord Walker, Lord Hoffmann and Lord Hope on the “mistake” issue, albeit his primary focus seems to have been on the judgment of Lord Walker: see paragraph [161].

However, Lord Scott, in his dissenting judgment, concluded that the imposition of ACT was not unlawful. Rather, what was discriminatory was the GIE system. Like Lord Hope he agreed with Park J’s analysis of the nature of the mistake but decided that the ACT was lawfully paid and not paid under a mistake in a sense that justified a restitutionary remedy, as opposed to a claim for compensation to recover a loss caused by the breach of Community law (paragraphs [82], [89] and [91]).

Writing in 2012, Professor Burrows suggested that the normal approach in the law of unjust enrichment that an unjust factor does not override a legal obligation on the claimant to make a transfer of value is subject to limited exceptions, which might be explained by the legal entitlement being “easily outweighed by the unjust factor”. He sought to explain the result in DMG as an exception to the rule that restitution should be denied where the defendant has a statutory right to a payment on the basis that the defendant’s right arose only “in one technical sense” (Professor Burrows, A Restatement of the English Law of Unjust Enrichment, p.34).

Professor Burrows’ explanation of DMG is queried in Goff & Jones, The Law of Unjust Enrichment (10^th ed.) at 2-11 by reference to the fact that the tax was strictly due under the ACT system. Goff & Jones goes on to discuss the reliance placed on DMG by Thornton J in Surrey CC v NHS Lincolnshire Commissioning Group [2020] EWHC 3550 (QB), [2021] QB 896 (“Surrey CC”), where unjust enrichment was found to have arisen in circumstances where the claimant council’s payments of accommodation and care costs were made under a legal obligation, but the defendant primary care trust should have assessed the relevant individual’s needs and, had it done so, would probably have concluded that he was eligible for NHS continuing healthcare. However, the only reason why the claimant had been required to pay was because of the defendant’s unlawful refusal to do so. Thornton J applied the approach suggested by Professor Burrows to find at [115] that the claimant’s legal obligation to pay was outweighed by the unjust factors at play. But she also drew an analogy with Gibb v Maidstone, discussed below, at [121], concluding that the claimant had discharged a liability that but for an unlawful decision would have been owed by the defendant.

DMG was obviously decided well before the Supreme Court’s explanation of the concept of unjust enrichment in Investment Trust Companies, and its further decision in Prudential to depart from Sempra Metals. It was the decision in DMG that had allowed Sempra to frame its claim for interest as being one in restitution for ACT paid under a mistake of law: Sempra Metals at [19], per Lord Hope. However, the approach taken by Lord Hoffmann and Lord Walker on the nature of the mistake is consistent with the application in Prudential of Lord Reed’s analysis in Investment Trust Companies, namely that the payment of money by mistake is normatively defective and gives rise to an obligation to reverse the benefit conferred (see [38] above).

I would derive the following points from DMG relevant to this case:

In contrast, in this case a statutory remedy has already been identified that gives effect to BT’s EU law rights, at least for the main period, in the form of the Old Scheme. Further, at least for the main period the tax was not collected unlawfully. The tax was properly due in the absence of a claim: see BT CA at [30] (discussed above at [95] and [96]). The requirement to claim bad debt relief has not been treated as “purely machinery”, to use Lord Hoffmann’s words. Rather, that requirement has been found to be of real significance and has been held to prevent a claim under s.80 VATA 1994. In the light of that, if the approach suggested by Professor Burrows were to be applied, I would not describe BT’s legal obligation to pay the money as merely “technical” and would not conclude that it was outweighed by any unjust factors. BT had the opportunity to make a claim under the Old Scheme up to 1997 and did not do so. I would also add that I consider Surrey CC to be clearly distinguishable on the facts.

Again the position is somewhat different for the 9-month period. As I have said, HMRC do not rely on the Old Scheme for the purposes of their strike out application. On that basis BT’s directly effective rights to claim bad debt relief could be argued to mean that £10 of tax paid at T3 in the example above was not properly due. It would logically follow that VAT was overpaid.

However, this is not a case like DMG where there was no available statutory remedy to recover unlawfully levied (or in that case prematurely levied) tax. Overpaid VAT could be recovered pursuant to s.80 VATA 1994 and its predecessor, including (as the facts of Iveco illustrate) in respect of periods well before those provisions were enacted. Although HMRC’s position in its dispute with BT has been that s.80 has no application, that is on the basis that the requirement for a claim under the Old Scheme cannot be ignored and (I infer) also reflects the fact that HMRC’s position in the proceedings as a whole is that the Old Scheme can be “read back” to the 9-month period. Once the Old Scheme is left out of account for the 9-month period, such that BT’s claim that it has overpaid tax – and the analogy with DMG – gains traction, it seems to me that it would also follow that s.80 would have been in point, just as it would be in relation to BT’s argument based on Iveco.

On that basis, BT must again be faced with the problem of s.80(7), which has the effect that s.80 provides an exclusive regime to recover overpaid VAT. However, as already discussed BT’s claim under s.80 has been finally determined against it, including for the 9-month period.

I should clarify that I do not consider that anything in BT CA or BT UT 2023 requires a contrary conclusion. There are obiter references in BT CA to a possible common law restitutionary claim at [86] and [118] in relation to the 9-month period, but given the conclusion that s.80 had no application that is unsurprising. The fact that s.80 was held not to apply in a discussion that did not distinguish between the main period and the 9-month period cannot, it seems to me, prevent this court from analysing BT’s arguments in favour of a common law restitutionary claim and concluding that, if accepted in relation to the 9-month period, they would necessarily engage s.80. Mr Cordara did not seek to submit otherwise when Ms Mitrophanous raised the final determination of BT’s s.80 claim in the context of Iveco, and I cannot see a material difference between that and BT’s argument based on DMG. I further note that, in the grounds for HMRC’s strike out application, the point was raised in the following terms:

“It remains HMRC’s position that BT had no claim under section 80. Had there been an overpayment made by BT, mistaken or otherwise, BT would have had a section 80 claim which, as is common ground, would have ousted any common law claim (see section 80(7)). That BT has no right to bring a claim under section 80, as established in BT/CA and GMAC/CA, defeats any claim that it has enriched HMRC by overpaying VAT.”

HMRC’s skeleton argument also raises s.80 in the context of DMG. It seems to me that the point was squarely raised.

Alternatively, and more simply, the conclusion that s.80 did not apply arguably creates an issue estoppel such that BT may no longer assert that there was an overpayment of VAT, and/or engages Henderson v Henderson: see [102] above.

In Gibb v Maidstone Ms Gibb sought to enforce a contract in which she had agreed terms of severance with the respondent NHS Trust. She had been paid approximately £75,000, whereas the contract provided for a payment of approximately £250,000. This court allowed Ms Gibb’s appeal against the judge’s conclusion that the compromise agreement was ultra vires, so the court’s comments about Ms Gibb’s alternative claim in restitution were obiter. Ms Mitrophanous did not however seek to challenge their correctness.

In that alternative claim, Ms Gibb had asserted that her entry into the compromise agreement had conferred benefits on the Trust comprising giving up an unfair dismissal claim and agreeing to terms of confidentiality, to immediate cessation of employment and to the avoidance of any requirement to conduct an internal grievance procedure. Laws LJ, who gave the leading judgment, referred at [29] to the submission of Counsel for Ms Gibb that the “foregoing of a valuable claim may confer a benefit – an enrichment – on the party not so pursued” and went on to say this at [30]:

“If everything else is equal I can see no principled distinction between a benefit consisting in money paid and a benefit consisting in a claim foregone.”

Laws LJ concluded that the unjust enrichment claim would have been made out if the contract had been ultra vires, and although the non-monetary elements of what Ms Gibb gave up would be valued very modestly the benefit to the Trust of releasing the unfair dismissal claim had been valued at nearly £70,000: see at [37]. Rimer LJ agreed with Laws LJ. Sedley LJ preferred not to express a view on the restitution issue.

Gibb v Maidstone is very different to the present case. Ms Gibb signed a document in which she gave up her right to pursue a claim for unfair dismissal. Laws LJ accepted at [36] that the fact that (on the alternative case) the contract turned out to be void did not affect the fact that the Trust had received and retained that benefit. In contrast, in this case BT merely failed to exercise its right to claim bad debt relief in time. It at no stage conferred a benefit on HMRC by giving up that right. A failure to take action before s.39(5) FA 1997 was enacted cannot be equated with “foregoing” a claim in the sense meant by Laws LJ.

As already mentioned, Mr Cordara further submitted that BT’s case wrongly overlooked the distinction between the procedures for assessing and collecting tax on the one hand (the “machinery”) and the underlying liability on the other. He relied on the classic description by Lord Dunedin, in Whitney v Commissioners for the Inland Revenue [1926] AC 37 at p.52, of the three stages of imposition of tax, namely declaration of liability, assessment and recovery, and an application of those principles by Cockerill J in XL Insurance Co SE v IPORS Underwriting Ltd [2021] EWHC 474 (Comm) at [34]-[37].

I have already explained that the procedure for claiming bad debt relief has been found in BT CA not to be a question of machinery. Rather, BT’s failure to claim bad debt relief meant that the tax it paid not only at T1 but also at T3 – and indeed in subsequent periods – was tax that was properly due as a matter of liability. It follows that even if the principle explained in Whitney can be translated from direct tax to VAT (a point for which no authority was cited), it cannot assist BT in this case. For the same reason BT can gain no assistance from s.81(3) VATA 1994, which provides for a set-off of VAT due against amounts owed by HMRC, on which Mr Cordara also relied. In respect of the main period, no amount was due from HMRC by way of bad debt relief in the absence of a claim. Although the position would have been arguably different for the 9-month period, it would also logically follow that BT would have overpaid VAT for the purposes of s.80, so engaging s.80(7): see above.

It follows that in my view none of Iveco, DMG or Gibb v Maidstone assist BT in establishing that there was the required normatively defective transfer of value to found a potential claim in unjust enrichment. The VAT paid at both T1 and T3 in HMRC’s example has been found to be properly due. If that was not the case then s.80 VATA 1994 would have been engaged, including the express exclusivity provision in s.80(7). BT has also not given up a right in a way analogous to Ms Gibb’s action in Gibb v Maidstone. It retained the right to claim bad debt relief until the enactment of s.39(5) FA 1997.

As explained in Investment Trust Companies, the purpose of restitution is to reverse a normatively defective transfer of value (see [36] above). In this case there has been no transfer of value in the sense explained by the Supreme Court. BT has simply failed to make a claim for VAT bad debt relief, whether by refund or by set off against VAT it owed, within the requisite period. Neither a failure to bring a claim for a refund nor a failure to exercise a right of set off can amount to a transfer of value, in circumstances where the right is not given up. The position is no different to that of a claimant who mistakenly omits to bring a valid claim before a limitation period expires. A mere omission of that nature cannot give rise to a claim in unjust enrichment.

Although the law of restitution remains a developing area, in my view this is a pure question of law. I cannot see that allowing the case to proceed to a fact-finding trial could make any difference to the arguments on this issue or to the court’s assessment of them. In my view, a claim based on unjust enrichment has no real prospect of success. This is therefore a case where the nettle referred to by Lewison J in Easyair should be grasped.

It follows that I respectfully disagree with the judge’s conclusion on this issue. In doing so, I should note that, unlike the judge, we have the benefit of the decision in BT UT 2023, which finally disposes of BT’s s.80 claim for both the main period and the 9-month period.

For the reasons I have already set out I cannot see how BT can make out a case in unjust enrichment for the main period. Although the argument that BT overpaid VAT is potentially different for the 9-month period, any analysis based on overpayment would necessarily engage s.80(7) VATA 1994. BT therefore has no claim in unjust enrichment in respect of that period either.

In those circumstances it is not necessary to consider the question of when the mistake was discoverable under the test established in FII 2020. However, I would observe that I consider that the judge was not only entitled to conclude that a trial of that issue would have been required, but could not properly have reached a different conclusion. The test established in FII 2020 requires a factual enquiry, which must be based in the evidence before the court. HMRC’s application did not include any witness evidence at all.

In those circumstances it would be wrong for the court to embark on what would essentially be its own judicial appraisal, unencumbered by evidence. This is so despite the apparent strength – at least with the benefit of hindsight – of aspects of HMRC’s case on this point, including the fact that the ECJ decided as long ago as 1997 that the power to derogate contained in Article 11C(1) was limited and that one element of the Old Scheme (the exclusion of barter transactions) contravened EU law: Goldsmiths (Jewellers) Ltd v Customs and Excise Commissioners (Case C-330/95) [1997] STC 1073. Although I was at one stage attracted by the argument that the position must be different for the 9-month period, in respect of which it should have been obvious to a well-resourced taxpayer like BT that the UK had not implemented any bad debt relief scheme at all, I was persuaded by Mr Cordara’s submission that the position could be affected by the subsequent introduction of the Old Scheme, because it could have induced a belief that anything available for the 9-month period would not be more favourable than that. In any event, however, the point already made about the requirement for evidence remains.

In conclusion, I would dismiss BT’s appeal in respect of the main period and allow HMRC’s cross-appeal in respect of the 9-month period, with the result that HMRC’s application for strike out or reverse summary judgment succeeds in full.