The Commissioners for HMRC v Blackrock Holdco 5, LLC

The Transfer Pricing Issue required the FTT to decide whether an arm’s length independent lender would have made the $4 billion of loans to LLC5 as LLC4 did and on the same terms. The FTT made further findings of fact in addition to the SOAF. The FTT heard expert evidence from Mr Timothy Ashley (“Mr Ashley”) on behalf of LLC5 and Mr Simon Gaysford (“Mr Gaysford”) on behalf of HMRC. They were quite different experts: Mr Ashley’s experience being in interest rate markets; whereas Mr Gaysford’s expertise was in economics.

The FTT began by recording the evidence of the two experts who gave oral evidence at [67]-[82]. It then identified the issue in dispute as follows at [87]-[88]:

It noted the agreed position between the parties at [89]:

The FTT outlined the parties’ respective cases at [90]-[92]. The parties were agreed that an independent arm’s length lender would require to be provided with covenants from various third-parties (including BGI and LLC6) in addition to those from LLC5 were it prepared to stand in the shoes of LLC4 and make a $4 billion loan to LLC5:

The FTT made findings regarding the experts’ evidence as to whether such third-party covenants would be forthcoming and the extent to which they were agreed at [96]:

The FTT then came to its conclusion at [101]-[103] which was that an independent lender would have entered into the loans subject to the covenants which would have been forthcoming:

The relevant provisions of the Taxation (International and Other Provisions) Act 2010 (“TIOPA”) are set out in the FTT Decision from [56]. They include the following.

Section 147 TIOPA provides:  

“147 Tax calculations to be based on arm's length, not actual, provision

(1)

For the purposes of this section “the basic pre-condition” is that— 

(a)

provision (“the actual provision”) has been made or imposed as between any two persons (“the affected persons”) by means of a transaction or series of transactions, 

(b)

the participation condition is met (see section 148), 

(c)

the actual provision is not within subsection (7) (oil transactions), and 

(d)

the actual provision differs from the provision (“the arm's length provision”) which would have been made as between independent enterprises.

(2)

Subsection (3) applies if— 

(a)

the basic pre-condition is met, and 

(b)

the actual provision confers a potential advantage in relation to United Kingdom taxation on one of the affected persons. 

(3)

The profits and losses of the potentially advantaged person are to be calculated for tax purposes as if the arm's length provision had been made or imposed instead of the actual provision…

…”

[Emphasis Added]

In respect of the predecessor legislation to section 147(1)(a), in DSG Retail Ltd v HMRC [2009] STC (SCD) 397 (“DSG Retail”), the Special Commissioners said at [65]:

“[The predecessor of s.147(1)] speaks of provision between two persons. For the [legislation] to apply those two persons must be identified. Whilst the transaction or series of transactions by which the relevant provision is made or imposed may encompass transactions between persons other than those two persons, the identified provision must be between two persons only. There is, in our view, no scope for reading 'two' as 'two or more'.”

Section 148 of TIOPA defines the participation condition for the purposes of section 147(1)(b):

“148 The “participation condition”

(1)

For the purposes of section 147(1)(b), the participation condition is met if—

(a)

condition A is met in relation to the actual provision so far as the actual provision is provision relating to financing arrangements, and

(b)

condition B is met in relation to the actual provision so far as the actual provision is not provision relating to financing arrangements.

(2)

Condition A is that, at the time of making or imposition of the actual provision or within the period six months beginning with the day on which the actual provision was made or imposed-

(a)

one of the affected persons was directly or indirectly participating in the management, control or capital of the other, or

(b)

the same person or persons was or were directly or indirectly participating in the management, control or capital of each of the affected persons ...

…

(4)

In this section “financing arrangements” means arrangements made for providing or guaranteeing, or otherwise in connection with any debt, capital or other form of finance.

…”

Section 151 expands upon the definition of “arm’s length provision” set out in section 147(1)(d) to include the position where no provision would have been made on an arm’s length basis:  

“151 Arm’s length provision

(1)

In this Part “the arm's length provision” has the meaning given by section 147(1). 

(2)

For the purposes of this Part, the cases in which provision made or imposed as between any two persons is to be taken to differ from the provision that would have been made as between independent enterprises include the case in which provision is made or imposed as between two persons but no provision would have been made as between independent enterprises; and references in this Part to the arm's length provision are to be read accordingly.”  

Section 152 applies where the actual provision relates to securities:  

“152 Arm's length provision where actual provision relates to securities

(1)

This section applies where– 

(a)

both of the affected persons are companies, and 

(b)

the actual provision is provision in relation to a security issued by one of those companies (“the issuing company”). 

(2)

Section 147(1)(d) is to be read as requiring account to be taken of all factors, including– 

(a)

the question whether the loan would have been made at all in the absence of the special relationship, 

(b)

the amount which the loan would have been in the absence of the special relationship, and 

(c)

the rate of interest and other terms which would have been agreed in the absence of the special relationship…

(3)

Subsection (2) has effect subject to subsections (4) and (5).

(4)

If—

(a)

a company (“L”) makes a loan to another company with which it has special relationship, and

(b)

it is not part of L's business to make loans generally,

the fact that it is not part of L's business to make loans generally is to be disregarded in applying subsection (2).

(5)

Section 147(1)(d) is to be read as requiring that, in the determination of any of the matters mentioned in subsection (6), no account is to be taken of (or of any inference capable of being drawn from) any guarantee provided by a company with which the issuing company has a participatory relationship.

(6)

The matters are—

(a)

the appropriate level or extent of the issuing company's overall indebtedness,

(b)

whether it might be expected that the issuing company and a particular person would have become parties to a transaction involving—

(i)

the issue of a security by the issuing company, or

(ii)

the making of a loan, or a loan of a particular amount, to the issuing company, and

(c)

the rate of interest and other terms that might be expected to be applicable in any particular case to such a transaction.”

Section 154 provides interpretation of defined terms such as ‘special relationship’ ‘guarantee’, ‘participatory relationship’ and ‘security’ in sections 148 and 152:

“154 Interpretation of sections 152 and 153

…

(3)

“Special relationship” means any relationship by virtue of which the participation condition is met (see section 148) in the case of the affected persons concerned.

(4)

Any reference to a guarantee includes—

(a)

a reference to a surety, and

(b)

a reference to any other relationship, arrangements, connection or understanding (whether formal or informal) such that the person making the loan to the issuing company has a reasonable expectation that in the event of a default by the issuing company the person will be paid by, or out of the assets of, one or more companies.

(5)

One company (“A”) has a “participatory relationship” with another (“B”) if—

(a)

one of A and B is directly or indirectly participating in the management, control or capital of the other, or

(b)

the same person or persons is or are directly or indirectly participating in the management, control or capital of each of A and B.

(6)

“Security” includes securities not creating or evidencing a charge on assets.

…”

Section 155 provides the definition of potential advantage in relation to UK taxation for the purposes of section 147(2)(b):

“155 “Potential advantage” in relation to United Kingdom taxation

(1)

Subsection (2) applies for the purposes of this Part.

(2)

The actual provision confers a potential advantage on a person in relation to United Kingdom taxation wherever, disregarding this Part, the effect of making or imposing the actual provision, instead of the arm’s length provision, would be one or both of Effects A and B.

(3)

Effect A is that a smaller amount (which may be nil) would be taken for tax purposes to be the amount of the person’s profits for any chargeable period.

(4)

Effect B is that a larger amount (or, if there would not otherwise have been losses, any amount of more than nil) would be taken for tax purposes to be the amount for any chargeable period of any losses of the person…”

Section 164 provides that Part 4 TIOPA “is to be read in such manner as best secures consistency between” the effect given to the legislation and the effect to be given in accordance with the OECD’s Transfer Pricing Guidelines (“the TPG”).  

HMRC advanced five grounds of appeal on the Transfer Pricing Issue for which permission was granted by the FTT.

The first ground of appeal was that the FTT erred in law by concluding that while an independent hypothetical arm’s length enterprise would not have entered into the loans on the same terms as the actual transaction it would have done so if certain covenants had been given by entities other than LLC5. The argument is that the FTT impermissibly altered the economically relevant characteristics by reading covenants from third parties (other than LLC5) into the arm’s length transaction. This was impermissible because section 147(1)(a) TIOPA limits the provision of the loan to one made between two parties i.e. the independent lender and LLC5. Further, these covenants did not exist in the actual transaction.

Second, even if contrary to the first ground, it were permissible to consider (1) covenants which did not exist in the actual provision when determining the terms of the arm’s length provision; and (2) such covenants would include those provided by other group companies, the FTT failed to properly consider the level of parental support required or other covenants regarding control of the borrower that the independent lender would need, and whether such covenants would be forthcoming.

Third, the FTT’s findings in relation to the covenants that would have been required by a hypothetical lender before entering into the transaction were unsupported by the evidence and its own findings in relation to that evidence.

Fourth, the FTT also erred in failing to consider whether a lender with the characteristics of LLC4 would have entered into the transaction, namely to have lent $4 billion to an independent borrower in the position of LLC5 for it to acquire preference shares.

Fifth, the FTT made an error at [86] where it decided not to take any account of Mr Gaysford’s evidence so far as it concerned the construction of the TPG. To the extent that the TPG set out economic, not legal, principles Mr Gaysford was qualified to comment, being an economist.

Mr Ewart QC concentrated on the first and third grounds in oral argument.

Mr Prosser QC, for LLC5, submitted that the FTT did not err in law in any of the ways suggested in the grounds of appeal. We address his arguments within the discussion section below.

The FTT found the “hypothetical” or “arm’s length” transaction to be a $4 billion loan by an independent lender to LLC5 having regard to the covenants which such an independent lender would have required.

Mr Ewart QC submitted that the FTT erred in law when it carried out the exercise of comparing the “actual transaction” with the “arm’s length transaction”, as required by s. 147(1)(d) TIOPA. Section 147(1)(a) refers to the “provision…made or imposed as between any two persons by means of a transaction or series of transactions” and he argued that that was necessarily limited, as per DSG Retail, to the provision between the borrower (LLC5) and the lender (LLC4). The hypothetical transaction has to be comparable to the actual transaction in terms of its economically relevant characteristics and therefore it is not permissible to add covenants from third parties such as LLC4, LLC6 and BGI that were not provided in the actual transaction. Mr Ewart QC submitted that the FTT had effectively reverse-engineered a fundamentally different transaction in a wholly different context so as to conclude that such a transaction would have taken place at arm’s length.

In reply, Mr Prosser QC submitted that the FTT was entitled to take into account covenants which did not exist in the actual transaction when determining whether there would be such a transaction at arm’s length. The relevant “provision” in s.147 TIOPA, is the loan from LLC4 to LLC5 and such a loan would have happened at arm’s length if LLC4 was an independent enterprise. The covenants that such a lender would require were, according to Mr Prosser QC, akin to services provided by third parties to the transaction in order to replicate the economically relevant characteristics of the actual transaction. The FTT was therefore allowed to take them into account and rely on such a hypothetical transaction as being sufficiently comparable to the actual transaction.

Mr Prosser QC further argued that s.152(5) TIOPA, which provides that no account is to be taken of any guarantee provided by a member of the borrower’s group, implicitly recognises that account may in principle be taken of services provided by another person, including by a member of the group. Moreover, he submitted that it was surprising that HMRC were advancing this ground, because if it is correct it would preclude the application of the transfer pricing legislation in many cases. Furthermore, there is no hint of HMRC’s approach anywhere in the TPG.

We begin our analysis by noting that, in fairness to the FTT, HMRC do not appear to have relied on the same arguments at first instance. Before the FTT, HMRC disputed whether the $4 billion of Loans to LLC5 would have been made by independent enterprises at arm’s length. HMRC submitted that the covenants that would have been required by an independent lender would not have been forthcoming. Nonetheless, permission was granted for HMRC to pursue this ground of appeal before us and no objection has been taken.

As correctly identified in the Decision at [101], the FTT considered that the exercise it had to carry out was to compare the “actual transaction” with an “arm’s length transaction”, as required by s. 147(1)(d) TIOPA 2010 and in a manner consistent with the TPG.

At [77] of the Decision, the FTT made an important finding of fact. It held that the arm’s length loan of $4 billion would not have been made by an independent lender on the terms and conditions of the Loans:

“[77] Both the Joint Statement and the Gaysford Statement (Footnote: ⁴) record that the experts agree that it would have been possible for LLC5 to execute a $4 billion debt transaction in December 2009 with an independent enterprise at similar interest rates to the actual transaction that took place between LLC5 and LLC4, but subject to different terms and conditions that independent lenders would have required to manage the credit risks appropriately.”

The FTT made it explicit at [89] of the Decision that the actual Loans would not have occurred at arm’s length with an independent lender:

“[89] It is clear from the evidence of the experts that the transaction that was actually entered into would not have taken place in an arm’s length transaction with an independent lender. It is therefore necessary to hypothesise a different transaction which independent enterprises would have entered into and, as it is for LLC5 to displace the closure notice and amendment made to its self-assessment, it can only succeed on the transfer pricing issue by positively establishing that there is a hypothetical transaction in which a hypothetical independent enterprise would lend $4 billion dollars to LLC5.”

As set out above, at [103] of the Decision, the FTT concluded that an independent lender would not have entered into the loans on the same terms as the actual Loans, and it would only have loaned $4 billion to LLC5 if the covenants (set out at [90] of the Decision) had been given by other entities in the BlackRock Group. The question is whether the FTT was entitled to conclude in those circumstances that the hypothetical provision with different terms is the appropriate comparator for the purposes of s.147(1)(d) TIOPA.

The focus of s.147(1)(a) TIOPA is on the “two persons” between whom the provision is made, in this case the lender and borrower. This is reinforced by Article 9 of the OECD’s Model Tax Convention, which is similarly worded to s. 147 (albeit it refers to “conditions” as opposed to “provision”) and refers to adjusting profits of “one of the enterprises” by reference to the conditions which would have been agreed between them if acting independently:

“…conditions are made or imposed between the two enterprises in their commercial or financial relations which differ from those which would be made between independent enterprises, then any profits which would, but for those conditions, have accrued to one of the enterprises, but, by reasons of those conditions, have not so accrued, may be included in the profits of that enterprise and taxed accordingly.”

The arm’s length principle embodied in s.147, and in particular the separate entity approach, is explained in para. 1.6 of the TPG, which the FTT referred to at [60] of the Decision:

“…arm’s length principle follows the approach of treating the members of [MNE, a multinational enterprise] group as operating as separate entities rather than as inseparable parts of a single unified business. Because the separate entity approach treats the members of an MNE group as if they were independent enterprises, attention is focused on the nature of the transactions between those members and on whether the conditions thereof differ from the conditions that would be obtained in comparable uncontrolled transactions. Such an analysis …is at the heart of the application of the arm’s length principle. [emphasis added]”

The hypothetical transaction must be sufficiently comparable with the actual transaction for the purpose of testing it. The 1995 version of the TPG, which was in force at the time of the transaction, discussed comparability at length (as do all subsequent versions). Paragraph 1.15 stated:

“Application of the arm’s length principle is generally based on a comparison of the conditions in a controlled transaction with the conditions in transactions between independent enterprises. In order for such comparisons to be useful, the economically relevant characteristics of the situations being compared must be sufficiently comparable. To be comparable means that none of the differences (if any) between the situations being compared could materially affect the condition being examined in the methodology (e.g. price or margin), or that reasonably accurate adjustments can be made to eliminate the effect of any such differences. In determining the degree of comparability, including what adjustments are necessary to establish it, an understanding of how unrelated companies evaluate potential transactions is required. Independent enterprises, when evaluating the terms of a potential transaction, will compare the transaction to the other options realistically available to them, and they will only enter into the transaction if they see no alternative that is clearly more attractive…[emphasis added]”.

The “economically relevant characteristics” are described at paragraph 1.17 of the TPG:

“…In order to establish the degree of actual comparability and then to make appropriate adjustments to establish arm’s length conditions (or a range thereof), it is necessary to compare attributes of the transactions or enterprises that would affect conditions in arm’s length dealings. Attributes that may be important include the characteristics of the property or services transferred, the functions performed by the parties (taking into account assets used and risks assumed), the contractual terms, the economic circumstances of the parties, and the business strategies pursued by the parties…[emphasis added]”.

In DSG Retail, the Special Commissioners stated at [78] that the similarly worded predecessor legislation to s.147(1)(d) TIOPA (paragraph 1(2)(a) of Sch 28AA of the Income and Corporation Taxes Act 1988) should be interpreted as requiring consideration of what provision independent enterprises sharing the characteristics of the actual enterprises would have made.

Risk is of particular significance in a lending transaction and is highly relevant to transfer pricing – see DSG Retail at [55]. It is discussed further at para. 1.23 of the TPG:

“…In the open market, the assumption of increased risk will also be compensated by an increase in the expected return. Therefore, controlled and uncontrolled transactions and entities are not comparable if there are significant differences in the risks assumed for which appropriate adjustments cannot be made. Functional analysis is incomplete unless the material risks assumed by each party have been considered since the assumption or allocation of risks would influence the conditions of transactions between the associated enterprises…[emphasis added]”.

The experts were agreed that the main concern of an independent lender of $4 billion to a company like LLC5 would be the risks around the fact that the borrower in the position of LLC5 had no control over the dividend flow to it and so its ability to repay the loan. In the actual transaction, LLC4 through its ownership of the LLC6 Common Shares, controlled LLC6 and so the dividend flow to LLC5. LLC4 did not therefore need covenants from LLC6 or BGI. In the hypothetical transaction however, the dividend flow would need to be secured so far as possible and this is what the covenants imposed by the FTT sought to do. Mr Prosser QC said that this was necessary to replicate the actual risks assumed by LLC4 in the Loans and that these were only services being provided by third parties that do not affect the substantive “provision” under consideration. By contrast, Mr Ewart QC submitted that not only did this materially alter the economically relevant characteristics of the transaction but also it changed the surrounding circumstances and context of the actual transaction, which is not permissible under the wording of the legislation.

We agree with Mr Ewart QC. Importing third-party covenants into a hypothetical transaction that did not exist in the actual transaction changes the nature of the provision which is to be compared. By permitting the introduction of third-party covenants, the FTT essentially compared a different transaction to the actual one. It enabled the FTT to reach the conclusion that the same Loans and terms including interest rates as between LLC4 and LLC5 would have occurred at arm’s length. However, we do not consider that new third-party covenants that were not present in the actual transaction and which materially affect the economically relevant characteristics of the transaction can be so imposed to provide a comparable arm’s length transaction.

In support of his argument that the provision of services by third parties can be introduced into the hypothetical transaction, Mr Prosser QC posited a number of examples that he said showed that HMRC’s so-called “two party rule”, that is that third parties cannot be introduced into the transaction as s.147 is only concerned with the “provision” between two persons, cannot be right. Indeed, he submitted that it may actually work against HMRC in certain respects. He gave three examples.

The first example was of a company, P, making a loan to its subsidiary S on terms that the interest rate is to be set by T, a third-party with no expertise in interest rate setting. In the arm’s length transaction, the experts agree that independent enterprises dealing at arm’s length would have agreed instead that the interest rates should be set by an expert interest rate setter rather than T. The expert setter would have set a higher rate of interest than T. Nonetheless, Mr Prosser QC submitted that if a fresh or different third-party could not be introduced in the arm’s length relationship, P could not be taxed on the higher interest rates it would have received.

The second example was of an interest free loan being made informally from a parent to its subsidiary with no detailed terms and without lawyers being utilised. However, in the arm’s length loan, independent enterprises would not have lent or borrowed at all without using third-party legal services to advise and draft the terms. Again, Mr Prosser QC submitted that if HMRC’s two party rule were correct, the parent could not be taxed on the interest which would have been charged on the arm’s length transaction.

Mr Prosser QC gave a further example in his oral submissions of an independent lender insisting on the borrower being credit-rated by an independent credit rating agency. Obviously, such would not be required in an intra-group transaction but at arm’s length there would be no transaction without the provision by a third-party of credit rating services.

Mr Ewart QC said that these examples are very different from the provision of third-party covenants that impinge on the very substance of the transaction and change the relevant risks and economic characteristics of the transaction. The first example would actually provide the evidence for what the interest rate would be at arm’s length. It also included a third-party in the actual transaction. And no one could seriously suggest that the provision of legal services in the second example or credit rating services in the third were akin to the provision of substantive third-party covenants.

We agree with Mr Ewart QC. The introduction of these types of third-party services in the hypothetical transaction are not a part of the substantive loan provision between the independent lender and the borrower. They do not affect the relevant economic characteristics of the parties or the transaction and may simply be helpful in determining the arm’s length price or interest rate that would have been agreed.

The other main strand to Mr Prosser QC’s argument was his reliance on s.152 TIOPA. Section 152(1)(b) applies where the actual provision is in relation to a security by an ‘issuing company’ (such as in this case, LLC5 which issued the loan notes to secure the loans). Section 152(5) requires, when conducting the analysis under sections 147(1)(d) and 152(6), that any guarantee provided by a company with which the ‘issuing company’ in the actual transaction has a participatory relationship (including membership of the same group of companies) should not be taken into account.

Mr Prosser QC argued that in this case there was no guarantee from LLC4 to LLC5 in the actual provision of the loan. However, ss.152(5) and (6) do not contemplate any guarantee from the parent company in the actual provision. The focus of ss.152(5) and (6) is on forbidding a guarantee to be taken into account in the hypothetical transaction because it would or might enable an arm’s length loan transaction to go ahead with the benefit of a deduction for interest which in the real world it would not. He submitted that it was most unlikely that Parliament was contemplating a situation where, and only where, there was a parental guarantee in the actual transaction which was then ruled out when considering the hypothetical transaction. He submitted that such a parental guarantee would be unnecessary in the intra-group transaction.

Mr Prosser QC’s main point was that the legislation specifically envisages or contemplates third-party involvement in the arm’s length transaction and this shows that there is no two party rule operating.

Mr Ewart QC disagreed with both the interpretation of s.152(5) and with its suggested effect. He submitted that Parliament did indeed contemplate the situation where, specifically for transfer pricing purposes, the group did provide parental guarantees to support the intra-group lending. It was precisely because Parliament did not want taxpayers to be able to argue that because a guarantee had been provided in the actual transaction, so it should be assumed that such a guarantee would be forthcoming in the hypothetical transaction. Therefore, the effect of s.152(5) is to remove consideration of such guarantees altogether when deciding whether the arm’s length transaction would have been made and if so on what terms. Such an interpretation is consistent with Mr Ewart QC’s overall thesis that third-party covenants can only be taken into account in the hypothetical transaction if they existed in the actual transaction.

Again, we agree with Mr Ewart QC’s interpretation of s.152(5). It is unclear why Parliament only chose to rule out guarantees (they are widely defined however in s.154(4)) and neither side was able to explain why the legislation is limited to guarantees. It could be, as Mr Ewart QC submitted, that this was a form of abuse that had been perceived in intra-group lending, and the provision of covenants to secure dividend flows was not something that was contemplated. It is clear to us that, whatever the reason for the specific exclusion of guarantees was, s.152(5) is intended to cover the situation where a guarantee has been provided, for whatever reason, in the actual transaction and it is then ruled out of the hypothetical transaction. This is because it would obviously have a potentially direct effect on price as a parental guarantee reduces the lender’s risk and therefore the interest rate to be charged.

We consider that what s.152(5) shows is that guarantees from third-parties affect the substance of the loan transaction, namely the “provision” between lender and borrower, and Parliament has therefore decided that they should be left out of account in determining the arm’s length provision. This strongly indicates that covenants from third-parties supporting the loan transaction similarly affect its substance. In the case of covenants that are not precluded by s.152(5), these can only be taken into account if they were provided in the actual transaction. If that were not so, the arm’s length transaction would not be sufficiently comparable to the actual transaction as the surrounding circumstances and context would not be the same. It is our view that Mr Ewart QC was correct to submit that the surrounding circumstances, such as the provision of third-party covenants, must be the same in both transactions to ensure that there is a comparison of like with like.

We understand that the effect of this is, as Mr Prosser QC argued, that LLC5 would not have been challenged if it had gone through the rather artificial exercise in the actual transaction of having covenants in place with LLC 4 and LLC 6 and BGI and so on as set out at [90] of the Decision. The BlackRock Group would not have needed to have such covenants in place in the actual transaction because these were intragroup companies, but if they went through that artificial process, then HMRC would have had to concede it would be possible to look at covenants provided by third-parties when considering the arm’s length transaction. This would therefore be open to abuse by a group determined to uphold the transaction for transfer pricing purposes.

We recognise that this is a possible effect. On our analysis, section 152(5) TIOPA was enacted in order to address this situation as regards the unnecessary introduction of guarantees in actual transactions. However, it does not cover covenants such as those provided in this case and it could therefore be used unscrupulously by groups seeking to gain a tax advantage.

However, we believe that this is unlikely to cause such a problem in practice and it will be very obvious if groups have sought to manipulate the actual transaction in that way by including wholly unnecessary covenants that attempt to anticipate what an independent expert might later decide would be required by an independent lender.

We accept that there is nothing in the legislation or the TPG that expressly rules out third-party involvement in the hypothetical transaction. But we consider that the focus of the legislation is clear and that is on the “provision…made or imposed as between any two persons”, which in this case is the Loans, and the only substantive changes that are tested by the “arm’s length provision” are to the terms of the Loans themselves, such as the interest rate. Third-party covenants that were not given as part of or in support of the actual transaction cannot be considered to be part of the hypothetical transaction as this materially changes the surrounding circumstances and alters the economically relevant characteristics of the transactions in question.

In conclusion, the FTT erred in law in permitting new third-party covenants absent from the actual transaction to be taken into account when considering whether an independent lender would make a $4 billion loan to LLC5. The FTT decided that such an independent lender would not have made a $4 billion loan to LLC5 without such covenants being in place and that important finding should itself have determined that there was no comparable arm’s length transaction. Having decided that, the FTT would have been bound to conclude that no “provision would have been made as between independent enterprises” (s.151(2)).

Therefore, we are satisfied that the FTT erred in law in a material manner and its decision should be set aside based upon this first ground of appeal. We will return to the consequences of this below, namely whether to remit or re-make the FTT’s decision.

In light of our decision on the first ground, the remaining grounds of appeal are no longer material, but we address them out of respect to the written and oral arguments we received.

Mr Ewart QC also concentrated upon the third ground of appeal in oral argument. This was that the FTT’s findings in relation to the covenants that would have been required by an independent lender before entering into the transaction were unsupported by the evidence and its own findings in relation to that evidence. He argued that the conclusion at [102] of the Decision that both experts were “agreed” that the covenants would be provided to an independent lender was wrong and based on a misunderstanding of the evidence given by both experts.

He took us through the transcripts of the evidence of the experts, Mr Gaysford and Mr Ashley, at the hearing before the FTT. He argued that the conclusion was wrong for two reasons. First, the experts were not agreed because they did not conceive of the covenants in the same terms. Secondly, Mr Gaysford did not accept, in either his written or oral evidence, that the covenants would have been provided or, that his concerns were not "deal-breakers”. At most, Mr Gaysford said he would rephrase the term “deal-breaker” as meaning there was a better commercial alternative. While he agreed with Mr Ashley that covenants would be required, he did not agree that they would be provided as it was his view that it would have been commercially irrational for the group to provide them, because of the cost and complexity involved in doing so.

Mr Prosser QC submitted that the FTT’s findings in relation to which covenants would have been required were supported by the evidence available to it. It made no error of law in finding so. In summary, it was entitled to find that the experts agreed that whatever covenants and other enhancements would have been required by an independent lender could probably have been achieved.

We are not satisfied that the FTT mischaracterised the evidence of Mr Gaysford in the way suggested. Ultimately, this was a challenge to the fact finding jurisdiction of the FTT. The FTT was entitled to reject HMRC’s interpretation of Mr Gaysford’s evidence. Even if there was a difference in the way the experts viewed the crucial covenant ensuring that preference dividends were paid, the FTT was entitled to decide that all the covenants would have been forthcoming at arm’s length. We accept Mr Prosser QC’s submissions that the FTT’s findings in relation to which covenants would have been required were supported by the evidence and its own findings in relation to that evidence.

Mr Ashley in his report stated that several covenants would be needed to provide protections to external bondholders including “Provision of additional covenants (e.g. preference share payment covenant)”.

This issue was further addressed in the experts’ joint statement set out at [78] of the Decision:

“f.

The preference share structure was unusual but not necessarily problematic given BGI US was already a successfully performing business. The preference shares carried an expectation that the Appellant should receive over USD700m annually in income which would have given it a sizeable debt capacity. The main issue was that the flow of value from BGI US to LLC6 and then to the Appellant via the preference shares was paid at the discretion of LLC4. Whilst a lender would probably be unlikely to accept this position, it should have been possible for BGI US, and LLC5 - with the explicit consent of LLC4 – to effectively ratify the legal and financial position to which the Appellant was entitled, that is the receipt of the value from BGI US. This ratification would most likely have been effected via inter-company agreements and covenants which would have formed part of the Appellant's borrowing transaction. Both experts agree that an independent lender would have required the protection described in this paragraph and that it probably could have been put in place. Mr Gaysford believes it would have been costly and complex to do so. Mr Ashley believes it would have been straightforward and that the associated 'cost' would have been an 'opportunity cost' (i.e. reduced flexibility to enter into further transactions) rather than a cash cost.

g.

In addition to the protections discussed in f above, the purpose of which would have been to secure the flow of value from BGI US and preference share dividends from LLC6, the experts agree that an independent lender would likely also have required other structural enhancements to the terms of the loans, to ensure the cashflow generation of BGI US could not be diverted in any way. Possible additional clauses would include (1) a negative pledge on further indebtedness within BGI US, LLC6 or indeed [LLC5], (2) a change of control clause and (3) a restriction on BGI US or LLC6 being able to lend money to any other entity- whether inside the BlackRock Group or not. These are well known standard clauses required in almost every external debt transaction- though to emphasise, one would not expect to see them in an inter-company transaction within a group.

h.

The experts cannot say with certainty whether all of the possible additional clauses listed on paragraph g would have been required to support a USD4bn loan or bond transaction by [LLC5]. However, in view of the structural subordination of LLC5 (being 2 entities away from the generation of cashflows), the experts agree that an independent lender would have required at least some of the enhancements discussed in paragraph g.

i.

Again, both experts agree that the enhancements discussed in paragraph g would have been necessary, and probably could have been achieved. Mr Ashley believes it would have been straightforward to do so and that the associated “cost” would have been an “opportunity cost” (ie reduced flexibility to enter into further transactions rather than a cash cost. In Mr Ashley’s experience, such enhancements are very common in debt transactions, including the BlackRock Group’s own revolving credit facility. Mr Gaysford believes it would have been costly and complex to do so, and that any “opportunity cost” would have been significant.” [emphasis added]

In summary, the experts agreed that whatever covenants and other enhancements would have been required by an independent lender, this probably could have been achieved. This issue was addressed further by Mr Ashley in his second witness statement at paragraphs 17-22 and he was cross-examined about this.

In cross-examination Mr Gaysford re-confirmed the view set out in the joint statement that any covenants required by the independent lender could have been given. Mr Gaysford, however, considered that the requisite covenants and other enhancements would have given rise to four particular costs and complexities on the borrowing side, namely (i) tax, (ii) the approval of the Office of the Comptroller of the Currency (“OCC”) [the US financial regulator], (iii) cost of debt, and (iv) loss of flexibility to move funds around.

These points were raised for the first time by Mr Gaysford in his evidence (the OCC arrangements in his report, and the other issues in oral evidence). Indeed, when he was questioned on these points, Mr Gaysford was unable to say whether any of them would actually be a problem. In any event, Mr Gaysford clarified in cross-examination that he was not suggesting that any of the four points was a “deal breaker”, merely that there was a “better commercial alternative” to the actual transaction, such that it would have been commercially irrational for the group to enter into it. To a certain extent, these points add to our concerns as to whether it is appropriate to take such third-party covenants into account in assessing the arm’s length transaction. But as findings of fact, they were open to the FTT.

Therefore, the FTT did not mischaracterise the effect of the evidence nor make unreasonable or unsupported findings at [102] of the Decision. It was entitled to conclude that an independent lender would have entered into the $4 billion loan arrangement made with LLC5 subject to it being able to obtain the necessary covenants and that the covenants would have been forthcoming.

HMRC had permission to argue further grounds of appeal. While he did not concentrate upon these orally, Mr Ewart QC did rely on three further grounds in his skeleton argument.

HMRC’s second ground of appeal was that the FTT failed properly to consider the level of parental support required and whether covenants would have been forthcoming, and in particular failed to consider why those covenants would have been given when, by so doing, they would enable a funding structure which was more expensive (circa $40m per year) than alternatives. By “alternatives”, HMRC had in mind in particular the alternative of the parent of the group, BRI, borrowing instead of a subsidiary with a lower credit rating, such as LLC5. This overlaps with Ground 3 above.

The fourth ground of appeal was that the FTT erred in failing to consider whether an independent lender with the other characteristics of LLC4 would have entered into the transaction, namely to have lent $4 billion to an independent borrower in the position of LLC5 for it to acquire preference shares. The independent lender’s most closely related option realistically available to the actual transaction would have been to invest directly in LLC6 or even elsewhere in the BlackRock Group, in order to avoid the added risk inherent in investing via an entity which did not control its income flows and required supportive covenants to mitigate that risk.

Both of these grounds of appeal suffer from the same problem that Mr Prosser QC identified. The transfer pricing legislation is concerned with whether, and if so on what terms, an independent enterprise would have lent $4 billion to LLC5 and whether a company with LLC5’s economic attributes would have been able to borrow $4 billion, not with whether it would be cheaper for the BlackRock Group to borrow elsewhere.

Mr Prosser QC was right to identify that the second and fourth grounds of appeal relied upon postulating a breach of the ‘separate entity approach’. It may be that a hypothetical arm’s length lender would have preferred to lend to a different entity in the BlackRock Group rather than LLC5 but that is not the point – that would be to compare a different transaction to the one that had actually taken place. The FTT conducted the necessary hypothetical exercise of examining the nature of the arms’ length lending to a company in the position of LLC5 and considering the question of the terms on which the arm’s length lender would lend to LLC5. As we have found however, it impermissibly took into account third-party covenants that such an independent lender would have required if it was to lend $4 billion to LLC5.

The fifth ground of appeal, regarding whether Mr Gaysford had the required expertise to express opinion on the TPG, was immaterial to the outcome of the case.

We conclude that there was no error of law by the FTT in relation to Grounds 2 and 4.

The FTT at [104]-[105] disregarded an alternative argument relied upon by Mr Prosser QC which was described as ‘the co-investor’ approach – this was whether an arm’s length independent lender, in addition to the loaning of $4 billion to LLC5, would also have subscribed for the LLC6 Common Shares to put it in the same position as LLC4.

We are satisfied that the FTT made no error in disregarding that argument for the reasons it gave at [105] of the Decision:

We therefore do not accept that the FTT erred in discounting the alternative co-investor hypothesis of Mr Prosser QC.

We have found that the FTT erred in law for the reasons set out above in relation to Ground 1 and we therefore allow HMRC’s appeal in this respect.

We are satisfied that we should remake the decision applying section 147(1)(d), (2) and (3) of TIOPA.  The actual provision of the loans from LLC4 to LLC5 differed from any arm's length provision in that the loans would not have been made as between independent enterprises. The actual provision conferred a potential advantage in relation to United Kingdom taxation. The profits and losses of LLC5, including the allowing of debits for the interest and other expenses payable on the Loans, are to be calculated for tax purposes as if the arm's length provision had been made or imposed instead of the actual provision. In this case, no arm’s length loan for $4 billion would have been made in the form that LLC4 made to LLC5 and hence HMRC’s amendments to the relevant returns should be upheld and confirmed.

The Unallowable Purpose Issue concerns whether there was a commercial purpose to the Loans or whether the purpose was to secure a tax advantage (or whether there were dual purposes). Both parties challenge some of the findings of fact made by the FTT on the Unallowable Purpose Issue so it is necessary to explore the findings in some detail.

Our conclusion on the Transfer Pricing Issue renders the Unallowable Purpose Issue immaterial to the outcome of this appeal. Nevertheless, we received full argument on this and deal with all the grounds raised by both HMRC and LLC5.

The FTT received witness statements and heard oral evidence from two witnesses of fact on behalf of LLC5: Mr Nigel Fleming (“Mr Fleming”) who had been the BlackRock Group’s EMEA head of tax at the time of the acquisition; and Mr J. Richard Kushel (“Mr Kushel”) who was a member and/or chair of several BlackRock Group entities. Mr Kushel was appointed to LLC5’s board on 30 November 2009. He presided over the board meeting which took place on that date, the minutes of which are addressed below, and in respect of which the FTT made findings at [42]-[53] of the Decision.

The FTT made further findings of fact relevant to the purpose of the transactions at [10]-[30] and [34]-[54] of the Decision.

The FTT began by making findings regarding the proposed acquisition of BGI US by the BlackRock Group as from June 2009. It found there was much internal discussion as to the structure of the acquisition which ultimately resulted in the proposals by which LLC4, LLC5 and LLC6 would be registered and utilised. This occurred in the period July to November 2009 – well before the board meeting and minutes of LLC5 on 30 November 2009 and the execution of the agreement for LLC6 to acquire the shares in Delaware Holding Inc, BGI’s parent, on 1 December 2009 (see [10]-[30]).

In particular, the FTT made findings regarding the start of the relevant period in June 2009 at [15]-[16]:

“15.

EY had been given, what Mr Fleming described as, a “very broad remit” to consider the whole structure of the acquisition of BGI taking into account a wide range of options. Mr Fleming understood that EY had initially suggested that BGI should be acquired through a UK resident entity taking on intercompany debt to take advantage of the “generous tax regime for interest deductions” operating at that time. He explained that he was already aware of the concept of including a UK resident entity in an otherwise US resident holding structure having previously discussed the possibility with HMRC and the then Financial Secretary to the Treasury, Stephen Timms MP, in connection with the “worldwide debt cap” rules subsequently introduced by Part 7 of TIOPA to limit the extent to which UK tax deductions can be claimed for finance expenses incurred on loan relationships entered into between group companies.

16.

However, rather than a UK limited company the decision was taken “relatively quickly” for BGI to be acquired by a US limited liability company (“LLC”) that was UK resident………”

It went on to find the following at [18]-[20]:

The FTT recorded the agreed fact that LLC5 had been registered in Delaware on 16 September 2009 (see [4(5)] of the Decision). It made findings regarding the advice received and preparations for its first board meeting on 30 November 2009. At [34] the FTT addressed the briefing note that was provided on 26 October 2009 to three employees of the BlackRock Group who were ‘potential members of the LLC5 Board’:

Mr Thomson, Mr Tooze and Mr DesMarais, to whom this note was sent, were to become three of the four members of the board of LLC5 (they were also to be joined by Mr Kushel, as set out below).

At [36], [38] & [39] the FTT went on to find that in November 2009 advice was given to the three proposed board members as to the tax implications of the proposed transaction in which LLC5 was to take part:

The FTT then addressed the evidence regarding the board meeting of LLC5 on 30 November 2009 by reference to the oral evidence of Mr Kushel, being the only one of the board members to give evidence before the FTT:

The FTT accepted Mr Kushel’s oral evidence at [49]-[53] as follows:

The board minutes for the meeting of LLC5 on 30 November 2009 were before the FTT and it noted at [43] that:

These documents were made available to the board. They set out the UK tax advantages to the BlackRock Group of structuring the acquisition of BGI through LLC4-6 which included the proposed loan transaction between LLC4 and LLC5.

Although they were not set out in the Decision, the LLC5 board minutes for the meeting on 30 November 2009 record relevantly as follows:

Having made these factual findings, the FTT came to its conclusions as to whether the loan transaction between LLC4 and LLC5 had as one of its main purposes an unallowable purpose, namely securing a tax advantage in the UK.

The FTT began by setting out its understanding of the law to be applied – that it was required to ascertain the subjective intention of the directors of LLC5 in entering the loan transaction - at [117]:

“117.

As a company is a legal construct, as is clear from Oxford Instruments, its intentions are, in the absence of evidence that they are acting as a puppet, being usurped or by-passed in the decision making process, those of its directors. In the present case there is nothing to suggest that, despite their significant involvement in the creation and advice on the implementation of the transaction, LLC5 ceded control to its advisers or other parts of the BlackRock Group. As such it is necessary to ascertain the subjective intentions and purpose of LLC5’s directors in respect of the transaction.”

It found that one of the main purposes of the transaction was securing a tax advantage for the following reasons at [118]-[121]:

In addition, the FTT also found that one of the main purposes of LLC5 entering into the loan transaction with LLC4 was also a commercial purpose as set out at [121]:

Having made a finding of there being both an unallowable tax advantage purpose and a commercial purpose to the transaction, the FTT then went on to decide the just and reasonable apportionment should be solely to the commercial purpose rather than the tax advantage purpose at [122]-[123]:

This abbreviated reasoning by the FTT was criticised particularly by HMRC but also on behalf of LLC5. The question for us is whether the factual conclusions as to there being two main purposes of the transaction were properly open to the FTT and if so whether it correctly apportioned all of the debits to the commercial purpose.

The relevant legislation in respect of the Unallowable Purpose Issue is set out in the Decision from [107]. By way of background, Part 5 of the Corporation Taxes Act 2009 (“CTA 2009”) sets out the provisions applicable to a company’s loan relationships. According to s. 302(1) CTA 2009, a company has a loan relationship “if it stands in the position of a creditor or debtor as respects any money debt… [which] arises from a transaction for the lending of money”.

Section 296 CTA 2009 provides: “Profits and deficits arising to a company from its loan relationships are to be calculated using the credits and debits given by this Part.”

The credits and debits arising from a company’s loan relationships in an accounting period are set off against each other and if the debits exceed the credits (or there are no credits), a deficit arises which can be surrendered as group relief to other group companies under Part 5 of the Corporation Tax Act 2010.

Chapter 15 of Part 5, CTA 2009 is headed, “Tax Avoidance”. Section 440 provides:

“Introduction

440 Overview of Chapter

(1)

This Chapter contains rules connected with tax avoidance.

(2)

In particular—

(a)

for rules about unallowable purposes and tax relief schemes and arrangements, see sections 441 to 443, …”

Section 441, as relevant, states:

“Unallowable purposes and tax relief schemes

441 Loan relationships for unallowable purposes

(1)

This section applies if in any accounting period a loan relationship of a company has an unallowable purpose.

…

(3)

The company may not bring into account for that period for the purposes of this Part so much of any debit in respect of that relationship as on a just and reasonable apportionment is attributable to the unallowable purpose.

…

(6)

For the meaning of “has an unallowable purpose” and “the unallowable purpose” in this section, see section 442.”

[Emphasis Added]

Section 442 provides:

“442 Meaning of ‘unallowable purpose’

(1)

For the purposes of section 441 a loan relationship of a company has an unallowable purpose in an accounting period if, at times during that period, the purposes for which the company—

(a)

is a party to the relationship, or

(b)

enters into transactions which are related transactions by reference to it,

include a purpose (“the unallowable purpose”) which is not amongst the business or other commercial purposes of the company.

(2)

If a company is not within the charge to corporation tax in respect of a part of its activities, for the purposes of this section the business and other commercial purposes of the company do not include the purposes of that part.

(3)

Subsection (4) applies if a tax avoidance purpose is one of the purposes for which a company—

(a)

is a party to a loan relationship at any time, or

(b)

enters into a transaction which is a related transaction by reference to a loan relationship of the company.

(4)

For the purposes of subsection (1) the tax avoidance purpose is only regarded as a business or other commercial purpose of the company if it is not—

(a)

the main purpose for which the company is a party to the loan relationship or, as the case may be, enters into the related transaction, or

(b)

one of the main purposes for which it is or does so.

(5)

The references in subsections (3) and (4) to a tax avoidance purpose are references to any purpose which consists of securing a tax advantage for the company or any other person.”

[Emphasis Added]

A “Tax advantage” is defined in s. 1139(2) Corporation Tax Act 2010 as including, “a relief from tax or increased relief from tax”. It is common ground that the deduction of loan relationship debits in respect of loan interest pursuant to Part 5 of CTA 2009 is a tax advantage.

In Travel Document Service & Ladbroke Group International v HMRC [2018] STC 723 (“TDS”), Newey LJ explained these provisions as follows at [39]-[47]:

HMRC had two broad grounds of appeal as follows.

First, the FTT’s finding that LLC5 had a commercial purpose in entering into the Loans was one which could not be reasonably entertained within the meaning of Edwards v Bairstow [1956] AC 14. The FTT found that LLC5 entered into the Loans in furtherance of the commercial purpose of its business of making and managing passive investments. However, the only investment made by LLC5 was in LLC6’s Preference Shares. There was no evidence before the FTT that the board had considered the likelihood of the cash flows it would receive from this investment given it had no control over the dividends received from LLC6. The “commercial purpose” found by the FTT was therefore nothing more than a consequence of the tax purpose.

Second, even if there were two main purposes to the Loans, a just and reasonable apportionment needed to be carried out pursuant to s.441(3) CTA 2009. In coming to the conclusion that all the debits ought to be apportioned to the commercial purpose, the FTT applied the wrong test. The only reason given by the FTT for its conclusion at [123] of the Decision that all the debits ought to be apportioned to the commercial purpose was that Mr Kushel’s evidence had been that LLC5 would have entered into the Loans even if there had been no tax advantage in doing so. The correct approach is to determine whether the reason the debits existed was in order to obtain a tax advantage on the basis of an objective consideration of all of the relevant facts and circumstances; the statutory test is thus to be applied without any gloss: see Versteegh Ltd and others v HMRC [2014] SFTD 547 at [166]. If the FTT had applied such a test, it would have found that the debits would not have arisen in the absence of the unallowable tax purpose.

LLC5 challenge the FTT’s finding that there was an unallowable main tax purpose to the Loans. The FTT was right to consider the subjective purpose of LLC5 in entering into the Loans and it did not find that LLC5’s directors had a subjective purpose of obtaining a tax advantage. Neither was there any evidence to support such a finding. LLC5 maintain that the FTT erred in law at [120] of the Decision in wrongly applying the test in Mallalieu v Drummond (Inspector of Taxes) [1983] STC 665, [1983] 2 AC 861 (“Mallalieu") on the inevitable and inextricable consequences of this type of arrangement.

Both parties have accepted that their respective challenges to the FTT’s findings as to the purpose of the Loans are challenges to the factual findings of the FTT. As such, they can only successfully challenge those findings if they satisfy the high threshold set out in Edwards v Bairstow [1956] AC 14.

The Upper Tribunal recently summarised the position in relation to such a challenge in HMRC v Anna Cook [2021] UKUT 15 (TCC), at [18]-[19]:

“18.

An appeal to this tribunal lies only on a point of law: section 11(1) of the Tribunals, Courts and Enforcement Act 2007 (“TCEA 2007”). While there cannot be an appeal on a pure question of fact which is decided by the FTT, the FTT may arrive at a finding of fact in a way which discloses an error of law. That is clear from Edwards v Bairstow [1956] AC 14. In that case, Viscount Simonds referred to making a finding without any evidence or upon a view of the facts which could not be reasonably entertained, and Lord Radcliffe described as errors of law cases where there was no evidence to support a finding, or where the evidence contradicted the finding or where the only reasonable conclusion contradicted the finding. Lord Diplock has described this ground of challenge as “irrationality” (Footnote: ⁵).

19… we have borne in mind the caveats helpfully summarised in Ingenious Games LLP & Others v HMRC [2019] UKUT 226 (TCC), at [54]-[69]. The bar to establishing an error of law based on challenges to findings of fact is deliberately set high, and that is particularly so where the FTT is called on to make a multi-factorial assessment. As stated by Evans LJ in Georgiou v Customs and Excise Commissioners [1996] STC 463, at 476:

… for a question of law to arise in the circumstances, the appellant must first identify the finding which is challenged; secondly, show that it is significant in relation to the conclusion; thirdly, identify the evidence, if any, which was relevant to that finding; and fourthly, show that that finding, on the basis of that evidence, was one which the tribunal was not entitled to make. What is not permitted, in my view, is a roving selection of evidence coupled with a general assertion that the tribunal's conclusion was against the weight of the evidence and was therefore wrong. A failure to appreciate what is the correct approach accounts for much of the time and expense that was occasioned by this appeal to the High Court.”

In FAGE UK Ltd v Chobani UK Ltd [2014] EWCA Civ 5 Lewison LJ stated at [114]:

“Appellate courts have been repeatedly warned, by recent cases at the highest level, not to interfere with findings of fact by trial judges, unless compelled to do so. This applies not only to findings of primary fact, but also to the evaluation of those facts and to inferences to be drawn from them.”

It follows that the question before us when exercising an “error of law” jurisdiction is not whether we would have made the same decision as the FTT. The appeal is not a re-run of the first-instance trial: as Lewison LJ also said in FAGE at [114], “the trial is not a dress rehearsal. It is the first and last night of the show”. The test is whether the FTT’s factual findings or evaluative judgment were within a reasonable range of conclusions that a properly directed tribunal could have made on the evidence available to it.

The FTT’s conclusion in relation to the commercial purpose of entering into the Loans was contained in one short sentence in [121]: “However, the evidence is that LLC5 entered into the Loans in the furtherance of the commercial purpose of its business of making and managing passive investments.” Mr Prosser QC submitted that the reference to the “evidence” is principally to the board minutes of the 30 November 2009 board meeting and to Mr Kushel’s oral evidence about that meeting, which was accepted by the FTT. He said that HMRC could not get close to passing the Edwards v Bairstow test for setting aside this factual finding.

Mr Ewart QC submitted that this finding was one that no reasonable tribunal properly instructed could have arrived at based upon the evidence before it. At most, the “commercial purpose” of LLC5 and the Loans, as identified by the FTT, was a consequence of the tax purpose as opposed to being a “purpose” in itself. The FTT appears to have assumed that as LLC5’s board was satisfied that entering into the Loans was a decision they could take in accordance with their fiduciary obligations, there was therefore a self-standing commercial purpose in entering into the Loans.

He submitted that the board minutes of 30 November 2009 do not show that LLC5 was considering any commercial purpose for the transaction. They mainly record the advice given to the board by Mr Fleming of EY on a range of matters, such as regulatory compliance and the debt cap which did not include the actual commercial purpose of the transaction being approved. He submitted that all the board minutes reveal is that LLC5’s board approved the transaction as being regulatorily compliant and not adverse to its commercial interests – that it was commercially viable. This was clear from paragraph 3 of the board minutes.

Likewise, Mr Ewart QC made the same points about the FTT’s finding about Mr Kushel’s evidence regarding the loan transaction at [119]:

He submitted that this was not evidence of the directors’ subjective purpose in receiving the loan of $4 billion from LLC4 as being for any commercial purpose, let alone the ‘making and managing of passive investments’ as was found by the FTT. It was merely a finding that the board considered the transaction to be commercially viable. This was further evidenced later in paragraph 3 of the board minutes: ‘it was necessary for the transaction to be considered by the board as viable for the Company without taking any UK tax advantage into account.’

The FTT’s finding that LLC5’s business was that of ‘making [and managing] passive investments’ appears to be based on the contents of Mr Fleming’s note of 26 October 2009 advising three of the potential board members (all those other than Mr Kushel) who were to be appointed just over a month later – as referred to in [34] of the Decision:

However, this part of the note merely describes the nature of LLC5’s business, after the Loans had been entered into, rather than its purpose. It also does not deal with the purpose of the Loans. It appears in a section of the note dealing with establishing LLC5’s business as being carried on in the UK.

At the beginning of the same note Mr Fleming described, ‘the purpose of the LLC[5] is to effect the acquisition of the BGI US business from Barclays.’ That could be said to be the real purpose of the Loans and the insertion of LLC5 into the acquisition structure. However, the FTT did not say that this was the purpose in the Decision.

The FTT accepted that Mr Kushel was a reliable and credible witness. It found that Mr Kushel: “considered his role as a board member was to be satisfied that a transaction was in the best interests of the entity concerned” ([49] of the Decision); “Provided the board felt the transaction was commercially advantageous for the company it “would not have been sensible or open to the directors to consider an alternative transaction” [(51]); and “reviewing the transaction he felt “comfortable” with it and had no concerns over its commercial availability” ([52]).

Mr Kushel’s witness statement, at [39]-[41], was to the effect that the Loans represented a good commercial deal and the primary or key purpose of LLC5 and the Loans was to acquire BGI US (referred to as BGINA, short for BGI North America):

Mr Ewart QC submitted that Mr Kushel was fully challenged on these passages in cross-examination. He argued that when Mr Kushel was asked what were the advantages that this structure may have been facilitating apart from the tax advantage, he replied that he knew that it addressed controlled foreign company (‘CFC’) issues and it was part of the broader step plan. Mr Kushel did not know that those CFC issues were only created as a result of the presence of a UK resident entity in the structure. Mr Kushel was therefore unable to provide an explanation of how LLC5 facilitated the acquisition despite the fact that he held various positions in the BlackRock Group including membership of the Global Executive Committee, which was described as the “highest level management body” in the Group.

Mr Ewart QC contended that this is a key point. As LLC5 was unable to put forward any reason why it was needed for the acquisition of BGI US, it follows that it did not serve any commercial purpose. By becoming a party to the Loans used to fund the acquisition of the LLC6 Preference Shares, LLC5 was simply furthering a tax purpose.

As he further submitted in relation to the Transfer Pricing Issue, it was common ground between the experts that an independent lender would not have lent $4 billion to LLC5 on the same terms as the Loans and the FTT made a finding to that effect: see [89] of the Decision. The FTT also found that the cost of borrowing would have been considerably lower (in the region of $40 million per annum) if the lending had been made to a different group entity [99]. These factors further illustrate that the tax purpose was the only purpose LLC5 had in entering into the Loans.

We are not satisfied that these criticisms take HMRC far enough to satisfy the high threshold required by Edwards v Bairstow in order to set aside the FTT’s finding that there was a commercial purpose to LLC5 and the Loans which was a main purpose.

Mr Ewart QC did not challenge the accuracy of the board minutes. Furthermore, the FTT found Mr Kushel to be a reliable witness and accepted his evidence. Ultimately Mr Kushel’s evidence as to a purpose of LLC5 and the Loans being a commercial one was accepted by the FTT. The FTT could have explained its conclusion in [121] more clearly and more fully. The compressed basis for the conclusion by reference to LLC5’s general business activities did not capture the evidence that had been accepted. Nonetheless, it was clear from [118] of the Decision that the FTT’s reference at [121] to “the evidence” included, in particular, the board minutes of the 30 November 2009 board meeting and Mr Kushel’s oral evidence about that meeting. There was no other direct evidence of what was in the minds of LLC5’s directors at the time.

Accordingly, while the FTT’s findings could have been better expressed, we do not consider that the Edwards v Bairstow threshold has been crossed. The FTT was entitled to find that one of the main purposes of LLC5 and the Loans, as subjectively held by its board members, was a commercial purpose. There was support in the evidence for that factual conclusion.

In addition to there being a main commercial purpose, the FTT also found there was a tax advantage purpose held by LLC5 in relation to the Loans. Mr Prosser QC argued that while the FTT was correct to consider the subjective purpose of LLC5 in entering into the Loans ([115] and [117] of the Decision), it wrongly sought to apply the test in Mallalieu to the subjective intentions of the LLC5 board members (see [120] of the Decision). In Mallalieu the taxpayer claimed to deduct the cost of buying black clothes for her court appearances. The Special Commissioners had found as a fact that preservation of warmth and decency was not a consideration which crossed her mind when she bought the clothes. The High Court and Court of Appeal held that, in the light of that finding of fact, the cost was wholly and exclusively incurred for the purposes of her profession. On appeal to the House of Lords, counsel for the Revenue (Mr Peter Millett QC) argued that the courts below had erred in attaching importance to the taxpayer’s conscious state of mind:

“The evidence of a conscious professional purpose did not displace the inference that there was another subconscious non-professional purpose. When the expenditure is on an everyday item such as clothing, and the pattern of use is consistent and well-established, it is easy as a matter of common sense to infer the taxpayer’s purpose without relying on any active mental process. The general purpose of clothing is taken for granted and no conscious thought is given to it. It did not cross the taxpayer’s conscious mind that she was going to be warm, clothed and decent when she laid out money on those clothes.”

A majority of the House of Lords agreed: see Lord Brightman at 370 B-D, where he said:

“Of course Miss Mallalieu thought only of the requirements of her profession…she would, if asked, have repeated that she was maintaining her wardrobe because of those requirements. It is the natural way that anyone incurring such expenditure would think and speak. But she needed clothes to travel to work and clothes to wear at work, and I think it is inescapable that one object, though not a conscious motive, was the provision of the clothing that she needed as a human being.”

In other words, Miss Mallalieu’s conscious and subconscious reasoning was as follows: (1) to fulfil my needs as a human being I must wear clothes, (2) to satisfy my professional needs those clothes must be black, (3) therefore I will buy black clothes. Her conscious professional purpose followed on from her subconscious private purpose.

In MacKinlay (Inspector of Taxes) v Arthur Young McClelland Moores & Co [1989] STC 898, [1990] 2 AC 239 (‘Arthur Young’), a partner was required to move house and the partnership incurred the removal costs, for which it claimed a deduction. In the House of Lords, Lord Oliver said at 255:

“The question in each case is what was the object to be served by the disbursement or expense? As was pointed out by Lord Brightman in Mallalieu's case, this cannot be answered simply by evidence of what the payer says that he intended to achieve. Some results are so inevitably and inextricably involved in particular activities that they cannot but be said to be a purpose of the activity. Miss Mallalieu's restrained and sober garb inevitably served and cannot but have been intended to serve the purpose of preserving warmth and decency and her purpose in buying cannot but have been, in part at least, to serve that purpose whether she consciously thought about it or not. So here the payment of estate agents' fees, conveyancing costs and so on, and the provision of carpets and curtains cannot but have been intended to serve the purpose of establishing a comfortable private home for the partner concerned even though his motive in establishing a home in that particular place was to assist him in furthering the partnership interests. Nobody could say with any colour of conviction that in purchasing new curtains he or his wife was acting upon partnership business.” [emphasis added]

In Vodafone Cellular Ltd v Shaw [1997] STC 734 (“Vodafone”), Millett LJ (as he then was) considered the meaning of “purpose” in the context of applying the test of whether expenditure had been wholly and exclusively incurred for the purposes of the trade at 742f-743a:

“The leading modern cases on the application of the exclusively test are Mallalieu v Drummond (Inspector of Taxes) [1983] STC 665, [1983] 2 AC 861 and MacKinlay (Inspector of Taxes) v Arthur Young McClelland Moores & Co [1989] STC 898, [1990] 2 AC 239. From these cases the following propositions may be derived. (1) The words for the purposes of the trade mean to serve the purposes of the trade. They do not mean for the purposes of the taxpayer but for the purposes of the trade, which is a different concept. A fortiori they do not mean for the benefit of the taxpayer. (2) To ascertain whether the payment was made for the purposes of the taxpayer's trade it is necessary to discover his object in making the payment. Save in obvious cases which speak for themselves, this involves an inquiry into the taxpayer's subjective intentions at the time of the payment. (3) The object of the taxpayer in making the payment must be distinguished from the effect of the payment. A payment may be made exclusively for the purposes of the trade even though it also secures a private benefit. This will be the case if the securing of the private benefit was not the object of the payment but merely a consequential and incidental effect of the payment. (4) Although the taxpayer's subjective intentions are determinative, these are not limited to the conscious motives which were in his mind at the time of the payment. Some consequences are so inevitably and inextricably involved in the payment that unless merely incidental they must be taken to be a purpose for which the payment was made.

Mr Prosser QC submitted that, when Lord Oliver and Millet LJ referred to results and consequences which are “inevitably and inextricably involved” with particular activities, both clearly had in mind personal benefits such as being warm and decent, or having a home, where common sense dictates that obtaining those benefits must inevitably form part of the individual’s conscious or subconscious purposes in carrying on those activities. They did not have in mind legal benefits, such as a tax deduction which by operation of a statute is available to a taxpayer as a result of carrying on a particular activity. Otherwise, no taxpayer who carries on the activity with knowledge of the statute would ever be entitled to the deduction

In the present case, the FTT referred to Vodafone at [113] of the Decision, and then at [120] held:

“Mr Kushel did not go so far as Ms Mallalieu, who 'had no thought of warmth and decency' when she bought her 'working clothes', and say that a tax advantage was not an object or purpose of LLC5. However, adopting the reasoning of the House of Lords in Mallalieu v Drummond as further explained in Vodafone to the present case it is necessary to look beyond the conscious motives of LLC5 and take account of the inevitable and inextricable consequences of it entering the loan relationship with LLC4. Having regard to all the circumstances of the case it is, in my judgment, clear that the securing of a tax advantage is an inevitable and inextricable consequence of the Loan between LLC4 and LLC5.”

Mr Prosser QC submitted that it was not clear what the FTT meant by “all the circumstances of the case”, other than the fact that LLC5’s directors knew of the tax advantage and believed that it would be available in respect of interest payable on the Loans. He submitted that, whatever other circumstances the FTT may have had in mind, in applying Millett LJ’s proposition (4) in Vodafone to that tax advantage it erred in law, both because the proposition can have no application to a benefit obtained by operation of a tax statute and also because the FTT overlooked the fact (which it had itself pointed out at [52]) that LLC5’s directors knew that LLC5 would not itself obtain any benefit (the tax benefit was to the BlackRock Group).

We consider that the FTT misapplied the test in Mallalieu to the situation before it and that it should have relied on the principles derived from TDS, the only authority on this legislative wording. The FTT should have considered LLC5’s main purposes in relation to the Loans from all the evidence before it rather than apparently focusing solely on the “inevitable and inextricable consequences” of entering into the Loans.

Nevertheless, we are not satisfied that there was any material error in the FTT’s finding that LLC5 also held a tax advantage main purpose in relation to the Loans. The FTT was entitled to look beyond the stated motives of LLC5’s board members when determining the purposes of LLC5 in entering into the Loans.

There are two important matters to bear in mind. First, Mr Kushel accepted in his witness statement at [40] that tax efficiencies were part of the purpose for the inclusion of LLC5 in the transaction, albeit he said that it was the secondary purpose and not the key purpose. The FTT made a similar finding at [120] that ‘Mr Kushel did not go so far as Ms Mallalieu, … and say that a tax advantage was not an object or purpose of LLC5.’

Second, it is undisputed, as evidenced in the minutes of the board meeting of 30 November 2009 and in Mr Kushel’s statement, that the board members of LLC5 were specifically advised to put any tax advantage out of their minds when considering the viability and hence whether to approve the Loans. As a result of this advice, the stated subjective intentions of the only director to give evidence were circumscribed and cannot represent the nature of the directors’ intentions had they been left freely and willingly to decide the main purposes of the transaction.

Therefore, it is necessary to look beyond the directors’ stated intentions. The effectiveness of anti-avoidance legislation cannot be undermined by tax advisers telling parties to ignore the tax advantage purposes of a transaction which has been planned by them or others for precisely that purpose. To hold otherwise would provide an easy way round the legislation.

We therefore conclude that the FTT was entitled to look beyond the stated motives or intentions of the board members to determine LLC5’s actual subjective purpose. This is supported by the approach in TDS where it is apparent that Newey LJ was prepared to infer a different purpose for Mr Turner using the shares in the swap (namely a tax advantage) than he stated in his evidence to be his subjective intention for continuing to hold the shares (which was an exclusively commercial purpose).

In our view there was ample evidence, as explained below, to support the finding that securing a tax advantage for the Group (which is a tax advantage to LLC5) was a main purpose of the creation of LLC5 and thereafter, its intention and purpose in entering into the Loans. These purposes were subjectively held by LLC5, even if the directors were told to disregard them in considering their approval to entering into the Loans.

In our view the FTT was entitled to examine ‘all the circumstances’ when deciding LLC5’s purposes in entering into the Loans. This included the evidence leading up to the creation and registration of LLC5 in September 2009 and the board meeting on 30 November 2009 which approved the Loans. The FTT was also entitled to take into account the material that was before the board members at that meeting together with what the directors were told in advance and advised at the meeting.

This approach is more justified in this case than in Mallalieu where no conscious thought was given to the tax advantages of the decision being made. LLC5’s directors were specifically told to put the obvious tax advantages of the proposed transaction out of their minds, thus proving in a sense that this was indeed their main purpose.

The FTT made findings that LLC5, a UK tax resident LLC, had been included in the acquisition structure for tax planning purposes. The evidence revealed that the LLC5 managers (referred to in the Decision as its directors) were aware of that: see [14-23], [34-38], [40] and [50-51]. The FTT noted at [14] that BlackRock’s Global Head of Tax, Harris Horowitz, first asked EY where the debt should be pushed down to, in an email of 27 June 2009. While the investment was taking place in the US, the debt was pushed down to the UK because of its “generous tax regime for interest deductions” (see [15]). There was no other reason given for involving the UK in the structure. At [24], the FTT noted the difficulties which arose as a result of the UK’s involvement in the structure and necessitated the creation of a third LLC. Mr Fleming’s evidence, as accepted by the FTT at [13], was that the planning for the acquisition of BGI US was undertaken at Group level by the Corporate Tax Group and EY.

We accept at this stage that this was evidence of the BlackRock Group’s purposes for LLC5’s business and the Loans because LLC5 itself did not yet exist. However, this was critical context and the material supporting the inclusion of LLC5 in the structure was later made available to LLC5 and was seen by its advisers and board members after it came into existence. By that point all of the tax planning had been finalised and agreed and all that LLC5 needed to do was to implement it by approving the Loans. The proposed transaction had been presented to the US regulator (the OCC) prior to the LLC5 board agreeing to enter into it: see [37]. Thus, by the time of the LLC5 board meeting, there was nothing to stop the tax advantaged structure, including LLC4 lending to LLC5, from going ahead.

At [49], the FTT referred to Mr Kushel’s evidence in his statement at [24] that he did not view it as part of his remit as a board member of an individual legal entity to begin to question or suggest changes to the underlying capital structure that it was proposed a transaction should follow. In his oral evidence, he stated that they were not considering whether or not it was the right business of LLC5 to invest in the LLC6 Preference Shares.

As to the LLC5 board meeting on 30 November 2009, when Mr Kushel initially indicated that he may be unavailable for it (as the date of the meeting was that of Thanksgiving in the US), Mr Fleming asked him to attend because it was “an important meeting for UK tax purposes”: [40] of the Decision.

Prior to the meeting, Mr Fleming had already briefed the other board members in the note he had prepared and which was circulated on 27 October 2009: [34] of the Decision. Mr Fleming also had a meeting with them on 10 November 2009 at which the main focus of discussions was how the UK tax rules around deductions for interest expenses worked: [36] of the Decision.

At [35] of the Decision, the FTT referred to an email from Mr Fleming to Mr Horowitz dated 10 November 2009 which stated that the transaction to acquire the LLC6 Preference Shares might be an “Achilles heel” that might concern HMRC and underlined the need for documentation supporting the fact that the directors made a nuanced and informed commercial decision to acquire the Preference Shares.

As found by the FTT at [31]-[51], at the board meeting itself, Mr Fleming’s evidence was that none of the attendees were coming to it “in a vacuum”. The meeting was held the day before the BGI transaction was to be executed for which a “detailed step plan was prepared which we were to follow”. He had added that “provided the board felt the transaction was commercially advantageous for the company” it was not “sensible” to consider an alternative transaction (if, say, the directors had decided not to proceed with the proposed transaction). He accepted during cross-examination that the board were presented with a “fait accompli”. Mr Kushel confirmed that there was no reasonable possibility that LLC5 would not enter into the transaction.

The reference to the directors not coming to the proposed transaction in a vacuum is presumably based not only on the briefing referred to above, but also on the fact that Mr Kushel and the other three LLC5 board members were directors of other BlackRock UK companies, as well as holding specific wide-ranging and senior posts in the international business (i.e. all businesses outside the US and Canada). Mr Kushel was the Chairman of International Business; Mr Roger Tooze was head of its Business Finance; Mr Colin Thomson was the head of its Financial Reporting Group; and Mr James Desmarais was its General Counsel. By virtue of those roles, they would have been well aware of the role LLC5 was playing in the wider Group and the acquisition structure. As noted above, Mr Kushel’s evidence was that he attended regular briefing sessions at which the transaction step plan was discussed. They would therefore not have been coming to the transaction solely from the point of view of LLC5. However, this does not mean the directors had abdicated their fiduciary duties by taking into account the benefit to the Group as a whole: see Oxford Instruments UK 2013 Limited v HMRC [2019] UKFTT 254 (“Oxford Instruments”) at [102].

Mr Fleming’s evidence was that the discussion of the tax aspects at the board meeting took place first and “was not very long.” The meeting took place “around about lunchtime” and lasted approximately 45 minutes ([50] of the Decision). According to his oral evidence, five minutes concerned him giving information on tax. During cross-examination, Mr Fleming described his role at the meeting as, “…to position the information to explain a little bit about the tax profile and then to leave them to get themselves comfortable that this was a commercial transaction”. This suggests that the purposes of commerciality and tax were intertwined.

There is no indication of how much time was given to going through and approving the 12 written resolutions before the board. Mr Fleming recalled that there was a “very robust consideration” of the commercial data. However, neither witness was able to explain precisely what was considered. No questions, or at least no questions important enough to be recorded in the minutes, were asked. There is no evidence as to what the other board members thought or did.

The evidence is that the BlackRock Group would not have used an acquisition structure with a UK resident LLC in the absence of the UK tax benefits of doing so. Absent those tax benefits, LLC5 would not have existed and so obviously would not have entered into the Loans to acquire the Preference Shares. LLC5 was aware of this when it approved the Loans.

The FTT’s findings therefore demonstrate that LLC5 was only included in the structure and thereby entered into the Loans so as to take the tax benefits for the Group. LLC5 contended that it merely being engaged in tax planning does not mean that it had a tax avoidance main purpose. It argued that in any case involving a substantial borrowing for commercial purposes, the borrower will take tax advice and will be told that the interest is deductible; this cannot mean that there is a tax main purpose. That might be true where the borrowing is needed for and driven by the commercial purposes. In this case, however, the borrowing by LLC5 specifically in the structure for that purpose was primarily motivated by securing a tax advantage.

The FTT was therefore entitled to find that LLC5 had a tax advantage purpose as one of its main purposes and as a main purpose of the Loans. The FTT did not make a material error of law in finding that LLC5 had an unallowable tax advantage purpose as a main purpose of the Loans it entered into.

We have upheld the FTT’s conclusions that the Loans had both commercial and tax advantage main purposes. The FTT decided that all of the debits that flowed from the Loans should be attributed to the commercial purpose for the, again very compressed, reasons it gave at [123]:

The FTT stated at [122] of the Decision that it was applying the test described by Judge Beare in Oxford Instruments, to the just and reasonable apportionment for the purposes of section 441(3) CTA 2009.

The approach taken by Judge Beare in Oxford Instruments was set out in full at [114] of the Decision. Judge Beare based that approach on the acceptance (albeit obiter) by the Court of Appeal in TDS at [50] to [54] of the proposition that, where a company has entered into a loan relationship for an unallowable purpose, as long as it can show that that purpose has not increased the debits arising in the company from those which would have arisen in any event even in the absence of the unallowable purpose, none of the debits should be attributable to that unallowable purpose.

Mr Prosser QC submitted that, were we to uphold the FTT and decide that there was a main tax advantage purpose to the Loans, the FTT applied the correct approach to the apportionment as set out in Oxford Instruments and Kwik-Fit Group Limited and ors v HMRC [2021] UKFTT 283 (TC) (“Kwik-Fit") at [122]-[123] of the Decision. Given that LLC5 would have entered into the Loans even if there had been no tax advantage in doing so, the unallowable purpose did not increase the debits, and therefore on a just and reasonable basis none of the debits should be apportioned to the unallowable purpose. The FTT based that finding on the unchallenged evidence that the board left any tax advantage out of account, as well as on the evidence, recorded at [53] and in the penultimate sentence of [119], that the transaction would have proceeded even if, at the last minute, the tax advantage had fallen away. Therefore, none of the debits were “attributable to” the tax purpose.

Mr Ewart QC submitted that the FTT erred in law and should have asked itself whether the Loans (and therefore the debits) would have existed at all if the benefit of the tax relief had never existed, and should have answered ‘No’ because there would have been no LLC5 and no Loans if the UK had not provided tax relief. Instead, he submitted, the FTT only considered the position once LLC5 was in existence and the Group was on the verge of acquiring BGI US despite it having made findings of fact in respect of the tax planning that resulted in LLC5 being included in the acquisition structure. The UK resident LLC5 included in the acquisition structure would not have existed absent the tax benefits because the overall commercial benefit (the acquisition of BGI) could have been realised without it.

Both parties filed supplementary written submissions dated 23 February 2022 on the apportionment issue. Mr Prosser QC made further points and an alternative argument in support of his primary submission. He argued that, as a matter of principle, it would be unjust and unreasonable to attribute any debits to a constructive tax purpose, and therefore disallow those debits, in circumstances where, although the company has knowledge of the relevant legislation and believes that a deduction will be available, it leaves that belief out of account when deciding to become a party to the loan relationship.

Further, he submitted that the amount of debits to attribute to the tax purpose, as opposed to the commercial purpose, on a just and reasonable basis should be determined by reference to the relative importance or significance of each purpose, that is, the tax and commercial advantages which were anticipated to be obtained when LLC5 decided to enter into the loans.

Alternatively, Mr Prosser QC submitted that LLC5 did not anticipate the tax advantage that actually arose as it anticipated a much lower tax benefit because of the expected operation of the worldwide debt cap. LLC5 cannot have had an actual tax purpose of obtaining the unanticipated tax advantage (which was greater). It would not be just or reasonable to apportion by reference to the tax advantage which was actually, but not anticipated to be, obtained. In particular, it would not be just or reasonable to ignore the important fact that LLC5 expected the worldwide debt cap to operate so as to deny relief for a very substantial proportion of the debits that it unexpectedly benefited from (indeed, until that issue was resolved in 2014, LLC5 submitted a number of tax returns on that basis).

We consider that the FTT was wrong to apply a subjective approach when making the just and reasonable apportionment. The statutory test is to be applied without any gloss: Versteegh Ltd and others v HMRC [2014] SFTD 547 at [166]. The correct approach is to determine whether the reason the debits existed was in order to obtain a tax advantage on the basis of an objective consideration of all of the relevant facts and circumstances. TDS confirms that the question of whether there is a tax avoidance main purpose is primarily to be decided by reference to the subjective intentions of the directors of a company. However, once such a purpose has been identified, the legislation simply requires there to be a just and reasonable apportionment. This test is to be applied objectively. The FTT erred by applying a subjective test.

If, as appears to have been found by the FTT, a wholly subjective approach were to be applied at the apportionment stage, it would enable a company to avoid the application of the unallowable purpose rules by simply relying on its stated purpose of entering into its loans, or indeed whether it would be able to halt the transaction at a late stage. Mr Kushel’s evidence that the transaction would have proceeded even if the tax benefits had fallen away at the last minute does not, however, indicate the existence of an additional commercial purpose (which meant the transaction was still viable absent the tax purpose). While the Group had considered whether the structure could be unwound if the tax benefits failed to materialise, as a practical matter pulling out of the transaction at such a late stage was impossible in those circumstances. As Mr Kushel had stated in cross-examination, the acquisition was a fairly complicated exercise comprising hundreds of individual transactions.

Depending on the facts, it may also be helpful to check the just and reasonable apportionment conclusion by applying a “but for” test: see Kwik-Fit at [129]. This is similar to the Court of Appeal’s conclusion in Fidex v HMRC [2016] STC 1920 at [74] where the company had a mixed purpose for being party to certain bonds but all of the debits arising in respect of them were allocated to the tax purpose because they would not have arisen absent the tax avoidance scheme entered into by the company.

We agree with Mr Ewart QC that the creation and insertion of LLC5 into the acquisition structure and the Loans approved by its board members, which gave rise to interest deductions, would not have been incurred but for the tax purpose. Its commercial purpose was a by-product or consequence of the tax-driven decision to place LLC5 in the acquisition structure which meant it received a commercial benefit from the deal. But this was all a consequence of the tax purpose without which no deal or benefit would have been obtained. The commercial purpose of LLC5 and the Loans would not have occurred but for the tax purpose.

Despite the FTT stating that it was applying the approach outlined in Oxford Instruments, if it had done so, it would have asked the question: would the Loans (and therefore also the debits) have existed at all if the benefit of the tax relief had never existed? The answer to that question is objectively “no”. There would have been no LLC5 and no Loans if the UK had not provided tax relief for interest in these circumstances.

In fact, the FTT appears to have asked itself a different, and in our view, wrong question: would the loans have existed if the tax relief had been withdrawn the day before the acquisition was due to complete? This relied on Mr Kushel’s evidence explained above but it is obvious that at that stage there would be no way that LLC5 would pull out of the transaction and potentially upset the huge and carefully negotiated acquisition of BGI US. The concentration must be on the reason why LLC5 was involved in the transaction and not on a hypothetical scenario that was not considered at the time. The FTT misapplied the ‘but for’ test to find that the debits, arising from a transaction that had been months in the planning and carefully structured to achieve a tax advantage, could be wholly attributed to the commercial purpose, simply because of the subjective belief that, if the tax benefits had at the last minute fallen away, the transaction would have still gone ahead, as held at [119].

This reasoning also applies to Mr Prosser QC’s alternative submission that at the time of the transaction the tax benefit which was expected to arise was much smaller than it turned out to be as a result of the application of the post-worldwide debt cap. However, this again depends on the subjective beliefs of LLC5 and not on the objective facts. Whatever the expected or actual outcome of the tax benefit it was entirely attributable to the unallowable purpose for the reasons set out above.

In conclusion, we are satisfied that the FTT erred in law in deciding it was just and reasonable to apportion all the debits arising from the Loans to the commercial purpose. The FTT should not have focused only on subjective beliefs of one of LLC5’s board members nor on the period just before the transaction completed in considering how to carry out a just and reasonable apportionment. This was a material error and we set the Decision aside in this respect.

We consider that we should re-make the decision by apportioning all the debits arising from the Loans to the unallowable tax purpose. The just and reasonable apportionment in this case is for all of the debits to be apportioned to the tax advantage (unallowable) main purpose so that they are not brought into account.

The FTT did not err in finding that LLC5 had both a commercial purpose and an unallowable tax advantage main purpose in entering into the Loans. However, it was wrong to decide that the just and reasonable apportionment was solely to the commercial purpose. But for the tax advantage purpose there would have been no commercial purpose to the Loans and all the relevant facts and circumstances lead inexorably to the conclusion that the loan relationship debits should be wholly attributed to the unallowable tax purpose and so disallowed.

The appeal is allowed on both the Transfer Pricing Issue and the Unallowable Purpose Issue.

The FTT’s Decision is set aside and re-made. HMRC’s amendments to LLC5’s returns are confirmed.