John Mander Pension Scheme Trustees Ltd v HM Revenue & Customs

The issue in this appeal is whether the Commissioners for Her Majesty’s Revenue and Customs have assessed the appellant for tax in the wrong tax year. The First Tier Tribunal [2011] UKFTT 686 (TC) and the Upper Tribunal [2013] UKUT 051 (TCC) concluded that the correct year of charge was the tax year to 5 April 2001. The appellant seeks to establish in this appeal that the correct tax year was to 5 April 1997.

The appellant is the sole trustee and administrator of the John Mander Limited Director’s Pension Scheme. On 19 April 2000 the Revenue gave notice of withdrawal of approval of the John Mander Scheme. The letter told Louvre Trustees Limited, the trustees at that time:-

“…it is the opinion of the Board of Inland Revenue that the facts concerning the administration of this scheme no longer warrant the continuation of its approval under s.591, Income and Corporation Taxes Act 1988. The Board, in exercise of its powers under s.591B ICTA 1988 has therefore decided to give notice that approval has been withdrawn with effect from 5 November 1996.”

On 27 July 2000 an assessment was made against those trustees, as administrators of the scheme under Case VI, Schedule D, in the sum of £475,200, being 40 per cent of the estimated value of Mr Mander’s pension fund as at 4 November 1996. The year of assessment was stated to be 2000/2001. The trustees appealed the assessment and sought, unsuccessfully, for permission to apply for judicial review of the Board’s decision to withdraw approval. Once Louvre Trustees Limited had resigned, the appellant was appointed in its place on 12 March 2002. On 22 January 2007 the Revenue issued a protective assessment. That second protective assessment was in respect of the same tax year as the earlier assessment, 2000/2001. This appeal turns on whether the assessment was in respect of the wrong year, 2000/2001, or should have been, as the appellant contends, 1996/1997.

The tax charge is imposed by s.591C TA 1988. The charge was introduced by s.61 of the Finance Act 1995. TA 1988 had previously made provision for the Board of Inland Revenue to approve certain pension schemes (see s.591 TA 1988). Approval had significant tax advantages (identified in s.592 and the following sections). The Finance Act 1991 introduced new sections in this part of the 1988 Act. Section 591A identified the effect on approved schemes of regulations made under s.591. By s.591A(2):-

“Any retirement benefits scheme approved by the Board by virtue of section 591 before the day on which the section 591 regulations come into force shall cease to be approved by virtue of that section at the end of the period of 36 months beginning with that day if at the end of that period the scheme -

(a)

contains a provision of a prohibited description, or

(b)

does not contain a provision of a required description,

unless the description of provision is specified in regulations made by the Board for the purposes of this subsection.”

It is worth noting that under s.591A approval ceases by virtue of the operation of the regulations. Under that section, (contrary to the description of Vos J in the Upper Tribunal at paragraph 20) approval is not withdrawn.

The Finance Act 1991 also introduced s.591B(1) which conferred power on the Board to withdraw approval if it was of the opinion that facts concerning an approved scheme or its administration ceased to warrant continuance of the approval. If the Board did reach that view it was required in the notice of withdrawal to specify a date from which approval ceased to run. Section 591B(1) reads:-

“If in the opinion of the Board the facts concerning any approved scheme or its administration cease to warrant the continuance of their approval of the scheme, they may at any time by notice to the administrator, withdraw their approval on such grounds, and from such date (which shall not be earlier than the date when those facts first ceased to warrant the continuance of their approval…), as may be specified in the notice.”

It should be noted that approval ceases under s.591B(1) only where the Board reaches an opinion that the facts warrant the withdrawal of approval and chooses to exercise its power to withdraw such approval by notice.

The Finance Act 1991 also introduced a third way in which approval might cease. Again, like s.591A(2), approval is not withdrawn. It merely ceases to apply once an alteration is made to a retirement benefit scheme (unless the alteration is itself approved by the Board or is within a specified class). Section 591B(2) reads:-

“Where an alteration has been made in a retirement benefits scheme, no approval given by the Board as regards the scheme before the alteration shall apply after the date of the alteration unless –

(a)

the alteration has been approved by the Board, or

(b)

the scheme is of a class specified in regulations made by the Board for the purposes of this paragraph and the alteration is of a description so specified in relation to schemes of that class.”

It is plain, therefore, that approval might cease by the operation of any one of the three methods introduced by the Finance Act 1991. But each of those methods is described in a different way. Under s.591A(2) operation of the Regulations, the scheme “shall cease to be approved”; under s.591B(1) the Board may “withdraw their approval” and under s.591B(2) “no approval shall apply”.

As I have already indicated, s.61 of The Finance Act 1995 introduced a tax charge in circumstances where approval had ceased in any one of the three ways identified in s.591A or s.591B. Section 591C provides:-

“(1)

Where an approval of a scheme to which this section applies ceases to have effect…, tax shall be charged in accordance with this section.

(2)

the tax shall be charged under Case VI of Schedule D at the rate of 40 per cent on an amount equal to the value of the assets which immediately before the date of the cessation of the approval of the scheme are held for the purposes of the scheme (taking that value as it stands immediately before that date).

(3)

Subject to section 591D(4), the person liable for the tax shall be the administrator of the scheme.

(4)

This section applies to a retirement benefits scheme in respect of which one or more of the conditions set out below is satisfied.

(5)

The first condition is satisfied in respect of a scheme if, immediately before the date of the cessation of the approval of the scheme, the number of individuals who are members of the scheme is less than twelve.

(6)

The second condition is satisfied in respect of a scheme if at any time within the period of one year ending with the date of the cessation of the approval of the scheme, a person who is or has been a controlling director of a company which has contributed to the scheme is a member of the scheme.

(6A) The third condition is satisfied in respect of a scheme if –

(a)

at any time within the period of three years ending with the date of the cessation of the approval of the scheme, the scheme has received a transfer value in respect of any person; …”

Section 591C(1) imposes the charge to tax. Mr Sykes, for the appellant, was disposed to accept that it was that sub-section which introduced the charge, although he had an alternative argument that it was imposed under sub-section (2). It seems to me clear that the tax is imposed in the circumstances identified in the first part of s.591C(1), that is, where an approval of the scheme ceases to have effect. Since the tax imposed was income tax, s.591C(2) identifies the Case under which tax is to be charged and the rate at which it is to be charged. But, there is no further identification of the year in which the tax falls to be charged other than the indication in the first part of s.591C(1) that it should be charged where an approval of a scheme ceases to have effect.

Section 591C(1) goes no further because it has to cater for the three different ways in which approval of a scheme ceases to have effect. Whilst the consequences of each of the three provisions in s.591A(2), in s.591B(1) and in s.591B(2) may compendiously be described as provisions under which approval ceases to have effect, the provisions themselves operate in three different ways to bring about the consequence that approval ceases to have effect.

For that reason, supplementary provisions in s.591D(7) provide:-

“(7)

The reference in section 591C(1) to an approval of a scheme ceasing to have effect is a reference to –

(a)

the scheme ceasing to be an approved scheme by virtue of section 591A(2);

(b)

the approval of the scheme being withdrawn under s.591B(1);

(c)

the approval of the scheme no longer applying by virtue of section 591B(2);

and any reference in section 591C to the date of the cessation of the approval of the scheme shall be construed accordingly.”

The first part of s.591D(7) recall the three different methods by which approval of a scheme ceases to have effect. Under s.591D(7)(a) approval ceases to have effect when the scheme ceases to be an approved scheme by the operation of the s.591 Regulations. Where the scheme was approved before the day on which those Regulations came into effect, that will be at the end of the period of 36 months beginning with the day on which the s.591 Regulations came into force provided the scheme contains a provision of a prohibited description or does not contain a provision of a required description. Accordingly, tax charged where approval ceases to have effect by virtue of the operation of s.591A(2) shall be for the year in which the 36 month period comes to an end. Under s.591B(2) the year of charge may similarly be identified by reference to the event which led to approval ceasing to have effect. The charge arises in the year in which an unapproved alteration is made in the retirement benefits scheme.

Section 591B(1) differs from s.591A(2) and s.591B(2) because it encompasses two different stages and thus two different dates. It refers firstly to the Board’s withdrawal of approval. This will usually be some time after the events which warrant such withdrawal. Secondly, it refers to the time from which the Board decides the cessation of approval will run. The earliest date will be the date when facts first ceased to warrant continuance of approval or 17 March 1987, or it could be from a later date as specified in the notice. The date of withdrawal of approval and the date specified in the notice are most unlikely to be the same, whereas under s.591A(2) and s.591B(2) there can only be one date when approval ceases to have effect, namely, the date when either of those two sub-sections bites.

Mr Sykes argued that under s.591(B)(1) the date when approval ceased to have effect was not the date when the approval of the scheme was withdrawn. It was the date specified in the notice of 5 November, 1996 in the tax year 1996/1997. Neither of the assessments charged tax in that year and, accordingly, they were either invalid or, if they had assessed the appellant to tax in that year, they were out of time.

His submission rested in part upon the impact of the opening full-out words of s.591D(7). The use of the expression “approval of a scheme ceasing to have effect” must refer, he said, to the date specified in the notice, that is, 5 November 1996 and not the date when the Revenue gave notice of withdrawal. That seems to me to ignore the plain words of s.591D(7)(b). Those words make it clear that the date, for the purposes of s.591C(1) when the charge to tax arises, is the date when approval of the scheme is withdrawn and not the date specified in the notice from which approval ceases to continue.

Mr Sykes then referred to the full-out words at the end of s.591D(7):-

“any reference in s.591C to the date of the cessation of the approval of the scheme shall be construed accordingly.”

The full-out words at the end of the sub-section reinforce, he submits, the contention that the date of charge is the date when approval of a scheme ceases. It is contrary, he submits, to any sensible canon of construction to suggest that the date to which the words at the end of the sub-section refers may differ from the date identified in the words which open the same sub-section. One date is referred to, it is the date of cessation and no other. That was 5 November 1996.

There was no dispute but that the date of cessation of approval means what it says. It was 5 November 1996. Neither the appellant nor the Revenue argued that the date of cessation of approval meant the date of withdrawal of approval. Indeed, the Revenue relied upon that provision to justify the quantification of the charge by reference to the value of the assets at midnight on 4 November 1996. But, in my view, it does not follow that the reference in s.591C(1) is to the same date.

It is necessary to draw attention to the contrast between the words “approval of a scheme ceasing to have effect” and “the date of the cessation of the approval” in s.591D(7).  It is no accident that that different wording is adopted.  There is only one place in s.591C where the expression “approval ceases to have effect” is used.  That is in s.591C(1).  But there are a number of references within s.591C to the “date of the cessation of the approval”.  

It is used in s.591C(2), for the purposes of quantifying the charge and valuing the assets. It is used in all three of the conditions which must be satisfied before s.591C applies. The first condition in s.591C(5) sets the maximum number of individuals who may be members of the scheme at a moment immediately before the date of cessation of approval. The second condition under s.591C(6) requires the controlling director of a company which has contributed to the scheme to be a member of the scheme within the period of one year ending with the date of the cessation of the approval of the scheme. The third condition is satisfied again by reference to whether the scheme has received a transfer value or contributions as identified in s.591C(6A) in a period of three years ending with the date of the cessation of approval of the scheme. As Mr Nawbatt pointed out, without demur from the appellant, the provisions could easily be avoided if the date at which the conditions must apply was later than the date specified in a s.591B(1) notice.

It is, accordingly, plain that the reference to the date of cessation of the approval of the scheme must be to the date from which approval ceased to have effect in the ordinary sense of the words. In the instant case that was, according to the notice given by the Board under s.591B(1), 5 November 1996. That is the date from which the Board has decided approval ceases to run. That was common ground. The dispute centres on whether it follows that the opening words of 591C(1) “Where an approval of the scheme to which this section applies ceases to have effect…” must be construed as referring to the tax year in which the date of cessation fell, namely in the year of assessment 1996-1997.

Were that the correct construction it would render the opening words in s.591D(7) unnecessary. Only one date would be relevant to s.591C, that is, the date when approval ceased to have effect. Under s.591A(2), that would be the date the scheme ceased to be an approved scheme by reason of the s.591 Regulations. It would be the date under s.591B(1) specified by the Board in its notice as being the date from which approval ceases to run. Under s.591B(2) it would be the date when the scheme was altered.

Such a conclusion wholly disregards s.591D(7)(b). As I have said, the reference to “the approval of the scheme being withdrawn” is not a reference to the date specified by the Board in its notice but to the date of withdrawal. That is the purpose of the reference to the three different regimes pursuant to which approval ceases to have effect. It is also the purpose lying behind the distinction between the opening words of s.591D(7) which refer only to s.591C(1) and the closing full-out words in their reference to s. 591C. The expression “date of the cessation of the approval” does not figure in s.591C(1). In those circumstances it only makes sense to construe the adverb ‘accordingly’ as it was construed by the First Tier Tribunal “that the date of cessation of approval must be construed according to the section numbers referred to set out in… s.591D(7)” [130]. The charge arises under s.591C(1) when approval of the scheme is withdrawn under s.591B(1). That was within the tax year 2000-2001.

The appellant sought to support its construction by reference to the provisions which introduced s.591C and s.591D. Section 61(3) FA 1995 provided:-

“This section (section 61) shall apply in relation to any approval of a retirement benefit scheme which ceases to have effect on or after 2 November 1994 other than an approval ceasing to have effect by virtue of a notice given before that date under s.591B(1) of the Taxes Act 1988.”

As the appellant rightly points out, that section must be a reference to the time when a scheme ceases to have effect and not to the time of the service of a notice withdrawing approval under s.591B(1). But s.61(3) has no relevance to the proper construction of s.591B which had been introduced in the earlier Finance Act 1991 nor to the proper construction of s.591C. The construction of s.591C, depends upon s.591D(7).

Mr Nawbatt, for the Revenue, additionally pointed out the unfortunate consequences if the appellant was correct as to the payment of interest. If the year of charge was the year the Board specified in its notice under s.591B then those charged to tax would be liable for interest long before the Board had exercised its discretion to withdraw approval; interest would be payable at a time when approval of the scheme was continuing. It seems to me unlikely, if not impossible, that there could be any intention to impose a liability to pay interest during a period, possibly long before, when approval had not been withdrawn but still applied. I reject the suggestion that s.591B(1) could be deployed prospectively. Such a consequence supports the Revenue’s contention that where s.591B(1) applies, s.591C(1) refers to the date of withdrawal of approval and not to the date which the Board specifies in its notice.

The appellant contended, in the alternative, that the charge was imposed by s.591C(2) and not by s.591C(1). But as I have already indicated, the charge is imposed under s.591C(1) in one of the three circumstances where approval of the scheme ceases to have effect. It is not imposed by s.591C(2) which refers merely to the Case under which the tax is imposed and the means by which the amount of tax is to be quantified. I can summarise my views by grateful reference to the decision of the First Tier Tribunal:-

“131.

S 591D(7) dealt with three different ways a scheme become unapproved and the dates for loss of approval were not the same in all three. In particular, cessations under S591A(2) and 591B(2) occurred automatically on a specified date or event and did not require the exercise of a discretionary power by HMRC. For s591D(7)(a) & (c) there was only one date: the date of cessation of approval. It makes no difference for these two sub-sections whether “construed accordingly” means the same date or in accordance with the specified section. They are the same.

132.

But it makes a difference for s591D(7)(b) and discretionary withdrawal of approvals: because if “construed accordingly” means ‘construed in accordance with s591B(1)’, then it brings in the two dates mentioned in that section: the date on which notice of withdrawal is actually given and the date from which such withdrawal is effective. So in using “construed accordingly” the draftsman intended to recognise that at least for s591(B)(1) the date of cessation of approval was not necessarily the same as the date of cessation.

133.

Consistent with our view the drafter must have meant separate dates, S 591C(1) goes on to provide that “tax shall be charged”. It does not say “tax shall be deemed to have been charged”. Tax is not retrospective without clear words and there is no suggestion here that the tax charge is imposed retrospectively. In 1998 and 1999 the trustees were not in fact liable to be assessed to this tax: the charge could not arise until HMRC took the decision to withdraw approval in 2000.”

For those reasons, I would dismiss this appeal.

I agree that the appeal should be dismissed for the reasons given by Moses LJ. The taxpayer’s argument that the correct year of assessment is 1996/97 rather than 2000/2001 depends upon treating the words (“where approval of a scheme … ceases to have effect”) in s.591C(1) as determinative not only of the circumstances in which the charge to tax arises but also of the year of assessment to which the charge relates. Only in this way can it be said that the tax arises not in the year in which approval is withdrawn under s.591B(1) but in the year specified in the notice of withdrawal as the date from which approval ceases thereby to have effect.

It is the Appellant’s case that the reference in s.591C(1) to approval ceasing to have effect is a reference in this case to 5 November 1996 but, as Moses LJ has explained, that provision is designed to encapsulate the three ways in which approval may come to an end (two of which are, so to speak, automatic) and not primarily with issues about the year of assessment. So far as s.591D(7) takes the matter any further, it merely confirms the inclusion of all three methods as giving rise to the charge under s.591C(1) and the application of the conditions set out in s.591C(5)-(6A) in all three cases.

The scheme ceases to be approved under s.591B(1) when approval is withdrawn by the notice given by the Board: see s.591D(7)(b). The date of the notice determines the relevant year of assessment. The fact that the notice may and often does specify an earlier date of cessation by reference to when the conditions for approval ceased to be satisfied is irrelevant for this purpose. The charge to tax is occasioned by the effect of the notice but it is the service of the notice which withdraws approval (from the date specified) and which therefore determines the year of assessment in which the charge to tax arises.

I agree with both judgments.