Stephen John Mullens v The Commissioners for HMRC

This is an appeal brought by Mr Stephen John Mullens against the decision of the First-tier Tribunal (the “FTT” - Judge Christopher McNall and Mr Julian Stafford) reported as [2021] UKFTT 131 (TC). So far as relevant to this appeal, the case concerned two separate matters: (1) the taxability or otherwise of six payments (amounting to about £40m in total) made to or for the benefit of Mr Mullens; and (2) whether penalties were correctly assessed as a result of Mr Mullens’ failure to declare those payments in his self-assessment returns as his income. Those payments were referred to by the FTT as “Payments 1 to 6” and we do the same in this judgment.

HMRC issued discovery assessments under s.29 of the Taxes Management Act 1970 (“TMA”) asserting a tax liability in relation to Payments 1 to 5 in various tax years from 1999/00 to 2008/09. Since Mr Mullens had submitted self-assessment returns for those tax years, HMRC had to establish the presence of one or more pre-conditions and they relied on the proposition that the loss of tax sought to be recovered by the assessments was attributable to careless or deliberate conduct of the taxpayer (under s.29(3) and (4) of TMA). We refer in this judgment to that conduct generically as “culpable conduct”.

In the case of Payments 1 to 4, HMRC relied on the extended time limits given by s.36(1) and (1A) of TMA for making the assessments (six years for careless conduct and 20 years for fraudulent or deliberate conduct). We refer to discovery assessments made in reliance on s.36(1) as “ETL assessments”. In the case of Payment 5, the discovery assessment was made within the normal four-year time limit (and, accordingly, s.36(1) and (1A) of TMA were irrelevant). We refer to an assessment of that kind as an “ordinary” discovery assessment.

In the case of Payment 6, HMRC were in time to enquire into the self-assessment return submitted by Mr Mullens. HMRC issued a closure notice in relation to the return amending it so as to impose additional tax by reference to that payment.

HMRC issued penalty assessments in relation to each of Payments 1 to 6. In the case of Payments 1 to 4, these were issued on the basis that the appellant had fraudulently submitted incorrect returns (s. 95 of TMA). In the case of Payments 5 and 6,penalties were issued on the basis that the failures to declare the payments were deliberate (Sch. 24 to the Finance Act 2007 (“FA 2007”)). The relevant culpable conduct relied on (fraudulent or deliberate) reflected the change in legislative language made by FA 2007 but nothing turns on any distinction in the meaning of those two expressions for the purposes of this appeal. Accordingly, we refer in the remainder of this judgment to “deliberate” conduct as a short-hand to include fraudulent conduct.

The taxpayer appealed against all of the discovery assessments (ETL and ordinary) as well as the closure notice, disputing, in relation to each payment, that any liability to tax arose. In the case of the ETL assessments, he also expressly appealed on the ground that the assessments were made out of time. He also appealed against all of the penalty assessments. In his grounds of appeal before the FTT in respect of the penalties, he referred to having a reasonable excuse for the relevant failures; but, self-evidently, if there was no tax due or no culpable conduct, the penalty assessments would fall away.

In the case of Payment 5 (an ordinary discovery assessment where no time limit point could be taken), it was common ground that the burden of showing the validity of the discovery assessments was HMRC’s but it was for the taxpayer to show that the tax assessed was wrong. And, in the case of Payment 6, it was not in dispute that HMRC had to show that there was a deliberate inaccuracy.

The FTT found that each of Payments 1 to 6 was taxable as Mr Mullens’ income and that penalties were correctly assessed as a result of his deliberate failure to declare those payments in his self-assessment returns as his income.

Mr Mullens applied for permission to appeal on 14 grounds but ultimately obtained permission, from the Upper Tribunal, on just the following:

(1)

Grounds 1 to 4 (the “Assessment Appeal”) relate to the ETL assessments only (dealing with Payments 1 to 4). It is common ground that, given the way that HMRC put their case, they bore a burden of proof in two respects. First, they had to establish that the pre-condition set out in s.29(4) of TMA was present (a “Section 29(4) Burden”). Second, they had to establish that the requirements of s.36(1) or (1A) of TMA were met so that they could make an ETL discovery assessment (a “Section 36 Burden”). Mr Mullens has not challenged the FTT’s decision so far as relating to the Section 29(4) Burden. However, he argues that the FTT erred by failing to realise that, for HMRC to discharge their Section 36 Burden, they had to show, in addition to culpable conduct, there was an actual loss of some tax in the years of assessment covered by the ETL assessments. Mr Mullens argues that to discharge their Section 36 Burden, HMRC needed to establish matters such as (i) the taxable source from which the payments derived; (ii) the status of the payments as income (rather than capital); and (iii) that the payments were taxable in the years specified in the ETL assessments, as distinct from other tax years (“Constituents (i) to (iii)”). Mr Mullens argues that the FTT erred by failing to recognise that HMRC bore this Section 36 Burden and/or by upholding the ETL assessments relating to Payments 1 to 4 when HMRC had not discharged that burden.

(2)

Ground 11 (the “Penalty Appeal”) covers penalties that HMRC imposed in connection with Payments 1 to 6. As regards Payments 1 to 4, Mr Mullens’ arguments are parasitic on those raised in the Assessment Appeal and he argues simply that, since the ETL assessments relating to Payments 1 to 4 must be set aside, so must the associated penalties. However, the Penalty Appeal also covers Payments 5 and 6. Mr Mullens accepts that his permission to appeal does not extend to challenging the taxability of Payments 5 and 6, but nevertheless argues that the FTT’s decision to impose a penalty in relation to those payments was vitiated by a failure to apply the correct burden of proof.

In the Assessment Appeal, Mr Mullens accepts that, although there was an additional Section 36 Burden on HMRC to show some loss of tax, they did not necessarily have to show the amount of tax actually assessed. Accordingly, the consequences of Mr Mullens’ position in the Assessment Appeal are as follows:

(1)

where HMRC made an ordinary discovery assessment relying on culpable conduct, it would be for the taxpayer to displace the assessment in all cases (whether the dispute concerned liability only, liability and quantum, or quantum only) while the burden on showing the culpable conduct would rest on HMRC. We took Mr Mullens to accept that, in such a case, to discharge their Section 29(4) Burden, HMRC would not need to establish Constituents (i) to (iii); but

(2)

where HMRC made a discovery assessment relying on culpable conduct under an ETL assessment, the burden on showing the culpable conduct would still rest on HMRC but:

(a)

if the dispute is, like the present one, “binary” in the sense that a particular payment is either entirely taxable or entirely free from tax, HMRC would have the burden of proving, potentially among other matters, Constituents (i) to (iii) in order to discharge their Section 36 Burden;

(b)

if there was a dispute about liability and quantum (for example if HMRC have made a general “back duty” assessment considering that a taxpayer has failed to declare a general category of taxable receipts), HMRC would have the burden of showing a liability to tax of at least some amount for the period covered by the assessment. They would need to establish the presence of Constituents (i) to (iii) in relation to that liability so identified but would not need to establish their presence in relation to all items covered by the ETL assessment. If successful in doing so, the burden would then shift to the taxpayer (under s.50(6) of TMA) to displace the amount actually assessed; and

(c)

if the dispute was only about quantum, HMRC would have no burden in respect of the tax assessed and it would be for the taxpayer to displace the amount actually assessed.

Mr Mullens was a solicitor and partner at Marriot Harrison LLP during which time he became professionally involved with the well-known Formula 1 figure, Mr Bernie Ecclestone. Mr Mullens resigned from Marriot Harrison LLP in 1999 to become a lawyer and adviser to the Ecclestone family interests. In 1999/2000 Mr Mullens received a payment of £1.2m (Payment 1). The FTT found that this was part of a remuneration package for his services to the Ecclestone family interests (see [150]). This payment was mentioned in a white space disclosure in Mr Mullens’ self-assessment return but, as the FTT put it at [282(2)], in terms that were “very artful and not genuinely revealing”.

In the tax year 2000/01 the appellant received two further payments both of which were made into an offshore account: £750,000 (Payment 2) and £300,000 (Payment 3). The FTT found that the payments were made as further financial rewards for Mr Mullens’ services (see [169]). Neither of those payments were declared in Mr Mullens’ self-assessment returns.

The appellant then received several large payments from Mrs Slavica Ecclestone, Mr Ecclestone’s wife. Mr Mullens contended that all of these payments were gifts. They included payments of the sum of $38m in 2006/07 (Payment 4), the sum of $19.5m (Payment 5)and the sum of £5m in 2012/13 (Payment 6). The FTT found that Payments 4 to 6 were made in return for services which Mr Mullens had provided to the Ecclestone family interests (see [176]). Again, none of these payments were declared in Mr Mullens’ self-assessment returns.

In making its findings of fact, the FTT made at [77] to [95] a series of adverse findings about Mr Mullens’ reliability and truthfulness as a witness. Having concluded ([90]) that Mr Mullens could not be safely regarded as a witness of truth, the FTT went on to make detailed findings in relation to each of the payments reaching the following conclusions:

“[150] … we are satisfied that Payment 1 was actually made in consideration of services which Mr Mullens had already rendered or would be rendering to the Ecclestone family interests. We are sure that Payment 1 was received by Mr Mullens as part of a remuneration package for his services to the Ecclestone family.

[170] We are satisfied that Payments 2 and 3 were made in consideration of services which Mr Mullens had already rendered or would be rendering to the Ecclestone family interests, and were therefore a receipt of his trade or profession under Schedule D Case 1.

[176] We are sure that the pattern was one of large sums of money being advanced to Mr Mullens, from the Ecclestone 'side', in return for services provided by Mr Mullens to the Ecclestone family interests.” [Payments 4 to 6 together]

[201] We have no doubt at all that the payment made on 7 April 2006 was in direct consequence of Mr Mullens' involvement in the sale and was in fulfilment of the terms set out in the letter of 14 October 2005. The payment of $38 million was referable to that letter. [Payment 4]

[237] We are satisfied that Payment 5 was received by Mr Mullens as a reward for services which he had carried out …

[265] We are satisfied that Payment 6 was not a gift, and insofar as need to do so we find that this payment was rather in connection with the retainer for his assistance with the Constantin Medien litigation.”

The FTT found that Mr Mullens knew that the payments should have been reported on his tax returns for the relevant periods and his conduct in not declaring them was deliberate.

The FTT made findings about Mr Mullens’ conduct in relation to Payment 1 as part of its conclusion that HMRC had discharged their Section 29(4) Burden in relation to the relevant assessment (see [282] and [283]). In respect of Payments 2 to 5, the FTT’s discussion of the validity of the discovery assessments was focused only on the issue of staleness (now irrelevant as a result of the decision of the Supreme Court in HMRC v Tooth [2021] UKSC 17). The FTT did not set out an express conclusion to the effect that HMRC had discharged their Section 29(4) Burden in relation to Payments 2 to 5. However, the FTT did set out its findings as to Mr Mullens’ culpable conduct in relation to Payments 2 to 5 in its discussion about the penalties assessed in respect of those payment (see [338] to [342] and [349]).

The FTT dealt with the different burdens relevant to the appeal before it in this way:

“20.

In relation to the discovery assessments, HMRC (as it accepts) bears the burden of proving that these were validly given, including that these were given in time, in accordance with sections 29 and 36 of the Taxes Management Act. Section 36(1) sets down the 6 year limit from the end of the year of assessment where the loss of tax has been "brought about carelessly". That is extended to 20 years where the loss of tax is brought about "deliberately": section 36(1A).

21.

If HMRC satisfies the discovery provisions (and if the discoveries have not gone 'stale') then the assessments (as is conventional) "stand good" and the burden shifts to Mr Mullens to show that the assessments should be cancelled or made in some different sum: section 50(6) of the Taxes Management Act 1970. He has to prove that the Payments were made for the reasons which he alleges. HMRC does not have to prove a negative.”

The FTT recorded at [23] that it had approached its analysis in the same sequence as the parties, namely by first considering the nature of the payments and whether they were taxable. Plainly, if the payments were not taxable, then, as the FTT noted, “there was nothing relevant for HMRC to discover, and hence the need to consider the statutory conditions falls away” – the reference here to the statutory conditions being a reference to the applicable conditions in ss.29 and 36 of TMA.

It is not disputed by HMRC that the FTT proceeded on the basis that discharging the Section 36 Burden in respect of the ETL assessments required HMRC to show nothing more than was necessary to discharge their Section 29(4) Burden. Mr Mullens challenges the FTT’s assumption. However, we took him to accept that, if the FTT was correct in its assumption, the direction made by the FTT as to the law would similarly be correct.

There were subsequent passages in its decision where the FTT recognised that, if the discovery assessments were not valid, the dispute about the liability to tax would be nugatory. That is most clearly revealed at [275] of its decision where the FTT described the discovery issue as the “anterior issue”. But it was not the only place where it was aware of that: see, for example, the reference at [154] to the liability to tax being “subject to any issue of discovery”.

Questions of statutory interpretation are at the heart of this appeal and it is with the relevant statutory provisions that we begin. We should note at the outset that the authority on which Mr Mullens principally relies (the Court of Appeal’s judgment in Hurley v Taylor [1999] STC 1) was decided by reference to earlier versions of the relevant discovery provisions of TMA which were in a materially different form from those applicable to this appeal. We consider, for the reasons given below, that the judgment in Hurley has nothing to say about the relationship between ss.29(3) and (4) and 36(1) and (1A) of TMA in their current form (although the judgment remains of relevance for other purposes). To understand the authorities and to place the current provisions of TMA in their proper context, it is necessary to set out briefly the changing nature of the discovery and time limit provisions as they have developed since the earliest of the cases (decided in 1965) that we discuss below.

HMRC have had since that time (and indeed before) a power to make discovery assessments. Until the changes made by Parliament in introducing the system of self-assessment (in 1994 for income tax and in 1998 for corporation tax), HMRC had merely to discover a loss of tax. There were no other pre-conditions: in particular, there was no need to prove any culpable conduct on the part of the taxpayer.

Section 29 of TMA was in the following form so far as relevant to the disputed assessments in Hurley:

“(3)

If an inspector or the Board discover—

(a)

that any profits which ought to have been assessed to tax have not been assessed, or

(b)

that an assessment to tax is or has become insufficient, or

(c)

that any relief which has been given is or has become excessive,

the inspector or, as the case may be, the Board may make an assessment in the amount, or the further amount, which ought in his or their opinion to be charged.”

It is worth noting that, in this version of s.29, the only threshold condition was an inspector or the Board “discovering” one of the situations mentioned in ss. 29(3)(a) to (c). The law considered at the time of Hurley contained no analogue of the Section 29(4) Burden and, for this reason alone, Hurley can provide no binding guidance on the relationship between the Section 29(4) Burden and the Section 36 Burden under the law applicable to the periods covered by Mr Mullens’ appeal. In addition, unlike the version of s.29(3) that came into effect after the introduction of self-assessment set out in paragraph 29 below, there was no reference at the end of s.29(3) to the assessment being one made to “make good to the Crown the loss of tax”. However, in our judgment that is of little significance: the Revenue could scarcely make an assessment of an amount which in their opinion “ought… to be charged” if the amount did not relate to the pre-conditions set out in paras. (a) to (c) of the subsection – which will inevitably lead to a loss of tax.

At the time of Hurley, the time limit for the making of a discovery assessment was, in the normal case, 5 years (see s. 34 of TMA). But that time limit was significantly extended in the case of culpable conduct on the part of the taxpayer. At the time of the Hurley judgment, s.36(1) of TMA was in these terms:

“36Fraudulent or negligent conduct

(1)

An assessment on any person (in this section referred to as “the person in default”) for the purpose of making good to the Crown a loss of tax attributable to his fraudulent or negligent conduct or the fraudulent or negligent conduct of a person acting on his behalf may be made at any time not later than twenty years after the end of the chargeable period to which the assessment relates.”

This subsection applied in the same way for the purposes of income tax and corporation tax (see the simple reference to “tax”) and, unlike s.29(3) of TMA as it stood at the time, did refer to making good to the Crown a loss of tax attributable to the culpable conduct. The provision worked by asking whether the tax assessed could be linked to fraudulent or negligent conduct.

The burden of proof in relation to a challenge brought against an assessment by a taxpayer was set out in s.50(6) of TMA. Its essential features at the time of the Hurley judgment were in materially the same form as the current version, which provides:

“(6)

If, on an appeal notified to the tribunal, the tribunal decides—

(a)

that, ... , the appellant is overcharged by a self-assessment;

(b)

that, ... , any amounts contained in a partnership statement are excessive; or

(c)

that the appellant is overcharged by an assessment other than a self-assessment,

the assessment or amounts shall be reduced accordingly, but otherwise the assessment or statement shall stand good.”

The introduction of self-assessment radically changed the way in which discovery assessments were made. Broadly stated and as explained in more detail by Moses LJ in Tower MCashback LLP v HMRC [2010] ATC 809 at [24], under the self-assessment system, a taxpayer who has submitted a tax return is at risk of a discovery assessment in only limited circumstances. Those circumstances include where a loss of tax discovered by HMRC is attributable to culpable conduct on the part of the taxpayer.

Section 29 of TMA provides (so far as relevant to this appeal) as follows:

“29Assessment where loss of tax discovered

(1)

If an officer of the Board or the Board discover, as regards any person (the taxpayer) and a year of assessment—

(a)

that any income […] which ought to have been assessed to income tax […] [has] not been assessed,

(b)

that an assessment to tax is or has become insufficient, or

(c)

that any relief which has been given is or has become excessive,

the officer or, as the case may be, the Board may, subject to subsections (2) and (3) below, make an assessment in the amount, or the further amount, which ought in his or their opinion to be charged in order to make good to the Crown the loss of tax.

(2)

[…]

(3)

Where the taxpayer has made and delivered a return under section 8 or 8Aof this Act in respect of the relevant year of assessment, he shall not be assessed under subsection (1) above—

(a)

in respect of the year of assessment mentioned in that subsection; and

(b)

... in the same capacity as that in which he made and delivered the return,

unless one of the two conditions mentioned below is fulfilled.

(4)

The first condition is that the situation mentioned in subsection (1) above was brought about carelessly or deliberately by the taxpayer or a person acting on his behalf.

(5)

to (7A) […]

(8)

An objection to the making of an assessment under this section on the ground that neither of the two conditions mentioned above is fulfilled shall not be made otherwise than on an appeal against the assessment.”

Thus, and by contrast with the version of s.29 considered in paragraph 24 above, to make a discovery assessment for a period for which a taxpayer had submitted a self-assessment return, it was no longer sufficient for an inspector or the Board to “discover” certain matters. Additional threshold conditions needed to be satisfied as well (see s.29(3)). The condition relevant to this appeal concerns culpable conduct on the part of the taxpayer, namely subsection (4). It is not in dispute that HMRC bear the Section 29(4) Burden of showing that the condition in s.29(4) is met.

The “situation” referred to in subsection (4) is a reference to what has been described as an “actual insufficiency” in the amounts charged to tax (see [33] to [34] of the judgment of Auld LJ in Langham v Veltema [2004] STC 544, which considered the meaning of “the situation” in the context of s.29(5)) or the “fact of the undercharge” in Hargreaves v HMRC [2014] UKUT 0395 (TCC) (“Hargreaves UT”) at [21(6)]). The “situation mentioned in subsection (1)”, therefore, is not a reference to HMRC’s making of the discovery, as specifically confirmed in Hargreaves UT at [21(6)].

More generally, and contrary to some of Mr Goldberg KC’s oral submissions, s.29(4) is not concerned with the officer’s subjective opinion but with objective fact (see [21] to [28] of Lewison LJ’s judgment in Hankinson v HMRC [2011] EWCA Civ 1566). It follows, therefore, that s.29(4) is asking whether the “fact of the undercharge” was brought about by a taxpayer’s careless or deliberate conduct: HMRC’s opinions on the taxpayer’s conduct, and the amount of the undercharge, are not relevant.

The general time limit for making a discovery assessment is four years from the end of the year of assessment to which it relates (see s.34(1) and (3) of TMA).

However, in the case of culpable conduct on the part of the taxpayer, that time limit is extended by s.36 of TMA the relevant provisions of which for this appeal are:

“36Loss of tax brought about carelessly or deliberately etc

(1)

An assessment on a person in a case involving a loss of income tax or capital gains tax brought about carelessly by the person may be made at any time not more than 6 years after the end of the year of assessment to which it relates (subject to subsection (1A) and any other provision of the Taxes Acts allowing a longer period).

(1A) An assessment on a person in a case involving a loss of income tax or capital gains tax —

(a)

brought about deliberately by the person,

(b)

to (d) [omitted]

may be made at any time not more than 20 years after the end of the year of assessment to which it relates (subject to any provision of the Taxes Acts allowing a longer period).”

The language of s.36(1) and (1A) (in particular, the references to a loss of tax “brought about [carelessly] [deliberately] by” the taxpayer) mirrors the language of s.29(4). As we have already explained, s.29(4) requires HMRC to show that the “fact of the undercharge” has been brought about by the culpable conduct. If that condition is met for the purposes of s.29(4), the similarity of the statutory language suggests that there would similarly be a “case involving a loss of income tax or capital gains tax … brought about [carelessly or deliberately]” for the purposes of s.36(1) and (1A). In short, there is a clear suggestion that s.36(1) and (1A) do not require HMRC to establish anything more to discharge their Section 36 Burden than they need to discharge their Section 29(4) Burden with the exception of showing, if there is any doubt, the date on which the assessment is actually made. We are not, of course, saying that the requirements of s.36(1) and s.36(1A) can always be met by reference to s.29(4). There will be occasions on which HMRC make a discovery assessment without needing to show that the conditions of s.29(4) are met. For example, HMRC may make an ETL assessment for a period for which a taxpayer did not deliver a self-assessment return. Even if a taxpayer has made a self-assessment return for a period, HMRC may rely on s.29(5) rather than s.29(4) and so not have to establish any deliberate or careless behaviour as a pre-condition to making the assessment. Our point is simply that, if HMRC have established that the requirements of s.29(4) are met, the language of the statutory provisions suggests that they need to do nothing more to discharge their Section 36 Burden than show that the assessment was made within the 6-year or 20-year period.

As we have explained, Hurley provides no authority pointing against this linguistic analysis set out in paragraph 35 above since it dealt with different statutory provisions.

In the case of corporation tax, the equivalent provisions to ss.29, 34 and 36 of TMA are currently found in paragraphs 42(1) and (3), 43 and 46(2), (2A) and (3) of Sch.18 to the Finance Act 1998 (“FA 1998”). Following changes made by the Finance Act 2008, they are in all respects in materially the same form as the TMA provisions applying to income tax. We consider it unlikely that Parliament has ever intended that the similarly-worded discovery provisions should operate in a substantively different way as between the two taxes. Indeed, it is assumed (often without more) that a decision in relation to one tax applies to the other. An example of that is the judgment of Henderson J as he then was in HMRC vHousehold Estate Agents [2007] EWHC 1684 (Ch) which concerned corporation tax and the reasoning of which was applied to income tax/ capital gains tax by the Court of Appeal in Hankinson (see [22] and [23]).

Although the current wording of the relevant provisions for corporation tax and income tax is aligned, there was a difference in wording with regard to the issue relevant to this appeal in the period between 1994 and 2008. Section 36(1) of TMA consistently included (in its various forms) in that period a reference to making good to the Crown a loss of tax but para. 46(2) of Sch.18 to FA 1998 was in these terms when originally enacted:

“(2)

In a case involving fraud or negligence on the part of—

(a)

the company, or

(b)

a person acting on behalf of the company, or

(c)

a person who was a partner of the company at the relevant time,

an assessment may be made up to 21 years after the end of the accounting period to which it relates.”

Unlike s. 36(1) of TMA, there was no reference in para. 46(2) of Sch.18 to FA 1998 to “making good to the Crown a loss of tax”. In our judgment, it is highly unlikely that Parliament intended, by referring to “making good … a loss of tax” in s.36(1) of TMA to impose a burden on HMRC in relation to ETL assessments to income tax and capital gains tax additional to HMRC’s burden in relation to ETL assessments to corporation tax. Rather, the corporation tax provisions reinforce the linguistic analysis set out in paragraph 35 above that where HMRC have already discharged their burden under s.29(4) of TMA or para. 43 of Sch.18 to FA 1998, neither s.36 of TMA nor para. 46 of Sch.18 to FA 1998 imposes a further burden on HMRC to establish a “loss of tax”.

As mentioned above, Mr Mullens relies heavily on the judgment of the Court of Appeal in Hurley approving various dicta of Park J at first instance. The case of Hurley cannot be understood without first understanding the cases cited by Park J as authority for the various propositions of law he set out. Accordingly, we start with an analysis of those cases before turning to Hurley itself and subsequent authorities.

Hudson v Humbles (Inspector of Taxes) (1965) 42 Tax Cas 380 was concerned with what the Revenue needed to prove in a case where they made an ETL assessment.

At the time of the judgment, s.47(1) of the Income Tax Act 1952 conferred power on the Revenue to make an ETL assessment and was (so far as relevant) in these terms:

"Subject to the provisions of this section and to any provision of this Act allowing a longer period in any particular class of case, an assessment […] may be […] made […] at any time not later than six years after the end of the year to which the assessment relates […]: Provided that where any form of fraud or wilful default has been committed by or on behalf of any person in connection with or in relation to income tax, assessments [...] on that person to income tax for that year may, for the purpose of making good to the Crown any loss of tax attributable to the fraud or wilful default, be [...] made as aforesaid at any time."

In his submissions before us, Mr Goldberg KC argued that the structure and effect of this provision was different from s.36 of TMA at the time of the judgment in Hurley. In particular, there was no reference in the precondition or proviso to a loss of tax. Mr Goldberg KC is correct to say that s.47(1) of the Income Tax Act 1952 required a threshold requirement, namely the presence of “fraud or wilful default”. However, we do not agree that the absence of a reference to a “loss of tax” is significant. Section 47(1) requires that the “fraud or wilful default” must be “in connection with or in relation to income tax”. It is difficult to see how a taxpayer could be guilty of fraud or wilful default in connection with or in relation to income tax unless the taxpayer was engaged in a course of conduct that prevented at least some tax properly due from being paid. Accordingly, the concepts of “fraud or wilful default” carry with them a link to a loss of tax. Section 47(1) permitted an assessment to be made for the purpose of making good to the Crown any loss of tax attributable to the fraud or wilful default.

The Revenue were relying in Hudson on capital statements to show culpable conduct on the part of the taxpayer. Capital statements were statements setting out the estimated income and expenditure of the taxpayer and, if the declared sources of income were less than the estimated expenditure of the taxpayer, the difference was assessed to tax. If the taxpayer’s explanation for the difference was rejected, the question in Hudson was whether the Revenue had done enough to meet their burden or whether they also needed to show that the relevant amounts were income and were from the relevant chargeable source (remuneration).

Before determining that issue in the negative, it is of note that Pennycuick J began his consideration of the case by saying at p. 384:

“It is well established that, where the Revenue makes an assessment which would be out of time apart from the proviso to Sub-section (1), the burden lies upon the Revenue to establish that some form of fraud or wilful default has been committed by the taxpayer in connection with or in relation to Income Tax. If the Revenue succeeds at this stage, the burden then shifts to the taxpayer to displace the assessment - for example, on the ground that it is excessive in amount […]”.

It is implicit in that passage that Pennycuick J was proceeding on the basis, consistent with our interpretation set out in paragraph 43 above, that the Revenue had to establish that there had been a loss of tax attributable to the conduct in question. If that were not so, it is hard to see how, having succeeded at that stage, the burden would shift to the taxpayer to displace the assessment. It necessarily had to be an assessment to income tax. It is also of note that no distinction was drawn between displacing the assessment in whole and displacing it in part: the latter was mentioned in terms simply as an example.

Pennycuick J then considered the taxpayer’s submission that the Crown had not made out a prima facie case for treating particular receipts as income or as income by way of remuneration. It was contended that the sums might be capital or, if they were not, might be income from some source other than remuneration. That submission was rejected (at p.386):

“There is nothing in the proviso which in any way restricts the nature of the evidence which the Revenue must give in order to establish prima facie a case of fraud or wilful default; and it seems to me that a statement of the kind which the Inspector produced in this case is sufficient to raise a prima facie case. I do not think it is necessary for the Revenue, in order to raise a prima facie case, to show the particular quality or source of the receipts which had not been accounted for.”

In our judgment, the effect of Hudson, when read together with s.47(1) of the Income Tax Act 1952 was that (i) the Revenue bore the burden of proving a threshold condition, namely the presence of “fraud or wilful default” in connection with or in relation to income tax; (ii) to discharge that burden, the Revenue necessarily had to establish that some income tax is unpaid; (iii) to discharge that burden, the Revenue did not need to establish Constituents (i) to (iii); but instead (iv) if the Revenue could show (for example, by way of capital statements) that there was a prima facie case of income tax not being paid as a result of fraud or wilful default which the taxpayer did not satisfactorily answer, that was sufficient for the Revenue to discharge their burden and the burden then shifted to the taxpayer to show why the assessment was incorrect.

Having concluded that was so as a matter of statutory construction, Pennycuick J went on (at p.387) to say that this outcome was in accordance with the justice and common sense of the matter:

“The taxpayer knows the full facts, and the Revenue does not. In the nature of things, it must often be the case that, even if the Revenue can show a prima facie case that receipts have not been satisfactorily accounted for, it has no material upon which to set up a prima facie case for bringing the receipts in question under one or other source of income. On the other hand, it is always open to the taxpayer to challenge the assessment, not only on the ground that there has been no wilful default but also on the ground that the receipts did not represent income from the particular source selected by the Revenue.”

That judgment was approved in James v Pope (Inspector of Taxes) (1972) 48 TC 142 in a judgment given by Ungoed-Thomas J. The limited nature of the burden on the Revenue was again emphasised: “‘prima facie case’ may in the present context be used in the sense of a case which requires explanation on the part of the taxpayer of the unexplained receipts or, alternatively, in the sense of a case which requires either such explanation or explanation why such explanation cannot be given”.

The next judgment cited in Hurley was of Walton J in Johnson v Scott (Inspector of Taxes) [1978] STC. The relevant discovery and ETL provisions were now contained in TMA and were in substantively the same form as they were at the time of Hurley.

In Johnson, Walton J said at p.52h to p.53d:

“Counsel for the taxpayer’s main ground of complaint has really been, however, that there was no evidence on which the commissioners could have arrived at their determinations. […]

Now in a case such as the present it is important at the outset to determine expressly where the onus of proof lies. It was correctly accepted by counsel on behalf of the Crown, that the onus lay on the Crown to show 'neglect' (which is the only default which is here material, in view of the commissioners' findings) on the part of the taxpayer [..]. The relevance of the finding of neglect is, of course, that it enables the Crown to make assessments for the purpose of making good the loss of tax thereby caused in cases where the assessments would otherwise be out of time (see the Taxes Management Act, s 39). However, if that onus be once discharged, then the onus of displacing the assessment actually made shifts to the taxpayer; it is the taxpayer who must adduce the evidence to show that the assessment is too large (see the Taxes Management Act, s 50(6)). I venture to think that on the part of the taxpayer in the present case this vital distinction has been overlooked or, very understandably, allowed to become blurred.”

We note that, in setting out the nature of the burden on the Crown, Walton J referred only to the need to show “neglect” (the relevant culpable conduct in that case). He then explained that, if the Revenue did show that conduct, it opened the door to the making of the assessment. Once it was shown that the assessment was competent, it was for the taxpayer to displace the amount of tax assessed.

Walton J went on to find that the Commissioners were entitled to come to the conclusion that the taxpayer’s income was considerably higher than the amounts which he had disclosed. Having held that, he went on to say at p.55d:

“Of course, once they have reached that conclusion, as they did, then, as I have already noted, the onus shifts; it is then for the taxpayer to show, by proper evidence, that the assessments made on him are incorrect in whole or in part. This, to some extent, the taxpayer in fact did.”

We note, again, that the shifting of the burden allowed the taxpayer to show that the assessment was incorrect “in whole or in part”. It seems clear that Walton J was not contemplating a distinction between the whole of the amount being in dispute (with no dispute as to quantum) of the kind that would flow if Mr Mullens’ position in the Assessment Appeal is correct.

Walton J went on to observe that, in cases of a kind such as the one before him where the Revenue were seeking to draw inferences as to the amount of the taxpayer’s income, “such inferences (which counsel for the taxpayer would undoubtedly dismiss as mere guesses) are, of necessity, all that the Crown can lay before commissioners”. He continued at p.56 h to j:

“Indeed, it is quite impossible to see how the Crown, in cases of this kind, could do anything else but attempt to draw inferences. The true facts are known, presumably, if known at all, to one person only, the taxpayer himself. If once it is clear that he has not put before the tax authorities the full amount of his income, as on the quite clear inferences of fact to be made in the present case he has not, what can then be done?”

In our judgment, James and Johnson simply reiterate the conclusions reached in Hudson which we summarise in paragraph 48 above together with the rationale for them given by Pennycuick J which we have quoted in paragraph 49 above.

We can now turn to the judgment of the Court of Appeal in Hurley itself, which was an appeal against a judgment of Park J (reported at [1998] STC 202) concerning capital statements relied on by the Revenue to justify discovery assessments. Some of those assessments were ETL assessments. Park J explained that there was a difference in burdens of proof between ordinary discovery assessments and ETL assessments. As explained above, this case was decided before the introduction of self-assessment and, in the case of ordinary discovery assessments, there was no need to show any culpable conduct on the part of the taxpayer.

In the case of ordinary discovery assessments, the burden rested on the appellant (Mr Hurley) throughout in accordance with  s. 50(6)  of TMA. In the case of ETL assessments, there was an initial burden which rested on the Revenue under s.36 of TMA. In Park J’s view the Revenue had failed to discharge that initial burden. He set out a number of propositions of law, which Aldous LJ in the Court of Appeal accepted in terms: see 9j of the report at [1999] STC 1. In formulating his propositions, it is plain that Park J considered that there was no material difference in meaning between the ETL provision applicable at the time of Hudson (s.47(1) of the Income Tax Act 1952) and s.36 of TMA. Otherwise, Hudson would not have been a relevant authority.

We set out below Park J’s propositions of law omitting the parts of his judgment in relation to which the Court of Appeal differed (and which have no relevance to the appeal before us):

“My propositions of law are as follows.

1.

By  s 36(1)  of the Taxes Management Act 1970 an assessment to income tax can be made on a person outside the normal six years period (but subject to a maximum 20 years cut-off) 'for the purpose of making good to the Crown a loss of tax attributable to his fraudulent or negligent conduct'.

2.

This requires the Revenue to show: (1) fraudulent or negligent conduct by the taxpayer; and (2) a loss of tax attributable to it.

3.

On appeal to the commissioners the burden rests on the Revenue of establishing para 2(1) and (2). If they do not discharge the burden the appeal should be allowed (see e g Hillenbrand v IRC (1966) 42 TC 617 at 623 per the Lord President (Clyde)). I will call this 'the s 36 burden'.

4.

The burden does not rest on the Revenue to any greater extent than the s 36 burden. If they establish some fraudulent and negligent conduct and some loss of tax attributable to it they have satisfied s 36. From then on s 50(6) takes over and applies as it does for in-date assessments: that is to say, thereafter the burden rests on the taxpayer to establish that the assessment is wrong (see eg Johnson v Scott (Inspector of Taxes) [1978] STC 48 at 53).

5.

Reverting to the s 36 burden which rests on the Revenue, it may or may not be discharged simply by capital statements which show deficiencies. Whether it is so discharged or not depends on whether the taxpayer tenders any explanation of the deficiencies, and if he does, on how the commissioners view his explanation. [There was a further sentence here in Park J’s judgment which is not repeated because it was rejected by the Court of Appeal].

Normally it makes no difference whether a tribunal says that it rejects some item of evidence or that it does not accept it, and the two expressions are often used indiscriminately. Where, however, the burden of proof is in issue the distinction between them can be important.

6.

To be precise about a case where the Revenue produce and prove capital statements which show deficiencies:

6.1

If the taxpayer advances no explanation for the deficiencies the capital statements by themselves can, and usually do, discharge the s 36 burden (see Hudson v Humbles (Inspector of Taxes) (1965) 42 TC 380 at 386 per Pennycuick J, James v Pope (Inspector of Taxes) (1972) 48 TC 142 at 150 per Ungoed-Thomas J).

6.2

If the taxpayer advances an explanation but the commissioners reject it (that is, they positively disbelieve it) the capital statements by themselves can, and usually do, discharge the s 36 burden. Commissioners often have cases where the taxpayer gives evidence seeking to explain the deficiencies by reference to betting winnings. The commissioners listen to the evidence, including the cross-examination, and in many cases they reject it: they find it to be untrue. That, taken with capital statements which show deficiencies, is enough for the Revenue to discharge the s 36 burden. This judgment should not be understood as indicating that in my view whenever a taxpayer alleges that he won money by betting, the Revenue must produce specific evidence that he did not.

What I have said in the above paragraph is subject to 7.1 below.

6.3

[This paragraph is not repeated because it was rejected by the Court of Appeal].

7.1

If the commissioners reject the taxpayer's explanation and therefore conclude that the capital statements are themselves sufficient for the Revenue to discharge the s 36 burden, their decision may be challenged by the taxpayer on appeal to the High Court but only on the Edwards v Bairstow ground that a decision positively rejecting the explanation (as opposed to one merely not accepting it) was one which no reasonable body of commissioners could possibly reach. …”

Having approved the dicta of Park J set out above, Aldous LJ said this at p.11:

“[…] the commissioners accepted the capital statements and therefore without a cogent explanation the Revenue would discharge the s 36 onus on them. It follows that the statements raised a prima facie case which was not displaced by an explanation that was not accepted.”

Mr Mullens relies strongly on the second of Park J’s propositions to the effect that the Revenue must show both fraudulent and negligent conduct and a loss of tax attributable to it. In our judgment, that emphasis is misplaced. As we have explained, establishing that a taxpayer has behaved fraudulently or negligently in relation to tax affairs necessarily requires it to be established that some tax is unpaid as a consequence of the culpable conduct. When Hurley is read as a whole, it is clear that Park J was concerned with the same issues that arose in Hudson and James,namely whether the Revenue needed to prove the taxability of particular items of income for particular years (for example Constituents (i) to (iii)) or whether they could discharge their burden by presenting a prima facie case, based on capital statements, that the taxpayer did not adequately answer. Once that is appreciated, the conclusions expressed by Park J as approved by the Court of Appeal are no different from those reached by the High Court in the cases of Hudson, James and Johnson.

In Hargreaves v HMRC [2016] 1 WLR 2981 the Court of Appeal dismissed an appeal brought by the taxpayer asserting that he had a right to a separate hearing (as a preliminary issue) putting HMRC to proof to show that the discovery conditions in s.29 of TMA had been met. That was a case concerning the residence of Mr Hargreaves and was one, like the case before us, where the taxpayer disputed the liability to tax on substantive grounds but also disputed that the discovery assessments were validly made. All of the assessments in Hargreaves were ordinary discovery assessments.

Unlike the case of Hurley, the relevant law at the time of Hargreaves was, in material respects, the same self-assessment system as is applicable to Mr Mullens’ appeal.

Various submissions were made seeking to justify why there was a right to a separate hearing. In relation to a submission made by Mr Goldberg KC (acting, as in the case before us, on behalf of the taxpayer) as to the penal effect of a discovery assessment (abbreviated to DA in the judgment), Arden LJ (as she then was) said this:

“[46] Mr Goldberg submits that the power to make a DA is penal in its effect. He submits that, if the taxpayer makes a small mistake, the door is open to HMRC to reopen the computation of all tax for the relevant period. This is because “the situation mentioned in subs (1) above” (used in subss (2) and (5)) is that “ any  income which ought to have been assessed to income tax” has not been assessed. Thus, if the taxpayer had treated income of £100 as not liable to tax, and HMRC assesses the full £100 to tax but HMRC can show that the conduct condition is met only in respect of £50, then on a literal reading of s 29 it would appear to follow that the whole of the assessment meets the conduct/officer condition and is validly made. This is a startling conclusion.

[47] I do not consider that this difficulty exists. I accept the submission of Mr Nawbatt that, once HMRC have shown that the conduct/officer condition is met, the taxpayer can show that the amount assessed is excessive. The position under s 29 is analogous to that where an assessment is made under s36 TMA on the grounds of the taxpayer's fraudulent or negligent conduct: see per Aldous LJ in  Hurley v Taylor  [1999] STC 1, 8”.

Arden LJ then set out para. [4] of Park J’s propositions of law which we have cited at paragraph 60 above. In our view, the reference to ‘analogous’ is a clear indicator that Arden LJ considered that the law was the same for ordinary discovery assessments under s.29(4) and ETL assessments made under s.29(4) but relying on s.36 for the extended time limit. The above passage emphasises, again, that issues relating to the correctness or otherwise of the sums assessed were issues in relation to which the taxpayer bore the burden. That was the very point made by HMRC in their submission which was accepted by Arden LJ.

Mr Nawbatt KC also submitted, on behalf of HMRC, that the Supreme Court had made clear in HMRC v Tooth [2021] UKSC 17 that the sole purpose of s.36 of TMA was to impose time limits in relation to discovery assessments. He cited various paragraphs of that judgment ([1], [25], [32], [38] to [40], [44] and [83]) in support of this submission. For our part, we consider that the following passages from [25] put the point most clearly:

“[25] The relevant parts of the statutory context against which section 29 and section 118(7) fall to be construed are not themselves subject to any dispute as to their interpretation. The principal points may be summarised as follows. […] the time limits for a discovery assessment depend upon which of the two enabling conditions (or which part of them) in section 29(4) and (5) is satisfied. The second condition attracts the general time limit for assessments of four years from the end of the relevant year of assessment: see section 34(1). If the carelessness part of the first condition is satisfied, the time limit is extended to six years: see section 36(1). If the deliberate limb of the first condition is satisfied, the time limit is further extended to 20 years: see section 36(1A)(a).”

While the Supreme Court did not need to address where the burden of proof lies in relation to discovery assessments, the passage just cited is fully consistent with our analysis of the law. Section 36(1) and (1A) of TMA are, in the context of discovery assessments made under s.29(4) by reference to culpable conduct, provisions concerned only with time limits.