The Commissioners for HMRC v Smith & Nephew Overseas Ltd & Ors

This appeal from the Upper Tribunal (Tax and Chancery Chamber) contains all the classic ingredients that make up the daily fare of the successful tax law specialist. First, it requires the untangling of interrelated statutory provisions, amended several times over the years and supplemented by statutory instruments, also amended. Secondly, the answer may ultimately lie, at least according to one of the parties, in what is meant by one or two perfectly ordinary English words found in one of those statutory provisions. Thirdly, an eye-wateringly large amount of tax is either due or not due depending on which of the two contending constructions is correct. In this appeal the Respondents’ entitlement to losses of about £675 million turns on the result and there is more at stake because of other cases which we are told are waiting in the wings for our decision in this case.

The appeal is against the decision of the Upper Tribunal (TCC) (Fancourt J and Judge Thomas Scott) of 29 November 2018 reported at [2018] UKUT 0393 (TCC), [2019] STC 116. They dismissed an appeal against the decision of the FTT (Judge John Brooks and John Agboola) at [2017] UKFTT 151 (TC), [2017] SFTD 678. Patten LJ granted permission to appeal to this court by order dated 24 May 2019. Before us the Appellants, HMRC, were represented by Mr Gibbon QC leading Mr Rivett QC and the Respondents were represented by Mr Ghosh QC leading Mr Bremner QC and Mr Bradley.

The dispute concerns each of the Respondents’ entitlement to set off foreign exchange losses against their liability to corporation tax. The exchanges loss arose as a result of the Respondent companies changing their functional accounting currencies from sterling to US dollars on 23 December 2008 at a time when the only asset on their balance sheets was a very substantial inter-company debt owed to them by their parent company. The debts were denominated in sterling but then had to be converted into dollars when the companies’ accounts were restated in dollars. The next day, the debts were disposed of as part of a group restructuring. The exchange losses arose from the Respondents’ ‘loan relationships’ as that term is used in Chapter 2 of Part IV of the Finance Act 1996 (‘Chapter 2’). All section numbers in this judgment refer to sections in that Act unless otherwise stated. I will need to examine the relevant provisions in more detail later but it is useful here to summarise the scheme set out in the Act as it applies to the exchange losses for which the Respondents claim relief from corporation tax in this case.

Section 80 provides that all profits and gains arising to a company from its loan relationships should be chargeable to tax as income in accordance with Chapter 2. It provides also that the Chapter has effect for the purpose of determining how any deficit on a company’s loan relationships is to be brought into account. Section 81 defines ‘loan relationship’ for the purposes of the Corporation Tax Acts. A loan relationship exists whenever a company stands in the position of a creditor or debtor as respects any money debt and that debt is one arising from a transaction for lending money. Section 82 sets out the method for bringing into account any gains or deficits arising from the company’s loan relationships and provides that those gains and

Section 84 then provides for what debits and credits are to be brought into account in respect of the company’s loan relationships. It provides that the credits and debits to be brought into account shall be the sums which when taken together ‘fairly represent’ all profits, gains and losses of the company arising from its loan relationships. As originally enacted, section 84 did not cover gains and losses arising from fluctuations in currency exchange rates as they affected a company’s loan relationships. The Finance Act 2002 introduced section 84A to deal with exchange gains and losses arising from loan relationships. Section 84A provides, broadly, that exchange gains and losses are included in the references in section 84 to profits, gains and losses arising from its loan relationships. The term ‘exchange gains and losses’ is defined by section 103(1A), which was also introduced by the Finance Act 2002.

Section 84A(3), however, excepted certain exchange gains and losses so that they were not included in the credits and debits covered by section 84. The category of exchange gains or losses to which section 84A does not apply because of section 84A(3) include those which fall within either section 84A(3)(a) or (b) and which are recognised in the company’s statement of total recognised gains and losses (‘STRGL’), rather than in its profit and loss account. Section 84A conferred on HM Treasury a regulation-making power to bring into account in prescribed circumstances amounts which are taken out of the regime by section 84A(3). HM Treasury exercised that power in 2002 making regulations which covered, amongst other things, the disposal of loan relationships in respect of which exchange gains and losses have been recognised in the company’s STRGL.

The Smith & Nephew group is a multinational group engaged in the development, manufacture and marketing of medical devices. The headquarters of the Smith & Nephew group is in the UK and the ultimate parent is Smith & Nephew plc. Prior to the restructuring, Smith & Nephew plc had two main trading groups:

a trading group which comprised the international operations of the Smith & Nephew group, the entities within which prepared their accounts using US dollars as the functional currency; and

a trading group that comprised the UK trading operations of the Smith & Nephew group. For periods before 23 December 2008 the companies in this trading group prepared their accounts using sterling as the functional currency.

All three Respondents were part of the UK trading group and have at all material times been resident in the United Kingdom for the purposes of UK corporation tax. Each of them was a subsidiary of Smith & Nephew Investment Holdings Ltd (‘S&N Holdings’) which is another UK company. As at 23 December 2008, each of the Respondents was entitled to an inter-company receivable (‘ICR’) owed to them by their parent, S&N Holdings. Smith & Nephew Overseas Ltd (‘Overseas’) was owed an inter-company receivable of about £1.63 billion; TP Ltd (‘TP’) was owed an intercompany receivable of about £524 million; and Smith & Nephew Finance Holdings Limited (‘S&N Finance’) was owed about £340 million. Although the ICRs were non-interest-bearing, their existence meant that the Respondents, which were otherwise dormant, had to prepare annual tax returns reporting notional interest income arising on the ICRs.

In December 2008 the UK trading group was restructured. The way in which the restructuring was carried out is relevant to the question of whether there was some ‘real world effect’ of the exchange rate losses on the finances of the group. Dealing first with the treatment of TP, there was a share purchase agreement entered into on 23 December 2008 under which Smith & Nephew plc agreed to buy from S&N Holdings all the share capital of Overseas and TP. The price that Smith & Nephew plc would pay S&N Holdings was £1.69 billion for Overseas and £500,000 million for TP. This consideration was expressed to be payable on completion in cash. However, TP did not stay for long as a subsidiary of Smith & Nephew plc because by an agreement dated 24 December 2008, Smith & Nephew plc agreed to sell its new subsidiary TP to Overseas (by then also a subsidiary of Smith & Nephew plc). Overseas agreed to pay Smith & Nephew plc $3.5 billion for the shares in TP. That $3.5 billion was payable by Overseas to Smith & Nephew plc partly in cash (as to about $2.4 billion) and partly by the issue and allotment by Overseas to Smith & Nephew plc of one ordinary share in Overseas at a premium equal to the balance of the consideration (that is about $1.05 billion). The effect of the restructuring on TP was that TP became a wholly-owned subsidiary of Overseas; the amount of the premium allocated to the share in Overseas was transferred to the share premium account of Overseas (or more accurately the group reconstruction account since this was an intra-group transaction) and the net inter-company receivable held by Overseas was reduced to $200,000.

TP also made an acquisition as part of the restructuring. Smith & Nephew plc owned a different subsidiary called Smith & Nephew USD Limited (‘USD’). By a further agreement dated 24 December 2008, TP agreed to buy the entire share capital of USD from Smith & Nephew plc for $3.5 billion. This was required to be satisfied partly in cash (in the amount of about $772 million) and partly by the issue and allotment to Smith & Nephew plc of one ordinary share in TP at a premium equal to the balance of the consideration, that is about $2.7 billion. USD therefore became a wholly-owned subsidiary of TP, the amount of the premium allocated to the share owned by Smith & Nephew plc in TP was transferred to the share premium account of TP and the intercompany receivable in TP was eliminated.

There was a separate agreement also dated 23 December 2008 by which Smith & Nephew plc agreed to buy all the share capital of S&N Finance from S&N Holdings for £344 million in cash. On 24 December 2008 the Board of Directors of S&N Finance resolved to pay a dividend amounting to an aggregate of £344 million to Smith & Nephew plc as its sole shareholder.

The effect of the restructuring was that Overseas, TP and S&N Finance ceased to be subsidiaries of S&N Holdings and became direct or indirect subsidiaries of Smith & Nephew plc and the intercompany receivables were substantially eliminated. Since the functional currency of Smith & Nephew plc was dollars, there was a change on 23 December 2008 in the functional currency of each of the three Respondents from sterling to US dollars for accountancy purposes. On that day, once the functional currency was changed to US dollars, each Respondent still held the ICR in its accounts denominated in sterling, albeit for only one day until the transfer agreements

Each of the Respondents produced statutory accounts for the accounting period ended 31 December 2008 prepared in US dollars. Accounting standards required them also to restate the previous year’s accounts in dollars. A note to each of the companies’ accounts for the year ended 31 December 2008 explained the effect of what had happened. Taking TP as an example, the sterling value of the intercompany receivable of course remained constant between 1 January 2007 and 23 December 2008 at £524,242,838. The value of that amount in dollars at the exchange rate on 31 December 2007 was higher than it had been at the exchange rate on 1 January 2007. Comparing those two figures, there was an exchange gain over the year 2007. But this gain was more than cancelled out by the weakening of sterling against the dollar over the course of 2008. The intercompany receivable expressed in dollars as at 23 December 2008 was worth substantially less than it had been worth at the start of the year. This created an exchange loss which, when netted off against the small exchange gain over the course of 2007, generated an aggregated exchange loss of $253,576,261. In effect, therefore, when the debt was repaid in sterling on 23

Overseas recognised in its accounts an exchange difference of dollar value $818,167,957 which converted to a sterling value of £445,868,096;

TP recognised an exchange difference of dollar value $253,576,261 which converted to sterling value of £138,188,698; and

S&N Finance recognised an exchange difference of dollar value $166,346,849 which converted to sterling value of £90,652,234.

The exchange differences were not reflected in the profit and loss accounts produced by each of the Respondents for the period ended 31 December 2008 but were in each instance included within the STRGL by each Respondent and described as ‘Revaluation (loss)/gain on change in functional currency’.

HMRC did not accept that the losses could be treated as debits for corporation tax purposes and on 16 April 2014 issued closure notices which disallowed the losses and made consequential amendments to the tax returns of each company.

HMRC has all along accepted that the restructuring of the Smith & Nephew group was undertaken for commercial reasons and not for tax avoidance reasons. It is also accepted that in November 2008 Smith & Nephew wrote to HMRC setting out their proposals then contemplated for waiving the ICRs and requesting clearances to the effect that the credits and debits arising from the release of the ICRs would not be brought into account for corporation tax purposes and that the debt waivers would be disregarded for capital gains tax purposes. A meeting was held in late November 2008 where HMRC refused to provide the requested clearance. It was following that meeting that the restructuring arrangements were revised to the series of transactions which then took place about a month later.

The relevant statutory provisions are found in the Finance Act 1996, as amended in 2002, 2004 and 2006. For our purposes the relevant provisions are those in force on 31 December 2008.

Section 80 introduces the ‘loan relationship’ regime. It provides:

This Chapter shall also have effect for the purposes of corporation tax for determining how any deficit on a company’s loan relationships is to be brought into account in any case, including a case when none of the company’s loan relationships falls by virtue of this Chapter to be regarded as a

Subject to any express provision to the contrary, the amounts which in the case of any company are brought into account in accordance with this Chapter as respects any matter shall be the only amounts brought into account for the purposes of corporation tax as respects that matter.”

The term ‘loan relationship’ is defined in section 81(1) as arising wherever the company stands in the position of a creditor or debtor as respects any money debt and that debt is one arising from a transaction for the lending of money. It is not disputed that the ICRs create a loan relationship for this purpose – there is no exclusion from the regime for intra-group lending.

Section 82 deals with how to compute the profits and losses arising from the loan relationship. Section 82 provides (so far as relevant for non-trading debits such as these):

the profits and gains arising from the loan relationships of the company, and

any deficit on a company’s loan relationships,

credits that are not brought into account under subsection (2) above (“non-trading credits”), and

debits that are not so brought into account (“non-trading debits”)

The reference in section 84(1)(a) above to the profits, gains and losses arising to a company from its loan relationships and related transactions includes a reference to exchange gains and losses arising to the company from its loan relationships.

Subsection (1) above is subject to the following provisions of this section.

Subsection (1) does not apply to an exchange gain or loss of a company to the extent that it arises–

in relation to an asset or liability representing a loan relationship of the company, or

as a result of the translation from one currency to another of the profit or loss of part of the company’s business,

The Treasury may by regulations make provision for or in connection with bringing into account in prescribed circumstances amounts in relation to which subsection (1) above does not, by virtue of subsection (3) or (3A) above, have effect.

The reference in subsection (8) above to bringing amounts into account is a reference to bringing amounts into account –

for the purposes of this Chapter, as credits or debits in respect of the loan relationships of the company concerned; or

for the purposes of the Taxation of Chargeable Gains Act 1992.

As I explain below, that provision was introduced in 2004 because a reference in the original text of section 84(1) to sums being computed in accordance with ‘an authorised accounting method’ was deleted from subsection 84(1) in 2004 and in effect replaced by the new section 85A(1) referring to generally accepted accounting practice (‘GAAP’). The words in parentheses ‘(including, in particular, section 84(1))’ were inserted by the Finance Act 2006 and are important for understanding the case law on this subsection to which I refer below. As to when amounts are regarded as being ‘recognised in determining a company’s profit or loss’ for the purposes of section 85A, section 85B provides that an amount is so recognised if it is included in, amongst other places, the company’s profit and loss account or its STRGL.

The terms ‘exchange gains’ and ‘exchange losses’ are defined in section 103(1A) and (1B), introduced in 2002 at the same time as section 84A was inserted into the Finance Act 1996. Those subsections provide:

profits or gains, or

losses,

This regulation applies in the circumstances prescribed by paragraphs (2) and (2A) below.

The circumstances prescribed by this paragraph are where there is a disposal of an asset by a company and the asset disposed of represents a loan relationship of the company in relation to which exchange gains or losses have fallen within subsection (4) of section 84A.

Where this regulation applies, an amount equal to the amount of any net gain or net loss shall be brought into account, for the purposes of Chapter 2, as a credit or debit (according to whether it is an amount of net gain or net loss) in respect of the loan relationship for the accounting period in which the disposal occurs.

For the purposes of this regulation, the amount of any net gain or net loss shall be calculated by finding the aggregate of the amounts representing the exchange gains and losses which fell within paragraphs (2) and (2A).”

Regulation 13(2A) was introduced in 2004 taking effect in respect of the same accounting period as the substitution of section 84A(3), that is for accounting periods beginning on or after 1 January 2005. The effect of it is that where the exchange gains or losses are recognised in the company’s STRGL rather than in its profit and loss account, they do not come into account in each tax year in which they are so recognised but the aggregate of the gains and losses recognised over the years is brought into account in the year in which the loan relationship is disposed of. It may at first sight seem strange that regulation 13(2) was not repealed in 2004 when paragraph (2A) was introduced since section 84A(4) to which paragraph (2) refers was repealed at that time. However both Mr Ghosh and Mr Gibbon agreed that regulation 13(2) would still be relevant to the recognition of exchange rate losses and gains in accounting periods occurring before the repeal of section 84A(4) took effect. Although that subsection is repealed, those figures are still to be taken into account in the aggregation exercise required by regulation 13(4).

The application of the loan relationships regime and in particular the meaning of the term ‘fairly represent’ in section 84(1)(a) was considered by this Court in GDF Suez Teeside Ltd (formerly Teesside Power Ltd) v HMRC [2018] EWCA Civ 2075, [2019] 1 All ER 528 (‘GDF Suez’). At the time the Upper Tribunal heard the appeal in this case, the hearing in GDF Suez had taken place but judgment was pending. Judgments were handed down on 5 October 2018, the lead judgment was that of Henderson LJ with whom Asplin and Kitchin LJJ agreed. Although there had been earlier cases on the meaning of the phrase, this was the first case to address the meaning after the important amendments made to the provisions by the Finance Act 2004 and the Finance Act 2006. GDF Suez did not concern exchange gains or losses and so did not require any analysis of section 84A or regulation 13. It did require an analysis of section 84(1)(a) taken together with section 85A(1), the latter provision having been amended in 2006. Section 84(1) as originally enacted had read like this (my emphasis):

Section 85 had set out two alternative accounting methods which were authorised for the purpose of that underlined phrase in section 84(1). In 2004 the words underlined above and section 85 were repealed and section 85A(1) was inserted but without the words ‘(including, in particular, section 84(1))’. Those words were inserted into section 85A(1) by the Finance Act 2006.

GDF Suez concerned the application of the ‘fairly represent’ test in section 84(1) to a tax avoidance scheme which had been notified to HMRC under the DOTAS rules. It concerned the transfer of claims that the taxpayer had against certain insolvent companies. Those claims were valuable in the sense that they were likely to result in a substantial payment at some point in the future and for that reason they had a market valuation of £200 million. In accordance with GAAP, however, the claims properly carried a value of nil in the taxpayer’s accounts. The taxpayer wished to avoid the situation in which it would be liable to corporation tax on the full amount of the sums received as and when the claims were paid out. It sought to achieve this by transferring the claims to a Jersey subsidiary and acquiring in return the shares in that subsidiary. The intention was that the shares in the subsidiary would still carry a value of nil in the taxpayer’s accounts but the base value of the claims in the hands of the subsidiary would be their market value of about £200 million. If the claims subsequently paid out more than that, any profit would be attributed to the taxpayer and taxed in the UK accordingly. But only profits arising from realisation of the claims in excess of the £200 million base value would be brought home to the UK and taxed in the hands of the taxpayer. The overall effect of the scheme if it worked was

One major issue before the court in GDF Suez was whether the ‘fairly represent’ test in section 84(1) added anything to the requirement now imposed by section 85A(1) that the credits and debits to be brought into account had been computed in accordance with GAAP. The Court of Appeal held that it did and that looking at the matter in the round there was no difficulty in concluding that a profit or gain arose to the taxpayer from the disposal of the claims. Such profit or gain could only be fairly represented by a credit in the hands of the taxpayer equal to the value of the claim at the date of the disposal. In that way the profit or gain was brought into charge to tax at the same value as was recognised for accounting purposes in the hands of the subsidiary. There was therefore a symmetrical outcome as between the two counterparties to the transfer of the claims. Conversely, if the base value of the claims in the accounts of the subsidiary were treated as £200 million (so that only the excess realised over that was taxed in the hand of the taxpayer) but the value of the shares acquired by the taxpayer was treated as nil, there was an asymmetry which did not fairly represent the outcome in the taxpayer’s hands even though it was compliant with GAAP.

In coming to the conclusion that the ‘fairly represent’ added an additional requirement over and above compliance with GAAP, Henderson LJ relied in particular on the introduction of the parenthetical cross-reference to section 84(1) inserted into section 85A(1) by the Finance Act 2006. That was a puzzling amendment given that the introductory words of section 85A(1) as inserted in 2004 already stated that the provision was subject to the provisions of Chapter 2. Why did Parliament think it appropriate to single out section 84(1) as a provision of Chapter 2 to which the computation rule in section 85A(1) was subject? That amendment, he said, was intended: (para. 43)

Henderson LJ went on to hold that this interpretation of the 2006 amendment to section 85A(1) was supported by the Explanatory Notes to the Bill which became the Finance Act 2006. Those Notes indicated that the provisions being inserted were to ‘close a number of loopholes and block a number of avoidance schemes’. The fact that there was no statutory guidance on the question of what the ‘fairly represent’ test meant did not point to a more limited purpose than the override for which HMRC contended: para. 93

This was shown by the fact that the proof of the pudding was in the eating of it – the FTT and UT in GDF Suez had not had any difficulty in applying the test to the facts even though it might be difficult to express its precise content in the abstract. They had found that the override should apply and that the accounting treatment of profits adopted by the taxpayer in that case although compliant with GAAP did not in the circumstances fairly represent its profits in the two accounting periods. The Court of Appeal agreed and an additional £200 million profit was therefore required to be brought into account for the purposes of the loan relationships regime.

Both parties accepted, for the purposes of the appeal before us at least, that GDF Suez was authority binding on us that the ‘fairly represent’ test in section 84(1) is intended to add an extra requirement over and above the requirement that the credits and debits have been calculated in accordance with GAAP. On the facts of GDF Suez,that extra requirement was not satisfied and the Court held that the amount of profit properly entered in the taxpayer’s accounts did not fairly represent the amount of profit in fact accruing. However, GDF Suez was not concerned with exchange gains and losses and so did not address the proper construction of regulation 13 or the interaction between section 84(1) and section 84A(1). It did not address how one should give content to the ‘fairly represent’ test in the context of exchange gains and losses arrived at in accordance not only with GAAP but with section 103(1A) and (1B). Whether the Court in GDF Suez was right in so far as it suggested that the content of the ‘fairly represent’ test was changed or reinforced by the amendments made by the Finance Act 2006 is not relevant to the determination of this appeal.

Before the FTT and the Upper Tribunal there were three principal issues raised by HMRC as reasons to deprive the Respondents of the tax relief they sought in respect of the exchange losses. The first issue was whether the accounts of each company for the relevant year had been drawn up in accordance with UK GAAP as required by section 85A. The FTT preferred the evidence of the Respondents’ expert witness Mr

The second issue before the FTT and the UT was whether the exchange differences calculated by the Respondents gave rise to ‘exchange losses’ within the meaning of section 103(1A). The FTT and UT held that they did. The UT held that the wording of the definition of exchange losses in section 103(1A) was clear and unambiguous and referred specifically to gains or losses which arise as a result of comparing at different times the relevant valuations. The UT said:

HMRC does not challenge that part of the UT’s decision in this appeal.

The Upper Tribunal then turned to the question whether the exchange losses did ‘fairly represent’ losses of the Respondents for the purpose of section 84(1). That is the issue which is live before us. The Upper Tribunal considered earlier authorities on the meaning and effect of the words ‘fairly represent’ including GDF Suez. They held at para. 81 that three propositions of general application could be drawn from GDF Suez in interpreting the ‘fairly represents’ requirement of section 84(1):

It is intended to operate as an override in the sense that it could take priority over and may override the GAAP accounting treatment mandated by section 85A(1).

It is not limited in its purpose or effect simply to issues relating to the attribution or allocation of gains and losses to accounting periods or to loan relationships and related transactions.

The concept of fairness needs no elucidation but provides a touchstone which is well-suited to application by accountants, lawyers and judges bringing their professional expertise and experience to bear.

The Upper Tribunal held at para. 87 that the Court of Appeal in GDF Suez had given a clear indication of the most significant factors to which a court should give weight when considering all the circumstances and determining whether to apply the ‘fairly represent’ override. Applying those factors, they held that there was no basis for holding that the exchange losses here did not fairly represent losses arising to the company from its loan relationships. The relevant factors were that there was no tax avoidance motive on the part of the Smith & Nephew group, unlike the position in GDF Suez; the acceptance of the exchange losses as being entitled to relief did not create a problem of ‘asymmetry’ in the treatment of the transaction for accounting purposes as there had been in GDF Suez; the application of the override was not needed to avoid a result which was clearly contrary to Parliament’s intention; and the result arrived at by the GAAP compliant accounts was not a manifest absurdity such that the need to apply the override was obvious. The UT therefore affirmed the decision of the FTT but for different reasons in relation to the third ground.

In HMRC’s appeal before this court it was common ground that:

the accounting treatment of the exchange losses arising from the restructuring of the Smith & Nephew group, including the restatement of the 2007 company accounts into dollar figures, was compliant in all respects with GAAP;

the amounts which the Respondents say should be treated as debits for the purposes of sections 82 and 80 are ‘exchange losses’ within the meaning of section 103(1A) for the purposes of the loan relationships regime;

the exchange losses are properly recognised in the STRGL rather than in the

the exchange movement was properly recognised in the Respondents’ STRGL for the year 2007, albeit that that recognition arose because of the restatement of those accounts in dollars following the change in functional currency to dollars on 23 December 2008;

the loan relationships giving rise to the exchange losses were disposed of by the Respondents on 24 December 2008 and so the circumstances prescribed in regulation 13(2A) apply here;

the computation of the exchange losses claimed by the Respondents was properly carried out in accordance with regulation 13(4).

Neither party to the appeal before us presented their case on the basis that the appropriate approach was to consider the factors which the UT identified as having

In their written submissions, HMRC concentrated on why they contended that the exchange losses did not reflect a ‘real world’ loss suffered by the Respondents as a result of the overall depreciation of sterling against the dollar over the period 1

However, the thrust of HMRC’s case as presented in oral argument over the two days of the appeal was still that the ‘fairly represent’ test could only be satisfied and the exchange losses only brought into account as a debit if they arose because of some ‘real world’ detriment caused to the Respondents as a result of the fluctuation of the dollar/sterling exchange rates in 2007 and 2008. Here the exchange loss arises only as a result of the restatement of the Respondents’ accounts into dollars on the change of their functional currency. The actual ICR was always a sterling debt and was paid off in sterling during the course of the series of restructuring transactions. There was no ‘realised’ loss suffered by the Smith & Nephew group and certainly not by the individual Respondents.

The Respondents argue that there were two real world consequences for the depreciation of sterling against the dollar over the relevant period, an increase in the value of capital that had to be found to fund the consideration for the purchases of TP and Overseas in the form of the premium shares in TP and Overseas and a concern over whether the losses created a shortfall in the distributable profits needed to support the dividend paid by S&N Finance to its parent. I explain these points in more detail later. Most of Mr Ghosh’s submissions related, however, to the two issues raised in the Respondents’ Notice served in June 2019. Those issues logically precede the question of whether the exchange losses ‘fairly represent’ their profits, gains and losses because they seek to establish that the ‘fairly represent’ test does not apply in this context at all or at least not in the way that HMRC contend. They argue that the exchange losses in this case are only brought into the regime at all by regulation 13 of the 2002 Regulations. That regulation provides simply that the exchange losses falling within its scope shall be brought into account. There is no need to pass them through the prism of section 84(1)(a) and the ‘fairly represent’ test at all; that test has no application to these losses. Alternatively, they submit that HMRC is asking the wrong question in examining whether the exchange losses ‘fairly represent’ the profits, gains and losses of the company. The question mandated by the way in which section 84A(1) interpolates exchange losses into section 84(1)(a) is whether the debits which the Respondents are seeking to bring into account fairly represent the exchange losses arising to them from their loan relationships. The answer to that question is clearly ‘yes’ since the debits exactly are the exchange losses. The inquiry should stop there and there is no need to explore what the effect on the affairs of the companies or the group of the exchange losses was in the ‘real world’ - whatever that means.

The issues raised by the appeal and the Respondents’ Notice are therefore threefold.

What is the effect of the 2002 Regulations bringing those exchange gains and losses which are recognised in the company’s STRGL into the regime, and in particular are they simply brought into account by virtue of those Regulations or do they also have to be shown to ‘fairly represent’ the profits, gains and losses of the company in accordance with section 84?

If the ‘fairly represent’ test does apply in this case, what is it that has to be fairly represented in order for the exchange gains and losses to be brought into account for the purposes of corporation tax?

What is the content of the ‘fairly represent’ test in this context and in particular does it require the Respondents to show that the exchange rate fluctuations giving rise to the losses for which relief is sought had some real world effect

The first point raised in the Respondents’ Notice arises from posing the question: how is it that exchange losses which are, in accordance with GAAP, recognised in the taxpayer’s STRGL (rather than in the taxpayer’s profit and loss account) brought into the loan relationships regime? Generally speaking, exchange gains and losses arising from loan relationships are brought into section 84(1)(a) through section 84A(1). Precisely how they are brought in is the subject of a further argument by the Respondents discussed later. Where an exchange loss arises in relation to an asset or liability representing a loan relationship of the company and it is recognised in the company’s STRGL, subsection (1) of section 84A does not apply because of section 84A(3). Subsection (8) of section 84A empowers HM Treasury to make regulations ‘for or in connection with bringing into account in prescribed circumstances’ amounts which are carved out of subsection (1) by subsection (3). It provides in subsection (9) that the Treasury can do this in one of two ways, either (a) by bringing them into account for the purposes of Chapter 2 as credits or debits in respect of the company’s loan relationships or alternatively (b) by bringing them into the Taxation of Chargeable Gains Act 1992 regime.

The power in section 84A(8) was exercised by the making of regulation 13 of the 2002 Regulations. The Treasury chose the first option in section 84A(9)(a), bringing the amounts into account as credits or debits in respect of the loan relationships of the company concerned. Regulation 13 sets out two different circumstances which are prescribed for the purposes of section 84A(8). Both circumstances involve where there has been a disposal of an asset which represents a loan relationship. The second circumstance described in paragraph (2A) is the one that applies here, namely where there has been a disposal of an asset representing a loan relationship in relation to which exchange gains or losses were recognised in the company’s STRGL. Paragraph (3) of regulation 13 then sets out what happens:

Paragraph (4) of regulation 13 then explains how to calculate the net gain or net loss to be brought into account. This is done by finding the aggregate of the amounts representing the exchange gains or losses which fall within paragraph 2A, in other words which were recognised in the company’s STRGL, over the years when the loan relationship was in existence.

Mr Ghosh argues that the effect of this is as follows. Exchange gains or losses arising from loan relationships are recognised in a company’s STRGL in the accounting years in which they arise. But they are not treated in those years as credits or debits for the purpose of Chapter 2 by virtue of section 84(1)(a) because they are carved out of

Mr Ghosh submits that this interpretation of regulation 13 does not create an opportunity for aggressive tax avoidance because of the provisions in Schedule 9 to the Finance Act 1996. In particular, he drew our attention to paragraph 13 of Schedule 9 which, very broadly, rules out the bringing into account of debits and credits arising from a loan relationship if the loan relationship of the company has ‘an unallowable purpose’. Unallowable purposes include a purpose which is not amongst the business or other commercial purposes of the company including where the purpose is a tax avoidance purpose.

HMRC describe Mr Ghosh’s argument as taking too literalist an approach to the construction of regulation 13 and its interaction with sections 84A and 84(1). They argue that the purpose of regulation 13 is simply to state when the exchange gain or loss arising from the loan relationship should be brought into account. It was intended to provide a route to bring additional exchange losses under the umbrella of section 84(1). It would, Mr Gibbon submitted, be incoherent for Parliament to have intended to permit this class of exchange gains and losses to bypass the governing concept of fair representation provided for in section 84(1). On the contrary, the overall coherence of the Chapter 2 scheme requires section 84(1) to be the pivot by which the accounting, GAAP world, moves over to the tax world. The aggregated amount arrived at on the application of regulation 13 can still be treated as a credit or debit to be brought into account if but only if it fairly represents for the accounting period in which the loan relationship was disposed of all profits, gains and losses of the company in accordance with section 84(1)(a).

Mr Gibbon describes regulation 13 as dealing with a timing point only. It provides that where the exchange loss is recognised in the STRGL, the netted off amount is brought into account only in the year of disposal of the loan relationship. Another way the Treasury could have done this might be, for example, to apportion the net loss or gain across all the years during which an exchange gain or loss was recognised in the company’s STRGL. That is not the way chosen by the regulation. Beyond that, there is nothing in the regulation to suggest that Parliament intended to allow these exchange gains and losses to avoid having to satisfy the important ‘fairly represent’ test as now interpreted by this court in GDF Suez.

This point was dealt with only very briefly by the Upper Tribunal since it was not placed centre stage in the parties’ arguments before them as it was before us. At para

The difficulty with HMRC’s construction of regulation 13 is that if its purpose is limited to specifying the accounting period in which the aggregated amount is to be considered, it is difficult to see how that aggregated amount in that accounting period falls within the loan relationships regime at all. It has never been HMRC’s case that the Respondents’ exchange losses fall outside the regime entirely. Section 84A(3) prevents the exchange losses falling within section 84A(1). The only way they can be brought within the regime is through the exercise of the power in section 84A(8). The power in section 84A(8) is to make regulations ‘for or in connection with bringing into account’ amounts in prescribed circumstances. That power is clearly not limited to a power to disapply section 84A(3) in prescribed circumstances – the regulations may directly bring amounts into account and, indeed, may do so either for the purposes of Chapter 2 or for the purposes of the capital gains tax regime. The wording of regulation 13 shows that the Treasury in exercising the power chose to bring the exchange gains and losses into the loan relationships regime and so subject them to corporation tax rather than capital gains tax. In my judgment the wording of regulation 13(3) stipulates that the amount computed in accordance with the regulation “shall be brought into account”. By stipulating that, the regulation did not stray outside the scope of the power conferred by section 84A(8) because that provision empowered HM Treasury to made regulations for bringing amounts into account as well as conferring a power to make regulations in connection with bringing amounts into account. That choice is indicated by the use of the words ‘shall be brought into account’ in regulation 13(3) and by the absence of any cross-reference to sections 84 or 84A.

HMRC have to rely on the inclusion in regulation 13(3) of the words ‘for the purposes of Chapter 2’ as routing the bringing of the amounts into account indirectly through section 84(1) rather than directly through regulation 13(3) itself. They submit that the purposes of Chapter 2 include the purpose of applying the ‘fairly represent’ test in section 84(1). I do not agree that those words can bear the weight that HMRC place on them. The words are intended to do two things. First to show that HM Treasury has chosen the corporation tax rather than the capital gains tax regime for bringing these amounts into tax. Secondly they show that although the net gain or net loss is brought into account, the relevance of that is limited to the application of the provisions in Chapter 2. In other words the net gain or loss is not to be treated as part of the company’s account for the purpose of any other aspect of corporation tax or for the purposes of applying any tax other than corporation tax. If therefore in some other statutory provisions a question arises whether the company has achieved a net gain or suffered a net loss arising out of the loan relationship, neither the taxpayer nor HMRC can point to regulation 13 to say that the amount has been brought into its account for that other purpose. It is brought into the account for the purpose, and only for the purpose, of applying the provisions in Chapter 2.

Further, the timing point raised by HMRC is one that the Respondents use to their advantage. They say that the ‘fairly represent’ test in section 84(1) does not work if it is applied to an aggregated figure computed in accordance with regulation 13(4). For the purpose of section 84(1) what must be fairly represented is the gain or loss in a particular accounting period. The aggregation of a series of gains and losses over a

That indicates to me that regulation 13 is intended to do what it says. It is the regulation itself which mandates the bringing into account of the aggregate amount as a credit or debit for the purposes of Chapter 2 without any reference back to section 84A or section 84(1)(a). Regulation 13 could itself have incorporated a further requirement that the aggregate amount arrived at under regulation 13(4) was brought into account only if it fairly represented the overall consequence of the loan relationship for the company. But there is no reference to such a test either directly or indirectly by reference back to the provisions in Chapter 2.

Mr Gibbon submits that the Respondents’ construction creates an odd distinction between the way that exchange losses which are recognised in the company’s profit and loss account are treated compared with the treatment of exchange losses which are recognised in the STRGL. Exchange rate losses recognised in the profit and loss account fall within section 84A(1) because they are not carved out by subsection (3). If, as HMRC assert, exchange gains and losses which fall within section 84A(1) are subject to the ‘fairly represent’ test in section 84(1), then it would be odd for the statute to treat them differently. The answer to this is that Parliament has decided to treat them differently by making exchange gains and losses which are recognised in the STRGL subject to the regime in regulation 13 rather than the regime in section

In my judgment therefore the Respondents are right to say that the way that regulation 13 works is that where a loan relationship has given rise to exchange gains or losses recognised in the company’s STRGL and the loan relationship is disposed of, a credit or debit calculated in accordance with the regulation is brought into account as a debit or credit for the accounting period in which the disposal takes place. For the purposes of section 82(1), therefore, it is a debit or credit given for the accounting period in which the disposal took place by the provisions of Chapter 2 and can therefore be used to compute any deficit on a company’s loan relationship for the purpose of section 80(4). The ‘fairly represent’ test does not apply.

The Respondents’ second argument is that if the fair representation requirement does apply (contrary to their first argument) then it is satisfied in this case. They have two routes to establishing this. The first, which is dealt with in this section, depends on an analysis of how section 84A(1) operates to incorporate exchange gains and losses into the loan relationships regime. Subsection (1) of section 84A brings exchange gains and losses in by expanding the reference in section 84(1)(a) to profits, gains and losses to include a reference to exchange gains and losses. The effect of section 84A(1) is therefore that section 84(1) reads something like this:

all profits, gains and losses of the company including exchange gains and losses and including those of a capital nature which (disregarding interest in any charges or expenses) arise to the company from its loan relationships and related transactions; ….”

This is different, the Respondents say, from an interpolation which might have been effected if section 84A(1) had said instead that the reference in section 84(1) to credits and debits to be brought into account shall include a reference to exchange gains and losses arising to the company from its loan relationships. Then the resulting text of section 84(1) would have looked something like this:

all profits, gains and losses of the company which arise to the company from its loan relationships and related transactions.”

It is also different from the result that would have been achieved if section 84A(1) had interpolated exchange gains and losses in section 82(1)(a) since that might arguably have made them subject to the section 84(1) test.

The way section 84A(1) in fact interpolates exchange gains and losses means that the question posed by section 84(1)(a) is not: do the exchange gains and losses fairly represent the profits, gains and losses of the company? Rather the question posed is: do the credits and debits to be brought into account fairly represent the exchange gains and losses arising to the company from its loan relationships? In this case, Mr Ghosh argues that there is no difficulty with concluding that the debits that the Respondents are seeking to bring into account do fairly represent exchange losses because the amount of the debits are exactly the same as the amount of the exchange losses. Now that HMRC have conceded that the claimed debit is an exchange loss within the meaning of section 103(1A), there is nothing for the ‘fairly represent’ test to do.

Initially, during the course of argument in the appeal, Mr Ghosh referred to other circumstances in which a debit to be brought into account would not fairly represent an exchange loss arising from a loan relationship or at least in which there might be some debate about it, emphasising that that was not the case here. Here, the debit is precisely the exchange loss (subject to the point discussed earlier about it in fact being an aggregate amount spanning several accounting periods). We invited the parties to suggest a scenario in which there might be some debate about whether a debit which a company was seeking to bring into account because of an exchange loss arising from its loan relationship might be said not fairly to represent a loss of the company. It proved very difficult for the parties to come up with a plausible example. The example proposed by the Respondents was a scenario adapted from the facts in two actual cases, Stagecoach Group Plc & anotherv HMRC[2016] UKFTT 120 (TC), [2016] SFTD 982 and Union Castle Mail Steamship Company Ltd v HMRC[2018] UKUT 316 (TCC) [2018] STC 2034. Those cases concerned a situation where Company A sold to Company B the rights to a percentage of the receipts expected to be generated by the loan relationship. As a matter of accounting, Company A was required to ‘de-recognise’ the loan relationship as to that percentage and enter an equivalent debit in its STRGL. Those two cases established that Company A was not entitled to relief for that debit because it did not fairly represent its losses arising from its loan relationships. One could adapt those facts to posit an exchange loss also having arisen at the time when the contracts which resulted in the derecognition are entered into. In those circumstances, by analogy with the two authorities cited, the debit would be disallowed on the basis that it did not fairly represent losses of the company from its loan relationships. HMRC argued that that was not a good example because the reason why relief was denied for the debit in that case was not because of the application of the ‘fairly represent’ element of section 84(1)(a) but because the element that the debit must be in respect of the loan relationship was not satisfied. The debit would arise from the contract transferring the entitlement to receipts from the loan relationship and not from the loan relationship itself.

In fact therefore neither party was able to envisage a situation in which a debit would not fairly represent for the accounting period in question an exchange loss of the company arising from its loan relationship. That, in my judgment, supports the conclusion I arrived at in relation to the application of regulation 13. It is less surprising that regulation 13 should bypass the ‘fairly represent’ test for exchange gains or losses recognised in the company’s STRGL if the ‘fairly represent’ test really has no immediately discernible application for exchange gains or losses recognised in the company’s profit and loss account which fall within section 84A(1).

Finally, the Respondents argue that the recognition of the debit in respect of the exchange losses did indeed fairly represent a loss that had real-world consequences for the Smith & Nephew group. The depreciation of sterling had a real and significant economic and commercial impact on the Respondents in two respects. First, so far as Overseas and TP are concerned, the group had to increase the capital that was needed as part of the restructuring. As I described before, the consideration for the sale of TP by Smith & Nephew plc to Overseas and the consideration for the sale of USD by Smith & Nephew plc to TP was $3.5 billion each. That valuation was, we were told, the result of an exercise carried out to arrive at a proper valuation of the assets of the part of the group being sold. Mr Gibbon confirmed that HMRC did not challenge the legitimacy of the prices paid for the two companies. In both cases the consideration was made up partly of cash payable on completion and partly of the premium on the single share in the buyer company which share was allotted to the seller. The cash consideration was in both cases equal to the inter-company receivable owed to the transferred company. Since the purchase price was expressed in dollars, the receivable had to be converted to dollars in order to determine the split of consideration between cash and share premium. The stronger sterling was against the dollar, the more dollars could be bought with the ICR, the higher the cash part of the consideration would be and the smaller the balancing payment in the form of share premium would need to be.

For the sale of USD by Smith & Nephew plc to TP the transaction worked like this. The consideration payable by TP to Smith & Nephew plc was fixed at US$3,500,000,000. The cash part of the consideration was fixed at US$772,367,000 because that was the dollar equivalent of the ICR of £524,242,838 owed by S&N Holdings to TP, converted at the appropriate spot rate. The balance represented by the issue of a single share in the buyer, TP, given to the seller, Smith & Nephew plc, was US$2,727,633,000. That was the amount of capital that had to be raised to support that share premium. If sterling had still been as strong against the dollar as it had been on the last balance sheet date of 31 December 2007, then the receivable (being £524,242,838) would have been worth more dollars, in fact it would have been worth $1,044,292,056. That larger sum then would have been the cash part of the consideration, leaving only a smaller balancing amount $2,455,707,944 to be made up by the value of the single share in the buyer, TP, given to the seller. The companies lacked the cash resources to make up the difference caused by the reduction in value in dollar terms of their sterling assets and therefore had to raise more capital to fund the shortfall. The additional capital that had to be raised and was reflected on the balance sheet was therefore $271,925,056 greater than it would have been but for the

The second real-world consequence relied on by the Respondents was more straightforward. It was the concern that there would need to be a capital reduction moving shareholder funds to distributable profits in order to fund the dividend paid on 24 December 2008.

Mr Gibbon’s response to this was to the effect that none of this represents a real loss like that suffered by a company which suffers a diminution in its assets because of a fluctuation in currency. The exchange losses are created by the retrospective recasting of the Respondents’ 2007 accounts into dollars for the purpose of comparing them with the 2008 accounts drawn up after the functional currency had changed. If the loans had been disposed of on 23 December 2008 before the change of functional currency rather than a day later, the ICRs would not have been restated in dollars and no comparison between their value on that date and their value at the start of the previous year’s accounting period would have been needed. The dollar and sterling exchange rate did not of course start moving only on 1 January 2007; there would have been fluctuations in 2006 and 2005 when the ICR were also included in the Respondents’ accounts. That date was chosen because of the accounting method but introduced a certain arbitrariness in the valuation of the losses calculated. Tax, he submitted, should reflect economic reality and should not tax something which is not an economic loss or benefit. The ‘fairly represent’ test was an open-textured phrase designed to achieve this.

I do not accept that the application of the ‘fairly represent’ test can or should lead to the kind of investigation into the reality of gains or losses for which HMRC contend. Section 103(1A) provides for how exchange gains and losses are to be computed. It is a purely arithmetical exercise carried out to arrive at ‘a result of comparing at different times the expression in one currency of the whole or some part of the valuation put by the company in another currency on an asset or liability of the company’. As to what those ‘different times’ should be in any given case, that is determined by GAAP. The FTT and UT by finding that the ‘foreign operations’ method was legitimate, thereby found also that the ‘different times’ used in the calculation of these exchange losses were the correct dates.

It was notable that HMRC did not at any stage of this appeal propose any other ‘different times’ which could be used to generate exchange losses which would more fairly represent the loss arising in this case. They did not, for example, say that a more ‘real’ calculation would be to compare the dollar value of the ICRs at the date the loans were first made with their value on 23 December 2008. HMRC’s case seems to be that there was no amount that needed to be brought into account as a fair representation of the exchange loss arising on the disposal of these loan relationships. The effect of that would, it seems to me, be that generally speaking, where an exchange gain or loss within the meaning of section 103(1A) arises because a

I agree with the UT’s description of section 103(1A) in paras 68 and 69 of their judgment, set out at para. 37 above. Section 103(1A) is a tax code provision and not an accounting provision. It mandates that two dates are chosen for making the comparison rather than, for example, using an average rate between those dates and it does not require those dates to be chosen in any particular way other than in accordance with GAAP. Section 103(1A) does not, in my view, leave room for any imported requirement of reality or genuine suffering. The search for some genuine or real loss arising for a company from exchange rate fluctuations would, in most cases, be illusory. The earlier discussion about the hedging of exposure to currency risk demonstrates how impossible it would be to identify a ‘real’ loss suffered by a company by reason of adverse currency movements. Companies do not operate like individuals buying a small amount of foreign currency to go on holiday. Such an individual may well suffer a ‘real’ loss when cashing in leftover dollars on their return home if the dollar has weakened against sterling whilst they were abroad. By contrast, hedging for foreign exchange exposure is a common and prudent step for any company, large or small, to take. It may be undertaken by going out into the financial markets and buying a swap corresponding to the particular exposure, or the particular exposure may be added to the company’s overall pool of currency exposure. In that pool, the sterling/dollar exposure may fortuitously be cancelled out by some equal and opposite dollar/sterling exposure that the company has arising from another transaction that it has entered into for the purposes of its business. To what extent would HMRC need to investigate how the hedge was achieved, at what price and whether it was strictly necessary before the company can show that there were ‘real world consequences’ for it of the currency fluctuation such that an exchange loss fairly represented some real world detriment? The HMRC approach risks opening up a very complicated investigation which is not appropriate or indicated by the wording of the legislation.

Mr Gibbon appeared at some points in his submissions to be relying on the fact that the exchange losses were only included in the Respondents’ STRGL and not in their profit and loss account. This, he seemed to suggest, meant that they were less the result of economic activity than would otherwise be the case, pointing out that the Companies Acts have described a company’s accounts as comprising the balance sheet and the profit and loss account. I do not accept that is a legitimate point. The statutory and accounting material we were shown shows that the STRGL was described as a primary financial statement that enables users to consider all recognised gains and losses of a reporting entity in assessing its overall performance. The accounting standard FRS 3 required the STRGL to be presented with the same prominence as other primary statements. The STRGL was described as contributing further to the purposes of financial reporting by providing information ‘that is useful

If, therefore, contrary to my conclusions on the application of regulation 13 and the interaction of section 84A(1) and section 84(1)(a), some real-world consequences are required before debits to the value of the Respondents’ exchange losses can be brought into account for the purposes of Chapter 2, I consider that the two consequences described by Mr Ghosh are sufficient for it to be said that debits fairly represent the exchange losses arising from their loan relationships.

I would therefore dismiss the appeal for reasons different from those given by the UT.

I agree with Rose LJ that this appeal should be dismissed. I am not, however, entirely in agreement with the route by which she reaches that conclusion. Since Rose and Coulson LJJ agree, I shall keep my judgment short. Nothing I say should be taken to detract from Rose LJ’s lucid analysis, for which I should express my unreserved admiration. I will mostly use the same abbreviations that she has adopted.

This second appeal has followed a slightly unusual procedure. The main issues that were argued before us were either not addressed, or not fully addressed, in the Firsttier Tribunal (“FtT”) or in the Upper Tribunal (“UT”). That is as a result of two factors: first, both parties have adopted new positions in this court, and secondly, the Court of Appeal’s decision in GDF Suez was only decided after the oral argument had concluded in the UT.

HMRC changed their position (a) by abandoning reliance on the first two issues decided by both the tribunals (namely whether the accounting treatment of the exchange losses was GAAP compliant, and the question of whether they were properly to be regarded as exchange losses at all), and (b) by placing most of their emphasis on the argument that the respondents’ accounting treatment of the exchange losses, whilst in accordance with GAAP, did not represent any economic or real world effect on the life or trading of the companies and, therefore, should not be regarded as fairly representing an economic loss. This latter point was bolstered by the submission that, if there had been an exchange gain, instead of an exchange loss, it would not have been taxable.

The respondents made their respondents’ notice arguments their central submissions on appeal. The first was that the legislation operated in such a way as not to impose any requirement that the accounting entries in question should “fairly represent” the exchange losses under section 84(1)(a), but rather that their inclusion for the purposes of calculating corporation tax was mandated by regulation 13(3). The second was that a proper reading of sections 84(1)(a) and 84A(1) led to the conclusion that the debits in question necessarily fairly represented the exchange losses, because the figures were identical.

The appeal was originally listed for hearing with an estimate of one day. It became apparent to the court by about 15:00 on 23^rd October 2019 that there would not be sufficient opportunity for the somewhat complex arguments to be properly elucidated. Accordingly, we suggested that the appeal should be adjourned after the first day for another full day of argument. Ultimately that could not be arranged until 15^th January 2020, by which time the parties had filed yet further written submissions with different emphasis even from those presented at the first day’s hearing, admittedly partly in response to questions from the court. I do not recite this history in order to make any criticism of the parties. They have addressed their arguments in a most measured and appropriate way. I have, however, been left with the feeling that the case has stood on somewhat shifting sands, which has not made it any easier to determine.

As [11]-[31] of the agreed facts record, the restructuring transaction was entirely driven by S&N’s desire to forgive the ICRs (inter-company receivables). On 26^th November 2008, HMRC told S&N that “the credits in SNIH and debits arising from the release of the inter-company liabilities (including those owed to the Appellants) should be ignored for tax purposes but provided no assurance as to the capital gains tax consequences of the proposal”. The restructuring, including the change of functional currency, was then devised by S&N’s advisers to achieve the purpose they could not be certain to achieve by simply releasing the ICRs. The suggestion, therefore, that there were real world consequences of the scheme must be viewed against that background, even though Mr Gibbon specifically submitted that no great weight should be placed on that fact. He accepted that this was not a case of a contrived or any tax avoidance scheme, nor did he attack the motivation for the change of functional currency.

In argument, the court asked Mr Gibbon to define HMRC’s reasons why in this case we should reach the conclusion that the suggested debits in respect of the foreign exchange losses did not fairly represent losses of the companies. His answer was threefold.

First, Mr Gibbon said that the entire transaction was ex post facto, in the sense that the sterling loans went back to 2005, but there was a currency exchange at the end of 2008 which required the 2007 figures to be restated.

Secondly, Mr Gibbon submitted that there was no suggestion that the currency exchange represented an economic loss or an economic exposure which existed prior to 23^rd December 2008. Admittedly, the companies could point to an economic effect being caused by the receivables having been repaid in sterling and the balance of the consideration for the transactions having been calculated by reference to the dollar value of the repayment, but the entries in the STRGL were only required because of the transactions. If the loans had simply been repaid in sterling, none of those consequences would have followed.

Thirdly, if debits were allowed without the taxpayers being able to point to an economic change which had created them, one would be giving effect to a mechanistic application of GAAP and the definition of exchange losses. That was the point at which, on these facts, fair representation should intervene.

Throughout the argument, I was inclined to deal first with the appeal and then only if necessary, with the points argued by the companies in their respondents’ notice. It seemed to me that, if HMRC were unable to show that the UT and the FtT had been wrong to hold that the debits in question fairly represented losses of the companies within the proper meaning of section 84(1), we would not need to consider regulation 13. Rose LJ has, of course, dealt with that point at the start of her judgment for reasons I completely understand. I shall, however, change that order, dealing first with the inter-action between sections 84(1) and 84A(1) (acknowledging that that too was a respondents’ notice point), then with the “fairly represent” question, before returning to regulation 13.

Rose LJ dealt with this question in section 8 of her judgment, concluding at [66] that it was less surprising that regulation 13 should bypass the ‘fairly represent’ test for exchange gains or losses recognised in the STRGL, if the ‘fairly represent’ test really had no immediately discernible application for exchange gains or losses recognised in the profit and loss account to which section 84A(1) applies. Her view is, I think, that the ‘fairly represent’ test would anyway have no discernible application here (if it applied, which she has held it does not), because the debits in question here do fairly represent exchange losses, being in exactly the same sums as those losses, HMRC having accepted that the debits were indeed exchange losses within section 103(1A) (see [64] above).

I agree that sections 84(1)(a) and 84A(1) are to be read in the way that Mr Ghosh submitted. That does, therefore, mean that, in respect of exchange losses covered by section 84(1)(a) the legislation provides that: the debits to be brought into account in respect of a company’s loan relationship are the sums which fairly represent for the accounting period in question the losses of the company including exchange losses arising from that loan relationship. Put even more simply, the legislation provides, in a situation of this kind, that the debits to be brought into account are those sums that fairly represent the exchange losses.

In my judgment, there is a reason why the legislation is cast in this way. That is because exchange gains and losses are of a different quality to other gains and losses incurred by a company in respect of its loan relationships. Exchange losses are a fact that a company cannot change. The only thing that can be changed by a company is the times at which comparisons are made. The definition in section 103(1A) makes this clear. Sections 85A and 85B provide that the amounts to be brought into account are to be in accordance with GAAP, and GAAP generally provides for the times at which comparisons are to be made. Thus, as I see the case, once HMRC gave up its argument that the debits in question did not accord with GAAP, it had, in the circumstances of this case and its other concessions, little left in its armoury.

It is easy to see why exchange gains and losses are of a different quality to other gains and losses arising from a company’s loan relationship. Loan relationships can normally be expected to produce interest, bad debts, redemption fees alongside many other types of gain and loss. None of these is quite so binary as an exchange gain or an exchange loss, which is a fact, dependent only on the times at which the comparison is made.

I need then to deal with whether or not section 84(1)(a), read in the way that I have suggested it must be, applies at all to exchange losses recorded only in the STRGL. I accept, of course, that section 84A(3) expressly disapplies section 84A(1) for those exchange losses. But regulation 13(3) brings those very same losses, where there is a disposal, back into account for the purposes of Chapter 2 as a debit. For the purposes of this argument, it does not seem to me to matter whether the exchange losses in question are mandatorily brought back in under regulation 13 as Rose LJ has held, or brought back in through section 84A(1) and 84(1)(a). In the first case, there is no “fairly represents” requirement at all. In the second case, the legislation provides that the debits to be brought into account are those sums that fairly represent the exchange losses. The third option that Mr Gibbon suggested to the effect that the exchange losses come back in directly to section 84(1)(a), by-passing section 84A(1), cannot be correct. The reference to “Chapter 2” in regulation 13(3) must include the whole of Chapter 2, and there must be some congruity between the tax treatment of exchange losses in the profit and loss account and in the STRGL.

In short, I conclude, as Rose LJ has done, that there is no real substance, in the circumstances of this case, to the “fairly represents” test in relation to these exchange losses. Once it was clear that the debits in question complied with GAAP, they had to be brought into account under section 84(1) because they were the sums that fairly represented the exchange losses which had arisen to the companies from their loan relationships. There was no place for a ‘real world’ overlay here as there had been in the circumstances of the GDF Suez decision.

This point is, in my view, determinative of the appeal, because, whether or not regulation 13 mandates that the debits in question are taken into account, the same result would, in the circumstances of this case, be achieved under section 84(1)(a).

Rose LJ dealt with this question in section 9 of her judgment, concluding at [75] that if some real-world consequences are required before debits to the value of the respondents’ exchange losses can be brought into account for the purposes of Chapter

For my part, I would not go so far as to hold that the two consequences relied upon by the respondents amounted to any necessary ‘real world’ element so as to pass the “fairly represent” test applicable in different circumstances to different kinds of gains and losses arising to a company from its loan relationships. I bear in mind the background features that I have referred to above, and do not think that HMRC’s arguments can be so easily dismissed. This was, after all, a scheme designed to get around HMRC’s refusal to give capital gains tax clearance to the waiver of the ICRs. The transactions were designed by S&N’s advisers to extinguish the ICRs without tax consequences. There was no improper motivation, but it was nonetheless a transaction which did not affect the net asset value of the group.

I do not think that we need to decide what the outcome would be if the legislation were different, and if it were not obvious as I have held it is, that the GAAP compliant debits had to be brought into account under section 84(1)(a) because they were the sums that fairly represented the exchange losses which had arisen to the companies from their loan relationships. We do not need to decide whether the debits in question really did, in real world terms, fairly represent losses of the companies on the facts of the case, and I would decline to do so.

In section 7 of her judgment, Rose LJ concludes that regulation 13 does indeed operate, as the respondents submitted, to exclude the requirement of fair representation contained in section 84(1). For the reasons I have already given, this appeal can be determined without the need to decide that issue, and I would, for myself, be disinclined to do.

My reasons can be shortly stated. Both Chapter 2 and the 2002 Regulations are complex and, in some ways, as HMRC pointed out, inconsistent. The issues raised by GDF Suez suffice to make that point good. I am, therefore, wary of adopting an ingenious, but perhaps binary, interpretation in order to impose coherence on statutory provisions which are themselves far from straightforward. In my judgment, this case can be resolved by looking at the very nature of an exchange loss. I take the view that the determination of how regulation 13 fits into the complex legislative framework can await a case in which a decision cannot otherwise be reached.

I do not think it is clear that HM Treasury intended to exercise its power under section

I accept that the words “shall be brought into account” in regulation 13 imply that the exchange loss is to be brought into account directly, rather than by passing through section 84(1)(a). But regulation 13(3) also provides that the exchange loss must be brought into account “for the purposes of Chapter 2, as a credit or debit”. These last words could be said to be significant. Section 82(1) provides that the “profits and

I fully understand Rose LJ’s reliance on the reference in section 84(1) to “the accounting period in question”. A debit representing an exchange loss brought into account by regulation 13 would, at first sight, not be able fairly to represent a loss for a particular accounting period, when the calculation is made by reference to gains and losses over a number of years. It could be, however, that the reference to “the accounting period in question” should be taken to refer in this context to the accounting period in which the disposal occurs and the net gain or loss is accounted for.

The regulation 13 argument was only dealt with perfunctorily in the UT. I would prefer not to decide the point in this case.