BlackRock HoldCo 5, LLC v The Commissioners for HMRC

The UT allowed HMRC’s appeal on the Transfer Pricing issue on the basis of an argument that was not raised in the FTT, namely that in determining whether an independent lender would be prepared to lend, the transfer pricing provisions do not permit the existence of third-party covenants to be hypothesised where those covenants are not present in the actual transaction. LLC5 challenges that conclusion as Ground 1 of its appeal.

HMRC defend the UT Decision and also rely on an additional ground by way of Respondent’s Notice, namely that the UT erred in holding that the FTT had been entitled to conclude on the basis of the expert evidence before it that an independent lender would have entered into the Loans subject to it being able to obtain the necessary covenants; and that the covenants would have been forthcoming (HMRC’s Ground 1).

The relevant legislation is contained in Part 4 of TIOPA.

Section 147 TIOPA relevantly provides:

“Transaction” is broadly defined by s.150(1) TIOPA to include “arrangements, understandings and mutual practices (whether or not they are, or are intended to be, legally enforceable)”. Section 150 also provides that “series of transactions” includes a number of transactions entered into in pursuance of the same arrangement, and expressly states that provision can be made between two persons by means of a series of transactions even if there is no direct transaction between them or the series includes transactions to which neither of them is a party. Section 148, read with ss.157 to 163, has the effect that the “participation condition” is met if, among other things, one of the persons controls the other. That is referred to as being a situation where one of the persons is “directly participating in the management, control or capital” of the other. The concept of “potential advantage” is defined in s.155 and includes the creation or increase in losses for tax purposes.

Section 151 elaborates on the concept of arm’s length provision, as follows:

It is common ground that the relevant “actual provision” referred to in s.147(1) is the Loans made by LLC4 to LLC5, that the “participation condition” is met by virtue of LLC4’s control of LLC5 (in the terms of the legislation, LLC4 was “directly participating in the management, control or capital” of LLC5) and that the Loans did confer a “potential advantage” in relation to UK taxation.

Section 152 TIOPA deals specifically with debt financing between corporate entities, in the following terms:

Section 154 defines some of these terms, as follows:

It is uncontroversial that s.152 is in point and that the necessary special relationship existed between LLC4 and LLC5 through the former’s control of the latter. It is also undisputed that the third party covenants that the FTT found would have been in place in an arm’s length transaction would not be “guarantees” within s.154(4).

Finally, but importantly, s.164 provides that Part 4 of TIOPA is to be read “in such manner as best secures consistency” with the transfer pricing guidelines issued by the Organisation for Economic Co-operation and Development (the “OECD guidelines”):

As discussed in oral argument, this is perhaps not drafted as clearly as it might be. We are not concerned with double taxation arrangements. The explanation is that the OECD guidelines are produced with reference to the OECD model double tax convention, and specifically Article 9 of the model, which deals with “associated enterprises”. Article 9 allows a contracting state to make adjustments to profits for tax purposes where, broadly, the terms of the actual arrangements between enterprises in the two contracting states differ from what would have been agreed between independent enterprises, and requires corresponding adjustments to be made by the other contracting state.

Section 164(1)(b) refers to “transfer pricing guidelines”. That term is defined in 164(4) by reference to guidelines published by the OECD. The definition changed over the relevant period. The effect in this case is that, for LLC5’s accounting periods ended 30 November 2010, 31 December 2010 and 31 December 2011 the relevant version of the OECD guidelines is that published in 1995, and for the periods ended 31 December 2012 to 31 December 2015 inclusive, the version published in 2010.

Later versions of the guidelines were published in 2017 and 2022. Although those later guidelines are not strictly applicable there was no dispute between the parties that we can consider them on the basis that (so far as they are relevant to this case) they simply elucidate or expand on points made in earlier versions. Indeed, both parties relied on the 2022 version on that basis.

Paragraph 1.6 of both the 1995 and 2010 versions of the OECD guidelines explains that what Article 9 of the model convention seeks to do is to adjust profits by reference to “the conditions which would have obtained between independent enterprises in comparable transactions and comparable circumstances” (a comparable “uncontrolled transaction”, as opposed to the actual “controlled transaction”). The 2010 version adds that this comparability analysis is at the “heart of the application of the arm’s length principle”, while explaining at para. 1.9 that there are cases, for example involving specialised goods or services or unique intangibles, where a comparability analysis is difficult or complicated to apply.

In its discussion of comparability analysis, para. 1.15 of the 1995 version states:

As can be seen from this, it is essential that the “economically relevant characteristics” are “sufficiently comparable”, in the sense of any differences either not having a material effect on the relevant condition (term) of the transaction, or being capable of being adjusted for with reasonable accuracy so as to eliminate their effect.

Paragraph 1.17 of the 1995 version expands on the concept of differences as follows:

Under a sub-heading “Factors determining comparability”, the guidelines discuss among other things the importance of a functional analysis, which “seeks to identify and to compare the economically significant activities and responsibilities undertaken or to be undertaken by the independent and associated enterprises” (para. 1.20 of the 1995 version). As part of this, the guidelines identify the relevance of risks assumed by the parties:

The OECD guidelines make clear that, in general, tax administrations should not disregard actual transactions. Both the 1995 and 2010 versions give two examples of when, exceptionally, this might be done (paras. 1.36 to 1.38 of the 1995 version; paras. 1.64 to 1.66 of the 2010 version). The first is where the economic substance of a transaction differs from its form, the classic example being debt that in economic substance amounts to equity. The second is a situation where the arrangements differ from those that would be entered into by commercially rational independent parties and the effect is to impede the application of transfer pricing principles, an example being a transfer of unlimited rights to intellectual property relating to future research. It was not suggested that either of these exceptions is in point.

A further point clarified by the guidelines is that the presence of a tax motive or purpose does not justify non-recognition of the parties’ characterisation or structuring of a transaction. Rather, the fact that a transaction has been entered into with tax savings in mind is not relevant for transfer pricing purposes: paras. 9.181 and 9.182 of the 2010 version.

Mr Prosser referred us to the 2022 version of the OECD guidelines in relation to two points. The first is the treatment of synergies available from membership of a group. The guidelines clarify that incidental benefits arising from group membership need not be adjusted for even if they are substantial. The example is given of an improved credit rating available to a borrower by virtue of membership of a group, as compared to what it would be on a standalone basis (paras. 1.177, 1.178 and 1.184 of the 2022 version). This point is also addressed in both the 1995 and 2010 versions at para. 7.13.

Secondly, in a section dealing with intra-group loans, para. 10.56 of the 2022 version notes that where a parent grants a loan to a subsidiary the grant of security is less relevant to its risk analysis because it already has control and ownership, such that the absence of contractual rights over the assets of the borrower “does not necessarily reflect the economic reality of the risk inherent in the loan”. The same section also addresses covenants, pointing out (in effect) that the drivers leading to covenants being required at arm’s length may not be present in an intra-group context and that it will be appropriate to consider whether there is in practice the “equivalent” of covenants (para. 10.86).

The FTT considered the evidence of two experts, Timothy Ashley for BlackRock and Simon Gaysford for HMRC. Mr Ashley had had significant experience in treasury management and the debt capital markets. Mr Gaysford did not have debt capital market experience. His expertise was as an economist.

The experts were asked to consider whether the Loans differed from those which would have been made between independent enterprises, including whether they would have been entered into at all. The FTT found that there was broad agreement between them on most issues, including that “an independent lender would, on the strength of the [BGI US] business, be willing to lend $4 bn to LLC5 but would have required covenants from LLC5 to do so” (para. 69 of the FTT Decision). This was reiterated at para. 77 in the following terms, by reference to a joint statement and Mr Gaysford’s report:

Mr Ashley’s report listed various types of covenant, but he agreed in oral evidence that the “critical ones” were:

As to the “preference share payment covenant” referred to in (2) above, the FTT recorded at para. 72 that:

The covenants that Mr Gaysford stated were required are set out at para. 73 of the FTT Decision. No doubt reflecting his different expertise, they are somewhat less specific and described more by overall effect rather than content, in particular covenants to “remove the existing uncertainty” and prevent “actions that could reduce the value of LLC5”.

The FTT went on to consider the experts’ evidence in further detail. On the question of the additional terms and conditions that an independent lender would require, the FTT set out at para. 78 the following extract from section 14 of the joint statement:

HMRC’s case before the FTT was, essentially, that the transaction would simply not have taken place at all at arm’s length. The FTT concluded that both HMRC’s submissions, and Mr Gaysford, were wrongly focusing on the position of the BlackRock group as a whole. Their objection was that there was a much simpler commercial alternative for a commercial lender to the group which did not involve a loan to LLC5, whereas the required comparison was between the actual lending transaction and a hypothetical loan to the same borrower (para. 101). The FTT went on:

As already explained, HMRC raised a new argument in the UT, namely that in determining whether an independent lender would be prepared to lend, the transfer pricing provisions do not permit the existence of third-party covenants to be hypothesised where those covenants are not present in the actual transaction.

In essence, the UT accepted Mr Ewart’s submission that s.147(1)(a) TIOPA imposes a “two party rule”, because it requires focus on the “provision…made or imposed as between any two persons”. That meant that in this case only the terms of the Loans could be considered. Third-party covenants not given as part of the actual transaction may not be taken into account because to do so “materially changes the surrounding circumstances and alters the economically relevant characteristics of the transactions in question” (para. 75 of the UT Decision).

In reaching that conclusion, the UT reasoned that the FTT had essentially compared a different transaction to the actual one, because importing third-party covenants changed the nature of the provision to be compared (para. 59). The UT also agreed with Mr Ewart’s submissions that s.152(5) TIOPA did not assist BlackRock. That only applied where a guarantee was present in the actual transaction (para. 70). Further, its existence indicated that not only guarantees but other types of third-party covenant “affect the substance of the loan transaction”. Those other sorts of covenant could only be taken into account if they were present in the actual transaction, because the actual and arm’s length transactions would otherwise not be the same (para. 72).

The UT recognised that the effect of this was that LLC5 would not have been challenged on transfer pricing grounds if it had gone through the “rather artificial exercise” of putting covenants in place with other group companies. However, this was unlikely to be a problem in practice because “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” (paras. 72-74).

On the basis of the FTT’s finding that an independent lender would not have advanced the Loans without third-party covenants, the FTT should therefore have concluded that no provision would have been made as between independent enterprises, within s.151(2) TIOPA (para.76). The UT accordingly re-made the decision to that effect (para. 100).

Given the UT’s conclusion on this point it was unnecessary to deal with other challenges to the FTT’s conclusions, but the UT went on to do so. Relevantly for our purposes, the UT concluded that the FTT did not mischaracterise the expert evidence or make unsupported findings in relation to it (paras. 79-88).

I consider that the UT erred in accepting HMRC’s argument that the transfer pricing provisions do not permit the existence of third-party covenants to be hypothesised where those covenants are not present in the actual transaction.

I agree that what the legislation requires is consideration of the “provision…made or imposed as between any two persons”, which must be compared to the provision that would have been made between two independent enterprises. As the Special Commissioners said in DSG Retail Ltd v HMRC [2009] STC (SCD) 397 (“DSG Retail”) at [65] (by reference to the predecessor legislation), two means just that, not “two or more”. However, nothing in the legislation or the OECD guidelines requires the position of third parties to be ignored if it is otherwise relevant.

The critical starting point is that, as the OECD guidelines make clear, any comparison requires the “economically relevant characteristics of the situations being compared” either to be “sufficiently comparable”, or that “reasonably accurate adjustments” can be made to eliminate the effect of any material differences.

In the real transaction, LLC4 had no need of any of the covenants considered by the experts because, quite independently of its ownership of LLC5, it had control of LLC6 and its subsidiaries, including BGI US (the “LLC6 sub-group”). LLC4 had no real world concern that members of the LLC6 sub-group might take on excessive additional debt or that they might grant unacceptable security to other lenders. It also had no real world concern that the preference share rights might be circumvented. As the UT recognised, for LLC4 to require covenants to guard against those (in practice unreal) risks would indeed be artificial. It was simply unnecessary. LLC4 was in a position to control those risks itself.

While I appreciate that this is not the way in which the case has been argued (at least on appeal), one way of approaching this is to remind oneself that, in hypothesising a comparable transaction and with the exception of their assumed independence, the actual characteristics of the parties are not adjusted. This was helpfully explained in DSG Retail at [78] in relation to the words “differs from the provision which would have been made as between independent enterprises” which now appear in s.147(1)(d) TIOPA. After stating that this might indicate enterprises not sharing the same attributes as the actual parties, the Special Commissioners said:

It is worth noting that this approach finds support in the domestic legislation, because it explains s.152(4) TIOPA, which requires the fact that it is not otherwise part of the lender’s business to make loans to be disregarded (see [27] above).

If the only change to the actual transaction is to break the group relationship between LLC4 and LLC5, LLC4 would still have its direct interest in LLC6 and would therefore still control both LLC6 and its subsidiaries. On that approach covenants from the LLC6 sub-group would be unnecessary. LLC4 could ensure that the preference share dividends were paid by LLC6 as anticipated.

As I say, the appeal was not argued on that basis, but it does underscore the artificiality of considering how LLC4 needs to be protected. In reality, and even if the control relationship was broken between LLC4 and LLC5, LLC4 does not need further protection. (I appreciate that the question whether LLC4 needs further protection is not the only one to ask, because for transfer pricing purposes it is equally necessary to consider whether LLC5 would be prepared to enter into the Loans as an independent borrower, without assurances e.g. that its preference share rights would not be frustrated. But despite that perhaps being more pertinent, it was not the focus of the debate.)

If we do put to one side LLC4’s actual control of the LLC6 sub-group, then there is a significant difference between the “economically relevant characteristics” in the actual and hypothetical transactions. The risks are quite different. In the actual transaction it is obvious that the only real risks to be assessed from LLC4’s perspective related to the performance of the BGI US business. In contrast, the lender in the hypothetical transaction would be exposed to an additional risk that something might be done by LLC4 or entities controlled by it to divert or otherwise frustrate the expected dividend flows on the preference shares to LLC5.

Similarly in the actual transaction, consistently with the unchallenged evidence before the FTT and reflecting the fact that LLC4 could receive no dividends on its common shares unless dividends were paid on the preference shares (see [9] above), LLC5 was not exposed to any real risk that the expected dividend flows to it would not materialise, beyond risks related to the performance of the BGI US business. In contrast an independent borrower would have no practical assurance that nothing else would be done to switch off or reduce the anticipated dividends on the preference shares.

There was an issue before us about the fact that evidence relevant to the point just made concerning the lack of risk for LLC5 was not reflected in findings of fact by the FTT. This point was raised in HMRC’s skeleton argument but was rightly not pursued by Mr Ewart in oral submissions. The evidence was unchallenged and it must also be borne in mind that the argument we are concerned with was raised for the first time in the UT. The absence of findings is therefore unsurprising. It would also be unfair for HMRC, having raised a new argument in the UT, then to rely in this court on a failure by BlackRock to secure relevant findings in the FTT.

As already explained, the OECD guidelines contemplate that adjustments may be made to ensure that material differences between “economically relevant characteristics” are eliminated. They also explain (see [37] and [38] above) that material risks must be identified and addressed.

That is precisely what the expert evidence sought to do. The evidence accepted by the FTT in effect adjusted the terms of the actual transaction to ensure that the “economically relevant characteristics” of the actual and hypothetical transactions – and in particular the risks assumed – were comparable. This was done by hypothesising the existence of certain covenants from other group members. In the real world those covenants were unnecessary: the risks did not exist. In the hypothetical world the existence of the covenants addressed the risks that might otherwise exist, so rendering a comparison possible on the basis that the “economically relevant characteristics” of the actual and hypothetical transactions were then sufficiently comparable. The reference in the joint statement to it being possible for BGI US and LLC6 “to effectively ratify the legal and financial position” (see [48] above, para. f.) reflects this.

Mr Ewart submitted that the parts of the OECD guidelines referred to above all concern comparability analyses, but in this case there was no comparable transaction because the FTT had found that an independent lender would not lend to LLC5. That finding reflected the expert evidence that there were no comparable transactions.

In my view that overlooks the FTT’s findings of fact. The FTT found that an independent lender would have entered into the Loans on the same terms, subject to third-party covenants being entered into. I do not think it matters that a precise equivalent transaction between parties acting at arm’s length is not identified. This is not a transaction involving specialised goods or services or unique intangibles of the kind referred to in para. 1.9 of the 2010 version of the guidelines, where a comparability analysis is problematic. It is a lending transaction.

What the OECD guidelines do require is that the “economically relevant characteristics” are “sufficiently comparable”, and that to the extent that there would otherwise be material differences “reasonably accurate adjustments can be made to eliminate the effect of any such differences”: see above. The economically relevant characteristics of the actual lender and borrower in this case include (a) LLC4’s actual control of LLC6 and (indirectly) BGI US, such that in practice it had no need, when deciding to make the Loans, of any covenants from those entities (or, of course, from itself), and (b) LLC5’s own position in the group, such that its board had no reason to be concerned about any possibility that the expected profit would not, if made, find its way up to it by way of preference share dividends, whether due to dividend flows being diverted or otherwise. The third-party covenants do just what the guidelines require, namely adjust for the absence of these risks in the actual transaction, so rendering the economically relevant characteristics comparable.

Mr Ewart also submitted that there is no mention in the OECD guidelines of introducing third parties. This includes the section in the 2022 version that addresses covenants in loan transactions, which only considers covenants from a borrower (see [42] above). However, the guidelines do not rule out reference to the position of third parties, and in my view a proper application of them may, for the reasons already given, require it to be taken into account. It is also noteworthy that the consideration of covenants in the 2022 version clearly reflects the reality that intra-group loans generally do not require the sorts of protections that an independent lender would need, and recognises that it is appropriate to consider whether there is the “equivalent” of covenants in practice. This is consistent with the approach that I have described. The hypothesised covenants are designed to put the independent lender in an equivalent position to the actual lender.

Another example also helps to illustrate the problems with HMRC’s position. As emphasised in the expert evidence, arm’s length lenders will always consider the source of the cashflows that will service any loan that they may make. Where the proposed borrower is a holding company rather than the entity that directly generates the cashflows, lenders will be concerned about structural subordination. This relates to the fact that the holding company’s own interest in the cashflows is limited to a direct or indirect interest in another entity, and the resultant risk that subsidiaries may take action, whether by additional borrowings, the grant of security or otherwise, that will have priority over the holding company’s interest, with the effect that the cashflows do not find their way up the group to service the loan.

As the experts jointly explained (see [48] above, para. g), this is commonly addressed by “structural enhancements” that restrict the actions of subsidiary entities. However, an arm’s length lender would not normally be satisfied by a commitment from the borrower to that effect. It would want direct commitments from the subsidiaries: indeed it would often seek security over their assets as well. As the experts indicate, the sorts of restrictions to which they refer are “well known standard clauses required in almost every external debt transaction”, but would not be expected in an intra-group loan transaction. That is because they are not typically needed in that case.

There is no principled distinction between that scenario and this one. In both cases an arm’s length lender will typically require some form of third party protection. Yet there is (rightly) no suggestion that a loan to a group entity which can properly be supported by the sub-group that it owns can be challenged under the transfer pricing rules because the borrower’s subsidiaries have not given negative pledges or granted security.

I should add that, on the facts of this case, I also respectfully disagree with the UT’s observations about manipulating transactions with artificial covenants ([53] above). While I agree that the covenants would have been unnecessary in the real transaction, and in that sense it would be artificial to include them, there would have been nothing wrong in doing so in fact.

There is also support for the relevance of third parties in the domestic legislation.

The first point is that it is worth recalling that the legislation applies not only to a transaction directly between two parties (A and B) but to a series of transactions by means of which a “provision” is made or imposed between A and B. That series will clearly involve third parties, and indeed may be such that there is no direct transaction at all between A and B: see [24] above. Nevertheless, that (indirect) provision must be compared to the arm’s length provision. While the “series of transactions” rule is not relevant on the facts of this case, there is an analogy because where it does apply the legislation must be contemplating that transfer pricing may involve considering the position of third parties in both the actual and hypothetical transactions. Similarly in the present case, third parties are both actually present and would be in the hypothetical transaction. The difference is that the practical comfort that they provide in the actual transaction in this case is not legally formalised, because it is unnecessary to do so.

Secondly and more significantly, I accept Mr Prosser’s submission that the existence of s.152(5) TIOPA materially undermines Mr Ewart’s main submission in oral argument on the transfer pricing rules, namely that the domestic legislation does not permit any reference to third parties because, as DSG Retail determined, “two means two”.

Section 152(5) is set out at [27] above. To recap, it provides:

The rationale for this provision is best understood by an example. Assume a UK subsidiary that forms part of a larger non-UK based group. Assume further that the UK subsidiary has limited assets or that it is already funded by a significant amount of debt rather than equity. The UK subsidiary takes out a loan from a non-UK resident sister company with the benefit of a guarantee provided by the non-UK parent. An arm’s length lender would also not have made the loan without the benefit of the guarantee, because it would view the subsidiary as thinly capitalised (see [19] above). Section 152(5) would have the effect that the guarantee is ignored in the hypothetical transaction, with the result that the financing costs under the loan would be disallowed.

HMRC’s position is that s.152(5) concerns a situation where a guarantee exists in the real world, whereas what BlackRock seek to do is to hypothesise something that does not exist in the real world. The UT accepted that argument.

With respect, that misses the point. If third parties were irrelevant then there would be no role for s.152(5). It illustrates that the “two means two” mantra cannot be applied without further analysis. Further, the fact that s.152(5) applies to guarantees provided by related parties provides a clear indication that, in contrast, guarantees from third parties who are not related can be taken into account. The same must apply to arrangements involving a third party that do not fall within the definition of guarantee in s.154(4) TIOPA (set out at [28] above), irrespective of whether that third party is a related party or not.

As to non-existence in the actual transaction, that overlooks what does exist in the real world. The appropriate comparison is not between the non-existence of covenants in the actual transaction and the covenants that a third-party lender would require, but between the actual risks in the real world and the risks in the hypothetical transaction. In the hypothetical transaction there are risks that third parties (specifically, the LLC6 sub-group) may take actions that prejudice the performance of the Loans. Those risks do not exist for the parties to the actual transaction. The covenants in the hypothetical transaction effectively bring the risks into line with each other, so that the transactions are comparable.

Thirdly, and related to this point, it is worth reiterating just how broad the concept of “actual provision” is (“provision … by means of a transaction or series of transactions”: s.147(1)(a) TIOPA). As noted at [24] above, “transaction” is very broadly defined to include “arrangements, understandings and mutual practices (whether or not they are, or are intended to be, legally enforceable)”. The concept of the “actual provision” is then extended further by the inclusion of “series of transactions”. The net is therefore intended to be cast very widely. Related parties are much more likely to be prepared to rely on informal understandings or non-binding arrangements than parties acting at arm’s length. There is no indication in the legislation that the mere fact that an understanding or arrangement is non-binding should prevent a comparison with an arm’s length arrangement that has a similar economic effect and which would, in practice, be legally binding.

A final point on the interpretation of the transfer pricing rules is that what the transfer pricing rules require is a consideration of the terms that would have been agreed between the two “affected persons” if they had been independent enterprises (see s.147(1) TIOPA). As the OECD guidelines make clear, this requires a “comparison of the conditions in a controlled transaction with the conditions in transactions between independent enterprises” (see [35] above). Where s.152 TIOPA applies, this is also spelt out by reference to the terms that would be agreed at arm’s length (s.152(2), see [27] above).

If a lender acting at arm’s length requires a third party to provide support for a proposed loan, whether by covenant, guarantee, security or otherwise, that is properly described as a term or condition that would be stipulated as between the lender and borrower. Put simply, the lender would be saying to the borrower that it will lend only on the condition that the support in question is provided, and no doubt normally only on the basis that the support remains in place during the period for which the loan is outstanding.

It is self-evident in those circumstances that the existence of that support will be a term or condition of the loan as between the lender and borrower. It is nothing to the point that there may also be a (separate) “provision” as between the borrower and the third party that may itself be subject to the transfer pricing rules, such as a fee payable by the borrower for the provision of a guarantee.

Of course, this does not mean that any intra-group loan to a thinly capitalised borrower can be treated as being on arm’s length terms by hypothesising some form of third party support that does not exist in the actual transaction. As already discussed, the economically relevant characteristics of the actual and hypothetical transactions, and in particular the risks assumed, must be sufficiently comparable.

As already explained, HMRC’s Ground 1 is that the UT erred in holding that the FTT had been entitled to conclude on the basis of the evidence before it that an independent lender would have entered into the Loans subject to it being able to obtain the necessary covenants, and that the covenants would have been forthcoming.

Despite the carefully put submissions of Ms Choudhury for HMRC on this point, I agree with the UT that the FTT was entitled to reach the conclusions that it did on this issue.

In summary, Ms Choudhury submitted that the FTT did not address the fact that the experts did not conceive of the covenants in the same terms, and also did not address the point that it became clear during oral evidence that Mr Ashley’s conception of the critical covenant in respect of the preference share dividends went no further than the preference share rights, and so could add nothing. In contrast, Mr Gaysford’s evidence made it clear that he considered that it would be necessary to go much further.

While it would have been clearer if the FTT had explicitly addressed differences between the experts about the content of the covenants and spelt out in more detail in its conclusions exactly what covenants it was referring to, it is clear that it preferred Mr Ashley’s evidence to the extent that the experts’ views differed. It was entitled to do so. Further, large parts of the experts’ evidence was agreed: see above. That included, critically, that an arm’s length lender would lend $4 bn to LLC5 subject to suitable covenants, and that those covenants could probably have been put in place. Mr Gaysford’s concern about “cost and complexity” was, as he clarified in cross-examination, really a point about there being a better commercial alternative, but as the FTT correctly identified Mr Gaysford was wrongly focusing on the position of the BlackRock group as a whole: see [49] above.

Ms Choudhury’s submission that the dividend related covenant proposed by Mr Ashley added nothing to the preference share rights ignores an important point, namely that it would provide the lender with direct contractual rights against LLC6 (and/or, potentially, LLC4) that it would do nothing to frustrate the payment of the preference share dividends. From a lender’s perspective that is different to the share rights held by LLC5. Further, Mr Ashley evidently understood that LLC6 could not be compelled to pay a dividend (see [46] above) but nonetheless considered that a covenant of that nature would be critical. If Ms Choudhury’s submission was right he would be insisting on something that he knew added nothing.

Mr Ashley’s position on this issue is in fact further, and helpfully, explained in his second witness statement, which the FTT must be taken to have taken into account. That describes three possible versions of what was described in section 14f of the joint statement ([48] above):

It was therefore clear that the covenants Mr Ashley had in mind were ones that would ensure that the preference share rights were adhered to. That is reflected in section 14f of the joint statement, which as already noted refers to covenants which “ratify” LLC5’s position.

In summary on the Transfer Pricing issue, I would conclude that deductions for interest on the Loans are not restricted under the transfer pricing rules. I would therefore allow LLC5’s appeal on Ground 1, set aside the UT Decision on that issue and re-make it by dismissing HMRC’s challenge to the conclusion reached by the FTT. Insofar as the conclusion I have reached relies on evidence before the FTT about the risks assumed in the actual transaction that was not reflected in a finding of fact (see [65] and [66] above) I would make additional findings of fact, pursuant to s.14 of the Tribunals, Courts and Enforcement Act 2007, that accept that unchallenged evidence. I would also dismiss HMRC’s challenge to the FTT’s findings about the expert evidence (HMRC’s Ground 1).

The unallowable purpose rule in s.441 CTA 2009 forms part of Part 5 of that Act, which contains the provisions governing the treatment of “loan relationships” for corporation tax purposes. The concept of a loan relationship includes any lending transaction, such as the Loans (s.302). In very broad terms, Part 5 provides for credits and debits from loan relationships to be determined in accordance with generally accepted accounting practice. For a non-trader like LLC5, net profits (that is, any excess of credits over debits) are taxed as non-trading profits. A net loss (being an excess of debits over credits) is a “non-trading deficit” (s.301). Among other things, non-trading deficits are available to be surrendered by way of group relief to offset UK profits of other group members.

Sections 441 and 442 CTA 2009 relevantly provide:

“Tax advantage” is defined in s.1139 Corporation Tax Act 2010. It is common ground that the deduction of loan relationship debits in respect of interest pursuant to Part 5 CTA 2009 is a tax advantage.

As already indicated, the FTT decided that, although there was an “unallowable purpose” in the form of a main purpose of securing of a tax advantage, there was also a commercial purpose. It further decided that none of the interest deductions should be attributed to the unallowable purpose on a just and reasonable apportionment, with the effect that they were deductible in full.

The UT decided that the FTT had erred in applying the reasoning in the well-known case of Mallalieu v Drummond [1983] 2 AC 861, 57 TC 330 (“Mallalieu”) in deciding whether there was an unallowable purpose, but nevertheless concluded that there was no material error in the FTT’s conclusion that there was a main purpose of obtaining a tax advantage. However, the UT also decided that the FTT did make a material error in its application of the just and reasonable apportionment test, and determined that all of the debits should be attributed to the tax advantage purpose.

BlackRock agree that the FTT erred in applying Mallalieu but say, as Ground 2 of their appeal, that the UT should not have substituted its own finding of a tax avoidance purpose in the absence of any finding by the FTT that LLC5 actually did have such a purpose. Ground 3 is that the UT erred on the apportionment issue.

In their Respondent’s Notice, HMRC say that the UT erred in concluding that the FTT had been entitled to find that one of the main purposes of LLC5 and the Loans was a commercial purpose (HMRC’s Ground 2), and that the UT was wrong to decide that the FTT had misapplied the test in Mallalieu and subsequent authorities (HMRC’s Ground 3).

There was a significant element of common ground between the parties in their interpretation of s.442. In particular, both parties accepted that this court should follow a summary by Newey LJ (with whom the other members of the court agreed) in Travel Document Service v HMRC [2018] EWCA Civ 549, [2018] STC 723 (“TDS”) at [41], in relation to the predecessor legislation in para. 13 of Schedule 9 to the Finance Act 1996:

It was also not disputed that, in deciding whether a loan relationship has an unallowable purpose, what matters is the company’s subjective purpose or purposes in being a party to the loan relationship in question. In this case that means LLC5’s purpose or purposes in being a party to the Loans. The focus can also be narrowed further to LLC5’s entry into the Loans, because it is common ground that its purposes in being a party to the loan relationship did not change thereafter.

The parties were quite right not to dispute the fact that what matters is the company’s subjective purpose or purposes in being a party to the loan relationship in question. The purpose or purposes for which a company is a party to a loan relationship may or may not be the same as, for example, the purpose or purposes for which the company exists, or the purpose or purposes of a wider scheme or arrangements of which the loan relationship forms part. Those other purposes may, for example, encompass the purposes of other actors. There is a contrast here between the unallowable purpose rule and the “targeted anti-avoidance rule” introduced by Finance (No.2) Act 2015 as ss. 455B-455D CTA 2009. That rule requires consideration of the main purpose or purposes of “arrangements”.

It was also common ground that for a corporate entity such as LLC5, which can only act through human agents, it is necessary to consider the subjective purpose of the relevant decision makers. Unless they have been bypassed or are effectively acting on instruction, that will normally be the board of directors. The same would apply to LLC5, although strictly its board was termed a “Board of Managers”, appointed pursuant to the terms of its LLC agreement. There was no suggestion that the board of LLC5 had either been bypassed or were acting on instruction when they agreed to enter into the Loans.

In contrast to the UT’s conclusion that the FTT should have relied solely on the principles derived from TDS, both parties (rightly) accepted that it was appropriate to consider other case law, and in particular Mallalieu, MacKinlay v Arthur Young McClelland Moores & Co [1990] 2 AC 239, [1989] STC 898 (“MacKinlay”) and Vodafone Cellular Ltd v Shaw [1997] STC 734 (“Vodafone”), However, they differed in the application of the principles to be derived from those cases to the facts of this case.

In Mallalieu, Baroness Mallalieu (then Miss Mallalieu) sought tax deductions for the cost of replacing and cleaning the (predominantly black) clothing that she wore in and on her way to court and chambers to work as a barrister. That clothing reflected Bar Council requirements about court dress. On her appeal against the Inland Revenue’s refusal to allow those amounts as expenses “wholly and exclusively laid out or expended for the purposes of” her profession, the General Commissioners found that she would not have bought the disputed items but for the requirements of her profession, and that considerations of warmth and decency had not crossed her mind when she did. However, and other than in respect of a collar designed to enable bands to be worn, they dismissed the appeal on the basis that she had a dual purpose, in the following terms (recorded in Lord Brightman’s speech [1983] 2 AC 861 at p. 872):

The full version of the Commissioners’ case stated appears in the Tax Cases report (57 TC 330). Apart from the comment – inappropriate at least to 21^st century eyes – that “Miss Mallalieu is an attractive blonde barrister”, paragraph 2 (referred to in the quoted passage above) recorded the Commissioners’ findings that the clothes in question were not to her taste, that she virtually never wore them except at or travelling to work, and that she wore them in chambers due to her busy court practice and the likelihood of being required to attend court at short notice. It was also the case, however, that the disputed items were of a kind that others might wear by choice.

The High Court reversed the Commissioners’ decision, holding that the only proper conclusion on the facts was that the sole purpose was to satisfy professional requirements. The Court of Appeal affirmed that conclusion. However, the House of Lords allowed the Inland Revenue’s appeal. Lord Brightman gave the leading speech, with which Lords Diplock, Keith and Roskill agreed. Lord Elwyn-Jones gave a short dissenting speech agreeing with the courts below, essentially on the basis that it was not open to the Commissioners to disregard the evidence that they had accepted of Miss Mallalieu’s “actual motive and purpose”, which was to carry on her profession, other benefits being “purely incidental”.

Lord Brightman described the “wholly and exclusively” test, then in s.130(a) Income and Corporation Taxes Act 1970 (“ICTA 1970”), in the following terms (at p.870):

Thus, in all but obvious cases it is necessary to “look into the mind” of a taxpayer, and object must be distinguished from effect. Both of those points apply equally to the unallowable purpose rules.

After considering the wider ramifications of a decision in the taxpayer’s favour, Lord Brightman went on to reject the lower courts’ conclusion that all that mattered was what was in Miss Mallalieu’s conscious mind, and that warmth and decency was merely an incidental effect. He said this (at p.875):

So while object and effect are not the same, object is not “inevitably limited” to conscious motives. The reality was that Miss Mallalieu obviously needed clothes for warmth and decency, and no doubt would have had to accept that if she was asked. This could not be described as merely an incidental effect or consequence. Her choice of a particular style and colour, which she would not otherwise wear, did not change that inescapable fact.

MacKinlay related to a policy of a large accounting firm to reimburse specified domestic removal expenses of partners who were required to move to work in a different office. Moves were made at the request of the executive committee of the partnership, which was also responsible for the policy and its operation. The expenses covered certain costs of the moving process, such as fees of estate agents, plus some additional costs such as relaying carpets and refitting curtains. A deduction was claimed in the partnership tax return for the removal expenses of two such partners but was denied. On appeal, the Commissioners found for the partnership. That decision was reversed by Vinelott J in the High Court. The Court of Appeal allowed the partnership’s appeal, but that was overturned in the House of Lords with the result that the deduction was denied.

As in Mallalieu, the issue was whether the expenditure was wholly and exclusively for the purposes of the business. However, in that case the business was the business of the partnership, considered as if it were a separate entity from the recipient partner (p.250A).

Lord Oliver gave the leading speech, with which Lords Bridge, Brandon, Templeman and Goff agreed. Lord Oliver concluded that the Court of Appeal had erred in segregating the purpose of the partnership, considered as if it was a separate legal entity, from the purpose of the individual partners (p.253C). What had to be considered was the purpose of the outlay, and whether it was exclusively for the purposes of the partnership business. That could not simply be answered by considering the motive behind the move. As Lord Oliver explained at p.255:

Therefore, motive is not necessarily the same as object or purpose. Further, “some” results are “so inevitably and inextricably involved” in an activity that they must be a purpose for it. An example of that was Miss Mallalieu’s sober garb. Another was the facts of the instant case, where the costs “cannot but have been intended to serve the purpose of establishing a comfortable private home for the partner concerned”.

Vodafone also concerned a claim to deduct under s.130(a) ICTA 1970, but in that case in a corporate context. It concerned an amount paid by the taxpayer to bring to an end a fee agreement related to the acquisition of know-how which it turned out was not required. The Special Commissioners accepted that the payment was of a revenue nature but dismissed the taxpayer’s appeal on the basis that the payment was made to benefit the trading position of the whole group, and not solely for the purposes of its own trade. Their decision was upheld in the High Court but the taxpayer successfully appealed to the Court of Appeal.

Millett LJ, with whom Sir John Balcombe and Hirst LJ agreed, pointed out that the question of whether a payment is made exclusively for the purpose of the taxpayer company’s trade or partly for that purpose and partly for another is a question of fact. He went on to say (at p.742):

In Vodafone, the Commissioners had found that the directors of the taxpayer had in mind its legal obligations under the fee agreement, rather than the position of its subsidiaries which had no such obligation but which it had been contemplated would reimburse the taxpayer for ongoing fees. The Commissioners had nonetheless inferred that the directors “must have had in mind” that the cancellation of the agreement would be of greater benefit to the subsidiaries because of the reimbursement arrangements. Millett LJ observed that this finding could not be supported, because it was based on a misunderstanding of their earlier findings that there was no agreement or formal arrangement in place, and in its absence the taxpayer would not have been properly able to seek reimbursement for the provision of know-how that was not required. However, he concluded that the appeal should be allowed on the simpler basis, by focusing on “what was the particular object which the directors were seeking to achieve?” (p.744h). Although it was self-evident that the directors’ purpose was to rid the group of a trading liability owed to a third party, the liability was that of the taxpayer alone, “ergo the directors’ intention, whether articulated or not, was exclusively to serve the purposes of the taxpayer company’s trade”. The impact on the wider group was a consequential or incidental effect (p.745).

Millett LJ’s useful summary brings out the points already referred to derived from Mallalieu and MacKinlay. For present purposes “object” can also be regarded as synonymous with purpose. So far as relevant to this case, and gathering the points together, I would summarise the key points as follows:

For reasons which will become apparent I will deal in some detail with the facts found by the FTT, elaborated to some extent by reference to the documentary evidence before it and also by reference to some oral evidence referred to by the UT. Cross-references are to paragraphs of the FTT Decision unless otherwise indicated.

BRI announced the proposed acquisition of BGI on 11 June 2009. The acquisition of BGI US completed on 1 December 2009. Between those dates a lot of work was done on structuring the acquisition, led (as such work usually was) by BlackRock’s corporate tax group. In particular, Ernst and Young (“EY”) were asked to consider where “debt push downs” should be done (in this context meaning “pushing” debt incurred to fund the acquisition down to other entities in the group through intra-group loans). EY, who had a “very broad remit”, suggested the use of a UK entity to acquire BGI US to take advantage of the “generous tax regime for interest deductions”. This fairly quickly developed into using an LLC, and by late July 2009 to a structure involving two LLCs, one of which was UK tax resident, with the UK resident LLC acquiring BGI US. (See paras. 9, 13-15 and 21-22.)

By mid-August the proposed structure changed to include a third LLC, for the reasons mentioned at [16] above.

Four UK-based BlackRock executives were identified as potential members of LLC5’s board. They comprised Mr Kushel, who at the time was Chairman of the BlackRock group’s international business, Colin Thomson (Head of BlackRock’s Financial Reporting Group for the international business), Roger Tooze (Head of BlackRock’s Business Finance) and James DesMarais (General Counsel for BlackRock’s international business). Mr Fleming, who was BlackRock’s head of tax for EMEA (Europe, Middle East and Africa) prepared a briefing note to be shared with Mr Thomson, Mr Tooze and Mr DesMarais (Mr Kushel’s senior leadership role meant that he was already involved in work on the acquisition). The note, dated 26 October 2009, explained among other things that “the purpose of [LLC5] is to effect the acquisition of the BGI US business from Barclays”, that its central management and control needed to be in the UK, and that:

Mr Fleming met with Mr Thomson, Mr Tooze and Mr DesMarais on 10 November 2009. The “main focus” of the discussions was to explain the UK tax rules regarding deductions for interest (paras. 36 and 38). By that stage the transaction had received regulator and internal stakeholder approval and was “largely final” and subject to revision only in “limited circumstances” (para. 37).

The board meeting was originally arranged for 27 November but was re-arranged to 30 November following a discussion with Mr Kushel. This was due to a clash with Thanksgiving which initially led to him intending to attend by telephone, before being told that the meeting was “important for UK tax purposes” (paras. 40-41).

The FTT deals with the 30 November board meeting at paras. 42 to 53. The meeting took place in London, with all four board members present and Mr Kushel acting as chairman. Those attending were not doing so “in a vacuum”, given the prior discussions. The meeting lasted around 45 minutes. Mr Fleming was present to explain LLC5’s role. Mr Kushel’s evidence, which was accepted (see [14] above), indicated that he clearly understood his fiduciary duty to act in the best interests of LLC5. However, provided that it was in the company’s best interests to enter into a transaction he did not consider it part of his remit to question or suggest changes to the “underlying capital structure” of a proposed transaction. The UT Decision notes at para. 172 that Mr Kushel’s oral evidence was that the board “were not considering whether or not it was the right business of LLC5 to invest in the LLC6 Preference Shares”, and at para. 176 that he “confirmed that there was no reasonable possibility that LLC5 would not enter into the transaction” and that Mr Fleming also accepted that the board was presented with a “fait accompli”.

Further, the meeting was very shortly before the completion of a complex transaction for which a detailed “step plan” had been prepared. Provided the board concluded that the transaction was commercially advantageous for the company it “would not have been sensible or open to the directors to consider an alternative transaction” (para. 52 of the FTT Decision). Rather, Mr Kushel saw it as his responsibility to satisfy himself that a proposed transaction had been:

The FTT also records that Mr Kushel regarded the other board members as “diligent and thoughtful” people who would not simply “rubber stamp” a transaction (para. 45).

Following initial references to the acquisition of BGI and to the step plan (prepared by EY), the most material section of the board minutes, which was accepted to be an accurate record, reads as follows:

The minutes record that the board went on to consider the terms of the first loan note and how it would be repaid, the relative values of the common and preference shares and unwind provisions (should they be required). The final comment before the formal resolutions reads:

As can be seen from the minutes, the board understood that LLC5 would obtain no benefit from the interest deductions. This was because – as well as LLC5 having no taxable income itself against which the expense could be offset – the group practice was to make group relief surrenders for no payment. Given that LLC5’s assets reflected an interest only in the US operations of BGI, it would also have been apparent to board members that no indirect benefit could be derived through its preference share investment.

Mr Kushel’s explanation that Mr Fleming’s advice to leave any UK tax advantage out of account would have been followed was evidently accepted by the FTT. Similarly, the FTT accepted Mr Kushel’s evidence that if the tax benefits had fallen away it would have been too late to alter the structure.

These points are also reflected in the following passages of Mr Kushel’s witness statement:

After referring to the need to ascertain the subjective intentions and purposes of LLC5’s directors, the FTT’s analysis is set out in the following three paragraphs:

The UT considered and rejected HMRC’s challenge to the FTT’s finding at para. 121 that there was a commercial purpose, on the basis that it did not meet the high threshold required under Edwards v Bairstow principles ([1956] AC 14). There was evidential support for the factual conclusion reached (paras. 150 -152 of the UT Decision).

The UT then considered Mr Prosser’s arguments that the FTT erred in relation to Mallalieu, saying at para. 161:

However, it then went on:

The UT went on to consider the evidence in more detail, at paras. 168-179, under a sub-heading “FTT’s findings and evidence in support of tax advantage main purpose”. It concluded as follows:

In my view there were errors in the approach of both the FTT and UT.

I disagree with Mr Ewart’s submission that the tax advantage was the sort of “inevitable and inextricable consequence” referred to in MacKinlay in discussing Mallalieu, and in Vodafone in discussing both of those cases. Rather, what can be drawn from those authorities are the points summarised at [124] above.

Purpose must be distinguished from effect. Even unavoidable effects are not necessarily the same as purposes. This is particularly clear from Lord Brightman’s example in Mallalieu of a medical consultant’s trip to the South of France (see [113] above). It cannot therefore be the case that any inevitable consequence can be a purpose. Indeed, the authorities spell that out: both Lord Oliver in MacKinlay and Millett LJ in Vodafone refer to “some” consequences or results being inevitably and inextricably involved in particular activities ([119] and [122] above).

As to what kinds of consequences fall into that category, the facts of Mallalieu and MacKinlay provide good examples of what Lord Oliver and Millett LJ would have had in mind. Miss Mallalieu obviously needed clothes for warmth and decency, and the clothes in question were ones that others could wear by choice. Miss Mallalieu would just not have chosen that particular colour or style. But, once acquired and worn, her black clothes performed the necessary functions of keeping her warm and decent. She would obviously have had to accept, if asked, that the items she was replacing or taking to the cleaners would not only fulfil those functions, but that she intended that they would do so. That was very far from the consultant in Lord Brightman’s example who benefits from incidental enjoyment during a work trip.

Similarly in MacKinlay, Lord Oliver explained that the expenditure “cannot but have been intended to serve the purpose of establishing a comfortable private home for the partner concerned”, and that the partner and his wife could not be said to have been acting on partnership business when they bought new curtains. That too is a clear example. In both cases the conscious business purpose necessarily encompasses another, non-business, purpose.

In contrast, in Vodafone any benefit to the wider group was merely consequential: see [123] above.

How then should this point be addressed in the context of s.442? The unallowable purpose rule forms part of a code, contained in Part 5 of CTA 2009, which governs the treatment of loan relationships for corporation tax purposes, and which among other things specifically contemplates tax relief for interest and other expenses of raising debt. The corporation tax relief available is obviously a valuable relief. It is unrealistic to suppose that it will not form part of ordinary decision-making processes about methods of funding a company. Indeed, it might well be wrong for directors to ignore that consideration in deciding what is in the best interests of the company concerned. I agree with Mr Prosser’s submission that it cannot have been Parliament’s intention that the inevitable consequence of taking out a loan should engage the unallowable purpose rules, subject only to consideration of whether the value of the tax relief is sufficient to make it a “main” purpose. Something more is needed.

I therefore agree with BlackRock that the FTT should not have applied the test in Mallalieu as it did. This is not for the reason given by the UT (to the effect that it should simply have applied TDS) because it is not the case that Mallalieu and the later cases that discuss it are irrelevant. Rather, the FTT made an error of law in proceeding on the basis that the “inevitable” consequence of tax relief was, without more, a main purpose.

I also consider that the approach that the UT then adopted contained both substantive and procedural errors.

First, the UT made references to the purpose of LLC5’s existence and inclusion in the transaction (see in particular at para. 163) without making it clear that the statutory test requires a focus on LLC5’s purpose or purposes for being a party to the Loans. While one may of course impact on the other, it is important to recognise that the purposes for which an entity exists and its purposes in entering into a transaction may be different.

Secondly, the UT were in my view not entitled to take the approach adopted at para. 164, in which subjective intentions were described as “circumscribed” and not what they would have been “had [board members] been left freely and willingly to decide the main purposes of the transaction”. Apart from the test of purpose not being answered simply by asking the decision maker (see [124] above), there was no evidence, let alone a finding, that the board was not free to consider the transaction properly. There is no suggestion that board members were acting on instruction. It is right that they were advised to leave any UK tax advantage out of account in assessing the viability of the transaction for LLC5, but that was for the entirely proper reason that the company would itself obtain no benefit from it. Given that last point, it was also inappropriate for the UT to make the unqualified comment that it did in para. 165 about there being an easy way round the legislation if tax advisers could tell parties to ignore tax advantages. In this case there was a sound reason for the board to leave tax out of account in assessing whether the transaction was in the best interests of LLC5.

Thirdly, the conclusion in paras. 165 and 166 that “it is necessary to look beyond the directors’ stated intentions” gives at least the appearance of concluding that the evidence of BlackRock’s witnesses should not have been accepted at face value. Instead, the UT needed squarely to address the point that the FTT did accept their evidence and that the only reason it gave for going beyond it to find that there was a tax avoidance purpose was Mallalieu. There is no factual conclusion beyond that to the effect that LLC5 had such a purpose in entering into the Loans.

This last point leads into the procedural error. Where the UT determines that the FTT has made an error of law, it may (but need not) set that decision aside: s.12(2)(a) Tribunals, Courts and Enforcement Act 2007 (“TCEA”). If, but only if, it does so then it must decide whether to remit the case or re-make the decision (s.12(2)(b)). It is only if it chooses to re-make the decision that it has power to make findings of fact: s.12(4).

In this case, it is apparent from para. 162 of its decision that the UT chose not to set the FTT Decision aside because it considered that there was no material error of law. In those circumstances it was not entitled to go on to make further findings of fact. However, I consider that that is what it then purported to do, both in the paragraphs just referred to and in the analysis in para. 168 onwards which considered not only the FTT’s findings of fact but further evidence.

In my view the only proper approach was to determine that, due to the FTT’s error, its conclusion that LLC5 had a tax advantage main purpose could not stand. There was not only a material error of law but the FTT Decision lacks the necessary factual findings to support the decision on a different basis.

It is then necessary to determine what the consequences of this should be. The thrust of Mr Prosser’s submissions was that HMRC have lost the chance to put their case in a different way. HMRC have consistently relied on the concept of consequences that are “inevitably and inextricably involved” and have defended the FTT Decision on that basis. More significantly, Mr Kushel’s evidence on the critical issue was never properly challenged. It was not put to him that, despite maintaining that the board members followed the advice that tax needed to be left out of account, they must have been thinking about tax, or that they intended or desired to achieve a tax advantage. Instead, the FTT accepted Mr Kushel’s evidence. That would have included what he said about purpose in his witness statement (see [138] above).

It would clearly be unfair to give HMRC another bite of the cherry by remitting the case in a way that would allow the evidence to be revisited and enlarged upon. An alternative would be to remit it on a basis that did not permit that, but inviting the FTT to make additional findings based on the existing evidence and findings of fact. However, I have come to the conclusion that remittal is not necessary. Rather, the appropriate course is for this court to re-make the decisions of both tribunals and, in doing so, to exercise this court’s own power (pursuant to s.14 TCEA) to make findings of fact to the extent required, based on the evidence before the FTT.

The FTT found that Mr Fleming’s advice to leave any UK tax advantage out of account was followed. It also found that the Loans would still have been entered into if the tax benefits had fallen away at the last minute. However, I agree with Mr Ewart’s submission that the factual enquiry is not limited to the factors that the board took into account in deciding that the transaction was in the best interests of LLC5, and that it is also not sufficient that the transaction would have proceeded without the tax benefits because it was too late to make changes.

As Nugee LJ suggested in argument, a simple starting point in ascertaining a person’s purpose for doing something is to consider “why” they did it. While this will not cover all the nuances – and in particular the potential distinction between purpose and motives discussed in MacKinlay – it is a sensible starting point.

There is an obvious answer to that question on the facts of this case. However it might be dressed up, LLC5 became a party to the Loans to obtain a tax advantage.

As has already been said, the board members of LLC5 were not operating in a vacuum. Mr Kushel was involved in the transaction in any event and the others had been briefed at an earlier stage. Much of the substantive discussion at the board meeting was about tax. The board obviously understood what the Loans were designed to achieve. Although, as already discussed, the purpose or purposes of being a party to a loan relationship cannot simply be elided with the purpose for which the relevant entity exists, in this case LLC5 had no other function. Its sole raison d’être was to enter into the Loans to obtain tax advantages for the BlackRock group.

That is obvious from the diagrammatic representation of the structure appended to this decision. Passing funds to LLC6 via a (UK tax resident) LLC5, rather than direct from LLC4, and doing so in part by means of loans rather than the pure equity contributions made at each other level in the structure, sought to achieve the objective of reducing the group’s UK tax bill. LLC5 gained no control over the BGI US group and therefore had no meaningful function in that respect. It was, and was structured to be, a mere passive recipient of preference share dividends without control or indeed any real influence. Its (UK based) directors would in any event have been unlikely to have any meaningful involvement with an entirely US based sub-group. (Mr Kushel was a senior executive, but Mr Fleming’s witness statement records that the intention was for him to “come off the board after closing”.)

It would be artificial to seek to divorce what occurred at the board meeting from its context. That context included:

I do not overlook that the board concluded that the transaction was commercially advantageous for LLC5. It is obvious that it was. The anticipated dividend flow on the preference shares was such that it would very comfortably exceed the cost of servicing the Loans. On that basis I have no difficulty with the FTT’s conclusion that LLC5 had a commercial purpose in entering into the Loans: it was set to make very significant profits from its investment. However, the fact that it would be able to make such profits was a consequence of the need to ensure that the transfer pricing analysis was robust, so that interest costs could not be denied on the basis that LLC5 was thinly capitalised. From the group’s perspective the commercial impact for LLC5 was in truth a by-product of the tax planning. Nevertheless, as far as LLC5 was concerned it was more than a (very welcome) consequence or effect, because it meant that the transaction made commercial sense from its perspective. As a result the board was able to conclude that it was in LLC5’s best interests to enter into the proposed transactions, including the Loans.

It makes no difference to the analysis that, by the time it got to the date of the board meeting, it would have been too late to change the structure if the hoped-for tax advantage had been no longer available. Quite apart from the fact that the board was presented with a fait accompli (see [131] and [132] above), the point was an entirely hypothetical one. In reality, it continued to be thought that a tax advantage would be available.

I do not consider that these conclusions involve an inappropriate attack on the unchallenged evidence of Mr Kushel or Mr Fleming, or on the FTT’s findings of fact. I agree with the UT that there is an analogy with TDS. In that case a conclusion that the company had a tax main purpose in using the relevant shares as it did was found not to be inconsistent with the honestly held view of the relevant director, a Mr Turner, that the shares continued to be held exclusively for the commercial purpose for which they were originally acquired. Similarly, in this case the fact that LLC5 had a tax avoidance main purpose is not inconsistent with board members properly putting the tax benefits out of their minds when deciding whether the transaction was in LLC5’s best interests on a standalone basis. The two questions are different.

Accordingly, I would conclude that LLC5 entered into the Loans with a main purpose of securing a tax advantage. In other words, I agree with the FTT and UT in the result.

I should emphasise that my conclusion that LLC5 had a tax main purpose is a conclusion reached on the particular facts of this case. It does not follow that other debt incurred in connection with a commercial acquisition – as the acquisition of BGI US undoubtedly was – would fall foul of the unallowable purpose rule even if the decision to borrow had regard, as it often would, to tax considerations. The facts of this case include, among other things, the use of a debt-funded UK resident entity in what is otherwise a wholly US-based, and equity funded, ownership chain, the related lack of any commercial rationale for LLC5, and the structure that then had to be put in place to ensure that LLC5 did not have control over the BGI US group, such that LLC5 not only had no commercial rationale but had no real commercial function.

It follows from what I have already said that I would further conclude that the FTT reached the correct decision in finding that LLC5 also had a commercial main purpose for entering into the Loans.

HMRC criticise the judge’s reference to LLC5’s business of “making and managing passive investments” as having no basis other than Mr Fleming’s briefing note (see [128] above), but it was an accurate description of what was anticipated to be a highly profitable activity. LLC5 could not legitimately have entered into the Loans except as part of a transaction that enabled it to deploy its newly acquired assets in its commercial interests, such that it was anticipated that it would not only meet its obligations under the Loans but would also make a profit.

The FTT’s reasoning on this issue was very brief. After the passage set out at [139] above, the FTT Decision went on as follows:

The UT considered that the FTT had wrongly adopted a subjective approach, and instead what was required was an objective consideration of all the facts and circumstances. A “but for” approach may be helpful as a check, but in this case the Loans would not have been incurred at all but for the tax purpose. The FTT had instead asked itself a different question, namely whether the loans would have existed if the tax relief had been withdrawn just before the transaction was due to complete.

Mr Prosser sought to support the FTT’s conclusion by submitting that the correct approach was one based on causation, and in particular a “but for” test. Further, although the exercise was objective in the sense that it was for the tribunal to determine what a just and reasonable apportionment would be, the apportionment was between the subjective purposes that had been identified (in other words, between the tax main purpose and the other, permissible, main purpose or purposes). In this case the board of LLC5 would have entered into the Loans even if the tax advantage had not existed. This “but for” approach was one that had been adopted in previous cases and gained some support from TDS at [50]-[54], a passage discussed at para. 124 of the Oxford Instruments decision relied on by the FTT.

Mr Yates also made submissions about the approach that should be adopted were we to conclude that it would not be correct to apportion 100% of the debits to LLC5’s commercial purpose. Two approaches were put forward. The first was to apportion by reference to the relative anticipated financial significance of the tax relief and the commercial advantage (being the excess of dividends over interest payable), taking account of the fact that at the time the Loans were entered into it was anticipated that tax relief for a substantial portion of the interest costs would be denied under the worldwide debt cap rules. These are the “debt cap rules” referred to in the board minutes (see [134] above). At the time they were contained in Part 7 TIOPA and, very broadly, restricted deductions in respect of financing costs of UK members of a group by reference to the group’s external financing expense. When the Loans were entered into a substantial disallowance was anticipated under those rules due to the way in which BlackRock expected that certain external liabilities would be characterised, but work in 2013 and 2014 resulted in agreement with HMRC that there would be no such disallowance.

The second approach was to allocate to the tax advantage purpose the proportion of the total interest expense that was anticipated to be allowable at the time the Loans were entered into, again after taking account of the anticipated restriction under the worldwide debt cap rules, with the balance being apportioned to LLC5’s commercial purpose.

I agree with Mr Prosser that what the legislation requires is a just and reasonable apportionment by reference to the relevant purposes. Those purposes are identified using a subjective approach: see above. The statutory test requires the identification and disallowance of “so much of” any debit that is “attributable” to the unallowable purpose: s.441(3). That is the enquiry that the tribunal must undertake. While the determination of a just and reasonable apportionment is an objective exercise, it is not necessarily the same as the consideration of “all the facts and circumstances” referred to by the UT, if by that the UT did not intend to have regard to the requirement to apportion by reference to the relevant purposes. The framework for the apportionment is the purposes that have been identified by the fact-finding tribunal. Subject to that point, however, I agree that all relevant facts and circumstances should be considered.

The position is straightforward if all the debits, or perhaps a defined part of them, are properly attributable solely to a tax avoidance main purpose. Conversely, if they are properly attributable to a purpose which is not an unallowable purpose then there will be no disallowance under s.441. Where debits are attributable to more than one purpose then an apportionment is required. As to the precise mechanism by which this is done, the legislation is not prescriptive. The answer to that question will inevitably be fact specific.

On the facts of this case I agree with the UT that the FTT was wrong to focus on what would have happened if tax relief had been withdrawn at the last moment. I would also concur with the result reached by the UT, namely that 100% of the debits should be apportioned to the unallowable purpose.

As already discussed, the purpose for which LLC5 was created cannot be divorced from its purpose in entering into the Loans. Further, the structure of the transaction was presented as a fait accompli to the board. The commercial advantage to LLC5 was of significance to it because it would not benefit from the tax advantage, but overall it was more in the nature of a by-product. On the facts there is no principled basis to identify any particular amount or proportion of the debits as being attributable to the commercial purpose.

The answer to the apportionment question is not altered by imagining what might have happened if the tax rules had changed at the last moment. That did not occur. In reality the board accepted and adopted the structure that had been devised to achieve a tax advantage.

While I understand Mr Yates’ submissions about alternative approaches and can see that some form of apportionment based on economic advantage could be appropriate in some cases, I do not consider that it would be appropriate to adopt either of the two approaches put forward in this case. Apart from the fact that the FTT was not asked to consider those alternatives, they both re-introduce subjective elements into what should be an objective exercise, namely LLC5’s expectations about the relative level of tax benefits as compared to other factors.

In any event, there is a simpler answer based on the “but for” approach that both parties were content that we should adopt on the facts of this case. If that approach is adopted then the only proper answer is not the one that the FTT gave. Rather, in the absence of the tax advantage the decision to enter into the Loans would never have been made.

Accordingly, I would dismiss LLC5’s appeal against the UT’s conclusion on apportionment.

In summary, I would:

The result of this is that tax deductions for the interest on the Loans are disallowed under the unallowable purpose rule.