Taxation logo taxation mission text

Since 1927 the leading authority on tax law, practice and administration

Arctic Killer?

MIKE TRUMAN looks at some past articles in Taxation on s 660A and asks whether the hopes raised in them have been killed off by the High Court in Jones v Garnett .

IT IS DANGEROUS to comment on the decision in a case before the full judgment is released. Nevertheless, I think it is worth giving a reaction to the decision in Jones v Garnett because it is such an important case, and s 660A has been the subject of so many articles in Taxation over the past couple of years.

Two questions

There are two key questions to be asked and answered in husband and wife company settlement cases:

  • is there a settlement at all, and
  • if there is, does TA 1988, s 660A(6) prevent the income falling within the section?

One of the first articles that Taxation published on the subject was by my predecessor as editor, Malcolm Gunn (' All Very Unsettling ', Taxation 13 March 2003, page 572. His view, in summary, was that the answer to the first question was 'yes', and the answer to the second was 'no, but …'. The 'but' was that the Inland Revenue thought the answer to the second question was 'yes', and it would therefore be wise to take further precautions. Over two years later I think that he was right. Unfortunately the Inland Revenue still disagree about the answer to the second question, and they now have a presiding Commissioner and a High Court judge on their side.

Is there a settlement?

In the opinion of Miss Judith Powell, the dissenting Special Commissioner, there was no settlement in the case of Jones v Garnett . The report of the case by David Smith and Jason Piper (' Poles Apart ', Taxation 7 October 2004, p 10) summarises her view as follows:

'At the time when the share was sold, Geoff was not bound to work for the company, or pay himself less than a market salary, or declare dividends. He was not worse off at the time merely because he intended in the future to provide bounty if the opportunity arose. As a result, when Diana bought the share, it was not part of an 'arrangement' that could qualify as a settlement.'

The argument to the contrary which was put by Malcolm Gunn in his article was based mainly in case law. From the limited information currently available, it seems that Park J took the same line, specifically relying on the Court of Appeal decision in the case of Crossland v Hawkins (1961) 39 TC 493 at 502. Malcolm's article refers to the later case of Mills v CIR [1974] STC 130 , but both were decided on very similar grounds.

Crossland v Hawkins

The facts, briefly, of Hawkins were that the actor Jack Hawkins entered into a common surtax avoidance scheme in the mid-1950s. Two shares in a new company were issued to clerks in his accountant's office as subscribers, and Mr Hawkins agreed to a service contract for the next three years with the company for which he would be paid £50 per week plus expenses. Normally the aim would have been to accumulate as much income as the old apportionment rules would allow, with a view to later liquidating the company and avoiding surtax on the capital amount, just as one might do now for taper relief. This only goes to prove my favourite comment that if you wait long enough, every tax rule (and therefore every tax scheme) from the past gets recycled.

However, some months later a settlement for the benefit of Mr Hawkins' children was created by their grandfather for £100, and £98 of this was used to subscribe for 98 shares in the company. When the company paid a dividend in 1956 almost all of it went to Mr Hawkins' children. The question was whether this was a settlement of which Mr Hawkins was the settlor.

The key finding by the Court of Appeal was that a combination of transactions, including one (the creation of the settlement) which was apparently not in contemplation when the company was formed, could form an arrangement that fell within the anti-avoidance provisions, and in particular within the definition of a settlement which was almost identical to the current definition in s 660G. Mr Hawkins was a settlor of the settlement by indirectly providing funds for it — his work generated income that was indirectly diverted into the settlement.

Whilst that successfully disposes of the argument that there was no arrangement, Kevin Slevin in a later article (' Muddled Thinking ', Taxation 31 July 2003, p 477) pointed out that treating all the transactions as part of the arrangement was not a discretion that the Inland Revenue had — it was an obligation imposed by the section. This, he said, prevented the Inland Revenue from treating the payment of each dividend in isolation as the property of the settlement, since they could not ignore the previous transactions, including the transfer of the shares. This has an impact on the answer to the second question, does it fall within s 660A(6)?

Does s 660A(6) apply?

Kevin's point was that an ordinary share gives no 'right' to income at all. There is at best a future expectation that dividends will be paid. Once the arrangement is interpreted broadly enough to include the transfer of the shares rather than the declaration of each individual dividend, it follows that there is no right to income under the arrangement and s 660A(6) is not excluded by its proviso.

The contrary view, based on the Inland Revenue's interpretation, was put in an article by Martin Riley (' A Sympathetic Approach ', Taxation 20 January 2005, page 361). Whilst Martin's article is the only one we have had which answers 'yes' to both questions, it has to be said that he seems to have put the line of argument which Park J has followed.

Martin's argument was that the arrangement is simply a way of carving up Mr Jones's income earning capacity. As such, the various steps within the arrangement are merely the vehicles for achieving that end, and since it is a right to income which is being transferred then the proviso to s 660A(6) prevents the exclusion from applying.

Park J, though he may prove to have been more detailed in his analysis, seems essentially to have followed a similar theme. Far more is involved here than merely a transfer of some shares, and the arrangement as a whole cannot be characterised as an outright gift. Specifically in this case Mrs Jones subscribed £1 for her share, so there is no gift, but it appears that even if the shares had been owned by Mr Jones alone and then transferred to Mrs Jones, Park J would have held that the whole arrangement was outside 660A(6).

This is, of course, a very different argument from the one used by Dr Brice. She carefully identified the source of the income as the dividend from the shares, but then proceeded to construe this as a right to income because Mr Jones, as the sole director, could cut off the flow of dividends whenever he wanted by refusing to declare one. This argument was comprehensively deconstructed by Peter Vaines ('Arctic Systems', Taxation 28 October 2004, page 90), and it is understandable why this line was not followed by Park J. However, there was a reason why Dr Brice took that line, and it seems to highlight a problem in Park J's argument, at least in the abbreviated form we currently have it reported.

Not a charging section

Park J is reported as saying that 'section 660A(6) did not take the case out of the charging scope of section 660A(1)'. But s 660A(1) is not a charging section at all, it is a deeming provision. It merely says that income from a settlement where the settlor retains an interest (as defined) 'shall be treated for all purposes of the Income Tax Acts as the income of the settlor'. Any charge to tax must be imposed by another section. If I were to set up a trust in which I retained an interest, and which gave the trustee a power to place bets with the trust capital, and if I were then to place a bet as trustee on a complete outsider winning the world snooker championships at 100-1, then in the highly unlikely event(!) of my bet proving successful, s 660A(1) would deem the income arising under the settlement to be my income for tax purposes. However, since betting winnings are only taxable in the hands of professional gamblers (who do not bet on 100-1 outsiders) there would still be no tax liability, because there is no head of charge to tax it under.

On the argument put forward by Martin Riley, which matches the view put forward in its guidance by the Inland Revenue, it is the earning capacity of Mr Jones that has been settled. But earning capacity is not taxable, only earnings are. Taking the argument to its logical extreme, the company is merely a part of the settlement 'arrangement' in which both Mr and Mrs Jones retain an interest, and it should therefore be ignored as part of the 'legal machinery' (an expression which comes from Crossland v Hawkins ) meaning that Mr Jones should have been taxed as a sole trader on all his earnings.

Not even the Revenue were arguing for such a result, and so the argument put to the Special Commissioners, and to which Dr Brice responded, was that the property which gave rise to the income was indeed the shares, and the dividends were taxable because of the particular circumstances in which they were payable in this case.

That now appears, at least in part, to have been overtaken by Park J, who wants to apply a more general approach. The transactions taken together amount to a settlement, the total package cannot be seen as an outright gift, and therefore a liability can arise.

Back to the legislation

There is no substitute here for a careful analysis of what the legislation says. I accept that there is an arrangement that constitutes a settlement under 660A(1). I think I probably also accept that the arrangement is more than just an outright gift. But let us assume for a moment (to make my case stronger) that Mrs Jones had not subscribed for her share, she had instead been given it by Mr Jones. Would Park J's argument still succeed in saying that the arrangement is wider than the gift and therefore the income is chargeable?

In my view it does not. You have to identify the income which you are charging to tax by means of a charging section. The only income that can be so identified is the dividend income arising from the shares.

Crucially, s 660A(6) does not operate on an all or nothing basis. It says that the reference to a settlement in 660A (1) does not include an outright gift by one spouse to another of property from which income arises, assuming the provisos do not come into play. Whatever else is not within s 660A(6), surely the ordinary shares are, as they are far more than just a right to income? You cannot simply say that the arrangement overall is more than an outright gift, because the overall arrangement does not disclose a chargeable head of tax. It may well be that Mr Jones' earning capacity is the income of the settlement for tax purposes; that still doesn't make it taxable. But as soon as you isolate the transfer of the shares as an individual element within that arrangement which does give rise to a tax liability, you crystalise the property which is excluded by virtue of s 660A(6).

Conduct of the case

We have also covered the conduct of this case in several articles in the past, most notably recently in Anne Redston's ' The Settlement Saga ' ( Taxation, 25 November 2004, page 201). The main problem for the PCG in fighting this case has not so much been the fees of those taking the case for the Joneses, who are all working well under their normal rates, but the refusal of the Inland Revenue to accept that this is a major test case on which they should at least have given an assurance that they would not seek their own costs. As soon as the case ended, the Inland Revenue applied for and were awarded their costs. It is this, more than anything else, which makes the prospect of an appeal to the Court of Appeal so daunting, and at the time of writing the taxpayers have yet to decide whether to do so. Whatever the need for funds turns out to be, whether to meet the existing costs or to fund a further appeal, we would urge readers to support any fundraising requests, and Taxation will gladly play its part in publicising them.

It is all the more surprising, then, that Park J apparently sees this case as a 'simple application of well-established principles'. It is not, and they are not, otherwise the case would not have aroused so much controversy, with most commentators coming down in favour of the taxpayer rather than the Inland Revenue. If it is simple, why did so many experienced and knowledgeable tax practitioners disagree?

The numbers game

Finally, Park J rightly points out that it will not affect all husband and wife companies, because those where a market salary (whatever that is) is taken by the fee-earner would probably not be caught. Despite the arguments that can be based on such cases as reductio ad absurdum , it is highly unlikely that practitioners were worried about such clients. They were worried about the ones where they had given spouses who did very little work a half share in the company, because the Inland Revenue's practice seemed to be that this was acceptable — by their own admission there were very few challenges.

In this context it is highly irritating to find that during the case the Inland Revenue apparently again repeated the claim that there are at most 30,000 cases affected. We published a clear and unambiguous explanation of why the Inland Revenue's figures were wrong at the end of my comment piece in Taxation 25 November 2004, page 198 — the databases they are using cannot capture dividend data from companies that only file a modified balance sheet. HMRC, as it is now, can legitimately dispute the estimates put out by the profession, but it is disingenuous to use the 30,000 estimate when the methodology has been shown to be completely wrong. I think it's time that they publicly withdrew the estimate, and I will be calling on them to do so, or to give a clear and corroborated justification for continuing it.

back to top icon