15 July 2024

Misunderstanding purpose in Osmond and Allen

Thomas Chacko’s article first appeared in Tax Journal on 3rd July 2024.

Speed read

The FTT in Osmond and Allen dismissed the taxpayer’s appeal in relation to the Transactions in Securities charge, finding a main purpose of obtaining an income tax advantage even though they were trying to avoid paying CGT. This misunderstands the test and is an example of a mistaken approach to anti-avoidance provisions seen in other cases. HMRC sometimes seek to avoid the factual difficulties with subjective purpose tests by arguing that a transaction fails them as a matter of definition, but this misunderstands how anti-avoidance provisions relate to the underlying tax legislation. The decision should hopefully be overturned.

In 1996 and 1998, Hugh Osmond and Matthew Allen (‘Osmond and Allen’) subscribed for shares in Xercise Ltd under the terms of the Enterprise Investment Scheme (EIS). One of the advantages of the EIS regime is that shares within it are generally exempt from CGT (TCGA 1992 s 150A).

Xercise ran a sports club in Great Yarmouth. The club was unsuccessful and the business was sold in 2002. However, Osmond and Allen were keen to hold on to the potential advantages of holding EIS-qualifying shares, and so persuaded the other shareholders not to dissolve or sell the company, but instead to purchase its own shares out of capital so as to leave Osmond and Allen the sole shareholders in Xercise Ltd. As the FTT found (in Osmond and Allen v HMRC [2024] UKFTT 378 at [11(4)]): ‘This action was taken precisely and only because the appellants appreciated that a future disposal of their shares in Xercise Ltd would be fully exempt from capital gains tax regardless of the nature of the future activity of Xercise Ltd or the size of the gains that might in future be generated.’

Accordingly, when Osmond and Allen sought (as part of a consortium) to take over the Pearl life assurance and pensions group in 2009, they persuaded their co-investors to use Xercise Ltd as a holding company. This meant that by the mid-2010s Osmond and Allen (though not the other investors) were holding CGT-exempt shares standing at a very large gain.

As they explained to the FTT, they became concerned that the EIS regime would be reformed so as to prevent them obtaining relief on their gains which, by this point, had little to do with the original business to which the EIS claim had related. They therefore hoped to crystallise the relief while it was available. One method would have been to sell their shares: but there was no obvious buyer.

Therefore they arranged for Xercise Ltd to buy back its own shares, for a total of £20m which was the maximum available as a return of capital.

HMRC did not challenge the application of the EIS regime, i.e. to argue that there should have been a CGT charge. Rather, HMRC sought to apply the Transactions in Securities (TIS) code.

At that time, one of the conditions in ITA 2007 s 684(1) was that ‘the main purpose, or one of the main purposes, of the person in being a party to the transaction in securities, or any of the transactions in securities, is to obtain an income tax advantage’.

Section 687(1) defined ‘income tax advantage’ as follows:

• ‘the amount of any income tax which would be payable by the person in respect of the relevant consideration if it constituted a qualifying distribution exceeds the amount of any capital gains tax payable in respect of it, or
• ‘income tax would be payable by the person in respect of the relevant consideration if it constituted a qualifying distribution and no capital gains tax is payable in respect of it.’

The first question the FTT addressed was whether Osmond and Allen had a main purpose of obtaining an income tax advantage when they entered into the reduction of share capital. It is well established that this is a subjective question, i.e. it depends on what they actually thought they were trying to do.

In terms of their intentions, the FTT held that they had not been trying to avoid paying income tax on an extraction of cash: they had no desire to extract funds from the company at all ([42]). They had not decided they wanted money and then decided how to extract it while paying as little tax as possible ([47]). Indeed, they undertook the share buyback ‘with some reluctance as it was the only way to crystallise the EIS disposal relief’ because they were concerned that ‘a change of government would affect its availability in their circumstances’ ([55]). If they had been able to crystallise it without extracting money they would have done so ([12(6)]).

One would therefore expect the taxpayers to have won: the FTT found they were not trying to avoid paying income tax on a dividend, not trying to extract value from a company, but were trying to secure a CGT exemption rather than risking (later) CGT treatment.

However, the FTT held that they did have a main purpose of obtaining an income tax advantage, because (as HMRC argued) to seek EIS relief was, by definition, to seek an income tax advantage (at [18(3)]): ‘EIS disposal relief generates an income tax advantage. This is because obtaining EIS disposal relief means the person is within the definition of income tax advantage as the CGT payable on the relevant consideration is less than the income tax payable if that relevant consideration had been paid to that person by way of a qualifying distribution.’

While (at [23]) the FTT noted that this was not a submission ‘seen made by HMRC in any previous case’, the FTT nonetheless accepted it, saying (at [27]) that ‘if you obtain EIS disposal relief, you must be within the definition of an income tax advantage as the CGT payable is necessarily less than the income tax which would have been paid had the consideration been paid as a qualifying distribution.’

It therefore did not matter that there was no prospect of Osmond and Allen taking a distribution. The FTT accepted, as explained above, their evidence that they did not in fact want the money at that time and so were not looking for a tax-efficient way to obtain it.

The FTT rejected the idea that there needed to be a ‘consciously considered comparable transaction’ – i.e. that the taxpayer had decided not to take a dividend and had chosen instead to use a reduction of capital – saying ‘the alternative transaction is already built into the definition of income tax advantage’ and that that definition ‘is a deeming provision limited only by the availability of distributable reserves. Whether or not the parties have any intention of carrying out a transaction in an alternative way, and in particular whether they consciously or subconsciously considered paying the consideration by way of a qualifying distribution, is when considering the statutory provisions, neither here nor there.’

The FTT decided that if a taxpayer wanted to obtain a CGT exemption (rather than pay full CGT), they were, by definition, aiming not to pay income tax. That is because they wanted to pay an amount of tax which was, by definition, lower than the income tax would have been, even if the possibility of their being liable to that income tax was entirely remote.

This led the FTT to the rather startling conclusion (at [37]) that: ‘We can see no principled reason why a main purpose of obtaining a CGT benefit or advantage cannot also be a main purpose of obtaining an income tax advantage.’

It is hard to see how this can be right. The definition of ‘income tax advantage’ is not a deeming provision and certainly does not seem intended to deem that a taxpayer had an intention that they did not in fact have: this seems at odds with decades of authority that the purpose test is a subjective test concerning the taxpayer’s actual intention, which in this case was to pay no CGT on a capital disposal rather than risk paying full CGT.

Further, in Allam v HMRC [2021] UKUT 291 (at [168] and [169]), the Upper Tribunal saw the need that there was an alternative transaction chargeable to income tax as ‘inherent in the requirement that a main purpose of being a party to the actual transaction was to obtain an income tax advantage’, and made clear that it was an important factor, when investigating purpose, to ask whether that alternative was considered by the taxpayer.

Purpose was not the only issue in Osmond and Allen: the FTT went on to reject the appellants’ arguments that a reduction of capital of this kind was, by definition, outside the scope of the TIS regime, and that there was a four year rather than a six-year time limit for HMRC to raise assessments. The appeals were dismissed.

However, if, as the FTT found, the taxpayers had no wish to extract funds from the company, were not trying to do so without declaring a dividend, but in fact were explicitly trying to obtain one form of CGT treatment (i.e. with EIS relief) rather than another, it seems that the taxpayers should have won.

The point of the purpose test

The TIS purpose test is specifically directed at avoiding income tax, in particular the income tax that would be charged when a dividend is declared. The legislation targets the case where a dividend could be taken, a capital receipt is obtained instead, and the taxpayer did this to obtain that tax saving. To read this as implicitly covering an attempt to avoid CGT misses the heart of the definition in s 687.

However, the FTT’s approach also shows a more interesting mistake as to how such purpose tests work, one which has appeared in HMRC’s submissions in other recent cases. This is to try to avoid the factual difficulties and messiness of investigating subjective intention by defining the transaction in a way that it must fail the purpose test. These arguments are attractive because they sidestep the evidential complexity of a purpose test: but I would say they involve a fundamental misunderstanding of what anti-avoidance provisions of this kind are doing.

HMRC unsuccessfully argued something similar in the recent Burlington Loan Management litigation ([2022] UKFTT 290, [2024] UKUT 152). A Cayman company held the right to interest arising out of the Lehman administration and sold it to an Irish company. The Irish company, unlike the Cayman company, could rely on article 12 of the UK/Irish tax treaty which provided that interest was only taxable in Ireland. However, article 12(5) disapplied this ‘if it was the main purpose or one of the main purposes of any person concerned with the creation or assignment of the debt-claim … to take advantage of this Article’.

Both tribunals (the FTT (at [193]) and the UT (at [68]) saw HMRC as arguing that ‘taking advantage’ in this context included mere planning on the basis that the relief would be available: because the Cayman company knew it would pay UK tax and the Irish company knew it would not, and they priced the sale to take account of that (effectively dividing the value of the relief between them), they were caught by article 12(5) and the UK could tax the interest.

The tribunals rejected this. The article included an anti-avoidance clause to prevent abuse of the relief: abuse did not necessarily involve artificial or contrived steps (see the UT at 59), but acknowledging the treaty and making arrangements so as to take the benefits it offered was not abuse. The UT referred at 79 to ‘improperly taking advantage’ of article 12: this is because taxpayers are supposed to take advantage of the Tax Treaty, or any other relief provided for them. If 12(5) removed protection just because a taxpayer was intending to rely on the treaty then the treaty would start to fall apart.

As the FTT said at [201] (and approved by the UT at [114]–[115]), for a taxpayer to fail a test like this there had to be ‘something more’ than merely taking account of the expected tax attributes of the parties to a transaction or the way the tax code is designed to tax it.

In this way, the FTT prefigured the reasoning of Falk LJ in BlackRock v HMRC [2024] EWCA Civ 330 when explaining how to approach the unallowable purpose test in CTA 2009 ss 441 and 442: this removes the ordinary entitlement of a company to take a tax deduction for its interest costs (or other loan relationship debits) where a main purpose, or one of the main purposes, of being party to the loan was to secure a tax advantage.

There, Falk LJ noted (at [150]) that as corporation tax relief is valuable, ‘it is unrealistic to suppose it will not form part of ordinary decision-making’ and 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.’ She went on to say (at [171]) that merely having regard to tax considerations when arranging borrowing is not what the unallowable purpose test is targeting: the legislation distinguishes from having appropriate, and inappropriate, regard to the tax effects of a transaction.

Purpose-based avoidance tests do not replace the tax code

The starting point, even for a transaction in securities, is that a capital receipt is taxed as a capital gain. The legislation does not aim to tax repayments of capital as if they were dividends, even if there are sufficient distributable reserves to declare a similar dividend: if the legislation intended that, this would be provided for in CTA 2010 Part 23, where distributions are defined, and where s 1000(1)B(a) specifically removes repayments of share capital. Whatever the TIS regime is doing, it is not supposed to replace or undermine this. Rather, it is, in some subjective circumstances, overriding it. The TIS includes a purpose test, and that means objectively similar transactions are taxed differently, because the purpose of the taxpayer matters. Capital receipts are taxed as capital unless the taxpayer fails the purpose test.

Similarly, the UK/Ireland tax treaty provides that Irish residents do not pay UK tax on UK source interest. That is its intended effect. There is nothing wrong with taxpayers knowing that, relying on that in designing their transactions, or for that matter seeking that out. They are supposed to do that!

In the loan relationships regime, companies are intended to get a deduction for their borrowing, and there needs to be some strong subjective evidence that they are misusing that before CTA 2009 s 441 takes that away from them. The fact that they expect that deduction and structure their transactions around it is not supposed to prevent the relief: the general legislative intention is that interest costs are deductible.

The factual difficulties are inescapable

The FTT in Osmond and Allen justified its wide reading of the purpose test (at [35]) because ‘the TIS regime is intended to be a freestanding anti-avoidance regime and … deliberately casts its net very widely’, but the constraint on that net is the requirement of a tax avoidance purpose. The TIS code prevents people from exploiting capital treatment to gain access (in some sense) to a company’s distributable profits without taking dividends: and whether that is what they are doing or not is always going to be subjective.

Attempting to get around subjective factors by defining transactions in such a way that they must have a purpose of tax avoidance misses the point of the various anti-avoidance overrides. They are not supposed to alter the underlying tax code, saying that reductions of capital should be generally taxed as if they were dividends or that the tax deduction for interest is something you can only have if you don’t really care about it. They are supposed to prevent abuse and manipulation of the underlying tax code, where taxpayers are setting things up so as to obtain a benefit that they aren’t really supposed to have.

Requiring the courts (and HMRC) to distinguish between ‘acceptable’ and ‘abusive’ attempts to rely upon a relief or upon capital treatment may make the test seem unpredictable, and dependent on the quality of the evidence. However, that is how a subjective test, overriding the normal tax treatment, is supposed to work. It targets cases where the system is, in some sense, being abused, so as to justify overriding the ordinary tax treatment.

This is why the courts have pointed out that the question is factual, depending on the particular facts of each case: in Allam, the fact that the capital transaction was simple and the alternative, chargeable to income, more complicated, was a factor (see the UT at [169]); in BlackRock the artificiality required to bring UK borrowing into the structure was relevant to deciding that the borrowing did have a tax main purpose (see [166] and [171]).

These purpose-based overrides can be contrasted with the equivalent in the SDLT legislation, FA 2003 s 75A, which deems a higher amount of SDLT to be payable when the transactions, taken together, produce a lower amount of consideration than would be paid on a ‘notional transaction’ as defined in the legislation. Unlike the TIS or unallowable purpose rules, there is no element of intention in s 75A and no requirement that the taxpayer is trying to exploit the system or avoid tax (see the Supreme Court in Project Blue v HMRC [2018] UKSC 30 at [42]).

This is important because the UK tax code is full of anti-avoidance purpose tests. The FTT in Burlington noted (at [195]) that HMRC’s wide approach to avoidance would undermine the secondary debt market: purchasers would not know whether they would pay tax on interest or not and it would become impossible to price debt. Arguments that purpose-based tests apply irrespective of whether a tax benefit is being obtained inappropriately, i.e. is being abused, move these purpose tests from being targeted at tax avoidance to making the basic structure of the legislation unpredictable.

Osmond is a surprising decision. Given that the FTT found that, subjectively, the taxpayers were not trying to avoid paying income tax, it is very hard to see how it can be correct in light of previous authority and I would expect to see it overturned.

For related reading, please see the following further articles in Tax Journal

X      Cases: H Osmond and another v HMRC (23.5.24)

X      Burlington in the UT: a clearer approach (K Rainsford, 20.6.24)

X Transactions in securities and counteraction assessments (P Miller, 14.12.22)