That might sound like a strange question, but it lurks just beneath the surface of the Upper Tribunal’s keenly-awaited decision in BlackRock which is the latest instalment in a series of “unallowable purpose” cases including Kwik-Fit and JTI.

The case concerns the structure used by BlackRock to purchase the North American investment management business of Barclays Global Investors in 2009. The acquisition structure entailed a Delaware-incorporated but UK-resident company (LLC5) borrowing $4 billion from its US-resident parent (LLC4) and contributing the funds, by way of subscribing for preference shares, to another US-resident company (LLC6), which acquired the BGI business. The witness evidence indicated that the UK filling for the sandwich had been chosen with a view to generating UK corporation tax deductions that could be surrendered to other group companies.

HMRC disputed LLC5’s attempt to claim non-trading loan relationship debits for the interest and other expenses payable in respect of the loan notes it had issued to LLC4, on two alternative grounds: first, as a matter of transfer pricing; and second, pursuant to the “unallowable purpose” rule which is now to be found in section 441 of the Corporation Tax Act 2009.

Whereas the First-tier Tribunal (FTT) held in favour of the taxpayer on (effectively) both of those issues, the Upper Tribunal (UT) has gone the opposite way.

Transfer pricing

For various reasons, including concerns about the application of the UK’s controlled foreign company rules, the voting rights had been structured such that LLC4 had control of LLC6, but LLC5 did not. Thus, LLC5 had no control over the flow of dividends on the preference shares issued by LLC6, but that did not matter to LLC4 in its capacity as LLC5’s creditor, as LLC4 itself controlled LLC6. Of course, an independent lender would not have had such control.

Consequently, at the FTT hearing, the expert witnesses agreed that an independent lender acting at arm’s length would be prepared to stand in the shoes of LLC4 by lending $4 billion to LLC5, but only if it was furnished with additional covenants from various third parties (including LLC6 and BGI) to protect the flow of dividends to LLC5. Based on the experts’ evidence, the FTT concluded that such covenants would have been forthcoming, and therefore held in favour of the taxpayer on the transfer pricing issue.

Before the UT, however, HMRC raised a new argument: namely, that the hypothetical arm’s length loan cannot take into account third party covenants which did not in fact exist. The UT agreed, deciding that this meant the FTT had failed to compare the actual transaction with an arm’s length transaction. Moreover, since the FTT had expressly found as a matter of fact that the actual transaction simply would not have taken place at arm’s length due to the lack of the covenants in question, the UT felt free to remake the decision on that basis — thus allowing HMRC’s appeal — rather than remitting the case to the FTT.

That is a somewhat brow-furrowing outcome — not least because, as the UT admitted, it implies that BlackRock could have supported its position if only it had undertaken the unnecessary and “rather artificial” exercise of obtaining the requisite third party covenants. As a matter of legal detail, the UT’s analysis is well-reasoned, and it clearly considered itself hidebound by the FTT’s finding of fact that, at arm’s length, the $4 billion simply would not have been advanced in the absence of the hypothesised covenants. Stepping back, however, it does seem odd that those covenants were considered to cause comparability problems, given that, at the macro level, their function in the hypothesis is simply to make the risks, assets and functions comparable by imitating contractually the control that LLC4 had over the dividend flows from LLC6.

Unallowable purpose

Having held in favour of HMRC on the transfer pricing issue, the UT’s decision on the unallowable purpose issue was, in its words, “immaterial to the outcome”. However, it will surely inform arguments in other cases, including the Kwik-Fit hearing in the UT in September.

The FTT had concluded that LLC5’s directors had, subjectively, two main purposes in borrowing the $4 billion: a main commercial purpose and also a main purpose of obtaining a tax advantage — the latter on the basis, following Mallalieu and its successors, that securing corporation tax deductions was an “inevitable and inextricable consequence” of issuing the loan notes and that, given its apparent importance, it must therefore be taken to have been one of the taxpayer’s main purposes. The FTT’s reasoning in that respect seemed weak, and, as many commentators and the taxpayer’s counsel pointed out, it would lade any innocent borrower with an imputed unallowable purpose. (One might also quibble that, in practice, tax consequences are rarely inevitable or inextricable — as illustrated by the UT’s reversal of the FTT’s decision!) Fortunately, the UT rejected the “inevitable and inextricable” approach, instead viewing the evidence in the round to uphold the FTT’s conclusion that LLC5 had an unallowable purpose of obtaining corporation tax deductions by playing its part in the overall plan.

Much more could be said about the attempts made by both sides to overturn the FTT’s conclusions on this issue, but for reasons of economy I’d like to focus here on the process of apportionment which the legislation requires where there is an unallowable purpose. The legislation states that a “just and reasonable apportionment” should be used to determine which debits, if any, are attributable to the unallowable purpose and must therefore be disallowed.

The FTT had followed in the footsteps of TDS and Oxford Instruments in adopting the view that the relevant debits should not be apportioned to the unallowable purpose to the extent that they would have been incurred even in the absence of any such purpose. Effectively, that is a “but for” test which asks whether the tax purpose increased the debits. Based on witness evidence to the effect that the structure would have been implemented as planned even if it had been discovered late in the day that the anticipated tax benefits would not materialise, the FTT attributed all of the debits to the commercial purpose and none to the tax purpose, such that no debits were disallowed.

The UT has effectively flipped that test on its head. In its view, apportionment is an objective exercise based on all the relevant facts and circumstances, and the FTT therefore erred by continuing to apply a subjective test even after reaching the apportionment stage. The UT concluded that, “but for” the tax advantage, there would have been no LLC5, no loan notes, and no debits in the first place. It followed “inexorably” that all of the loan relationship debits must be attributed to the unallowable purpose and so disallowed.

Because… reasons

That approach has its merits, but I wonder how the two tribunals have ended up deploying contrary “but for” tests when interpreting legislation that contains no such wording. Of course, the concept of a “just and reasonable apportionment” is a broad one, which affords judges considerable leeway to do justice in varied and unpredictable scenarios. However, as the FTT warned itself in Kwik-Fit, judges must be careful not to re-write the legislation. Parliament clearly chose not to enact a simple “but for” test, and it is worth asking why. It is now trite law that tax statutes should be interpreted “purposively” — that is, with a view to identifying and giving effect to the intentions of the legislature. Yet nowhere in its decision does the UT expressly attempt to delineate the purpose of the unallowable purpose rule.

Hence the question: what is its purpose? How far can its long arm reach? Is it really intended to disallow all debits in cases where the taxpayer has a main commercial purpose in being party to the loan, but both the taxpayer and the loan owe their very existence to group tax planning? The legislature’s decision not to enact the turbo-charged version of the unallowable purpose rule which was published in draft form in 2008 (which would have extended the rule to cases where the loan forms part of a wider tax avoidance arrangement) might suggest not.

Given the amounts at stake, and the FTT’s receptiveness to its arguments, BlackRock may well seek permission to appeal. Perhaps the Court of Appeal will take the opportunity to delve into the history of the unallowable purpose rule in greater detail, and thereby shed more light on its — still surprisingly inscrutable — purposes.