Astall & Anor v Revenue & Customs

These are appeals by Mr Astall (“Mr Astall”) and Mr Edwards (“Mr Edwards”) (collectively “the Appellants”) from a decision released on 14th August 2007 of the Special Commissioner (Dr Avery Jones).

He dismissed the Appellants’ appeals against the Revenue’s amendment of their Self Assessment returns for the year of assessment 2001/2002 whereby the Revenue refused the Appellants’ claim to have sustained a loss in that year from the discount on a relevant discounted security for the purposes of the Finance Act 1996 Schedule 13 (“Schedule 13”). He consequently rejected a claim to be entitled to relief from income tax on the amount of income for that year equal to the amount of the loss.

It was common ground before the Special Commissioner (as set out in paragraph 10 of the decision) that each of the Appellants sustained a “loss” within the meaning of Schedule 13 paragraph 2 the amount of Mr Astall’s loss being £1,989,464 and the amount of Mr Edwards’ loss being £4,976,098.

It was also common ground that each of the Appellants sustained that loss from the discount on a “Security” within the meaning of Schedule 13.

The sole issue before the Special Commissioner was whether that security was a “relevant discounted security” as defined by Schedule 13 paragraph 3.

The Special Commissioner decided that the security was not a “relevant discounted security” as defined. The Appellants submitted that his decision was wrong in law.

On 28th January 2002 Mr Edwards paid £7,000 to establish a Settlement. One of the Trustees of the Settlement was a company called Linkfast Industry Ltd (“the Issuer”). On 31st January 2002 (“the Issue Date”) the Issuer acting as a Trustee of the Settlement issued to Mr Edwards a loan note instrument constituting £6,238,366 zero coupon loan notes due 31st January 2017 (“the Security”).

The issue price in respect of the Security was £5,278,276 which Mr Edwards duly paid to the Issuer on the Issue Date. He financed that purchase partly by a loan from Kleinwort Benson Private Bank of £3,278,276 at interest of 2% over base rate, and the balance out of his own resources.

By Schedule 2 Condition 3.1 unless previously redeemed the Security was to be redeemed on the Final Redemption Date which was 15 years after the Issue Date. However by Schedule 2 Condition 3.9 upon the occurrence of the Transfer Event the Final Redemption Date would become 65 years after the Issue Date. The amount payable on redemption on the Final Redemption Date (whether in 15 or 65 years) was an amount equal to the principal sum of the Security £6,228,366. However if a Transfer Event occurred the Security (which carried no interest) would fall substantially in value.

By Schedule 2 Condition 3.2 a window was given between one and two months after the Issue Date but in any event before the occurrence of the Transfer Event whereby the note holder was entitled to redeem the Security. By condition 3.3 the amount payable on redemption was 100.1/118 of the Principal amount of the note namely £5,283,554. In order to redeem a note holder would have to give at least 7 days notice. The latest date for the giving of early notice of redemption by Mr Edwards was 24th March 2002. However in the event of early redemption after the Transfer Event condition 3.10 provided that the transferee of the Security was entitled to redeem upon giving notice to the Issuer and by condition 3.11 the amount payable on such early redemption was the open market value of the Security or if greater 5% of the principal amount of the Security namely £311,418. The Third Party could elect for the amount payable to be simply the 5%.

Therefore after a Transfer Event the redemption price was reduced by 95% of the original principal advanced.

A Transfer Event was defined in Schedule 2 condition 1 as occurring in the following circumstances:-

The Security was issued to Mr Edwards as part of an admitted tax avoidance scheme: his sole purpose for subscribing for the Security was to sustain an income tax loss in 2001-2002 by selling it before 6th April 2002 to a Third Party after it had fallen in value that is following the occurrence of the Market Change and the service of a Transfer Notice.

On 4th February 2002 KPMG acting on behalf of Mr Edwards and 63 others met Mr Stuart Gower of Hambros to discuss whether Hambros would buy the Security. Negotiations then ensued for some weeks thereafter. On 5th February 2002 KPMG approached 6 other banks by telephone to see if they were interested. A Market Change occurred on 28th February 2002. On the same day Hambros sought approval for the purchase from their head office in Paris and written approval was given on 7th March 2002. Hambros were not definite about purchasing until then. On the same day Hambros was given details about Mr Edwards and his Security and made a non binding offer in principle to acquire the Security. The next day he signed a Transfer Notice proposing to transfer the Security to Hambros. On 18th March 2002 the Security was sold to Hambros by Mr Edwards for £302,545 (decision paragraph 4 (31) CP Appellants’ skeleton paragraph 21 (viii)). Seven days later on 25th March 2002 Hambros redeemed the Security for £311,418. Thus Hambros made a profit on buying and redeeming the Security of £8,873.

This left £4,966,858 (plus any accrued interest and less any costs) in the Edwards Trust. In August 2002 Mr Edwards requested from the Trustees an interest free loan of £3,280,000. The Trustees resolved to make that loan on 9th August 2002 and he received the money 3 days later which was then used to repay his Kleinwort Benson loan (£3,284,742.73). On 21st July 2003 the Trust purchased the residential property Mill Waters in Buckinghamshire.

His position was the same as Mr Edwards. The Kleinwort Benson facility in his favour was dated 21st January 2002 which enabled a loan of £1,477,000 to be paid subject to an arrangement fee of £7,385. The loan note in his case was issued on 25th January 2002 with a redemption date of 25th January 2017 in the sum of £2,489,851. A Market Change took place at 3pm on 25th February 2002. He gave a notice of intention to transfer the Security on 8th March 2002 and it was duly sold on 18th March 2002 to Hambros for £120,946 and 7 days later it redeemed the Security for £124,493 (profit £3,547).

That left £1,985,550 plus accrued interest less any costs in the trust. He repaid his loan to Kleinwort Benson early in the amount of £1,477,000 and interest of £17,966.79 was debited to his account with Coutts & Co on 9th April 2002 who took a charge over the Astall Trust assets in consideration of Coutts & Co giving credit facilities to Mr and Mrs Astall. Coutts & Co’s charge was limited to £1,500,000.

As the Appellants freely admitted the transactions took place solely as a tax avoidance scheme promoted by KPMG. They targeted individuals with incomes of at least £1,000,000. A total of 64 people having such an income subscribed to the schemes and claimed to submit total losses of about £156,000,000 to offset against other income which they have earned during that tax year.

In addition to the banker’s fees KPMG’s fees were initially 1% of the estimated amount of his income (plus VAT) on signing the engagement letter, a further 1.5% (plus VAT) payable on the sale of the Security and a success fee of 7.5% (plus VAT). If the scheme were successful in relation to income of £2,000,000 (Mr Astall’s position for example) he would gain £545,262 after taking fees of £239,288 into account (that includes fees for establishing a Trust and administration). If he were unsuccessful he would lose £72,708 having taken fees of £66,038 into account. If the scheme failed because of the Market Change conditions or a failure to find a purchaser he would lose £29,632 (this figure includes tax on profit on redemption and winding up the trust). Mr Astall was prepared to lose such a sum since he could lose that amount in an afternoon on movements in the share price of Celestica Inc a company listed on the New York and Toronto Stock Exchanges of which he was a member of Senior Management.

The relevant statutory provisions are set out in Schedule 13 of the Finance Act 1996 as follows:-

“The statutory provisions relevant to this appeal in Schedule 13 to the Finance Act 1996 as they were in 2001-02 are:”

Charge to tax on realised profit comprised in discount

1—(1) Where a person realises the profit from the discount on a relevant discounted security, he shall be charged to income tax on that profit under Case III of Schedule D or, where the profit arises from a security out of the United Kingdom, under Case IV of that Schedule.

(2)

For the purposes of this Schedule a person realises the profit from the discount on a relevant discounted security where—

(a)

he transfers such a security or becomes entitled, as the person holding the security, to any payment on its redemption; and

(b)

the amount payable on the transfer or redemption exceeds the amount paid by that person in respect of his acquisition of the security.

(3)

For the purposes of this Schedule the profit shall be taken—

(a)

to be equal to the amount of the excess reduced by the amount of any relevant costs; and

(b)

to arise, for the purposes of income tax, in the year of assessment in which the transfer or redemption takes place.

(4)

In this paragraph “relevant costs”, in relation to a security that is transferred or redeemed, are all the following costs—

(a)

the costs incurred in connection with the acquisition of the security by the person making the transfer or, as the case may be, the person entitled to a payment on the redemption; and

(b)

the costs incurred by that person, in connection with the transfer or redemption of the security;

and for the purposes of this Schedule costs falling within paragraph (a) above shall not be regarded as amounts paid in respect of the acquisition of a security.

Realised losses on discounted securities

2—(1) Subject to the following provisions of this Schedule, where—

(a)

a person sustains a loss in any year of assessment from the discount on a relevant discounted security, and

(b)

makes a claim for the purposes of this paragraph before the end of twelve months from the 31st January next following that year of assessment,

that person shall be entitled to relief from income tax on an amount of the claimant's income for that year equal to the amount of the loss.

(2)

For the purposes of this Schedule a person sustains a loss from the discount on a relevant discounted security where—

(a)

he transfers such a security or becomes entitled, as the person holding the security, to any payment on its redemption; and

(b)

the amount paid by that person in respect of his acquisition of the security exceeds the amount payable on the transfer or redemption.

(3)

For the purposes of this Schedule the loss shall be taken—

(a)

to be equal to the amount of the excess increased by the amount of any relevant costs; and

(b)

to be sustained for the purposes of this Schedule in the year of assessment in which the transfer or redemption takes place.

(4)

Sub-paragraph (4) of paragraph 1 above applies for the purposes of this paragraph as it applies for the purposes of that paragraph.

Meaning of “relevant discounted security”

3—(1) Subject to the following provisions of this paragraph and paragraph 14(1) below, in this Schedule “relevant discounted security” means any security which (whenever issued) is such that, taking the security as at the time of its issue, the amount payable on redemption—

(a)

on maturity, or

(b)

in the case of a security of which there may be a redemption before maturity, on at least one of the occasions on which it may be redeemed,

is or would be an amount involving a deep gain, or might be an amount which would involve a deep gain.

(1A) The occasions that are to be taken into account for the purpose of determining whether a security is a relevant discounted security by virtue of sub-paragraph (1)(b) above shall not include any of the following occasions on which it may be redeemed, that is to say—

(a)

any occasion not falling within sub-paragraph (1C) below on which there may be a redemption otherwise than at the option of the person who holds the security;

(b)

in a case where a redemption may occur as a result of the exercise of an option that is exercisable—

(i)

only on the occurrence of an event adversely affecting the holder, or

(ii)

only on the occurrence of a default by any person,

any occasion on which that option is unlikely (judged as at the time of the security's issue) to be exercisable;

but nothing in this sub-paragraph shall require an occasion on which a security may be redeemed to be disregarded by reason only that it is or may be an occasion that coincides with an occasion mentioned in this sub-paragraph.

(1B) In sub-paragraph (1A) above “event adversely affecting the holder”, in relation to a security, means an event which (judged as at the time of the security's issue) is such that, if it occurred and there were no provision for redemption, the interests of the person holding the security at the time of the event would be likely to be adversely affected.

(1C) An occasion on which there may be a redemption of a security falls within this sub-paragraph if—

(a)

the security is a security issued to a person connected with the issuer; or

(b)

the obtaining of a tax advantage by any person is the main benefit, or one of the main benefits, that might have been expected to accrue from the provision in accordance with which it may be redeemed on that occasion.

(1D) In sub-paragraph (1C) above “tax advantage” has the meaning given by section 709(1) of the Taxes Act 1988.

…

(3)

For the purposes of this Schedule the amount payable on redemption of a security involves a deep gain if—

(a)

the issue price is less than the amount so payable; and

(b)

the amount by which it is less represents more than the relevant percentage of the amount so payable.

(4)

In this paragraph “the relevant percentage”, in relation to the amount payable on redemption of a security, means—

(a)

the percentage figure equal, in a case where the period between the date of issue and the date of redemption is less than thirty years, to one half of the number of years between those dates; and

(b)

in any other case, 15 per cent.;

and for the purposes of this paragraph the fraction of a year to be used for the purposes of paragraph (a) above in a case where the period mentioned in that paragraph is not a number of complete years shall be calculated by treating each complete month, and any remaining part of a month, in that period as one twelfth of a year”.

The only issue was whether or not the Security was a “Relevant Discounted Security” the issue being whether on redemption of the Security either on maturity or redemption before maturity the amount payable is or would be an amount involving a deep gain or might be an amount which would involve a deep gain.

The powers on an appeal are not unfettered. It is limited to a review and can only be allowed if the Special Commissioner’s decision is either wrong or there is some other reason as to why the appeal should be allowed. It is limited to questions of law save questions of fact as determined by Edwards v Bairstow [1956] AC 14 which requires an Appellant to show that the Special Commissioner acted without any evidence or on any view of the facts could not be reasonably entertained.

After a hearing which covered 3 days the Special Commissioner delivered a careful decision which extended to 31 pages. There is one error in paragraph 6 (14) of the determination. He wrongly stated that Mr Astall and Mr Edwards had until 5th April 2002 (the end of the tax year) to find another purchaser if Hambros failed to find one. In fact he failed to take into account the Appellants would have had to give a notice to redeem their Security under condition 3.2 so that Mr Edwards would have had to serve a notice of redemption by 24th March 2002 at the latest and Mr Astall by 18th March 2002. Neither party drew this error to the attention of the Special Commissioner when he released his determination in draft. That is unfortunate. However in my view I do not think it makes any difference. When one analyses the judgment as a whole he would have come to the same conclusion in my opinion had the correct shortened days been identified to him as opposed to the year ending 5th April 2002. Further if the parties wished to have achieved a redemption by 5th April 2002 and less than 7 days was available I have no doubt given the compliant nature of the relationship between all the parties a waiver would have been obtained to abridge time so as to achieve an effective redemption by the shorter date if required.

Upon review of the Special Commissioner’s decision for the reasons I have set out in this judgment and on the basis of his judgment I conclude the only possible result was that a new purchaser would have been found in that shortened period or the notice period would have been abridged so as to achieve an effective redemption before 5th April 2002.

The Special Commissioner concluded that the security was not a “Relevant Discounted Security” because there would never be a deep gain (as required by paragraph 3 (1)) as defined in paragraph 15 (1) of Schedule 13. The whole purpose of the regime was to charge a deep gain on a Relevant Discounted Security and therefore give effect the other way to any losses that might accrue on any “Relevant Discounted Security”. It was summarised in paragraph 22 of his judgment as follows:-

“In my view, the Ramsay approach as explained in Barclays Mercantile entitles and requires me to construe legislation aimed at considering all possibilities in such a way as to limit those possibilities to real ones. A purposive construction of the definition of relevant discounted security must have regard to real possibilities of redemption, not ones written into the document creating the Security that the parties know, and any reasonable person having the knowledge available to the parties knows, will never occur. Mr Prosser’s argument that a purchaser of a security must be able to determine from its terms whether it is a relevant discounted security carries no weight in these circumstances. There will only ever be one purchaser of it who is fully aware of the scheme, who will redeem it within seven days so that the Security will be outstanding for a maximum of two months. The purpose of the legislation is to tax gains on securities that are issued at a deep discount and conversely to relieve losses on such securities. The difference between the issue price and the redemption price must give rise to a possibility of making a gain that can be objectively seen to exist. This Security never had this possibility; it is a practical certainty that there will be a loss of 94. To decide otherwise would be to return “tax law [to] being ‘some island of literal interpretation’”

This is echoed in paragraph 25 of the Respondents’ skeleton argument as follows:-

“Applying these principles, the Court must identify the purpose of the RDS definition in the context of the statutory RDS provisions as a whole, and construe the definition accordingly. It was clearly intended that a deep gain on a security should fall within the charge to income tax, and losses should be allowable where the security fails to make the taxable deep gain but instead makes a loss. Put differently, there is no discernable intention in the legislation that a taxpayer should be enabled to manufacture allowable losses by the expedient of inserting mechanisms giving the theoretical possibility of deep gains (but in circumstances where the parties intended that instead far deeper losses should be made).

It would be artificial and incorrect on the above principles to limit the investigation to a consideration of whether there is a bare possibility on the face of the security that there “is or would or might be” a deep gain. Importantly, para 3(1) of Sch 13 says “taking the security as at the time of its issue”. That must mean taking the security (i.e. the bundle of rights and obligations as created by the taxpayers as part of the scheme, and evidenced in the security document) as a whole, not single provisions out of context”.

After setting out the documentation and the evidence he then set out his findings in paragraph 6 as follows:-

“6.

I find the following further facts in relation to the actions for finding a purchaser for the Securities:

(1)

KPMG deliberately did not seek a purchaser for the Security until after it was issued. It was originally intended that they would not do so until two months after the issue of the Security with the result of the Market Change condition being known one month after issue, but because of the shortness of time KPMG approached banks after the issue of the Securities.

(2)

The decision of KPMG not to obtain a purchaser for the securities until after their issue was (as is conceded) purely an intended anti-Ramsay device: ie the only purpose of inserting it was in the hope of repelling an anticipated argument by the Revenue based on WT Ramsay v IRC [1980] STC 300. The Appellants relied on KPMG for this decision and understood that it had been inserted by KPMG as part of the scheme, were duly warned about the risk which the two-part fee arrangement (see paragraph (…) below) takes into account, and they did not ask for this uncertainty to be reduced or avoided by approaching prospective purchasers at an early stage. They were content to rely on KPMG in finding a purchaser.

(3)

Mr Stuart Gower of Hambros was first approached by telephone on 30 or 31 January 2002 by Mr Mark Patterson of KPMG not giving any details of the scheme and requesting that he brought a lawyer to a meeting on the following Monday 4 February 2002. (There had been a previous approach to Mr Gower by Mr Kilshaw of KPMG trying to interest Hambros’ clients in the scheme, also without giving any details, on 10 September 2001 which Mr Gower referred to his London office but no further action was taken. No mention of a possible involvement of Hambros as purchaser of the Securities was made then.) Mr Gower duly attended the meeting accompanied by a lawyer from Simmons & Simmons. He signed a confidentiality letter and was told that by 4 February 2002 10 clients had Securities where the Market Change condition was satisfied, and for the remainder of the 64 cases it would be determined whether it was satisfied by mid to late February 2002. He was sent a confidential information memorandum of 7 February 2002 with sample documentation. Mr Gower expressed interest and was asked to quote his level of fees. In a telephone conversation with Mr Patterson on 13 February 2002 he was informed about the other banks’ quotes (see paragraph 6(11) below) and asked if he could reduce his quote. Mr Gower suggested £220,000 to £250,000. On 19 February 2002 he finally quoted a discount of £235,066 (2.85%) based on a total principal amount of the 64 securities of £164,958,530, 5% of which is £8,247,927. (The discount figures quoted below for other banks are all based on an issue price of £129m, so that the redemption price would be £152,220,000 and on that basis Hambros’ discount would be £216,914.) Further documentation was provided to Hambros on 11, 18 and 20 February 2002.

(4)

Hambros relied on KPMG for “know your client” money laundering checks. They performed their due diligence on the documents in a data room at KPMG’s offices from 26 February 2002.

(5)

Hambros made an offer subject to due diligence for the first batch of 37 cases, which did not include the Appellants, on 25 February 2002, stating that a number of formalities, including due diligence, needed to be completed and so they reserved the right not to proceed at any time before the contract was completed. Due diligence of these cases was completed on 28 February 2002.

(6)

Hambros emailed their banking division in Jersey and London for approval of the first 37 transactions on 26 February 2002 saying that one transaction was over their unsecured lending limit (the purchase of the Securities being treated as a loan for internal risk purposes) and was being referred to their head office in Paris. The credit application included the following:

“Credit risk

The underlying assets supporting the loan note consist entirely of cash in the same currency. As such there is no exposure to market movements of any nature. The cash is held with Lloyds and as such there is a risk upon this institution however this is considered entirely acceptable. The cash balances held will be directly confirmed by the trustees. Furthermore as the Bank will only be entitled to claim 5% of the principal value of the Loan Note there will be 20 times cover held by the Trust in respect of the loan note. The client is presently borrowing against the cash deposit the sum of [£3,278,276 for Mr Edwards; £1,477,000 for Mr Astall] and our margin cover is therefore reduced to a level of [13 times for Mr Edwards; 14 times for Mr Astall] as opposed to 20 times.

Performance risk

We rely upon KBTL [Kleinwort Benson Trustees Limited] in their capacity as the directors of the corporate trustees to meet their obligations under the terms of the loan note. KBTL are liable under the terms of the loan notes and hence have a corresponding right over the funds held by the trust. In view of the fact that KBTL control the funds (they are the sole signatories of the Bank account and the quality of them as the Trust arm of a major international bank, we have no concerns as to their fulfilling their obligations.”

The 20 times cover (before taking the borrowings into account) is slightly exaggerated because the 5% is of the redemption value, which is 118% of the issue price, and not of the issue price as suggested here, but they were still well covered. I suspect that the reduction to 13 and 14 times covered should have been a reduction by these figures. Hambros could rely on KBTL being in control of the funds because a trustee could under the terms of the trust be removed only on 14 days notice, which the Appellants warranted in the contract for the sale of the Security had not occurred. The Appellants signed a statement to KBTL that they would not object to the application of the trust funds to meet the redemption, and that they would authorise the payment as co-signatory of the bank account (the signatories were KBTL and either the settlor or his wife). The credit application also said there were no problems in relation to documentation and structural risks, tax risk, or reputational risks.

(7)

Hambros first approached their head office in Paris on 28 February 2002 explaining that they did not have all the facts available until 26 February 2002. Further information was provided on 1 March and verbal approval was obtained on 5 March 2002 with written confirmation of this approval being provided on 7 March 2002. Mr Gower said (and I accept) that if that case has been turned down by his head office he would not have gone ahead with any of the others. The approach to the head office was not known to KPMG, who were not told of the delays in obtaining approval.

(8)

Hambros received details of the second batch of 22 cases, including the Appellants, on 7 March 2002 and on the same day made a non-binding offer in principle to purchase them. I notice from paragraph (…) above that Mr Astall was told on 4 March 2002 that Hambros had expressed an interest in buying the Security, which was before they had made the offer, and he was asked to sign a transfer notice and let KPMG have it by noon on 7 March 2002.

(9)

Hambros were given details of the identities of the clients in the third tranche on 14 March 2002 and made an offer in principle to purchase them on 15 March 2002.

(10)

In all cases Hambros elected to redeem the Security at 5% of the redemption value instead of relying on the condition enabling this to be at market value (see condition 3.13 in paragraph above).

(11)

KPMG also approached a number of other banks (none being UK High Street banks) on 5 February 2002 by telephone to see if they were interested in purchasing the securities. This was followed by a confidentiality letter which the bank was to sign and return by fax. A Confidential Information Memorandum was then sent and the bank was asked to make an indicative offer by 11 February 2002 (extended to 12 February 2002). If the offer was satisfactory KPMG said they would send a set of documents for approval by the bank’s advisers. The reactions of the six banks approached, which I have anonymised to protect their confidentiality since they did not take part in the transaction, were:

(a)

Bank A (UK: the country stated is the office dealt with, not necessarily the head office or parent company) were not interested.

(b)

Bank B (Guernsey) quoted a total purchase price of £6,375,000 on 12 February 2002 which a note of a telephone conversation on 11 February 2002 shows was mistakenly based on 5% of the issue price and on that basis represents a discount of 1.2%. They also quoted 5 to 10 days to obtain credit committee approval. Mr Patterson said in evidence that he thought that the discount was too low and suspected that they had not understood the proposal properly, which is borne out by their miscalculation. He turned them down on 13 February 2002 on the basis that the 5 to 10 days was too long. They replied saying that such transactions needed head office approval and they should talk about mechanics if a future opportunity arose, which demonstrates that they were interested in such business in principle.

(c)

Bank C (not clear which office) said it was outside the scope of their normal activities.

(d)

Bank D (Isle of Man) quoted a discount of £370,000 (4.8%) based on an issue price of £130m on 11 February 2002 which they reduced to £300,000 (3.9%) on 12 February. KPMG must have tried for a further reduction as they declined to match the price quoted by KPMG and pulled out of the transaction on 13 February 2002. Mr Patterson suspected that they would have needed a lot of approvals and would have not managed to satisfy the time limit.

(e)

Bank E (Jersey) quoted a discount of £115,050 (1.5%) based on an issue price of on 12 February 2002 but wanted a secured instrument, saying in a telephone conversation on 13 February 2002 that if it was unsecured the timescale was too challenging.

(f)

Bank F (UK) quoted a discount of £967,500 (15%) on 11 February 2002 which KPMG turned down as being too high.

(g)

I also record that Mr Astall said (and I accept) that Coutts, who were advising him, told him, without being prompted, that they would not be interested in purchasing the securities. They were never approached by KPMG.

(12)

In summary, KPMG treated Hambros as the favoured purchaser by inviting Mr Gower to a meeting, which they did not do for the other banks. Four of the other banks were interested but were turned down by about 13 February 2002, one on the basis of their quoted time scale but this was the excuse rather than the reason. Hambros were not definite about purchasing until after obtaining their head office approval on 7 March 2002 and even then their offer was not legally binding.

(13)

The question for me is the likelihood at the date of issue of the Security (25 January 2002 for Mr Astall; and 31 January 2002 for Mr Edwards) of a purchaser being found for the Security within the time limit which was ideally before budget day and by 5 April 2002 at the latest. Budget day was originally expected to be in the second week in March but it was announced on 28 January 2002 that it would be on 17 April 2002; accordingly by 25 January 2002 the time limit was then known to be 5 April 2002. The two expert witnesses gave evidence about the likelihood, based on different instructions. Mr Milligan was asked for an opinion as at the time of issue; and Mr Cunnell on the assumption that the Market Change condition had been satisfied. Mr Milligan concluded that it would be very challenging within the proposed timescale, and Mr Cunnell concluded that Hambros would have been a willing buyer. They met with a view to agreeing a joint statement but despite extensive discussions there was too much difference in the questions put to each of them to be able to agree a joint statement. I am grateful to them for trying but it seems to me that as Mr Milligan could not make the assumption that the Market Change condition had been satisfied before any sale was to take place his approach was more complicated than it need have been. For example, he considered that a purchaser would want to include an analysis of the pricing of the security before satisfaction of the condition, which he described as “non-trivial” which I can well understand. Mr Cunnell, who had, in my view, more realistic instructions to consider the likelihood of a purchaser being found on the assumption that the Market Change condition had been satisfied, had less relevant experience as he had worked almost entirely with Midland Bank and its successor HSBC, which were not the type of banks that would have considered purchasing the Securities. He was essentially commenting that he did not find the actual dealings by Hambros surprising, which Mr Milligan criticised as being based on hindsight.

(14)

I am left on my own to draw a conclusion from the facts found. My view is considerably influenced by (a) Hambros’ credit application from which I have quoted, in effect taking a realistic view that there were no risks, (b) that four out of the six banks came up with a price at which they would have been prepared in principle to buy the securities, three of which represented a discount of well under 10%, and (c) Hambros obtained approval from their Head Office in Paris in 7 days, which is within the 5 to 10 days quoted by Bank B. Hambros’ view about the risk is what I would have expected of a bank prepared to do this type of business, which I accept would not have included a High Street Bank. They are paying a discount based on a price of 5 when 85 was sitting in a bank account controlled by KBTL as sole director of a newly-formed trustee company (although in the case of both the Appellants the Security was charged to secure some borrowings that ranked ahead of Hambros), in circumstances where all parties were knowingly involved in a tax-avoidance scheme and were not likely to take actions that would prejudice its success. They were essentially buying cash at a discount. KPMG had given figures to the Appellants based on a discount of 10% on the basis of which they went ahead with the scheme and so a discount of this amount was acceptable to them. Even if Hambros had been refused permission by their head office on 7 March, KPMG still had until 5 April to find another purchaser. It had taken Hambros from 4 February to 7 March, which does not suggest that it would be impossible to do so within just under a month.

(15)

My conclusion is that while the risk of not finding a purchaser existed and the Appellants were warned about it, it was so small as to be a practical certainty that at the time of issue of the securities (and assuming that the Market Change condition had been satisfied when the purchaser was considering the purchase) KPMG would succeed in finding purchasers willing and able to purchase the securities within the time scale at a discount of not more than about 10%, and that this was known to KPMG”.

As the Bairstow case determines, those findings can only be challenged if there was either no evidence to support them nor any reasonable Special Commissioner could have come to the conclusion expressed. I have already referred to and disregarded his error as to the need to find a purchaser by 5th April 2002. His key conclusion in his judgment in my view is paragraph 6 (15). The reality he found (and I do not see how this is open to challenge) is that the risk of not finding a purchaser was insignificantly small. I accept he used different expressions at different parts in his decision but the language used (paragraph 6 (15) is a good example) shows that he concluded that a purchaser would be found and that whilst there is always a possibility of a purchaser not being found it was so unlikely that it would be discounted by the Appellants. It will be seen that sub paragraph (15) decided that on the basis that the Market Change conditions had been satisfied. The attractiveness to the banks is summarised in paragraph 6 (14). They were essentially buying cash at a discount. The Appellants of course were not unhappy about that because they wished a buyer to be found who would purchase the Security at the low price created by the structure of the arrangements so as to create a loss. The Appellants of course wished to use the loss to set off against their other tax liabilities.

In paragraph 7 he referred to further uncertainties that affected Mr Astall and made findings there. In paragraph 7 (3) he calculated that Mr Astall would lose £72,708 if the scheme were unsuccessful against a “gain” of a loss of £545,262 to set against his other income of £2,000,000 (after taking into account the fees of £239,288 payable to KPMG and for setting up the Trust and for administration). If the scheme failed because of the Market Change conditions or the failure to find a purchaser he would lose £29,632. This figure included tax on the profit on redemption of winding up the Trust. Crucially the Special Commissioner found “he was prepared to lose such a sum since he could lose that amount in an afternoon on movements in the Celestica share price. He saw this as an acceptable risk and he was very used to taking risks….”

Next he analysed the submissions of the competing parties and then gave reasons for his decision (paragraph 13 and following). He considered the reasoning in 3 areas (a) the Market Change condition, (b) the decision not to seek purchasers until after the issue of the Security, (c) the terms of the Security including in particular the terms for redemption.

Here he applied what he called the Scottish Provident principle (from IRC v Scottish Provident [2005] STC 15). In respect of the Market Change Condition he considered the pessimistic 15% chance that it would not be satisfied against the more optimistic analysis namely 85% chance it would be satisfied. In paragraph 16 he concluded (correctly) that the condition was inserted purely as an anti-Ramsay device. The dollar/sterling exchange rate was irrelevant to the Security which was in sterling. The 85% chance of the conditions being satisfied was favourable enough to make the risk one which the Appellants were willing to accept. Accordingly he concluded that the Market Change Condition was not inserted for any commercial reason whilst he acknowledged that it could be said that the 15% chance of it not being satisfied created a real commercial risk. However the odds were favourable enough to make it a risk the parties were willing to accept in interest of the scheme. Thus on the basis of Scottish Provident it could be ignored and the conditions of it not being satisfied and the possibility of the early redemption option being exercised would also be ignored. In the alternative in paragraph 17 he determined that if he was wrong about the Scottish Provident principle then he would apply the reasoning that he applied to Mr Astall’s case.

I understand that to mean that if he was wrong on the Scottish Provident points all of the cases will fall for the same supplemental reason that he attributed to the failure of Mr Astall’s case which I shall deal with further in this judgment.

In essence his rejection of this being relevant is because of his factual finding that it was a practical certainty at the time of issue of the Securities KPMG would succeed in finding purchasers willing and able to purchase the Securities within the timescale at the discount of not more than 10%. Once again he expressed the view that the odds were favourable enough to make a risk which the parties were willing to accept and he therefore concluded that the possibility of a purchaser not being found was something he could ignore.

Here after reference to Scottish Provident, Ramsay and Barclays Mercantile Business Finance Ltd v Mawson [2005] STC 1 he reminded himself in his review of the authorities of the fact that the Ramsay case did not introduce a new doctrine operating within the special field of revenue statutes and that it was important to avoid sweeping generalisations about disregarding transactions undertaken for a purpose of tax avoidance (referring to MacNiven (Inspector of Taxes) v Westmoreland Investments Ltd [2001] STC 237, [2003] 1 AC 311.

After that review he turned to review the facts of the transactions realistically and his appreciation of the facts was summarised in paragraph 21 as follows:-

“Accordingly I turn to viewing the facts of the transaction realistically. My appreciation of the facts is as follows:

The scheme is entirely artificial and the Appellants had no commercial purposes in entering into it other than generating an artificial loss to set against taxable income.

The terms of the Security were, in the words of the KPMG memorandum given to clients interested in the scheme, “structured so that it falls within the definition of a relevant discounted security for tax purposes.” The premiums of 0.1% under the early redemption option, and 18% on final redemption after either 15 or 65 years, and in particular the same premium for both periods, do not reflect a market return for a zero-coupon security being the last two being about 1.1% and 0.25% respectively compounded annually. The three alternatives given to the Appellants after the Market Change condition had been satisfied that is set out in paragraph (…) above amounted to a choice between (a) redeeming the Security at 100.1 of the issue price (the 0.1 necessarily being paid out of the initial capital of the trust) and accordingly losing the benefit of the fee of 1% plus VAT that had been paid; (b) selling the Security with the hope of a tax loss of 94; and (c) holding the zero-coupon Security for 15 years with an effective annual yield of about 1.1%. Apart from Mr Astall’s particular circumstances which I consider below, it was a practical certainty that nobody was going to choose (a) or (c). (This choice reminded me of what in the 1970s became known as the “X sham option.” Mr X was a highly respected member of the tax bar who had very reasonably advised the promoter of a tax scheme that it would stand a better chance of success if the taxpayer had a choice of action so that it was not clear that one result was preordained. The promoter took the advice somewhat too literally and wrote an option into the scheme that when one analysed the figures nobody would choose, the name being coined by counsel for the Revenue to pull Mr X’s leg about it.)

The choice offered to the purchaser of the Securities was between redeeming them on 7 days notice and realising the discount as a profit, or holding the zero-coupon Securities for 65 years with an effective annual yield of about 4.76% (assuming annual compounding) which Mr Patterson said (and I accept) was about the market rate. It was also a practical certainty that any purchaser would immediately redeem them at 5% of their redemption price, and this was known to KPMG. No purchaser would have considered holding the Securities for 65 years.

The Trustees would not have wished to have the Securities remaining in issue for 65 years. This would have seriously inhibited the investment of the trust funds and might have involved personal liability of the trustees for the whole of this period. They regarded it as a practical certainty that this would not happen, and a reasonable person with the knowledge available to the parties would conclude that the Securities would be immediately redeemed by the purchaser.

It was a practical certainty that the securities would cease to exist within two months of their issue: either (a) the Market Change condition would be satisfied and the securities would be sold and redeemed by the purchaser as above; or (b) in the event of the Market Change condition not being satisfied (the possibility of which I have decided should be ignored, except in relation to Mr Astall), the Appellants would redeem their securities.

Ignoring the existence of the Market Change condition and the possibility of a purchaser not being found, and the consequences of their not being satisfied, as I have already decided, the essence of the scheme is that each of the Appellants subscribes 100 for a Security the terms of which on giving notice to sell it suddenly change so that it is now worth only about 6 (5% of 118). It is then sold to a bank for 6 less a turn for the bank, which redeems it for 6 a week later. The remaining 94 remain in a trust for the benefit of the Appellant (matched by an equivalent tax-free gain in the trust which has had its liability to repay the Security removed)”.

It will be observed that in sub paragraph (3) he referred to a practical certainty that any purchaser would immediately redeem the Security on buying them at 5%, further that the Trustees would not have wished to have the Security remaining in issue for 65 years and that a reasonable person with the knowledge available to the parties would conclude that Securities would be immediately redeemed by the purchaser (4). Finally it was a practical certainty that the Securities would cease to exist within two months of their issue: either (a) the Market Change Condition would be satisfied and the Security sold or (b) the Market Change Condition would not be satisfied (which he concluded should be ignored) and the Appellants would redeem the Securities (5). He summarised the essence of the scheme as follows:-

“the essence of the scheme is that each of the Appellants subscribes 100 for a Security the terms of which on giving notice to sell it suddenly change so it is now worth only about 6 (5% of 118). It is then sold to a bank for 6 less a turn for the bank which redeems it for 6 a week later. The remaining 94 remain in a trust for the benefit of the Appellants (matched by an equivalent tax free gain in the Trust which had had its liability to repay the Security removed”.

As I have said above his conclusion in paragraph 22 was that the purpose of the legislation was to tax gains on securities that are issued at a deep discount and conversely to relieve losses on such securities but that the difference between the issue price and the redemption price must give rise to a possibility of making a gain that can be objectively seen to exist. He then said “this Security never had this possibility; it is a practical certainty that there will be a loss of 94…”

He accepted as a fact that it could not be said that there was no reasonable likelihood of the early redemption option being exercised enabling him to redeem the Security between one and two months after issue being exercised. This might arise because Mr Astall might not proceed and would then redeem. If he exercised, this would give rise to a gain of 0.01% on the issue price but which being in excess of the relevant percentage of 0.08% referable to the period of 2 months constituted a deep gain. This would then mean that the Security would be a relevant discount security allowing him to claim a loss of 95% of the redemption price.

He rejected this for the reasons he set out in paragraph 26. His conclusion there was that the only realistic view of the fact was that the scheme had been designed to fix the capital of the Trust so as to meet the premium on early redemption and that was the only possible source of funds and that the parties intended that it should be used for that purpose and that was the only possibility when considering the matter as at the date of issue of the Security. He concluded therefore that early redemption would always require a circular transaction using the capital of the trust. He asked the question whether or not such an artificial arrangement was within the relevant statutory provisions construed purposively and he said that the answer was no.

Mr Prosser QC who appeared for the Appellants submitted a comprehensive and detailed attack on the Special Commissioner’s decision and supplemented it with a detailed Reply in response to the somewhat briefer skeleton argument of the Respondents.

It seems to me however that the starting point of the attack arises out of paragraph 140 of Mr Prosser’s skeleton argument where he contends that the Special Commissioner’s conclusion that a sale within the time limit was “practically certain” was one which no reasonable tribunal could have reached and therefore was erroneous in law relying on the Bairstow case.

However in my view looking at the findings in paragraph 6 in the light of the evidence summarised in that paragraph I agree with the conclusion of the Special Commissioner and I do not think any other conclusion would have been possible. It was in the light of those findings and his consideration of the legal principles (which I shall deal with below) that he was led to the conclusion that the purpose of the legislation was to tax gains on securities that are issued at a deep discount and conversely to relieve losses on such securities. On the facts he found that there was never a possibility of these Securities making a gain looked at objectively. Thus it was he said a practical certainty that there would be a loss. Given that analysis his decision was that the Security in question is not within the regime for relief because it is not a “relevant discounted security”. His conclusion on that is based on his findings in paragraph 6 and his conclusion that there would never be a deep gain in reality.

The Special Commissioner has come to that conclusion on the facts and I do not see that there is any basis for challenging his factual findings.

This fatally flaws the Appellants’ appeals because it shows that the Special Commissioner has come to his conclusion as to whether or not the Security is within the legislation by determining the facts that were applicable to that Security. His conclusion cannot be faulted and therefore his decision cannot be faulted that on the facts of the Security and in the light of all the evidence the Appellants have not established that the Security falls within the definition “relevant discounted security”.

In my view that is enough to dispose of the Appellants’ appeal. As there was no realistic possibility of a deep gain there can be no basis (to put it in the words of Mr Ewart QC for the Respondents) to allow the tax payer to manufacture allowable losses by the expedient of inserting mechanisms giving a theoretical possibility of deep gains but in circumstances where the parties intended that instead far deeper losses should be made. The Special Commissioner was plainly alive to that prospect and on his findings that was an inevitable conclusion.

Mr Prosser QC in his skeleton argument set out a number of occasions where he submitted that the Special Commissioner misapplied the law (for example paragraph 59, 108, 113 and 119).

The theme of the criticisms is that the Special Commissioner understood that the House of Lords decision in SPI laid down a rule or principle that any tax statute must be construed in the way laid down in the SPI case (see paragraph 59 of the Appellants’ skeleton for example). It is thus submitted on behalf of the Appellants that the Special Commissioner failed to follow the guidelines in Barclays Mercantile as a result.

In order to understand that observation it is necessary to start with what the Special Commissioner said.

In paragraph 11 (4) he analysed Mr Prosser QC’s submissions in the context of Ramsay, SPI, MacNiven and Barclays Mercantile.

In paragraph 12 he dealt with the submissions of Mr Ewart QC which were to the effect that the general approach was explained in Barclays Mercantile but that the SPI case was more relevant. It is from here that the expression SPI principle came from. Thus in a section of his decision where he gives reasons (paragraphs 13-26) he picks up Mr Ewart QC’s contention that the Market Change Condition was directly within the SPI principle and he accepts that (paragraph 17). He came to the same conclusion in respect of the decision not to seek a purchaser for Security (paragraph 18). In respect of the terms for redemption he follows SPI but analyses it by reference to Ramsay, Barclays Mercantile and what was said by Lord Nichols of Birkenhead in MacNiven. In paragraph 21 and following he then viewed the facts of the transaction realistically and the application of the facts led to the conclusions set out in that paragraph. He rejected Mr Prosser QC’s submission that the only relevant facts were the terms of the Security (paragraph 22) observing:-

“that [i.e. only the relevant facts in the term of the Security] presupposes that the possibility of redemption written into the Security are within the facts when viewed realistically.

He then went on to conclude as I have set out earlier that Ramsay as explained in Barclays Mercantile entitled and required him to construe the legislation aimed at considering all possibilities in such a way to limit those possibilities to real ones.

Ramsay came under close scrutiny in the MacNiven case see Lord Nichols of Birkenhead paragraphs [1]–[8] (especially paragraph 8), Lord Hoffman [30]–[35], paragraph [48], and paragraph [58]-[62]. Lord Hutton emphasised that the essential element of the transaction to which the Ramsay principle was applicable was that it should be artificial paragraph [93].

The House of Lords then reviewed Ramsay again in Barclays Mercantile [2005] 1 All ER 97 and the SPI case [2005] 1 All ER 325. The judgments were handed down on the same day. The composition of the House of Lords was the same but separate hearings had taken place. In the Barclays Mercantile case the Ramsay case was again considered in paragraph [26] – [38]. The new approach (i.e. Ramsay’s approach) is set out in paragraph [32] namely to give a commercial construction in order to determine the nature of the transaction to which it was intended to apply and then to decide whether the actual transaction (which might involve considering the overall effect of a number of elements intended to operate together) fell within the statute description. The difficulties of the application of the principle are clearly summarised in paragraph [34] by virtue of the fact that tax is imposed by reference to economic activities or transactions which exist “in the real world” and second that a good deal of effort is devoted to structuring transactions in a form which will have the same or nearly the same economic effect as a taxable transaction but which it is hoped will fall outside the terms of the taxing statute. As is said in paragraph [37] sweeping generalisations about disregarding transactions undertaken for the purposes of tax avoidance need to be avoided. That was a reference to the MacNiven case. Once again their Lordships reinforced the need to focus carefully upon the particular statutory provision and identify its requirements before one can decide whether circular payments are elements inserted for the purpose of tax avoidance that should be disregarded or treated as irrelevant for the purpose of statute (ibid paragraph [38]).

Finally in SPI their Lordships referred to the Ramsay case again at paragraphs 18 and 19 on the particular facts as follows:-

“THE QUESTION OF CONSTRUCTION”

[18]

SPI is entitled to treat the loss suffered on the exercise of the Citibank option as an income loss if the option was a 'qualifying contract' within the meaning of s 147(1) of the Finance Act 1994. Section 147A(1) (inserted by the Finance Act 1996) provides that a 'debt contract' is a qualifying contract if the company becomes subject to duties under the contract at any time on or after 1 April 1996. By s 150A(1) (also inserted by the 1996 Act) a 'debt contract' is a contract under which a qualifying company (which means, with irrelevant exceptions, any company: see s 154(1)) 'has any entitlement … to become a party to a loan relationship'. A 'loan relationship' includes a government security. So the short question is whether the Citibank option gave it an entitlement to gilts.

[19]

That depends upon what the statute means by 'entitlement'. If one confines one's attention to the Citibank option, it certainly gave Citibank an entitlement, by exercise of the option, to the delivery of gilts. On the other hand, if the option formed part of a larger scheme by which Citibank's right to the gilts was bound to be cancelled by SPI's right to the same gilts, then it could be said that in a practical sense Citibank had no entitlement to gilts. Since the decision of this House in WT Ramsay Ltd v IRC, Eilbeck (Inspector of Taxes) v Rawling [1981] 1 All ER 865, [1982] AC 300 it has been accepted that the language of a taxing statute will often have to be given a wide practical meaning of this sort which allows (and indeed requires) the court to have regard to the whole of a series of transactions which were intended to have a commercial unity. Indeed, it is conceded by SPI that the court is not confined to looking at the Citibank option in isolation. If the scheme amounted in practice to a single transaction, the court should look at the scheme as a whole. Mr Aaronson QC, who appeared for SPI, accepted before the Special Commissioners that if there was 'no genuine commercial possibility' of the two options not being exercised together, then the scheme must fail”.

It is important to appreciate that the House of Lords in all 3 decisions emphasised the need to consider the language of the statute, the facts of the transactions and to determine whether those analyses of the facts fall within the language of the taxing statute as construed. There is in my view no assistance from seeing how different statutes were construed and applied to different facts in different cases. In the SPI case as paragraph [19] shows on the facts of that case if there was no genuine commercial possibility of the two options not being exercised together then the scheme must fail.

When their Lordships came to apply the construction they said this:-

“[22]

Thus there was an uncertainty [in Craven v White] about whether the alleged composite transaction would proceed to completion which arose, not from the terms of the alleged composite transaction itself, but from the fact that, at the relevant date, no composite transaction had yet been put together. Here, the uncertainty arises from the fact that the parties have carefully chosen to fix the strike price for the SPI option at a level which gives rise to an outside chance that the option will not be exercised. There was no commercial reason for choosing a strike price of 90. From the point of view of the money passing (or rather, not passing), the scheme could just as well have fixed it at 80 and achieved the same tax saving by reducing the Citibank strike price to 60. It would all have come out in the wash. Thus the contingency upon which SPI rely for saying that there was no composite transaction was a part of that composite transaction; chosen not for any commercial reason but solely to enable SPI to claim that there was no composite transaction. It is true that it created a real commercial risk, but the odds were favourable enough to make it a risk which the parties were willing to accept in the interests of the scheme.

[23]

We think that it would destroy the value of the Ramsay principle of construing provisions such as s 150A(1) of the 1994 Act as referring to the effect of composite transactions if their composite effect had to be disregarded simply because the parties had deliberately included a commercially irrelevant contingency, creating an acceptable risk that the scheme might not work as planned. We would be back in the world of artificial tax schemes, now equipped with anti-Ramsay devices. The composite effect of such a scheme should be considered as it was intended to operate and without regard to the possibility that, contrary to the intention and expectations of the parties, it might not work as planned”.

It will be seen that paragraph [22] addresses both the points that the Special Commissioner considered namely the lack of commercial reason for choosing the Market Change Condition by reference to changes in the dollar/sterling exchange rate and the fact that whilst there was a commercial risk that that would not be achieved the odds were favourable enough to make it a risk which the parties were willing to accept in the interest of the scheme. He was equally in my view entitled to take into account as part of the facts that the creation of the Market Change Condition was inserted as an anti-Ramsay device just like KPMG’s decision not to seek purchase for the Security until after it had been issued. As he said in paragraph 12 (2) both those should be ignored for the same reason as the uncertainty created in the SPI case.

This is not as Mr Prosser QC submits in his skeleton argument on the part of the Appellants that the Special Commissioner considered that SPI created a new test or principle which he was bound to apply. To the contrary it is plain that the Special Commissioner fully appreciated the derivation of this principle from Ramsay as commented on by the House of Lords in the later decisions. He is not merely because he refers in a shorthand way to the SPI principle intending to suggest that a new principle was to be found there which required him to make the determination in the way he did. When one looks at his careful judgment as a whole in my view it cannot be faulted.

Equally the circularity argument which he decided in the Astall appeal follows from a careful consideration of the authorities and an application of them. It too cannot be faulted and were his decision as regards the Edwards appeal to be wrong (which I do not think is the case) his alternative basis by applying the same reasoning on the Astall appeal would be equally applicable.

Further in the Astall case simply because there was a possibility that he might not enter into the scheme does not affect the consideration of the scheme as a whole although on the facts of Mr Astall’s case it introduced a further uncertainty. That uncertainty however was not relevant to the Special Commissioner’s determination for the reasons he set out in his decision that the scheme did not attract relief for losses because the Securities were not “relevant discounted securities”.

There are no grounds for suggesting that the Special Commissioner came to factual conclusions which he was not entitled to do and there is nothing to suggest that he incorrectly applied the law.

Accordingly I dismiss the Appellants’ appeals.