The Commissioners for HMRC v Sintra Global, Inc & Anor

This appeal raises an important question of principle about the burden of proof when a civil penalty is imposed by the Commissioners for His Majesty’s Revenue and Customs (“HMRC”) on a taxpayer, and the taxpayer wishes to challenge the penalty on the ground that the underlying liability to tax which underpins the penalty is wrong. The question, shortly stated, is whether the legal burden of proof rests on the taxpayer to establish in the penalty proceedings that he is not liable to the underlying tax, in the same way as it normally would on an appeal by the taxpayer against the relevant assessment; or whether the legal burden of proof rests on HMRC to establish the correctness of the underlying liability to tax, in the same way as HMRC generally have to prove all other aspects of liability to pay a penalty. In particular, does the fact that civil tax penalty proceedings in the UK are treated as giving rise to “criminal charges” within the meaning of Article 6(1) of the European Convention for the Protection of Human Rights and Fundamental Freedoms (“the ECHR” or “the Convention”) require the conclusion that the legal burden rests on HMRC to establish the correctness of the underlying liability to tax if it is put in issue by the taxpayer in the penalty proceedings?

In the present case, this critical point has been determined in favour of the taxpayers both by the First-tier Tribunal (“FTT”) and, on a first appeal, by the Tax and Chancery Chamber of the Upper Tribunal (“UT”). HMRC now pursue a second appeal to this court, with permission granted by the UT. There are three grounds of appeal, of which the first (“Ground 1”) raises the important question of principle which I have identified.

In so concluding, the UT (Edwin Johnson J, sitting with Tribunal Judge Ashley Greenbank) declined to follow the decision of a differently constituted UT (Tribunal Judges Andrew Scott and Jennifer Dean) in HMRC v Mohammed Zaman [2022] UKUT 00252 (TCC) (“Zaman”) which was released on 20 September 2022, shortly before the FTT (Tribunal Judge John Brooks and Gill Hunter) released their decision in the present case (“the FTT Decision”, [2022] UKFTT 0365 (TC)) on 5 October 2022, after a three-month hearing which had run from 4 April to 28 June 2022. The decision in Zaman was released too late for the FTT to take it into account, but it was considered at some length by the UT in their decision (“the UT Decision”, [2024] UKUT 00346 (TCC)) at [108] to [114], leading them to conclude at [114] that “the decision may be regarded as per incuriam” and was “wrong, and so we will not follow it.”

As I shall seek to explain, I have respectfully come to the opposite conclusion, and I consider that Zaman was in substance correctly decided. If the other members of the court agree, I would therefore allow HMRC’s appeal on Ground 1.

The two taxpayers with whose affairs we are directly concerned on this appeal are the second respondent, Mr Parul Malde (“Mr Malde”), and the first respondent, Sintra Global Inc (“Global”). Mr Malde is a businessman resident in the UK who at all material times was actively involved in the alcohol trade. He has carried on his business through corporate entities which he controls. One such entity was a company incorporated in Belize called Sintra SA (“SA”), which the FTT expressly found Mr Malde to have controlled during the period from 2004 to 2011. In 2011, SA transferred its business to Global, which was incorporated in Panama on 16 February 2011. In view of the conclusions it had reached on other issues, it was unnecessary for the FTT to find whether Global was also controlled by Mr Malde, but it is now common ground that the FTT would inevitably have made such a finding had it needed to do so.

It has throughout been HMRC’s basic case that SA and then Global were involved in the fraudulent diversion of alcohol into the UK from the EU between 2004 and 2014 inclusive, making use of a process generally known as “inward diversion fraud”. Like both Tribunals, I will gratefully adopt the description of this process given by Judge Falk (now Falk LJ), when sitting in the FTT, in Dale Global Ltd v HMRC [2018] UKFTT 363 (TC) at [50] to [52]:

“50.

In outline, alcohol diversion fraud is used to evade excise duty and VAT through abuse of the Excise Movement and Control System (“EMCS”), which permits authorised warehouse keepers to move excise goods from warehouse to warehouse within the EU on behalf of account holders, in duty suspense. Any movement requires the generation of an Administrative Reference Code (“ARC”) within the EMCS, which must travel with the goods. The system has operated in electronic form since January 2011. An ARC number will typically last for a few days, and expires when the load is recorded on the system by the receiving warehouse as having been being delivered.

51.

Inward diversion fraud, which is the type of fraud potentially relevant in this case, operates as follows. Alcohol originating in the UK is supplied under duty suspension to tax warehouses on the near continent, principally in France, the Netherlands and Belgium (what follows uses the example of France). Once in the tax warehouse they will usually change hands a number of times and will often be divided up before being reconstituted. A supply chain is set up with a purported end customer based in France. Some of the goods will be consigned back to the UK in duty suspense using an ARC number. This is the “cover load”. Within the lifetime of the ARC number further consignments of goods of the same description will purportedly be released for consumption in France, attracting duty at low French rates, but will in fact be smuggled to the UK using the same ARC number. These are the “mirror loads”, and this will carry on until the ARC number expires or one of the loads is intercepted by Customs, following which a new ARC number will be generated in a similar manner.

52.

Mirror loads are typically sold immediately following their arrival in the UK for cash. This process is known as “slaughtering”. The UK customers may create false paper trails to generate the impression that the goods were supplied to them legitimately.”

According to HMRC, after they had thoroughly investigated the activities of SA and Global, those two companies, when under the control of Mr Malde, had evaded UK VAT and excise duty on alcohol supplied in the UK by falsely declaring that the goods were either destined for other EU countries or were in duty suspension. It has always been common ground that neither company was ever registered for VAT or (with one exception) submitted any VAT returns, and that neither company ever accounted for any VAT or excise duty arising from its trade in the UK.

Between July 2015 and December 2017, HMRC issued a series of decisions and assessments to SA, Global and Mr Malde. These are helpfully summarised in the UT Decision at [17] in relation to SA, [18] in relation to Global, and [19] in relation to Mr Malde.

We are no longer directly concerned with SA, but it is still relevant to note the steps taken by HMRC against SA. First, by a letter dated 16 July 2015, HMRC notified SA that it was liable to be registered for VAT for the period from 1 December 2004 to 26 March 2012 under the relevant provisions of the Value Added Tax Act 1994 (“VATA 1994”). Secondly, by the same letter, HMRC made a “best of judgment” assessment of the VAT payable by SA in respect of that period under section 73 of VATA 1994, in the amount of approximately £11.75 million. Thirdly, on 20 July 2015, HMRC issued an assessment to unpaid excise duty for the same period under section 12(1) of the Finance Act 1994, in the amount of approximately £19.6 million. Fourthly, by a letter dated 8 December 2016, HMRC informed SA that they intended to charge a civil evasion penalty under section 60 of VATA 1994 in the amount of £11,162,180, as a result of SA’s dishonest failure to register for VAT and to submit VAT returns over the same period (“the civil evasion penalty”). HMRC further stated that they intended to recover 100% of the civil evasion penalty from Mr Malde by a notice under section 61, with the consequence that HMRC would not be seeking to recover any of the civil evasion penalty from SA.

The relevant provisions of sections 60 and 61 of VATA 1994, as they stood at the material time, are conveniently set out in the UT Decision at [28] and [29]. The key points to note at this stage are:

Pursuant to these provisions, HMRC duly issued a director’s liability notice (“DLN”) for 100% of the civil evasion penalty to Mr Malde on 8 December 2016.

In relation to Global, HMRC took the following steps. First, on 16 July 2015 they informed Global that it was liable to be registered for VAT between 1 April 2012 and 30 June 2015, and they made a “best of judgment” assessment to VAT in respect of that period under section 73 of VATA 1994 in the sum of approximately £8.9 million. Secondly, on the same day, they issued a penalty assessment to Global under section 123 of, and paragraph 1 of schedule 41 to, the Finance Act 2008 (“FA 2008”) in the sum of approximately £8.7 million, in relation to the failure of Global to notify HMRC of its liability to register for VAT for the same period (“the registration penalty”). Thirdly, HMRC issued an assessment to unpaid excise duty for the same period of approximately £14.18 million. Fourthly, on 11 October 2017, HMRC issued a penalty assessment to Global under schedule 24 to the Finance Act 2007 (“FA 2007”) in the same amount as the registration penalty, and as an alternative to it, in relation to an inaccurate VAT return submitted on 12 October 2016 (“the inaccuracy penalty”). Finally, on 21 December 2017, HMRC issued a penalty assessment to Global pursuant to para 4 of schedule 41 to FA 2008 in the sum of approximately £13.83 million, for handling goods subject to unpaid excise duty (“the excise duty penalty”).

It will be seen that the penalties against Global were issued pursuant to provisions contained in FA 2007 and FA 2008, which replaced the regime previously contained in VATA 1994 sections 60 and 61 which were repealed (subject to transitional provisions) by FA 2007. In what follows, references to provisions contained in the replacement regime are to the provisions in force at the relevant times.

As to the registration penalty, para 1 of schedule 41 to FA 2008 provided that a penalty is payable by a person who fails to comply with a “relevant obligation” listed in a table. The table includes, in relation to VAT, the obligation under para 1 of schedule 1 to VATA 1994 to notify HMRC of liability to register for VAT, which arises (broadly speaking) when a person who is established and makes taxable supplies in the UK has an annual turnover which exceeds a specified limit, which was then £82,000.

As to the alternative inaccuracy penalty, this was imposed under para 1 of schedule 24 to FA 2007, which applies where (a) a person (P) gives HMRC a document of a kind listed in the table in sub-para (4); (b) the document contains (broadly) a material inaccuracy which might have affected the amount of tax payable; and (c) the inaccuracy was “careless (within the meaning of paragraph 3) or deliberate”. Degrees of culpability are then set out in para 3, which provides that an inaccuracy is “careless” if it “is due to failure by P to take reasonable care”. The table in sub-para (4) includes VAT returns under regulations made pursuant to VATA 1994.

As to the excise duty penalty, this was imposed under para 4 of schedule 41 to FA 2008 which provides that a penalty is payable by P where:

“(a)

after the excise duty point for any goods which are chargeable with a duty of excise, P acquires possession of the goods or is concerned in carrying, removing, depositing, keeping or otherwise dealing with the goods, and

(b)

at the time when P acquires possession of the goods or is so concerned, a payment of duty on the goods is outstanding and has not been deferred.”

In relation to Mr Malde himself, apart from the DLN under section 61 of VATA 1994 which I have already mentioned, HMRC issued three personal liability notices (“PLNs”) making him personally liable for (respectively) the registration penalty, the inaccuracy penalty and the excise duty penalty levied on Global.

The PLN for the registration penalty was issued on 16 July 2015 under paragraph 22 of schedule 41 to FA 2008, which provided that:

“(1)

Where a penalty under any of paragraphs 1, 2, 3(1) and 4 is payable by a company for a deliberate act or failure which was attributable to an officer of the company, the officer is liable to pay such portion of the penalty (which may be 100%) as HMRC may specify by written notice to the officer.

(2)

Sub-paragraph (1) does not allow HMRC to recover more than 100% of a penalty.

…”

In the case of a body corporate, “officer” was defined in sub-para (3) as including a director or a shadow director within the meaning of section 251 of the Companies Act 2006.

The PLN for the inaccuracy penalty was issued on 11 October 2017 under similar provisions contained in para 19 of schedule 24 to FA 2007, which apply where a penalty is payable “for a deliberate inaccuracy which was attributable to an officer of the company”. The PLN for the excise duty penalty was issued on 21 December 2017, again under para 22 of schedule 41 to FA 2008.

SA did not appeal to the FTT against the decision that it was liable to be registered for VAT, the related VAT assessment, the excise duty assessment or the civil evasion penalty. Mr Malde, however, did appeal against the DLN issued to him in respect of the civil evasion penalty.

Global did not appeal to the FTT against the inaccuracy penalty or the excise duty penalty, but it did appeal against the other decisions and assessments. However, as the UT explained at [24], its appeals against the assessments to VAT and excise duty could not be entertained unless either Global deposited the disputed amounts of tax or duty with HMRC, or HMRC were satisfied (or the FTT decided) that the requirement to deposit that amount would cause Global to suffer hardship: see VATA 1994 section 83(3) and (3B) and section 16(3) of the Finance Act 1994. Global applied to the FTT for a decision that it would suffer hardship if required to deposit the tax or the duty, but this application was refused by the FTT (Judge Falk) in a decision dated 26 October 2016: see [2016] UKFTT 0726 (TC). It should be noted that the hardship provisions did not apply to Global’s appeal on the issue of liability to be registered for VAT.

Thus, the only appeals of Global which proceeded to a determination by the FTT were its appeals against (a) the decision of HMRC that Global was liable to be registered for VAT between 1 April 2012 and 30 June 2015, and (b) the registration penalty issued by HMRC on 16 July 2015.

For his part, Mr Malde appealed to the FTT against each of the three PLNs and the DLN issued against him.

It is also relevant to note, as the UT recorded at [23], that in their statements of case for each of the relevant appeals HMRC pleaded that both SA and Global “were involved in alcohol diversion fraud; selling large quantities of beer and wine in the UK without accounting for VAT and excise duty”. The sample pleading in our bundles, relating to the liability to be registered appeal, also says it is HMRC’s case that Mr Malde “set up and controlled both companies”.

The FTT Decision, produced after a three-month hearing, is unavoidably lengthy: it runs to 668 paragraphs and 130 pages. I will not attempt to summarise it in any detail, especially as that task has been usefully performed by the UT in the UT Decision at [38] to [50]. For present purposes, it will be enough to highlight a few significant points.

First, the FTT was critical of some of the evidence of the 24 officers of HMRC who gave oral evidence and were cross-examined, including in particular the evidence of Mr Dean Foster (a member of HMRC’s Fraud Investigation Service, Organised Crime – Civil MTIC/Alcohol Team) who was responsible for most of the decisions under appeal. The FTT found that his answers to questions were “frequently evasive, often obstructive, and on occasions inconsistent, contradictory and misleading.” On the other hand, the FTT was also critical of Mr Malde’s evidence, finding that much of it was “inconsistent with statements that he had previously made in interviews and/or correspondence and as such casts doubt on its veracity”.

Secondly, the FTT gave a description of evidence drawn from various criminal investigations, including evidence from “Operation Rust”, relating to the sale of alcohol by York Wines to SA; evidence from “Operation Banjax”, which resulted in the convictions of ten individuals for money laundering the proceeds of diversion frauds, in the context of payments made to and by Global and other companies controlled by Mr Malde; and evidence from “Operation Epsom”, concerning a fraud based on the sale of illicit alcohol in which Global was again implicated.

Thirdly, the FTT gave a detailed description of the methodology which Mr Foster had adopted in computing the sums in the various assessments and penalty notices in issue.

Fourthly, after summarising the procedural issues and the applicable legislation, the FTT addressed two preliminary issues: (1) the burden of proof, and (2) the order in which it would deal with the specific issues before it. The FTT discussed the burden of proof at [593] to [597], starting from the general proposition (which was not in dispute) that in tax appeals “it is for the taxpayer to establish the correct amount of tax due and … this burden of proof does not change merely because allegations of fraud may be involved”. In support of this proposition, the FTT referred to passages in three decisions of this court to which I will need to return. In reverse chronological order, they are Awards Drinks Limited v HMRC [2021] EWCA Civ 1235, [2021] STC 1576 (“Awards Drinks”); Khan (trading as Greyhound Dry Cleaners) v HMRC [2006] EWCA Civ 89, [2006] STC 1167 (“Khan”); and Brady (Inspector of Taxes) v Group Lotus Car Companies plc [1987] STC 635 (“Brady”).

The FTT then turned to the position in penalty proceedings, saying at [595]:

“In contrast to the general rule for tax assessments, it has also long been accepted that HMRC bears the burden of proving that a person is liable to a penalty (see e.g. King v Walden [2001] STC 822 at [71] and Massey v HMRC [2016] STC at [58]). In penalty proceedings, which are punitive and do not concern liability to tax, and which engage Article 6 ECHR (right to a fair trial), the normal common law on burden of proof applies, i.e. that the person who makes the allegation must prove it. It is therefore for HMRC to prove the default which is the trigger for the penalty.”

Having so directed itself, the FTT concluded at [596] that:

“Accordingly, in general terms, given the nature of these appeals, the burden of proof falls on HMRC to establish the allegations before the Tribunal and the liability to penalties.”

The FTT then noted, at [597], that the standard of proof was the civil standard, namely the balance of probabilities. This too has always been common ground.

In relation to the second preliminary issue (the order in which the issues should be determined), the FTT accepted the submission of Mr Alistair Webster KC (appearing then, as now, for the taxpayers) that they should first consider the “place of supply issue”, i.e. whether SA and Global diverted and sold alcohol in the UK, before considering whether Mr Malde was the controlling mind behind the companies: see the FTT Decision at [598] to [605]. For reasons which I confess I find it difficult to follow, the FTT agreed with Mr Webster that it would effectively reverse the burden of proof if they considered the issue of control first, notwithstanding the cogent argument of Mr John McGuiness KC (then appearing for HMRC) that if Mr Malde were found to be in control of SA and Global it would substantially undermine his primary defence that he did not control the companies and therefore could not produce their trading records.

The FTT then proceeded to consider the place of supply issue in detail, reaching conclusions which, in the event, were sufficient to dispose of all the live issues on the appeals. This may be seen from the UT Decision at [48]:

“As regards the main issues before it, the FTT reached the following conclusions.

(1)

SA was the owner of alcohol that was smuggled into the UK and sold in the UK between 2004 and 2011 (FTT [637]).

There is no express finding to this effect, but we take it as implicit in this conclusion that supplies of alcohol were made by SA in the UK in relevant periods.

(2)

In “the absence of evidence that Global was the owner of goods that were supplied in the UK”, Global was not liable to be registered for VAT in the relevant periods (FTT [644]).

As we understand it, by this finding, the FTT decided that HMRC had not discharged its burden of proof to show that Global made supplies of alcohol in the UK in the relevant periods. It followed from this conclusion that the appeals against assessments that HMRC had issued to Global (in respect of both VAT and excise duties) and the related assessments and liability notices issued to Mr Malde were allowed. This dealt with all the decisions, assessments and PLNs with the exception of the DLN.

(3)

Given its conclusion on the question of place of supply it was not necessary for the FTT to determine whether Mr Malde was the controlling mind behind Global (FTT [645]).

(4)

Mr Malde controlled SA in all relevant periods (FTT [649]).

(5)

HMRC and, in particular, Mr Foster did not “fairly consider” the evidence before them and accordingly the assessment made against SA was not made to the best of their judgment as required by section 73 VATA. The FTT considered that the failings of HMRC and Mr Foster in this regard were of such a degree that, had the assessment been appealed by SA, it would have been necessary to set it aside in its entirety “in the interests of justice” (FTT [664]).

The effect of this conclusion was that the related DLN issued to Mr Malde also fell away.”

Accordingly, the FTT allowed the appeals of Global and Mr Malde against all the relevant decisions and assessments: see the FTT Decision at [666]. Perhaps conscious that this might appear a surprising conclusion to have reached, the FTT appended this comment at [667]:

“Finally, we would adopt the following observation of Mr Webster and Mr Gurney from their closing written submissions on behalf [of]the appellants that:

‘… there can be no criticism of the fact that the Respondents decided to investigate Mr Malde, given his role in the formation of the offshore entities and their bank accounts. They generated suspicion, and that suspicion was amplified by Mr Malde’s reluctance to volunteer information (born, as it was, out of distrust of HMRC resulting from previous problems with them). The problem is that much of the above demonstrates – and clearly demonstrates, in our submission – that suspicion generated a fixed view as to the involvement of Mr Malde and a determination to make him pay which blinded the officers to the defects in their analysis. A fixed view was arrived at, despite the difficulties with the evidence, and has been persisted with from relatively early in the investigation.’

To this we would add that had HMRC, and Mr Foster in particular, taken a less myopic approach to this case, particularly with regard to Mr Malde, from the commencement of their investigations we may well have reached entirely different conclusions.”

With permission granted by the FTT, HMRC appealed to the UT on six grounds which are conveniently summarised in the UT Decision at [51], as follows:

“(1)

the FTT erred in law when it concluded ‘in general terms’ that the burden of proof fell on HMRC to establish the allegations before the tribunal and the liabilities to penalties (FTT [593]-[596]);

(2)

the FTT erred in law in its approach to the issues and evidence by compartmentalizing factors rather than examining the totality of the evidence;

(3)

the FTT’s conclusion (at FTT [643]) that Adrena ‘supplied’ the alcohol in the UK (and that Global did not) was demonstrably inconsistent with the underlying evidence;

(4)

the FTT erred in law in concluding that there was a breach of the “best of their judgment” requirement in section 73 VATA (in relation to the liability of SA, which underlies the civil evasion penalty);

(5)

in the alternative to (4), even if there had been a breach of the “best of their judgment” requirement in relation to some elements of the assessment, it was an error of law to set aside the whole assessment rather than correcting the amount to a fair figure (FTT [662]-[663]); and

(6)

the FTT failed to give any or adequate reasons with respect to a number of key matters set out in the application.”

For present purposes, the challenges to specific findings of fact, and the alleged lack of reasons, in grounds (3) and (6) are no longer relevant. It should also be remembered that an appeal from the FTT to the UT, like an appeal from the UT to this court, lies only on points of law arising from the decision under appeal: see the Tribunals, Courts and Enforcement Act 2007 (“TCEA 2007”) sections 11(1) and 13(1).

The UT began their discussion of ground (1) by observing, correctly, at [54] that “the question of where the burden of proof lies assumes some significance in this case. This is because of the manner in which the FTT reached its conclusion that Global was not liable to be registered for VAT in the UK.” Referring to the relevant passage in the FTT Decision at [642] to [644], which they set out, the UT pointed out that the FTT “relied upon HMRC’s failure to prove that Global owned any of the goods that were sold in the UK in determining that Global had not made taxable supplies in the UK and so was not liable to register for VAT in the UK. That conclusion was sufficient to determine all the appeals in favour of Global and Mr Malde, with the exception of Mr Malde’s appeal against the DLN.”

After describing the reasoning of the FTT on the burden of proof, the UT recorded at [62] the submission of Mr Hayhurst (also appearing then, as now, for HMRC) that there is “a strand of authorities which supports the proposition that in all tax cases (including appeals against penalty assessments) the burden is on the taxpayer except where: (i) the statute expressly or impliedly places the burden of proof on HMRC or (ii) where the liability for which HMRC contends requires proof of particular knowledge or state of mind on the part of the taxpayer”, e.g. the requirement on HMRC to show that the taxpayer knew or should have known of a connection with fraud in relation to cases within the Kittel principle (see the decision of the European Court of Justice in Axel Kittel v Belgian State (C-439/04) [2008] STC 1537). Support for this strand of authorities, submitted Mr Hayhurst, could be found in the cases of Awards Drinks, Khan and Brady, as well as in my decision, sitting in the UT, in Ingenious Games LLP v HMRC [2015] UKUT 0105 (TCC), [2015] STC 1659 (“Ingenious”).

The rival submissions of Mr Webster KC for the taxpayers, recorded at [64], accepted that, in relation to tax assessments, the burden of proof is on the taxpayer to show that the relevant assessment is wrong and to establish the correct amount of tax due, but he argued that, where dishonesty or fraud is expressly alleged by HMRC, they must plead, particularise and prove it in the same way as they would have to in civil proceedings: see E Buyer UK Ltd v HMRC [2017] EWCA Civ 1416 at [98] per Sir Geoffrey Vos C. Further, where a penalty is imposed on a taxpayer, HMRC bear the burden of proving liability to it unless an exception to Article 6 ECHR can be justified: see Euro Wines (C&C) Ltd v HMRC [2018] EWCA Civ 46, [2018] 1 WLR 3248 (“Euro Wines”). In the present case, however, said Mr Webster, it would not be appropriate to make an exception.

The UT then reviewed the case law, beginning with the burden of proof in tax appeals which they discussed at [67] to [90]. At [69], they observed that the rationale for the general rule “is variously expressed in the cases”. Sometimes, it follows from a statutory rule that an assessment will stand good unless it is successfully appealed, such as section 50(6) of the Taxes Management Act 1970. In other cases, it is justified on the basis that the taxpayer should be in possession of the evidence needed to prove his case. Further, in Khan Carnwath LJ suggested at [70] that the rule may be the product of a broader principle that, subject to exceptions, the burden is on the appellant in the case of an appeal against an enforcement action taken by a public authority.

The UT next considered a “more general principle” advanced by Mr Webster, consistent with the FTT’s conclusion that HMRC have the burden of proof in any case where “fraud or dishonesty is pleaded with full particularity” (see the FTT Decision at [594]). Mr Webster submitted that in the present case HMRC had pleaded fraud and thereby assumed the burden of proof, even if the relevant liability to tax did not require proof of fraud or dishonesty. The UT rejected this submission, holding at [77] that the cases relied upon by Mr Webster “fall short of establishing the principle for which [he] contends”. The UT were similarly unimpressed by the reliance Mr Webster placed on Khan, for the reasons they gave in [78]. The UT there expressed their own view “that, unless one of the well-established exceptions applies, the burden remains on the taxpayer in tax appeals and is not reversed merely because allegations of fraud or dishonesty are involved (whether as a result of HMRC pleading fraud or dishonesty as part of its case or by suggesting that a taxpayer or witness is guilty of fraud or dishonesty in meeting the taxpayer’s case).” The UT also rightly pointed out that Carnwath LJ had said much the same thing in Khan at [69]:

“There is no problem so far as concerns the appeal against the VAT assessment. The position on an appeal against a ‘best of judgment’ assessment is well−established. The burden lies on the taxpayer to establish the correct amount of tax due …

It should be noted that this burden of proof does not change merely because allegations of fraud may be involved (see e.g. Brady (Inspector of Taxes) v Group Lotus Car Companies PLC [1987] STC 635 at 642 … per Mustill LJ).”

The UT then reviewed the cases relied on by HMRC (Brady, Ingenious and Awards Drinks) before stating their conclusions on the burden of proof in tax appeals at [89] to [90], as follows:

“89.

We regard all these cases (Brady, Ingenious, Awards together with Khan) as supportive of our view. We accept that none of them directly addresses the question of whether HMRC assumes the burden of proof in relation to the essential elements of the tax liability in a case where fraud or dishonesty is not an essential element, but HMRC pleads fraud or dishonesty. However, in our view, the burden does not change in such cases; the burden remains on the taxpayer to show that the closure notice is wrong and to establish the correct amount of tax. If for any reason, HMRC plead fraud or dishonesty in such cases – for example, because HMRC are seeking such a finding from the court or tribunal because of the part that it might play in future penalty proceedings – HMRC should plead their case with appropriate particularity, and prove it. However, that does not reverse the burden of proof in relation to the essential elements of the underlying assessment.

90.

For these reasons, we do not agree with the FTT’s broad conclusion or with Mr Webster KC’s general principle.”

The UT next considered the burden of proof in penalty appeals, from [91] to [114]. I will need to consider various aspects of their reasoning on this part of the case later in this judgment. At this stage, it is enough to say that the UT:

Those paragraphs read as follows:

“112.

Having reviewed the authorities, in our view, the key cases (principally King v Walden and Euro Wines) are clear that the question of the scope of the burden of proof in penalty appeals must be determined by reference to the application of the presumption of innocence in Article 6(2) ECHR. On that basis, the burden must fall on HMRC in penalty appeals unless an exception to the presumption can be justified. Whether or not an exception can be justified will depend upon all the facts and circumstances of the case. A penalty appeal in which a penalty is challenged on the basis that the underlying assessment is wrong – once questions of estoppel or abuse of process have been addressed – remains a penalty appeal. As such, and as a general rule, the principle that the burden is upon HMRC in penalty appeals includes a penalty appeal of this kind; that is to say a penalty appeal where the challenge is made on the basis that the underlying assessment is wrong.

113.

As we have mentioned above, it is arguable that Carnwath LJ’s comments regarding the burden of proof in his judgment in Khan are strictly obiter. However, they are approved by Henderson LJ in Awards (Awards [37]). In any event, we do not regard our conclusion as inconsistent with Carnwath LJ’s judgment in Khan. In our view, his judgment falls within the exception. The relevant statute in Khan – section 60(7) VATA – defined the matters that were for HMRC to prove on any appeal. The clear implication was that the burden fell upon the taxpayer in relation to other matters. Carnwath LJ considered the implications of Article 6(2) ECHR and decided that they were adequately addressed.

114.

As regards the decision of the Upper Tribunal in Zaman, which falls within the same statutory context as the appeals against the PLNs in this case, for the reasons that we have given, in our view the decision may be regarded as per incuriam given that there is no reference in the decision to the leading cases on penalty appeals or to Article 6(2) ECHR. In any event, we respectfully disagree with the conclusion in that case. We acknowledge that, although a decision of the Upper Tribunal is not binding on a later Upper Tribunal, as a tribunal of coordinate jurisdiction we should normally follow the decision of the earlier tribunal unless we are satisfied that the earlier decision is wrong (see Gilchrist v HMRC [2014] UKUT 169 (TCC) at [94]). However, on this issue, we are satisfied that the decision in Zaman is wrong, and so we will not follow it.”

It can be seen, therefore, that the UT declined to follow Zaman, considering it to have been decided per incuriam, and the UT instead held that “as a general rule, the principle that the burden is upon HMRC in penalty appeals includes a penalty appeal of this kind; that is to say a penalty appeal where the challenge is made on the basis that the underlying assessment is wrong” [112].

In considering the application of that general rule to the facts, the UT began with the DLN issued to Mr Malde under section 61 of VATA 1994, in relation to the civil evasion penalty imposed on SA for SA’s dishonest failure to register for VAT. For the reasons given in [116] to [122], the UT found no error in the FTT’s approach to the DLN, holding at [122] that: “All the issues before the FTT were for HMRC to prove apart from the quantum of the underlying liability on which the penalty was based”. Although the underlying liability was that of SA, not Mr Malde, the UT held in [118] that the same rules applied to Mr Malde’s appeal as would have applied to an appeal by SA, with section 60(7) VATA 1994 expressly imposing the burden of proof on HMRC in relation to the matters there specified: see sub-paragraphs [10 (a) and (c)] above.

Turning to the penalties involving Global, imposed under the successor regime in FA 2007 and FA 2008, the UT at [124] set out the key issues (apart from quantum) in relation to the registration penalty charged on Global, and the three PLNs issued to Mr Malde in respect of the registration penalty, the inaccuracy penalty and the excise duty penalty respectively. In relation to the registration penalty charged on Global, the key matters to be proved were said by the UT to be that “(a) Global was required to register for VAT (i.e. its taxable supplies exceeded the VAT threshold) but failed to register; and (b) the failure was due to relevant conduct on the part of Global.”

I would draw attention to the explanatory comment which the UT added at [125]:

“125.

At this point, we should note that we have treated Global’s appeal against the decision of HMRC that it was required to register for VAT as subsumed within the appeal against the registration penalty. Although at first sight, it would appear that this is a separate appeal (and not a penalty appeal), in this case, the only practical manifestation of the decision that Global was required to be registered for VAT purposes is the registration penalty. The issues that are relevant – in particular, whether Global made taxable supplies in the UK in excess of the VAT threshold – are the same as those which form the basis of the registration penalty. In our view, any decision on the imposition of the burden of proof in relation to the decision that Global was required to be registered must therefore follow the decision on the burden of proof in relation to the registration penalty.”

At first blush, this comment seems (if I may say so) rather strange, because, as the UT recognised, Global’s successful appeal to the FTT against the decision to register it for VAT was not a penalty appeal, but a separate appeal going to liability; so it may seem paradoxical, and risk putting the cart before the horse, to subsume it within Mr Malde’s personal appeal against the registration penalty. I will need to return to this point later, which is raised by HMRC’s second ground of appeal to this court (“Ground 2”).

The UT then referred briefly to various extra-statutory materials to which Mr Webster had referred in support of his submission that Parliament did not intend to make any alteration to the burden of proof when it introduced the revised penalty regime in FA 2007 and FA 2008. The materials included Explanatory Notes for some of the legislation and extracts from HMRC’s internal manuals. The UT preferred not to base their decision on such material, however, and they explained their approach and conclusions as follows in a passage which I will set out in full:

“127.

We prefer not to base our decision on the extra-statutory materials. In our view, we can determine this issue by reference to the principles drawn from the case law to which we have previously referred. These are penalty appeals. On all of the issues relating to the penalty, the burden should fall on HMRC as they are all issues to which Article 6(2) ECHR potentially applies unless there is some good reason to justify a departure from that principle. We have considered whether the presumption of innocence in Article 6(2) ECHR should apply in these cases. This is not a case (such as Euro Wines) of a statutory exception to the general rule which can be tested by reference to whether the exception is a justifiable and proportionate response to a legitimate legislative aim. In the absence of an express statutory provision to the contrary, in our view, the general rule – that on penalty appeals the burden falls on HMRC – should apply. We can see no good reason to depart from the general principle in this case. These are significant penalties. They are dependent on proof of ‘deliberate’ conduct on the part of Global and/or Mr Malde, which even if it does not equate directly to the concept of dishonesty carries similar connotations of moral turpitude.

128.

The question for us is whether that general rule should also apply to those aspects of the appeals where the appellants’ case challenges the penalties on the grounds that call into question essential elements of the relevant tax assessment – principally, in relation to the VAT penalties, whether or not Global made taxable supplies in excess of the threshold, and, in relation to the excise duty penalties, whether Global held excise goods in the UK on which excise duty had not been paid.

129.

As we have described, those points were decided by the FTT’s conclusion that there was an “absence of evidence” that Global was the owner of the goods that were supplied in the UK. That conclusion which was clearly based on its decision on the burden of proof, in effect, decided all the penalty appeals in the appellants’ favour with the exception of the appeal against the DLN.

130.

This is therefore a similar question to that which arose in Zaman, where the Upper Tribunal decided that, in circumstances where no appeal against the underlying liability had been contested, the burden in relation to those elements should remain on the taxpayer.

131.

We will not follow that approach in this case. These are penalty appeals. The general rule is that the burden is on HMRC to prove all aspects of these appeals that relate to the imposition of the penalties consistent with the presumption of innocence in Article 6(2) ECHR. We should only depart from that rule where a justifiable and proportionate exception to the presumption can be justified on the facts and circumstances of the case. We cannot justify an exception on the facts and circumstances of this case for the following reasons:

(1)

The penalties in this case are not of a regulatory nature. They are substantial penalties that are only payable if it is shown that there has been “deliberate” conduct on the part of Global and/or Mr Malde that has caused a loss of tax.

(2)

The relevant issue in each case is whether Global was the owner of the goods in the UK. That issue is integral to the “failure” on which each of the penalties is based – the failure to register, the failure to provide an accurate return, and the failure to ensure that duty was paid on excise goods that are held in the UK.

(3)

That issue has not been the basis of a prior decision of a court or tribunal (as Global was unable to appeal following the failure of its hardship appeal) nor is it the subject of an agreed settlement.

(4)

HMRC has not raised any question of issue estoppel or abuse of law.

132.

For the sake of completeness, we note that the FTT decided each of the appeals in relation to the PLNs that were issued under the regime in FA 2007 and FA 2008 by reference to the place of supply issue. As a result, the question of the burden of proof in relation to the quantum of a penalty assessment under each of these provisions does not arise in relation to the appeals before us. We do not need to determine that question to decide these appeals and we do not do so.

133.

For all of these reasons, in our view, the FTT was right to conclude that the burden of proof fell on HMRC in relation to all relevant aspects of the penalty appeals with the possible exception of the quantum of the penalty assessment. Although it is not directly relevant, we take comfort that our view is in line with the guidance of Carnwath LJ in Khan in relation to the previous penalty regime.

Conclusion

134.

We dismiss this first ground of appeal.”

Although the UT dismissed HMRC’s appeal on ground (1), they went on to allow the appeal on grounds (2), (4) and (5), but they nevertheless declined to set aside the FTT Decision save in relation to the DLN.

The nub of ground (2) was that the FTT had erred in law in its approach to the issues and the evidence by compartmentalising factors rather than examining the totality of the evidence [135]. The focus of this ground was again on the way in which the FTT decided the place of supply issue. HMRC argued that the FTT took a wrong step when it decided to determine the place of supply issue before deciding whether Mr Malde was the controlling mind behind both SA and Global, and the FTT was thus in breach of its obligation to consider the totality of the evidence relevant to that issue [136]. This argument was accepted by the UT at [141] to [147], where they rejected Mr Webster’s arguments (recorded by the FTT at [603] to [604] of the FTT Decision) that to proceed in this way would have had the effect of reversing the burden of proof, and would have been unfair or inappropriate in all the circumstances: see [33] above. As the UT pertinently observed, at [146]:

“What the FTT was not entitled to do, given its duty to consider the totality of the evidence relevant to the place of supply, was to place the issue of control into a separate compartment, which was left unconsidered when the FTT came to consider the place of supply issue. The place of supply issue was not suitable to be treated as, in effect, a preliminary issue. There was no clear separation between the evidence relevant to place of supply issue and the evidence relevant to the issue of control. The evidence in relation to each issue, at the least, was capable of overlap.”

Nevertheless, having correctly diagnosed this error of law in the FTT’s approach, the UT declined to set aside the FTT Decision on this ground. The UT gave two reasons for this refusal in [148]. In summary, the first reason was that the FTT had decided the place of supply issue, in relation to both SA and Global, on a basis which had little connection with the issue of control. The second reason was that the FTT clearly regarded Mr Malde as an unreliable witness, and they must have taken that unreliability into account in relation to the place of supply issue. Accordingly, there was, in the view of the UT, “no reason to believe that the FTT’s failure to consider the control issue in advance of the supply issue had a material effect on the FTT’s view of the credibility of Mr Malde’s evidence in relation to the place of supply issue.”

The UT considered grounds (4) and (5) together. In outline, both grounds concerned the “best of judgment” assessment under section 73 of VATA 1994 made by HMRC against SA. Ground (4) contended that the FTT erred in law when it concluded that there was a breach of the “best judgment” requirement. Ground (5) contended that, even if there was such a breach, it was an error of law to set aside the whole assessment rather than correcting the amount to a fair figure. As the UT explained at [192] and [193], these grounds related only to Mr Malde’s appeal against the DLN and they concerned only the quantum of the underlying assessment made under section 73 on SA, which (as we have seen) SA did not appeal. In relation to all the other live appeals which had succeeded before the FTT on the place of supply issue, there had been no need for the FTT to consider the validity and quantum of the underlying assessments, including the section 73 assessment which underlay the registration penalty issued to Global and the corresponding PLN issued to Mr Malde.

Having reviewed the main authorities on “best of judgment” assessments made under section 73, and considered the parties’ submissions, the UT identified various errors of law made by the FTT which required both grounds of appeal to be allowed: [214] and [227]. The UT then considered whether to set aside the FTT Decision pursuant to section 12(1) of TCEA 2007, concluding that the errors could not be regarded as immaterial and that the FTT Decision should therefore be set aside “in so far as it relates to the DLN” [228]. That left the question of whether the case should be remitted to the FTT, or the UT should remake the decision. The UT understandably said that they were in no position to consider the validity or quantum of the assessment themselves, leaving them with no option but to remit the issues about the validity and quantum of the assessment on SA to the FTT [229].

The present position in relation to the remitter is that the remitted proceedings have been stayed pending the outcome of HMRC’s appeal to this court, to which I now turn.

HMRC pursue three grounds of appeal, for all of which permission was granted by the UT in a reasoned decision released on 17 December 2024 (“the PTA Decision”).

As I have already said, Ground 1 raises the important point of principle identified at the start of this judgment. Ground 1 reads as follows:

“the UT erred in law when it determined, save for quantum, that the burden of proof fell on HMRC with respect to ‘all relevant aspects of the penalty appeals’. HMRC submits that where a taxpayer challenges a Schedule 24 or 41 penalty assessment on the grounds that the underlying tax liability is wrong, save for the exceptions at [11] of HMRC’s skeleton argument, the burden of proof remains on the taxpayer with respect to the underlying tax liability component of the penalty.”

The exceptions allowed for in paragraph 11 of HMRC’s skeleton argument in support of the appeal are where (a) “the statute expressly or impliedly provides otherwise” or (b) “the essential elements of liability involve dishonesty or fraud in which case HMRC has the burden with respect to those specific issues”.

Ground 2 is that the UT erred in law when it subsumed Global’s “liability to be registered” appeal, which was an underlying tax liability appeal, within the registration penalty appeal (see [48] above). When a tax appeal is heard concurrently with a schedule 24 or 41 penalty appeal, then subject to the same exceptions in paragraph 11 of HMRC’s skeleton argument “the burden is always on the taxpayer with respect to both the tax appeal and the liability component of the penalty appeal.”

Ground 3 challenges as an error of law the refusal of the UT to set aside the FTT Decision when the UT decided that the FTT’s error of law in compartmentalising the evidence was not material to the place of supply issue (see [52] above). When considering whether an error of law is material, “the higher court needs to ask itself whether the error is objectively capable of being material rather than focusing on whether it was material to what was an already tainted decision.”

I find it helpful to begin my consideration of Ground 1 with a review of some of the case law of the European Court of Human Rights (“the Strasbourg court”) on the application of Article 6 to penalty proceedings in the field of taxation.

Article 6 is headed “Right to a fair trial”. For present purposes, the key provisions are the following:

“(1)

In the determination of his civil rights and obligations or of any criminal charge against him, everyone is entitled to a fair and public hearing within a reasonable time by an independent and impartial tribunal established by law.

…

(2)

Everyone charged with a criminal offence shall be presumed innocent until proved guilty according to law.

(3)

Everyone charged with a criminal offence has the following minimum rights:

…”

In the seminal case of Ferrazzini v Italy (44759/98) [2001] STC 1314 (“Ferrazzini”), the Grand Chamber of the Strasbourg court decided, by a majority, that ordinary tax disputes between the state and a taxpayer fell outside the scope of the “civil rights and obligations” limb of Article 6(1). It was common ground that such disputes did not engage the “criminal charge” limb of Article 6(1). The context was an application by the taxpayer, F, to set aside various assessments to tax arising from a transfer of assets made by him to a company of which he was the majority shareholder. He complained that the duration of the relevant proceedings had exceeded a “reasonable time” for the purposes of Article 6(1). The state (Italy) argued that Article 6 was inapplicable, because “[t]he existence of an individual’s tax obligation vis-à-vis the state belonged … exclusively in the realm of public law” (para 21 of the judgment).

The core of the court’s reasoning is in para 29:

“In the tax field, developments which might have occurred in democratic societies do not, however, affect the fundamental nature of the obligation on individuals or companies to pay tax. In comparison with the position when the convention was adopted, those developments have not entailed a further intervention by the state into the 'civil' sphere of the individual's life. The court considers that tax matters still form part of the hard core of public authority prerogatives, with the public nature of the relationship between the taxpayer and the tax authority remaining predominant. Bearing in mind that the convention and its protocols must be interpreted as a whole, the court also observes that art 1 of Protocol 1, which concerns the protection of property, reserves the right of states to enact such laws as they deem necessary for the purpose of securing the payment of taxes (see, mutatis mutandis, Gasus Dosier-und Fördertechnik GmbH v Netherlands (1995) 20 EHRR 403 at 434, para 60). Although the court does not attach decisive importance to that factor, it does take it into account. It considers that tax disputes fall outside the scope of civil rights and obligations, despite the pecuniary effects which they necessarily produce for the taxpayer.”

Accordingly, Article 6 is not engaged by tax matters which fall within the “public” sphere of a taxpayer’s life and “still form part of the hard core of public authority prerogatives, with the public nature of the relationship between the taxpayer and the tax authority remaining predominant.” It is also important to note that the concept of “civil rights and obligations” in Article 6 is an autonomous one which “cannot be interpreted solely by reference to the domestic law of the respondent state” (para 24).

Ferrazzini was decided on 12 July 2001, a few days after this court had given judgment in Han and others v HMRC [2001] EWCA Civ 1040, [2001] 1 WLR 2253 (“Han”). In Han, it was decided by a majority (Potter and Mance LJJ, Sir Martin Nourse dissenting) that in appeals brought by taxpayers in civil penalty proceedings under section 60(1) of VATA 1994 and (in one case) similar provisions relating to excise duty in section 8 of FA 1994, the “criminal charges” limb of Article 6(1) was engaged, and the taxpayers were entitled to the protection of the Article, including the minimum rights specified in Article 6(3). In his judgment, Potter LJ explained how this conclusion followed from the existing Strasbourg case-law on the “criminal charges” limb, including the influential case of Engel v The Netherlands (No1) (1976) I EHRR 647: see his judgment at [55] to [64], and [65] to [79].

One of the most recent Strasbourg cases reviewed by Potter LJ in Han was Georgiou (trading as Marios Chippery) v United Kingdom (40042/98) [2001] STC 80 (“Georgiou”), where the court ruled that a civil penalty imposed for dishonest evasion of VAT under section 13(1) of the Finance Act 1985 (the terms of which were identical to section 60 of VATA 1994) amounted to a criminal charge within the (autonomous) meaning of Article 6(1): see Han at [64]. Georgiou was a decision of the Third Section of the court. I draw attention to the following passage on p 88 of the STC report, both for what the court said about the factors which, taken together, indicated that the penalty in question was a “criminal charge”, and for their recognition that it is sometimes not possible to separate those parts of the proceedings which determined a criminal charge from those which did not:

“… The court notes that the penalty proceedings in the present case were classified as civil, rather than criminal, in domestic law. However, as in Bendenoun v France (1994) 18 EHRR 54 at 74–76, paras 44–48, the penalty was intended as a punishment to deter re-offending, its purpose was both deterrent and punitive and the penalty itself was substantial. These factors taken together indicate that the penalty imposed in the present case was a 'criminal charge' within the meaning of art 6(1).

As to whether the assessments themselves should also be seen as 'criminal charges' for the purpose of the art 6 guarantees, the applicants argue that since the penalty procedures rely on the assessments for their validity, it would be wrong not to look at the proceedings as a whole. The court accepts that it is not possible, given the various matters which were being determined by the tribunal, to separate those parts of the proceedings which determined a 'criminal charge' from those parts which did not. It will consider the proceedings to the extent to which they determined a 'criminal charge' against the applicants, although that consideration will necessarily involve the 'pure' tax assessments to a certain extent.”

The solution adopted by the court, therefore, was to consider the proceedings to the extent to which they determined a criminal charge, although such consideration would of necessity involve the “pure” tax assessments “to a certain extent”. This passage shows that, at least in some contexts, it is permissible to sever domestic penalty proceedings into parts which do, and parts which do not, involve the determination of a criminal charge within Article 6(1).

A good example of a case where such severance would probably have been impossible is provided by King v United Kingdom (No 2) (13881/02) [2004] STC 911 (“King”), which was the European sequel to the protracted and complex proceedings involving the taxpayer, Mr King, before the Special Commissioners and then (on appeal) the High Court (Jacob J) and (at one stage) the Court of Appeal. The proceedings before the Special Commissioners and Jacob J are reported as King v Walden (Inspector of Taxes) at [2001] STC 822 (“King v Walden”). They defy short summary, but they included various penalty assessments and out-of-time income tax assessments which all shared the common feature that, in one way or another, the burden was on HMRC to establish different combinations of fraud, wilful default or neglect on the part of the taxpayer. This led Jacob J to observe, at [81], that “there is no matter giving rise to penalty determinations which the Revenue have not at one point or another had to prove”. Jacob J also held (ibid) that:

“The true legal position is that the ultimate (so-called “legal”) burden at all times lay on the Revenue.”

Against this background, Jacob J held that Mr King’s Article 6 rights were engaged by the penalty proceedings, but they had not been breached. On Mr King’s subsequent complaint to the Strasbourg court, he argued that there had been a breach of his right under Article 6(1) to a “fair and public hearing within a reasonable time”, while the UK government argued that the penalties were civil and administrative in nature, did not create a criminal record and were concerned with negligence which was not typically criminal conduct, with the result that Article 6 was not engaged. It was held by the Fourth Section of the court that the complaint could not be dismissed as inadmissible, and the question whether the delay (stretching over almost 14 years) was unreasonable would have to be examined on the merits. With regard to the existence of a “criminal charge”, the court noted (at p 920 of the STC report) that the procedures concerning the assessment of tax owing by Mr King fell outside the scope of Article 6(1), applying Ferrazzini, but the penalties, calculated as a percentage of the unpaid tax, did engage the Article because “their main purpose is to exert pressure on taxpayers to comply with their legal obligations and to punish breaches of those obligations”. As in previous cases, the court did not consider it determinative that “a sentence of imprisonment was not at stake in the proceedings.”

Another significant decision of the Strasbourg court, about a year after Ferrazzini, was Janosevic v Sweden (34619/97), (2004) 38 E.H.R.R. 22 (“Janosevic”), which mainly concerned a non-criminal tax surcharge regime in Sweden. After a tax audit of the applicant’s taxi firm, the relevant Tax Authority increased his tax liabilities to reflect the audit findings and also ordered him to pay surcharges of between 20% and 40% of the increased liability. Various enforcement proceedings ensued before, having exhausted his domestic remedies, the taxpayer applied to the Strasbourg court, alleging various violations of his Article 6 rights.

The court began its assessment of the case by summarising the current state of the court’s jurisprudence in relation to tax disputes:

“64.

The Court has consistently held that, generally, tax disputes fall outside the scope of ‘civil rights and obligations’ under Art.6 of the Convention, despite the pecuniary effects which they necessarily produce for the taxpayer. The facts of the present case do not give reason to review that conclusion.

65.

Having regard to the fact that tax surcharges were imposed on the applicant, the question arises whether the proceedings in the present case instead involved a determination of a ‘criminal charge’. The Court reiterates that the concept of ‘criminal charge’ within the meaning of Art.6 is an autonomous one. In determining whether an offence qualifies as ‘criminal’, three criteria are to be applied: the legal classification of the offence in domestic law, the nature of the offence and the nature and degree of severity of the possible penalty.”

The court then examined the relevant features of the tax surcharge regime, noting at para 66 that the surcharges were treated as administrative in character under the Swedish legal system, but concluding at para 71 that they nevertheless fell within the “criminal charge” limb of Article 6(1). The court then said, at para 79:

“It is further to be noted that the relevant domestic proceedings have concerned both taxes and surcharges. In the light of its conclusion that Art. 6 does not apply to the dispute over the tax itself, the Court will consider the proceedings to the extent to which they determined a ‘criminal charge’ against the applicant, although that consideration will necessarily involve the ‘pure’ tax assessments to a certain extent.”

Clearly, this approach echoed the approach already taken by the court in Georgiou: see [66] and [67] above.

The court then considered the presumption of innocence in Article 6(2), stressing at para 96 that it “is one of the elements of the fair trial that is required by [Article 6(1)]”. After a detailed examination of the relevant Swedish legislation and procedure, the court concluded at para 104 that “the presumptions applied in Swedish law with regard to surcharges are confined within reasonable limits”, with the consequence (para 110) that “the applicant’s right to be presumed innocent has not been violated in the present case”.

On 23 November 2006, the Grand Chamber of the Strasbourg court delivered its important judgment in Jussila v Finland (73053/01) [2009] STC 29 (“Jussila”). This was another case involving tax surcharges, with the difference that they were of modest amount and imposed for relatively minor errors and deficiencies in the taxpayer’s book-keeping for his VAT returns. The court held that the surcharges were nevertheless “criminal” within the autonomous meaning of Article 6, and (following the approach in Janosevic) that the minor nature of the penalty was irrelevant: see [35]. The court added, at [36]:

“Furthermore, the court is not persuaded that the nature of tax surcharge proceedings is such that they fall, or should fall, outside the protection of art 6. Arguments to that effect have also failed in the context of prison disciplinary and minor traffic offences (see, variously, Ezeh and Connors and Öztürk, cited above). While there is no doubt as to the importance of tax to the effective functioning of the state, the court is not convinced that removing procedural safeguards in the imposition of punitive penalties in that sphere is necessary to maintain the efficacy of the fiscal system or indeed can be regarded as consonant with the spirit and purpose of the Convention. In this case the court will therefore apply the Engel criteria as identified above.”

The court also recognised, in its discussion of compliance with Article 6, that the obligation to hold an oral hearing, although of fundamental importance, is not absolute, and that national authorities “may have regard to the demands of efficiency and economy”: see [41] and [42]. The court further recognised, at [43], that “Tax surcharges differ from the hard core of criminal law”, with the consequence that “the criminal-head guarantees will not necessarily apply with their full stringency”. Finally, the court again repeated in [45] the guidance previously given in Georgiou and Janosevic about the approach to be followed in cases where “it may not be possible to separate those parts of the proceedings which determine a ‘criminal charge’ from those parts which do not.” The court must consider such proceedings “to the extent to which they determined a ‘criminal charge’ against the applicant, although that consideration will necessarily involve the ‘pure’ tax assessment to a certain extent.”

In the result, the court held by fourteen votes to three that there had been no violation of Article 6: [49].

We have been referred to a wealth of domestic case law which may throw light on Ground 1, at all levels from the FTT to the Supreme Court, so I will need to be selective in my treatment of it.

I begin with the important decision of the House of Lords in Sheldrake vDirector of Public Prosecutions [2004] UKHL 43, [2005] 1 AC 264, (“Sheldrake”) which remains the leading case on the imposition of a reverse burden of proof on the defendant in UK criminal proceedings, and the compatibility of such a burden both with Article 6 and with the common law. The House was concerned with two separate appeals, which involved the imposition of reverse burdens in very different contexts: in the case of Mr Sheldrake, under section 5(2) of the Road Traffic Act 1988; and in the other case, which came before the courts on a reference by the Attorney General (No 4 of 2002), section 11(2) of the Terrorism Act 2000. As Lord Bingham of Cornhill (who gave the leading judgment, with which Lord Steyn and Lord Phillips of Worth Matravers MR agreed) lucidly explained at [1]:

“My Lords, sections 5(2) of the Road Traffic Act 1988 and 11(2) of the Terrorism Act 2000, conventionally interpreted, impose a legal or persuasive burden on a defendant in criminal proceedings to prove the matter respectively specified in those subsections if he is to be exonerated from liability on the grounds there provided. That means that he must, to be exonerated, establish those matters on the balance of probabilities. If he fails to discharge that burden he will be convicted. In this appeal by the Director of Public Prosecutions and this reference by the Attorney General these reverse burdens (‘reverse’ because the burden is placed on the defendant and not, as ordinarily in criminal proceedings, on the prosecutor) are challenged as incompatible with the presumption of innocence guaranteed by article 6(2) of the European Convention for the Protection of Human Rights and Fundamental Freedoms (1953). Thus the first question for consideration in each case is whether the provision in question does, unjustifiably, infringe the presumption of innocence. If it does the further question arises whether the provision can and should be read down in accordance with the courts’ interpretative obligation under section 3 of the Human Rights Act 1998 so as to impose an evidential and not a legal burden on the defendant. An evidential burden is not a burden of proof. It is a burden of raising, on the evidence in the case, an issue as to the matter in question fit for consideration by the tribunal of fact. If an issue is properly raised, it is for the prosecutor to prove, beyond reasonable doubt, that that ground of exoneration does not avail the defendant.”

Lord Bingham went on to examine the Convention and the Strasbourg jurisprudence under Article 6, having introduced this part of his analysis with the following valuable observations at [9]:

“The right to a fair trial has long been recognised in England and Wales, although the conditions necessary to achieve fairness have evolved, in some way quite radically, over the years, and continue to evolve. The presumption of innocence has also been recognised since at latest the early 19th century, although (as shown by the preceding account of our domestic law) the presumption has not been uniformly treated by Parliament as absolute and unqualified. There can be no doubt that the underlying rationale of the presumption in domestic law and in the Convention is an essentially simple one; that it is repugnant to ordinary notions of fairness for a prosecutor to accuse a defendant of crime and for the defendant to be then required to disprove the accusation on pain of conviction and punishment if he fails to do so. The closer a legislative provision approaches to that situation, the more objectionable it is likely to be. To ascertain the scope of the presumption under the Convention, domestic courts must have regard to the Strasbourg case law. It has there been repeatedly recognised that the presumption of innocence is one of the elements of the fair criminal trial required by article 6(1): see for example, Bernard v France (1998) 30 EHRR 808, para 37.”

After reviewing the main Strasbourg authorities, including Salabiaku v France (1988) 13 EHRR 379, which he described at [11] as “the leading Strasbourg authority on the presumption of innocence”, and Janosevic, which he discussed at [20], Lord Bingham summed up the position, at [21], in an influential passage which has often been quoted and applied in later cases:

“From this body of authority certain principles may be derived. The overriding concern is that a trial should be fair, and the presumption of innocence is a fundamental right directed to that end. The Convention does not outlaw presumptions of fact or law but requires that these should be kept within reasonable limits and should not be arbitrary. It is open to states to define the constituent elements of a criminal offence, excluding the requirement of mens rea. But the substance and effect of any presumption adverse to a defendant must be examined, and must be reasonable. Relevant to any judgment on reasonableness or proportionality will be the opportunity given to the defendant to rebut the presumption, maintenance of the rights of the defence, flexibility in application of the presumption, retention by the court of a power to assess the evidence, the importance of what is at stake and the difficulty which a prosecutor may face in the absence of a presumption. Security concerns do not absolve member states from their duty to observe basic standards of fairness. The justifiability of any infringement of the presumption of innocence cannot be resolved by any rule of thumb, but on examination of all the facts and circumstances of the particular provision as applied in the particular case.”

Lord Bingham then reviewed the leading UK cases since the Human Rights Act 1998, from [22] to [33], before turning to the facts of the two cases and concluding that (a) in the case of Mr Sheldrake, the reverse burden did not infringe the presumption of innocence, but (b) in the case referred by the Attorney General, which concerned the offence of membership of a proscribed terrorist organisation under section 11(1) of the 2000 Act, and where the defence enacted in subsection (2) required the defendant to prove (in summary) that the organisation was not proscribed when he joined it, and that he had not taken part in its activities at any time while it was proscribed, the reverse burden on the defendant could not be justified as it stood, and it therefore had to be read down under section 3 of the Human Rights Act and given effect as imposing on the defendant an evidential burden only: see [51] to [53].

In Euro Wines, a decision of this court in January 2018, it was held that a reverse burden of proof placed on a defendant to excise duty penalty proceedings under para 4(1) of schedule 41 to FA 2008 (for which see [16] above) was compatible with Article 6. The reverse burden in question was imposed by section 154(2) of the Customs and Excise Management Act 1979 (“CEMA 1979”), which so far as material provided that:

“Where in any proceedings relating to customs or excise any question arises … as to whether or not – (a) any duty has been paid or secured in respect of any goods; … the burden of proof shall lie upon the other party to the proceedings.”

As David Richards LJ (who delivered the leading judgment, with which Patten and Gloster LJJ agreed) explained at [6], this provision and its predecessors had for many years formed part of the legislation governing proceedings relating to customs or excise. These provisions, he said, “reflect the particular difficulties in collecting duty on moveable goods, as well as the importance of the collection of duty to the public finances.” Thus, the effect of the partial reversal of the burden of proof in section 154(2) was that, while the burden would still be on HMRC to establish that the person in question had acquired possession of the goods or been concerned in their carriage etc, that the goods were chargeable with duty and that the excise duty point for the goods had passed, the burden would be on the appellant to prove that the duty had in fact been paid: [7].

The rigours of this regime were, however, mitigated not only by the opportunity provided to the person concerned to establish that the duty was paid, but also by various other provisions including a defence of reasonable excuse: [9] and [10].

On the law, the court directed itself by reference to Salabiaku v France, Janosevic and Sheldrake, including the key passage in Lord Bingham’s speech at [21]. This led the court to the “clear conclusion” that the reverse burden was compatible with Article 6(2): [32]. David Richards LJ then set out the factors that led him to this conclusion, at [33] to [43]. They included: (a) the essentially regulatory nature of the penalty, which did not require proof of any fault; (b) the fact that traders in duty goods carry on business in a sector where the payment of duty is of the highest importance, and they know they may be exposed to the risk of penalties in certain circumstances; (c) the opportunity for the trader to show that the duty had been paid, or to undertake due diligence and rely on the “reasonable excuse” defence; and (d) the fact that evasion of duty is a longstanding problem, with serious consequences for the public finances. As he said at the end of [43], “It is always a question of examining the reverse burden in its particular context and determining whether it is reasonable and proportionate”.

I come now to Khan, which was decided by this court (Buxton, Carnwath and Timothy Lloyd LJJ) in February 2006. The case concerned a dry-cleaning business of which the appellant, Mr Khan, was the sole proprietor and which was not registered for VAT. Following an investigation, HMRC issued a notice of compulsory registration for VAT with effect from 1 July 1997, and they later made “best of judgment” assessments for the period to 30 July 1999. A penalty under section 60 of VATA 1994 was also imposed, which (as we have seen) required proof by HMRC of dishonest evasion of tax by Mr Khan. On appeal to the VAT Tribunal, Mr Khan was represented by his accountant and neither he nor the accountant gave evidence. The tribunal found that the registration notice had been correctly issued, that the best of judgment assessments to VAT had been properly made (subject to a point on quantum) and that HMRC had discharged the burden of proving to the civil standard that Mr Khan had been dishonest, so the penalty assessment was also upheld in a reduced amount A subsequent appeal to the High Court was dismissed by Hart J.

On Mr Khan’s further appeal to this court, the leading judgment was given by Carnwath LJ. He recorded at [23] that the appeal raised four groups of issues, which he proceeded to consider under the headings of: (i) unfairness at an interview with Mr Khan during the initial investigation; (ii) breach of Article 6 at the hearing before the tribunal; (iii) burden and standard of proof; and (iv) irrationality. Having rejected the appeal under the first two headings, Carnwath LJ began his consideration of the burden of proof at [63]. At [66], he recorded a submission by counsel for Mr Khan, relying on the presumption of innocence under Article 6(2), that the burden was on HMRC to prove both that Mr Khan’s taxable supplies exceeded the threshold, and that the assessment to VAT was correct. Counsel submitted that this was particularly important where the assessed amount was used to support an inference of dishonesty. For their part, counsel for HMRC (Christopher Vajda QC leading Nicola Shaw) conceded that the burden of proof lay on them, not only to show that Mr Khan had exceeded the registration threshold, but also to establish the quantum of tax evaded: [67]. HMRC submitted that they had discharged the burden on both issues, for the reasons given by Hart J. They also agreed with the view expressed in his judgment (see [2005] STC 1271 at [70]) that:

“70 … in a case such as this, where the tribunal is also considering the quantum of the assessment to which the penalty is referable, the effect of that burden is likely to be academic.”

Carnwath LJ then discussed the matter in an important passage which, despite its length, I need to set out in full:

“[68] In spite of the common ground, which thus emerges from the judgment below and the submissions of the parties, I find some difficulty with this analysis. In view of the potential importance of this issue for other cases, I am reluctant to allow this judgment to rest simply on concessions, although I acknowledge that understandably the issues may not have been fully explored in argument. I will summarise my own understanding of the principles derived from the relevant statutory provisions and case law.

[69] There is no problem so far as concerns the appeal against the VAT assessment. The position on an appeal against a 'best of judgment' assessment is well-established. The burden lies on the taxpayer to establish the correct amount of tax due:

'The element of guess-work and the almost unavoidable inaccuracy in a properly made best of judgment assessment, as the cases have established, do not serve to displace the validity of the assessments, which are prima facie right and remain right until the taxpayer shows that they are wrong and also shows positively what corrections should be made in order to make the assessments right or more nearly right.' (See Bi-Flex Caribbean Ltd v The Board of Inland Revenue (1990) 63 TC 515 at 522-523 per Lord Lowry.)

That was confirmed by this court, after a detailed review of the authorities, in Customs and Excise Comrs v Pegasus Birds Ltd [2004] EWCA Civ 1015, [2004] STC 1509. We also cautioned (see [2004] STC 1509 at [38]) against allowing such an appeal routinely to become an investigation of the bona fides or rationality of the 'best of judgment' assessment made by Customs:

'Evidence to the tribunal

[38] … (i) The tribunal should remember that its primary task is to find the correct amount of tax, so far as possible on the material properly available to it, the burden resting on the taxpayer. In all but very exceptional cases, that should be the focus of the hearing, and the tribunal should not allow it to be diverted into an attack on the Commissioners' exercise of judgment at the time of the assessment …'

It should be noted that this burden of proof does not change merely because allegations of fraud may be involved (see e.g. Brady (Inspector of Taxes) v Group Lotus Car Companies plc [1987] STC 635 at 524, [1987] 3 All ER 1050 at 1057–1058 per Mustill LJ).

[70] The other rights of appeal under s 83 of the 1994 Act have not been given such detailed examination by the courts. However, the general principle, in my view, is that, where a statute gives a right of appeal against enforcement action taken by a public authority, the burden of establishing the grounds of appeal lies on the person appealing.

[71] That principle is well-established in other statutory contexts, particularly where the relevant facts are peculiarly within the knowledge of the person appealing. For example, a local planning authority may serve an enforcement notice if it 'appears' that there has been a breach of planning control. The owner can appeal against the notice on various grounds, which may, for example, include a denial of the acts complained of, or a claim that permission is not required. It has long been clear law that the burden of proof rests on the appellant. That was confirmed recently in this court in Hill v Secretary of State for Transport, Local Government and the Regions [2003] EWCA Civ 1904, [2003] All ER (D) 339 (Nov). Buxton LJ said (at [43]–[44]):

‘[43] The appellant accepted that there is a longstanding decision in planning law, Nelsovil Ltd v Minister of Housing and Local Government [1962] 1 WLR 404, [1962] 1 All ER 423, which has been generally regarded as placing the burden of proof on the appellant in an enforcement notice appeal. That view was developed in the leading judgment of Widgery J and pungently summarised by Slade J at page 409 of the report: “It is a novel proposition to me that an appellant does not have to prove his case.”

[44] … The general principle that the appellant must prove his case seems to be unassailable …’

[72] It is true that both Nelsovil and Hill were planning cases, but the statements in the former were expressed quite generally. There may of course be something in the nature of the appeal, or the statutory context, which requires a different approach. For example, under the jurisdiction of the Transport Tribunal, it was held that, whereas on appeal against refusal of a licence the burden lay on the appellant, that was reversed on an appeal against revocation of a licence (see Traffic Comr v Muck It Ltd (Secretary of State for Transport, intervening) [2005] EWCA Civ 1124, (2005) The Times, 13 October). That decision turned on the construction of the particular regulations and the European Directive on which they were based.

[73] The ordinary presumption, therefore, is that it is for the appellant to prove his case. That approach seems to me to be the correct starting-point in relation to the other categories of appeals with which we are concerned under s 83 of the 1994 Act, including the appeal against a civil penalty. The burden rests with the appellant except where the statute has expressly or impliedly provided otherwise. Thus, the burden of proof clearly rests on Customs to prove intention to evade VAT and dishonesty. In addition, in most cases proof of intention to evade is likely to depend partly on proof of the fact of evasion, and for that purpose Customs will need to satisfy at least the tribunal that the threshold has been exceeded. But, as to the precise calculation of the amount of tax due, in my view, the burden rests on the appellant for all purposes.

[74] This view is reinforced by a number of considerations: (i) it is the appellant who knows, or ought to know, the true facts; (ii) s 60(7) makes express provision placing the burden on Customs in relation to specified matters. This suggests that the draftsman saw it as an exception to the ordinary rule, and seems inconsistent with an implied burden on Customs in respect of other matters; (iii) the distinction is also readily defensible as a matter of principle. Mr Young relied on 'the presumption of innocence' under art 6 of the Convention, but he was unable to refer us to any directly relevant authority. The presumption clearly justifies placing the burden of proof on Customs in respect of tax evasion and dishonesty; but once that burden has been satisfied, a different approach may properly be applied (compare R v Benjafield [2002] UKHL 2, R v Rezvi [2002] UKHL 1 (on appeal from R v Benjafield, R v Leal, R v Rezvi, R v Milford) [2003] 1AC 1099, [2002] 1 All ER 815, in relation to confiscation orders in criminal proceedings); (iv) in relation to the calculation of tax due the subject-matter of the assessment and penalty appeals is identical. This link is given specific recognition by s 76(5) (allowing combination in one assessment). It would be surprising if the Act required different rules to be applied in each case; (v) s 73(9) provides that the assessed amount, subject to any appeal, is 'deemed to be an amount of VAT due …'. In a case where either there was no appeal against the assessment, or the penalty proceedings followed the conclusion of any such appeal, this provision would appear to preclude any attempt to reopen the assessment for the purpose of assessing the penalty. The subsection does not apply directly where, as here, the penalty appeal is combined with an appeal against the assessment, and the assessment has not therefore become final, but it indicates another link between the two procedures. (I do not see the provision as necessarily confined to enforcement, as Mr Young argues. Nor in the present context do I need to spend time on his argument that this interpretation could cause unfairness in proceedings against a third party under s 61, although I note that under that provision there appears to be a general power to mitigate the penalty.); (vi) to reverse the burden of proof would make the penalty regime unworkable in many cases. In a case such as the present, a 'best of judgment' assessment is needed precisely because the potential taxpayer has failed to keep proper records, so that positive proof in the sense required in the ordinary civil courts is not possible. The assessment may be no more than an exercise in informed guesswork. Indeed to put the burden on Customs would tend to favour those who have kept no records at all, as against those who have kept records, which are merely inadequate, but may be enough to give rise to an inference on the balance of probabilities.”

Having thus directed himself, and after dealing with a further authority which I need not mention, Carnwath LJ concluded at [77] that the tribunal’s decision was “unimpeachable”. He continued (ibid, with my emphasis):

“There was ample material to support a prima facie case that there had been evasion of tax, and that this had been intentional and dishonest. In the absence of evidence from Mr Khan to the contrary, the tribunal was clearly entitled to reach its conclusion. As to the precise amount of tax evaded, I do not see how it was possible in the absence of better records for Customs to do more than they did. Whether or not this would have been sufficient if the burden of proof had truly been on them, on the view I take of the law it was enough to support the tribunal’s conclusion.”

It seems to me clearly implicit in the words which I have emphasised that, in Carnwath LJ’s view, the legal burden of proof rested firmly on the appellant taxpayer to displace the assessments to VAT which underlay the penalty; or as he had put it at the end of [73], quoted above and again with my emphasis:

“But, as to the precise calculation of the amount of tax due, the burden rests on the appellant for all purposes.”

The reasoning of Carnwath LJ on the issue of burden of proof had the agreement of both Lloyd LJ at [85] and of Buxton LJ at [86], the latter of whom expressly endorsed “what Carnwath LJ has said about the issue of burden of proof”. I conclude that his reasoning on this issue forms part of the ratio decidendi of the appeal, and as such is binding on us sitting in this court. I acknowledge that in Awards Drinks I said at [37] that the guidance on the burden of proof given by Carnwath LJ in Khan “may have been technically obiter” in the context of an appeal against “best of judgment” assessments to VAT under section 73 of VATA 1994. With the benefit of the much fuller argument we have had in the present case, I now think that I was wrong to make that caveat; and, in any event, Awards Drinks was not a case about penalties.

In Doorstep Dispensaree Ltd v Information Commissioner [2024] EWCA Civ 1515, [2025] 1 WLR 2635 (“Doorstep Dispensaree”), this court (Newey, Asplin and Stuart-Smith LJJ) followed Khan, Brady and Awards Drinks in holding that on an appeal under section 162 of the Data Protection Act 2018 against a civil penalty notice imposed pursuant to section 155 of that Act, the burden of proof was on the appellant to demonstrate that the penalty could not stand. In so holding, Newey LJ (who gave the leading judgment, with which the other two members of the court agreed) regarded Khan as the “most helpful authority” on the burden of proof ([31]) and was influenced by the consideration that, as in Khan, it was the appellant who knew, or ought to know, the true facts: [39]. Nor was Newey LJ deterred from so concluding by the fact that, on an appeal against the penalty, section 163 of the 2018 Act would require the FTT to consider the matter “afresh” on a full merits review: see [21] and [40]. He added, at [41]:

“In my view, the burden of proof on an appeal against a penalty notice is throughout on the appellant”.

It is only fair to mention, however, that there is no indication in the report of Doorstep Dispensaree that any reliance was placed by the appellant, at least in this court, on Article 6 of the Convention or the reverse burden of proof under Article 6(2).

By contrast, full consideration was given to Article 6 in the next case I wish to mention, Omagh Minerals Ltd v HMRC [2018] UKFTT 697 (TCC) (“Omagh Minerals”). This was a decision of the Tax Chamber of the FTT (Tribunal Judge Guy Brannan). It provides a good illustration of circumstances where the kind of separation exercise envisaged by the Strasbourg court in Jussila and Georgiou may be appropriate. I would also wish to pay tribute to the high quality of the decision. The decision was on a preliminary issue in an appeal by the taxpayer company from an assessment by HMRC to aggregates levy and an associated civil penalty levied under paragraph 9(1) of schedule 6 to the Finance Act 2001. By virtue of paragraph 9(4), conduct falling within sub-para 9(1) “shall not give rise to liability to a penalty under this paragraph if the person concerned satisfies the Commissioners or, on appeal, an appeal tribunal that there is a reasonable excuse for his conduct.” The paradigm case for the imposition of a penalty under sub-para (1) was where a return was made for an accounting period which understated the taxpayer’s liability to aggregates levy, with the amount of the penalty being fixed at 5% of the understatement. Larger civil penalties could be imposed under paragraph 7 of the schedule in cases of dishonest evasion, and directors could also be made liable for penalties under the latter paragraph.

The preliminary issue was whether the substantive proceedings on the taxpayer’s appeal were of a criminal nature such as to engage Article 6, and (if so) whether Article 6 would apply to the whole proceedings, or only to the part which concerned the penalty. After discussing Ferrazzini, Georgiou and Jussila, the FTT concluded at [49] (rightly, in my respectful opinion) that the penalty under paragraph 9(1) was criminal in nature for the purposes of Article 6, despite its relatively small size: that was the critical point which the Grand Chamber had settled in Jussila. However, on the second limb of the preliminary issue, the FTT held that it was only the penalty part of the proceedings which engaged the Article: see [54] to [64].

In so holding, the FTT observed at [60] that the burden of proof usually falls on the taxpayer to displace an assessment: see, for example, Brady. The judge then said he would be reluctant to disturb such a well-established rule, which in his view was “eminently sensible”, by “holding that the criminal head of Article 6 applied to assessments as well as to penalty proceedings whenever assessment appeals are heard with penalty appeals”. It would be odd, he added in [61], if “the application of Article 6 to an assessment appeal depended fortuitously on whether an assessment appeal was heard separately or together with a penalty appeal”. The judge went on to accept at [62] that there may be cases where the proceedings cannot be separated because the two parts are “so intimately linked”, as in Georgiou where the penalty proceedings “required a finding of dishonesty after the consideration of extensive evidence at the hearing spanning 49 days – evidence which related to the assessment and the penalty”. But the appeal in the instant case was very different. The main issue was whether the rock extracted by the taxpayer from its opencast gold mine was exempt from aggregates levy. The penalty was contingent on the answer to that question, but it was a separate issue: [63]. Accordingly, Article 6 applied only to the part of the proceedings which concerned the penalty.

I come finally to the decision of the UT in Zaman. This was an appeal to the UT by HMRC in July 2022 against a decision of the FTT in 2021 which had allowed Mr Zaman’s appeal against a PLN issued to him on 26 October 2018, following an assessment to VAT against Zamco Limited (“Zamco”) of which Mr Zaman was the sole director. The assessment on Zamco was for under-declared VAT of approximately £1.9 million in periods from 2016 to 2017 on the basis that Zamco’s supplies of alcoholic goods in the relevant period took place in the UK. Zamco did not appeal against the assessment. On 23 October 2018, HMRC charged Zamco to a penalty in the sum of approximately £1.7 million under paragraph 1 of schedule 24 to FA 2007 (for which see [15] above). The penalty was likewise not appealed by Zamco. The decision of the FTT, on Mr Zaman’s appeal against the PLN, was that his evidence was inconsistent and lacking in detail, and it was likely that there had been illicit activity in his supply chains, but that Zamco itself was not the initiating or driving force in the illicit activity. Accordingly, the FTT felt driven to conclude that the appeal had to succeed on the basis that it had not been proved by HMRC, on the balance of probabilities, that the alcoholic goods in question were removed to the UK by Zamco or under its direction, or that the place of supply of all Zamco’s supplies during the relevant period was the UK.

HMRC’s grounds of appeal in the UT were that (a) the FTT erred in its approach to the burden of proof, holding that it rested solely on HMRC, and (b) the FTT erred in its evaluation of the evidence, because it failed to draw the correct inferences from its findings of fact. On the first ground, HMRC accepted that they bore the burden of proof in relation to the validity of issuing the PLN, and that Mr Zaman could appeal the PLN on the basis that the underlying assessment and the penalty were wrongfully issued: [21]. They submitted, however, that the FTT erred in treating the appeal as a penalty appeal in which the legal and evidential burden of proof remained with HMRC throughout in relation to all issues. Once HMRC had established that the PLN was validly issued, it should then have been for Mr Zaman to establish that the assessment to VAT on Zamco was wrong, in the same way as if Zamco had decided to appeal the assessment: [23].

Mr Zaman represented himself at the hearing of the appeal, which was conducted remotely by video-link. HMRC were represented by a departmental officer from the Office of the Advocate General. Mr Zaman’s case was that all Zamco’s trading transactions fell outside the scope of UK VAT, and that HMRC (upon whom the burden lay) had been unable to prove their case: [29].

The core reasoning of the UT on the first ground is contained in [34], where they said:

“However, in our judgment, HMRC are plainly right that, as per their submissions that we have set out above, if the challenge to the PLN was brought on the basis that the assessment to VAT on Zamco was wrong, the legal rules relating to the way in which the assessment could have been challenged by Zamco if it had appealed the assessment remain in play in any appeal against the PLN. As we set out above, it is well-established law that it is for the taxpayer to prove, by evidence, that an assessment to VAT issued by HMRC is incorrect. HMRC do not have that evidential burden and that cannot sensibly be affected by the fact that the challenge to the assessment occurs in satellite litigation where, as in this case, a penalty charged on Mr Zaman is sought to be defended on the basis that the assessment to VAT on Zamco was wrong. In our judgment, it is clear that the FTT lost sight of the fact that after establishing whether the PLN was validly issued, the evidential burden in relation to the assessment to VAT on Zamco shifted to Mr Zaman when he sought to positively challenge the assessment as the sole basis on which the PLN was invalidly issued: see the FTT’s overall conclusion at [96] as to whether HMRC had discharged the burden of proof: (emphasis added)

‘We thus find that it has not been proven, on the balance of probabilities, that the alcoholic goods in question were removed to the UK by Zamco or under its directions; and so, for the same reason, it is not proved that the place of supply of all of Zamco’s supplies in the relevant period was the UK, such that its VAT returns in that period contained inaccuracies. Given the burden of proof on HMRC, this means that we have to allow the appeal…’”

Accordingly, the UT held that the FTT had misdirected itself in law on the burden of proof, and the error had affected its approach in a materially relevant way: [35]. HMRC’s first ground of appeal therefore succeeded, and there was no need for the UT to address the second ground: [36]. The UT declined, however, to remake the decision in favour of HMRC, and remitted the case to the same panel of the FTT for reconsideration in the light of the UT’s conclusion: [39].

Before leaving Zaman, I would observe that, in my respectful opinion, the UT’s reasoning in [34], and their recitation of some of HMRC’s submissions in [29] and [30], appear to exhibit some confusion between the legal burden of proof and the so-called evidential burden. As Lord Bingham explained in Sheldrake at [1], an evidential burden is not a burden of proof. It is, rather, “a burden of raising, on the evidence in the case, an issue as to the matter in question fit for consideration by the tribunal of fact”. See too the helpful observations of Mustill LJ in Brady at [1987] STC 635, 643-4 which are quoted by Newey LJ in Doorstep Dispensaree at [35]. The burden of proof which normally lies on a taxpayer to displace an assessment to tax is a legal one, which means that the assessment will stand unless the taxpayer discharges the burden by adducing the necessary evidence before the tribunal of fact. The same must be true, in my judgment, of the burden which was recognised by the UT in Zaman, and of the burden which HMRC invite us to recognise in the present case. Such a legal burden may properly be described for short as a “reverse burden of proof” where it is imposed on a person who would otherwise have the protection of the presumption of innocence in Article 6(2). But the fact that a legal burden is reversed does not turn it into an evidential burden, except perhaps in the obvious and non-technical sense that the person who bears the legal burden of proof on any issue will in practice normally be obliged to adduce relevant evidence in order to discharge it.

HMRC

Mr Hayhurst prefaced his submissions on behalf of HMRC on Ground 1 by making some preliminary points. He reminded us that the five appeals now before us comprise: (a) the appeal of Global on the registration issue, which he submits is the only appeal which directly involves an underlying tax liability; and (b) the four penalty appeals, of which one is brought by Global against the registration penalty and three are brought by Mr Malde against the PLNs issued to him in respect of the registration penalty, the inaccuracy penalty and the excise duty penalty respectively levied on Global. Mr Hayhurst confirmed that HMRC do not dispute that the four penalty appeals are, at least in part, to be regarded as “criminal” for the purposes of Article 6, but without prejudice to the argument that this classification does not necessarily apply to the whole of the subject-matter of the penalty appeals if the parts of them relating to the underlying liability to tax can be separated out.

HMRC further submit that there are in principle three different types of proceedings in which penalties may be litigated. The first type is where the underlying liability to tax has already been determined in separate proceedings before the penalty is imposed and then appealed. In proceedings of that type, the appeal will be against the penalty alone and it will not generally be open to the appellant to seek to reopen the question of liability to the underlying tax, on application of the principles laid down by the House of Lords in Johnson vGore Wood & Co (a firm) [2002] 2 AC 1: see the discussion by this court of the relevant principles in play in HMRC v Kishore [2021] EWCA Civ 1565, [2022] 2 All ER 90 (“Kishore”), especially at [30] and [31] per Newey LJ.

The second type of proceedings is where the appeal on liability to the underlying tax and the appeal against the related penalty are heard concurrently in the same proceedings. In cases of this type, the normal procedure is for the issue of liability to be taken first, with the burden of proof resting on the taxpayer, and if that issue is determined in favour of HMRC the tribunal then decides the penalty issue, with the burden normally being on HMRC to establish the specific conditions which have to be satisfied for its imposition, but without revisiting the question of underlying liability to the relevant tax which has already been decided.

The third type of case is where the appeal is against the penalty alone, and there has been no previous determination of the underlying tax liability. This can happen, for example, where the taxpayer is a company which has gone into liquidation and the liquidator decides not to appeal the relevant assessment; or where the taxpayer in principle wishes to appeal the assessment, but is unable to deposit the amount of tax at stake or to satisfy the alternative “hardship” test to the satisfaction of HMRC or the FTT: see [22] above.

With that basic classification in mind, the question of principle which we have to decide is where the burden of proof lies if the appellant against a penalty, whether imposed directly on him or indirectly via a PLN or DLN, wishes to contend that the underlying liability to tax which grounds the penalty is wrong. HMRC accept that it is generally open to an appellant in penalty proceedings to challenge the underlying liability, if it would not be an abuse of process to do so on Johnson v Gore Wood & Co principles. But if such a challenge is made, does the burden lie upon the appellant to prove that the underlying liability is wrong, or is the burden upon HMRC to prove it is correct as one of the conditions which have to be satisfied for imposition of the penalty?

The correct answer to this question, say HMRC, is that in principle the burden should be on the appellant, as it would be on an appeal against the underlying assessment, and that Article 6 does not, on the facts of the present case, require the opposite conclusion. HMRC’s primary submission is that the questions of underlying liability raised by Global and Mr Malde can easily be separated from the other penalty conditions which have to be satisfied, and Article 6 has no application to the former pursuant to the principles laid down by the Strasbourg court in Ferrazzini, Jussila, King and Janosevic. If that is wrong, HMRC’s secondary submission is that, even if the presumption of innocence in Article 6(2) does apply to the tax liability component of the relevant penalties, it would not “unjustifiably infringe the presumption of innocence” (see Sheldrake at [31]) to hold that the burden of proof remains on the taxpayer with respect to the tax liability component, with the consequence that there is again no infringement of Article 6.

I will not rehearse here the detailed arguments which HMRC advance in support of their primary and secondary submissions, but I will consider them where necessary in the course of my discussion later in this judgment.

The taxpayers

The taxpayers’ basic submission is that HMRC’s Ground 1 has no merit, and there were no errors of law in the decision of the UT. In particular, they submit, the FTT was right to conclude at [596] that “the burden of proof falls on HMRC to establish the allegations before the Tribunal and the liabilities to penalties”; and the UT was likewise right to conclude that the burden of proof fell on HMRC to establish that the conditions for liability were met in relation to Mr Malde’s penalty appeals, and that the FTT Decision involved no error of law.

Although the procedural background is not in dispute, the taxpayers emphasise that the FTT was not required to consider any “underlying tax liability”, and they submit that the recurrent use of that phrase by HMRC is a fundamental mischaracterisation of the true issues in the appeals, which do not concern liability to tax, but rather liability to penalties for wrongful conduct. Further, HMRC’s allegations of inward diversion fraud were at the heart of HMRC’s case, and without proof of that fraud there was no proper basis for the penalties which were issued.

In his oral submissions to us, Mr Webster KC said several times that there were “foundational facts” which had to be proved by HMRC before any burden of proof could fall on the taxpayer. This was reflected, he said, in the fact that HMRC had agreed without argument or discussion to open the appeals, to which the taxpayers then responded, in the same way as would ordinarily happen in any other case where HMRC bore the burden of proof.

In their written submissions, Mr Webster KC and Mr Gurney argue that both tribunals were right to conclude that the burden of proof was on HMRC to establish that all the conditions for liability are met in relation to penalty appeals. They identify the “essential reasoning” of the UT, which they submit is “well supported by principle and authority”, to the effect that:

For clear statements that the burden of proof in penalty appeals lies on HMRC, the taxpayers point to the judgments of Jacob J in King v Walden at [71] (to which one might add [81]: see [69] above) and of David Richards LJ in Euro Wines at [7] (see [84] above). They also endorse what I said, albeit in a dissenting judgment, in R (PML Accounting Ltd) v HMRC [2018] EWCA Civ 2231, [2019] 1 WLR 2428, at [101]:

“an appeal … against the imposition of a penalty … engages different considerations because the proceedings are of a penal character and involve the exaction of money by the state from the taxpayer. For those reasons it is well established (and is common ground) that the burden of proof in penalty proceedings lies on HMRC, not on the taxpayer.”

The taxpayers also support the conclusion of the UT that Zaman was decided per incuriam. They point out that the judgment in that case contained no reference to Article 6(2) or to any of the leading cases on the burden of proof in penalty proceedings. They submit that the decision “reveals a paucity of reasoning”, perhaps because the taxpayer was unrepresented and the UT did not have the benefit of full argument on the issue. They say that the UT referred only to authorities concerning tax assessments, and that it failed to recognise the important distinction between an assessment to tax and a punitive penalty. Moreover, a penalty appeal can have no effect on the underlying assessment which stands and is liable to enforcement accordingly, whatever the outcome of the penalty appeal may be.

In relation to HMRC’s reliance upon Omagh Minerals, apart from the obvious point that it was only a first-instance decision by the FTT, the taxpayers submit that it concerned concurrent appeals against both an assessment to tax and a penalty. Similarly, the case of Jussila concerned an appeal against both a tax assessment and a 300 Euro surcharge, and the Strasbourg court decided that Article 6 applied to the latter, but not the former. The present proceedings, however, do not contain elements of both penalties and tax disputes, and the principles in Ferrazzini are not engaged. Furthermore, even if the possibility of dividing the proceedings into parts which do, and parts which do not, determine a “criminal charge” were potentially available, it would not be possible to perform such an exercise in the present case, any more than it was in Georgiou. Liability for the penalties under schedule 24 to FA 2007 and schedule 41 to FA 2008 depended on the central allegation that the taxpayers fraudulently smuggled alcohol into the UK and then sold it. As the UT held, that issue was integral to the “failure” on which each of the penalties was based.

In relation to the registration penalty under schedule 41, a penalty is only payable where a taxpayer has failed to comply with the obligation to register for VAT. HMRC seek to characterise that as an underlying tax issue, but it is rather a condition for imposition of the penalty which had not been determined in any other proceedings, and the resolution of which had no impact at all on any substantive tax liability. Article 6 plainly applied to resolution of that issue, and the same considerations applied to the inaccuracy penalty and the excise duty penalty.

As to HMRC’s secondary argument, the taxpayers argue that the imposition of a reverse burden of proof on them would be neither justifiable nor proportionate, but it would be both unfair and unreasonable. As the Explanatory Notes to the relevant Finance Acts and HMRC’s own internal manuals make clear, HMRC must have the evidence in their possession to justify the penalty before it is imposed, and they should then be best placed to deploy that evidence to make good their case. Requiring a taxpayer to disprove criminal allegations made by the state is inherently unfair. The penalties exceeded £30 million in total, and they were dependent upon allegations of dishonesty and fraud which are likely to cause substantial moral and reputational damage. The UT was correct to conclude that such penalties are not of a regulatory nature.

The taxpayers contrast the position in Euro Wines, where this court concluded, after a careful examination of the European and domestic authorities, and in the context of a modest regulatory penalty, that it was reasonable and proportionate to expect the trader to take reasonable steps to satisfy itself that its stock was duty-paid, but not to disprove its alleged liability to excise duty which HMRC still had to establish.

In considering the arguments ably presented to us on both sides, I find it helpful to begin with the key principles established by the Strasbourg court. Ferrazzini has stood for over 24 years as authority that ordinary tax disputes between the state and a citizen or other resident taxpayer do not attract the protection of Article 6, because they do not involve the determination of either “civil rights and obligations” or “any criminal charge” within the autonomous meaning of those expressions in Article 6(1). Thus, neither limb of Article 6(1) is engaged by ordinary disputes about liability to VAT, or about ancillary matters such as the obligation of a taxpayer to register for VAT when his taxable turnover exceeds a specified limit. By contrast, the relevant Strasbourg authorities show that tax penalties, even of modest amount, and even if they would be classified domestically as civil and/or administrative in character, are likely to fall within the “criminal charge” limb of Article 6(1) and to attract the presumption of innocence under Article 6(2).

This basic classification gives rise to an obvious problem where the imposition of a penalty under a domestic tax system is linked in some way with an underlying tax liability issue, which may or may not be determined in the same proceedings as the penalty. For example, the taxpayer who is assessed to a penalty may wish to contend by way of defence that an underlying tax liability upon which the penalty depends for its validity or amount is wrong. Does this then mean that the resolution of the underlying tax issue must be treated procedurally in the same way as the criminal part of the proceedings, even though Article 6 would not be engaged at all if separate proceedings were brought to resolve that issue?

The guidance given by the Strasbourg court on this problem is to be found principally in Georgiou, Janosevic and Jussila. I have already referred to the key passages in those authorities: see especially [67] to [68], [73] and [76] above. The guidance indicates that, where possible, the parts of the proceedings which determine a “criminal charge” should be separated from those parts which do not; and, if that is not possible, the court must consider the proceedings “to the extent to which they determined a ‘criminal charge’ against the applicant, although that determination will necessarily involve the ‘pure’ tax assessment to a certain extent.” Where severance of the proceedings is possible, the clear implication in my judgment is that Article 6 should have no application at all in relation to the non-criminal part of the proceedings. Where severance is not possible, the position is less clear-cut but I read the guidance as requiring the court to apply Article 6 to the extent necessary to determine the criminal part, but no further; and not to be deterred from doing so by the fact that it may be necessary, in the course of this exercise, to consider the non-criminal part of the proceedings to a certain extent. I find no support in these passages for the proposition, which at times the taxpayers in the present case seemed to embrace, that where severance is not possible, the court is bound to treat the whole of the proceedings as subject to Article 6. I think the true position is more nuanced than that, and the task of the court in such cases is to do the best it can to ensure that the admittedly criminal element in the case attracts the protection of Article 6, but with the minimum of prejudice to the non-criminal element which, viewed in isolation, would fall outside the scope of the Article altogether. I also find some support for this approach in the express recognition by the court in Jussila that “tax surcharges differ from the hard core of criminal law” and that “the criminal-head guarantees will not necessarily apply with their full stringency”: see [76] above. This recognition implies, in my view, that tax penalties typically occupy a position at the lower end of the criminal spectrum, and that it may be appropriate to reflect this in considering what safeguards are required by Article 6 to protect the taxpayer in such cases.

Translating these principles into the circumstances of the present appeals, my starting point is that the underlying best of judgment assessments to VAT and excise duty on Global (and, for the earlier periods, on SA) clearly fall outside the ambit of Article 6 and within the principle of Ferrazzini. So much is not disputed by the taxpayers. It has also, rightly, been common ground throughout that, on any appeals to the FTT from those assessments, the legal burden lay on the taxpayer to displace the assessment and to establish, to the civil standard of proof on the balance of probabilities, what the correct amount of VAT or duty payable was. Further, the burden of proof on the taxpayer was not affected by the fact that HMRC fully pleaded fraud in their statements of case: this is abundantly clear from Brady, Ingenious, AwardsDrinks and Khan, and was correctly recognised by the UT in rejecting the taxpayers’ argument that, by pleading fraud with full particularity, HMRC had thereby assumed the burden of proof: see [41] and [42] above, and note that the UT’s conclusion on this point has not been challenged in this court by a respondent’s notice.

I also have no doubt that similar principles (a) apply to Global’s appeal against HMRC’s decision that it was liable to register for VAT, and (b) would have applied to HMRC’s similar decision relating to SA, had SA appealed against it. The issue here was essentially one of liability to VAT, since the basic question for decision was whether the taxpayer company had made taxable supplies in the UK during the specified period above the annual turnover threshold of £82,000. I am unable to accept the taxpayers’ argument that Global’s appeal on this issue should be treated any differently, so far as concerns the burden of proof on the taxpayer, from Global’s appeals against assessments to VAT and excise duty. There was, it is true, one relevant procedural difference between the assessment appeals and the registration appeal, in that Global was entitled to pursue an appeal on the latter without having to deposit the tax at stake or establish hardship (see [22] above); but this was a difference which operated solely in favour of the taxpayer, and it did not reflect any special feature of the underlying liability to tax which triggered the obligation to register.

Turning now to the penalty appeals, I emphasise that HMRC now accept, in both their written and their oral submissions, that the burden lies on them to establish any special or bespoke features of the relevant legislation upon which liability to the penalty depends, apart from any question about liability to the underlying tax itself whether on the part of the individual taxpayer or on the part of a company of which the taxpayer is a director. So, for example, I understand HMRC to agree that in the case of Global’s appeal against the registration penalty, it would be necessary for HMRC to establish the basic conditions of liability under para 1 of schedule 41 FA 2008, including Global’s failure to notify HMRC of its liability to register for VAT; and in the case of Mr Malde’s appeal against the linked PLN making him personally liable for the whole of the registration penalty, the burden would likewise be on HMRC to establish the further conditions specified in para 22 of the schedule, namely that there had been a “deliberate act or failure” which was attributable to him as an officer of the company, his status as such an officer (as defined) and the reasons why HMRC decided to specify the maximum of 100% as the portion of the penalty which he should be liable to pay. All of this follows, in my judgment, from the long-established general principle that the legal burden rests on HMRC to establish the ingredients of liability to a penalty.

To take another example, in the case of the excise duty penalty levied on Global, against which Global did not appeal, I consider that on Mr Malde’s appeal against the linked PLN issued to him the burden was again on HMRC to establish the further conditions contained in para 22 of the schedule which had to be satisfied in relation to the PLN. If, however, a question were to arise whether the excise duty in question had been paid, there would then be a partial reverse legal burden on Mr Malde under section 154(2) of CEMA 1979 to establish the fact of payment: see generally Euro Wines, in particular at [4] to [7]. I should also mention here an argument advanced to us by Mr Hayhurst in his oral submissions that David Richards LJ mis-stated the law in Euro Wines at [7], where he said that the effect of the partial reversal of the burden of proof under section 154(2) was that the burden would still be on HMRC to establish the basic ingredients of liability to the penalty under para 4 of schedule 41, namely that the person in question had acquired possession of the goods or been concerned in their carriage etc, that the goods were chargeable with duty and that the excise duty point for the goods had passed: see [84] above. I will return to this criticism at [139] below.

On the one hand, it may at first sight seem intuitively reasonable that the burden should remain with HMRC on these matters if they wish to uphold the imposition on Mr Malde personally of a penalty in excess of £13.8 million, in addition to the unappealed penalty of the same amount for which Global was (by default) already liable and which HMRC were in principle able to enforce against Global. On the other hand, it seems clear that if Global had pursued an appeal against the original assessment to unpaid excise duty of approximately £14.18 million (see [12] above), the burden would have been on Global to displace the assessment in the usual way. So why, it may be asked, should Global be relieved of bearing the burden of proof on those matters merely because the question arises in the context of a linked PLN issued to Global’s controlling director? This question brings us to the issue at the heart of the appeal, and to HMRC’s core submission that where in penalty proceedings under schedules 24 or 41 of FA 2007 and FA 2008 respectively a taxpayer wishes to challenge the underlying tax liability on which the penalty is based, the legal burden should rest on the taxpayer to show that the underlying tax liability is wrong, in the same way as it would on an appeal by the taxpayer (or a related company) against the relevant assessment.

I have not found this an easy question, but on balance I have concluded that HMRC’s core submission should be accepted. The factor which weighs with me most strongly in so concluding is the arbitrary and anomalous results which would follow if the legal burden of proof on the issue of underlying liability admittedly rests on the taxpayer if the question is determined in separate proceedings, typically before imposition of the penalty, or if it is determined in concurrent proceedings at the same time as the penalty; but not if it is determined in separate penalty proceedings where, if the taxpayers are right, and as both the UT and the FTT have held, the burden lies on HMRC to establish every aspect of the penalty, including any challenge to the underlying tax liability. The taxpayers seek to justify this result by arguing that the imposition of penalties differs in principle from the resolution of ordinary tax disputes, and it is therefore unsurprising if the burden rests on HMRC to establish every element of liability to the penalty (subject to any express statutory provision to the contrary). But this argument, although superficially attractive, does not in my opinion meet the objection of principle that such an approach would indirectly subvert the long-established and salutary rule that the legal burden of proof normally lies on the taxpayer to displace an assessment to tax.

It cannot be right, in my view, that the burden should shift to HMRC merely because the taxpayer wishes to raise the issue of an underlying liability to tax as a defence to penalty proceedings. Not only would such a result be inconsistent with the principles which justify the normal rule, the most important of which is that the taxpayer should normally have access to all the information needed to resolve his tax affairs, but it would give rise to the risk of inconsistent decisions if the same question of underlying liability were determined differently depending on the stage at which it arises. Moreover, it would risk creating a perverse incentive for the taxpayer to avoid appealing an underlying assessment, when the burden would on normal principles be on the taxpayer to displace the assessment, but then to raise the issue as a defence to related penalty proceedings, when the burden would be on HMRC to justify the assessment. That risk is all the more apparent when the underlying assessment is on a company controlled by the taxpayer, which may well have no or insufficient assets to satisfy a tax judgment, and a penalty of approximately the same amount is then imposed on the taxpayer via a PLN or DLN.

The right solution in principle to this conundrum, in my judgment, is that in penalty proceedings the burden should normally lie on HMRC to establish the primary facts which need to be proved to justify the imposition of the penalty, but if the taxpayer wishes to contend in his defence that an underlying liability to tax which underpins or is reflected in the penalty was wrong, a separate legal burden rests on him to prove it, in the same way as it would on an appeal to the FTT against the relevant assessment or decision.

Such a solution would have the advantage of not engaging Article 6 in relation to this part of the taxpayer’s defence, because (as I will explain) I would also accept HMRC’s primary submission that a defence of this nature can usually be severed without difficulty from the rest of the penalty proceedings. Furthermore, even if that is wrong, I would accept HMRC’s secondary submission that the imposition of a reverse burden of proof on the taxpayer in relation to the underlying tax liability would be both reasonable and proportionate, and would satisfy the principles derived by Lord Bingham from the relevant case law in Sheldrake at [21]: see [81] above.

In a little more detail, I agree with HMRC that in relation to the penalties under the regime contained in schedule 41 FA 2008 and schedule 24 FA 2007 which are the subject of this appeal, there should usually be no difficulty in separating the underlying tax liability components which do not engage Article 6 from the bespoke penalty conditions that do. Thus, in the case of the registration penalty, the underlying tax component in para 1 of schedule 41 is that a trader failed to register for VAT. This component can be isolated, it mirrors the liability issue in the underlying tax dispute, and it is not intertwined with any other conditions. By contrast, penalties under schedule 41 are allocated different levels of culpability under para 5, which classifies failures to comply under various headings as either “deliberate and concealed” or “deliberate but not concealed”. HMRC accept that these elements, like other bespoke conditions, are for them to prove because they do engage Article 6.

Similarly, the inaccuracy penalty was issued under schedule 24 FA 2007, which was the same penalty regime as was considered in Zaman. The underlying tax liability components are set out in paras 1(a) and 2, namely that a trader submitted a VAT return containing an inaccuracy and the inaccuracy led to a loss of tax. These components can be isolated, and they follow from determination of the underlying tax liability: it is inaccurate to submit a nil VAT return, as Global eventually did, and there will inevitably have been a loss of tax if the trader was liable to be registered on the basis of taxable supplies made in the UK but failed to do so. These elements can easily be separated, and Article 6 does not apply to them. By contrast, para 1(3) of schedule 24 sets out the further requirement (“Condition 2”) that the inaccuracy was either “careless (within the meaning of paragraph 3) or deliberate”. This test of culpability does not form part of the underlying tax liability, and it can easily be separated. It does engage Article 6, and the burden of proof on it lies on HMRC.

A possibly more problematic example is provided by the excise penalty under schedule 41 FA 2008, which had the more complex underlying tax component set out in para 4(1) of the schedule: see [16] above. HMRC submit, without contradiction, that this component of liability to the excise penalty mirrors in substance (although the language is not identical) the liability element of an excise assessment made under section 12(1) FA 1994 and associated regulations. On that footing, HMRC submit that this component does not engage Article 6, and it can be separated from the levels of culpability specified in para 5 which do engage the Article. However, while it is clear that this would have been the position on a direct appeal by Global against the assessment of unpaid excise duty, it is less obvious that this was also the position in relation to the excise duty penalty, and in Euro Wines David Richards LJ expressly held at [7] that, on comparable facts, the burden rested on HMRC to establish the ingredients of liability set out in para 4 in linked penalty proceedings. I will return to this point at [139] below, but I will say now that I agree with HMRC’s analysis in relation to the excise penalty. In my judgment, if Mr Malde wished to challenge the ingredients of underlying liability specified in para 4, a legal burden rested on him to make good the challenge and that part of the proceedings should be severed with the consequence that it did not attract the protection of Article 6. This result would also harmonise with the reverse burden expressly imposed on him by CEMA 1979 section 154(2) to show that the relevant duty had been paid.

HMRC make the further point, in this context, that the establishment of an underlying tax liability does not necessarily lead to the imposition of a penalty: in schedule 24, this follows from the fact that the taxpayer’s conduct must at least be “careless” if it is to be penalised; while schedule 41 provides a statutory defence of “reasonable excuse” in para 20 in relation to any act or failure which is “not deliberate”, the burden being on the taxpayer to establish the defence.

Since I would accept HMRC’s primary case on severability, I will deal with their alternative secondary case more briefly. In support of their secondary case, HMRC rely (among other matters) on the recognition by the Strasbourg court in Jussila that tax penalties “differ from the hard core of criminal law” and may justify a less stringent approach to protection of the taxpayer. They also rely on the recognition by this court in Khan of the general principle that where a statute gives a right of appeal against enforcement action taken by a public authority, the burden lies on the appellant to prove his case, and on the approval of that principle by this court in Doorstep Dispensaree. HMRC further rely on the good reasons why the burden of proof on appeals against assessments generally lies on the taxpayer; on the legitimate aim of the state to collect tax; on the features of the alcohol trade emphasised by this court in Euro Wines; and on the anomalies which I have already identified if the burden were to lie on HMRC to establish the underlying tax liability in penalty proceedings. HMRC also point to the protections and safeguards for the taxpayer which remain in place, including the right to a determination by an independent tribunal (the FTT) which will give a fully reasoned decision after consideration of all the evidence.

It is enough to say that I find HMRC’s secondary case convincing, if it is necessary to go that far, and that I am unpersuaded by the counter-arguments advanced by the taxpayers.

I recognise that the taxpayers can find some apparent support in the authorities for the general proposition that the burden of proof in penalty proceedings lies on HMRC in relation to all issues, subject to any statutory provision to the contrary. But the issue raised by Ground 1 has not previously been considered in any cases above the level of the UT, and such general statements must be read and understood in their context. For example, the taxpayers place considerable reliance on some of the statements made by Jacob J in King vWalden (see [69] above), but the penalties and out-of-time assessments there in issue all expressly or impliedly placed the burden of proof on HMRC in one way or another, and the underlying elements of liability involved either fraud or dishonesty. That case is therefore of no assistance in the present context. Similarly, my general statement in the PML Accounting case (see [114] above) was not only obiter in a dissenting judgment but also did not purport to do more than record a general understanding which was common ground. The possible existence of the exception for which HMRC now contend was not in issue and was not raised in argument, so in the present context my statement clearly has no value as a precedent.

I must, however, return to what David Richards LJ said in Euro Wines at [7] and HMRC’s criticism of it: see [126] above. The first point to be made is that the case concerned the validity of the reverse burden of proof imposed on the taxpayer by section 154(2) of CEMA 1979, and the court’s reasoning was therefore principally directed to that issue. It was unnecessary for the court to decide where the burden of proof lay in relation to the other elements of liability to the penalty under para 4(1) of schedule 41 to FA 2008. The judge’s observations on that subject in [7] were therefore obiter. Secondly, the judge’s observations merely reflected the conditions of liability as set out in para 4(1) of the schedule, and there is no indication that the question in the present case was raised before the court in any shape or form. The nearest approach to it was in the appellant taxpayer’s first ground of appeal, which was that the UT had “failed to recognise or to attach appropriate weight to the fact that pursuant to paragraph 4(1) of Schedule 41 to the 2008 Act the revenue had to conduct an investigation and confirm that duty was outstanding on the goods and so was the appropriate party to prove that duty had not been paid”: see [2018] 1 WLR 3248 at 3249G and the judgment at [29]. David Richards LJ no doubt had this ground of appeal and counsel’s submission in support of it well in mind when he said what he did in [7], and I need hardly add that any statement of his must have strong persuasive force, especially as the other members of the court (Patten and Gloster LJJ) agreed with his judgment without qualification. However, with the benefit of the detailed argument in the present case I have to say that, in my respectful view, the dicta of David Richards LJ about the burden of proof on HMRC in [7] were mistaken and should not be followed.

In the light of what I have already said, I repeat that in my view Zaman was in substance correctly decided, although the misleading references to an evidential burden of proof should be read as referring to a legal burden: see [102] above. I would reject the taxpayers’ submission, accepted by the UT, that it was decided per incuriam, although I accept that the decision did not deal with the Article 6 aspects of the case and did not refer to many of the cases to which we have been referred. In my judgment, however, the decision well captured at [34] the central points which show why HMRC’s core argument in the present case should be accepted. We were also informed that the decision has been followed and applied, without apparent difficulty, in a number of later decisions of the FTT.

I do, however, question whether the concept of “per incuriam” is a helpful one to deploy in the present case, where as a matter of precedent the decision in Zaman was not binding on the UT, although as they rightly recognised at [114] they should normally follow a decision of an earlier UT “unless we are satisfied that the earlier decision is wrong”. The UT were so satisfied and decided not to follow Zaman, as they were fully entitled to do. But there was no need to stigmatise the earlier decision as having been reached per incuriam, which is a stringent doctrine normally invoked to justify a departure from the principle of stare decisis. There is a good deal of case law on the subject to which we were not referred, so I prefer to say no more about it.

Finally, I hope that the members of the UT and the FTT will forgive me if I do not go through the relevant parts of the UT and FTT Decisions to explain why I have taken a different view from theirs on Ground 1. The question is ultimately a short but important one of principle, and I have done my best to explain why I consider that HMRC’s approach to it is correct in law. The court is very grateful to both Tribunals for the care and clarity with which they examined the question, but it would not be helpful to prolong this judgment with a detailed critique of their reasoning.

In the result, if the other members of the court agree, I would allow HMRC’s appeal on Ground 1 and make a declaration in terms which reflect [130] above.

The focus of Ground 2 is on the reasoning of the UT in [125] of the UT Decision, which I have set out at [48] above and briefly commented upon at [49]. The UT there explained why they had decided to treat Global’s appeal against the decision of HMRC that Global was required to register for VAT (“the registration decision”) as subsumed within Global’s appeal against the registration penalty. Both the registration decision and the registration penalty were issued and communicated to Global by separate letters dated 16 July 2015, together with the linked “best of judgment” assessment to VAT under section 73 of VATA 1994 which was included in the same letter as the registration decision. In each letter Global was duly informed of its right to appeal to the FTT within 30 days. Global exercised this right, and (as I have already explained) did not have to pay the tax in dispute or establish hardship in order to pursue its appeal against the registration decision and the registration penalty, although it did have to (but failed to do so and was unable to establish hardship) in relation to the VAT assessment. The basis of calculation of the registration penalty was set out in a schedule to the penalty assessment letter, on the footing that the behaviour which led to the failure to register was “deliberate and concealed”.

In the light of my decision on Ground 1, I consider that the burden of proof lay on Global to establish to the civil standard of proof on appeal to the FTT that (a) the registration decision was wrong; (b) the section 73 assessment to VAT was wrong; and (c) the penalty assessment was also wrong, because Global had no underlying liability to VAT and no obligation to register for VAT during the relevant period from 1 April 2012 to 30 June 2015. Since Global was unable to pursue its appeal against the VAT assessment, the burden on it in the penalty proceedings was in practice the same as it would have been had the VAT appeal been pursued. Such a result seems to me to accord both with principle and with the authorities correctly understood.

By contrast, the burden of proof lay on HMRC in the penalty appeal to satisfy the FTT, again to the civil standard of proof, that the basic ingredients of liability to the penalty (apart from the issue of underlying liability) had been satisfied, including (if it was not admitted) the fact of Global’s failure to register for VAT, that Global’s conduct had been “deliberate but concealed” within the meaning of the relevant paragraphs of schedule 41 (see in particular paras 5(1)(a) and 6(2)(a)) and that the amount of the penalty had been appropriately calculated as set out in the schedule to the relevant letter, including the small reduction allowed for “prompted disclosure” which reduced the level of the penalty from 100% of the potential lost revenue to 97.5%. See too my comments at [125] above about the burden on HMRC in relation to Mr Malde’s linked appeal against the corresponding PLN issued to him, again on 16 July 2015, under para 22 of schedule 41.

Against this detailed procedural background, I think that the UT clearly erred in principle when it decided to assimilate Global’s appeal against the registration decision with its appeal against the penalty. They were separate appeals against separate decisions, taken in the context of the elaborate new regime for penalties enacted by Parliament in the Finance Acts of 2007 and 2008. The ingredients of each appeal should have been analysed and dealt with separately when considering where the burden of proof should lie, and the assimilation of the two appeals ran the risk of imposing on HMRC a burden of proof in relation to the underlying tax liability which cannot in my judgment be justified.

None of the reasons given by the UT for taking this course appear to me to withstand scrutiny. In particular, I can see no principled basis for saying that the burden of proof in relation to the registration decision must follow the burden of proof in relation to the registration penalty merely because the issues which arise in each case may be the same.

For these short reasons, I would allow HMRC’s appeal on Ground 2.

The gist of Ground 3 is summarised in [60] above. It challenges the refusal of the UT to set aside the FTT Decision on the ground that the FTT’s unduly compartmentalised approach to the evidence, which the UT had correctly diagnosed, and against which there has been no further appeal, was not in their view material to the place of supply issue.

As a preliminary comment, I would observe that the impact of Ground 3 may in practice be small in view of my conclusion on Ground 1, because the errors of both Tribunals in relation to the burden of proof, if I am right on Ground 1, are so substantial that I can see no realistic alternative to remitting all the open appeals to the FTT for a rehearing, at least in relation to the place of supply issue. On such a rehearing, the question of burden of proof will have to be revisited in the light of this court’s decision, and the evidence will have to be led and considered afresh, subject to any directions which may be given to reduce the scope of the new evidence in the interests of case management, the avoidance of duplication, and the saving of resources. The practical significance of the point raised by Ground 3 may therefore be lessened, because the approach taken to it by the UT in the UT Decision will no longer have any binding force. At best, it would be of persuasive value only.

Nevertheless, I will deal briefly with Ground 3, because the point of principle which it raises may become relevant on the remitted hearing and here too I am persuaded by HMRC’s argument that the UT took a wrong step. The appropriate test to apply on a question of materiality can usually be encapsulated by asking whether the tribunal of fact might, not would, have reached a different conclusion if it had not misdirected itself in law. If that test is applied to the FTT Decision, it seems to me all but self-evident that the FTT might have reached a different conclusion on the all-important place of supply issue if it had not decided to treat it without reference to the question whether Mr Malde was in control of SA, Global and other key corporate entities at the material times. If nothing else, his control, if established, would have seriously, and perhaps fatally, undermined the contention which he placed at the forefront of his defence that he knew nothing about the disputed transactions and was unable to produce any documentary evidence from his companies relating to them. It is elementary that the evidence must be considered as a whole by the tribunal of fact, but the FTT in effect disabled themselves from doing this by their misguided decision to examine the place of supply issue in isolation from the question of control, regardless of where the burden of proof on it lay.

The taxpayers’ oral arguments on Ground 3 were skilfully and attractively presented by Mr Gurney, who argued this part of the case for them; but in my view he had no answer to the simple point at the heart of HMRC’s appeal on this issue. The basic error made by the UT, if I may say so, was to attempt to assess the question of materiality on a subjective rather than an objective basis, and to try to make it fit in with the basic approach of the FTT which was already tainted both by their mistaken view of where the burden of proof lay and by their over-compartmentalised treatment of the evidence.

I therefore conclude that HMRC’s appeal should also be allowed on Ground 3.

For the reasons I have given, I would allow HMRC’s appeal on all three grounds. Apart from making a suitable declaration to give effect to my conclusion on Ground 1, I can see no escape from the need to remit all the open appeals to the FTT for a rehearing, at the very least on the place of supply issue. It is also my provisional view, in agreement with brief submissions made to us by Mr Hayhurst after he had opened the appeal, that the rehearing should take place before a differently constituted panel of the FTT. I would however invite the parties to make brief written submissions on that question, as well as on the scope of the remitter and on the question whether any further directions about the rehearing should be given at this stage, bearing in mind the more limited remitter which has already been ordered by the UT and which is currently stayed: see [56] above. I would propose that the written submissions should be filed and served within 7 days after our draft judgments are circulated to counsel and other authorised recipients, without prejudice to the other steps which the parties are also directed to take before our judgments are handed down.

I agree.

I also agree