Pegasus Birds Ltd. v Customs and Excise

The term “best of their judgment” is derived from section 73(1) of the 1994 Act:

“Where a person has failed to make any returns required under this Act ... or to keep any documents and afford the facilities necessary to verify such returns, or where it appears to the Commissioners that such returns are incomplete or incorrect, they may assess the amount of VAT due from him to the best of their judgment and notify it to him.” (emphasis added)

It should be noted that the shorthand “best judgment”, as used in some of the cases, may be misleading, if it is taken to imply a higher standard than usual. The statutory words “to the best of their judgment” are used in a context where the taxpayers’ records may be incomplete, so that a fully informed assessment is unlikely to be possible. Thus the word “best”, rather than implying a higher than normal standard, is a recognition that the result may necessarily involve an element of guesswork. It means simply “to the best of (their) judgment on the information available” (Argosy Co v IRC [1971] 1 WLR 514, 517 per Lord Donovan).

Section 83 provides:

“Subject to section 84, an appeal shall lie to a tribunal with respect to any of the following matters...”

There is then set out a series of actions, decisions, and other matters arising under the Act listed under paragraphs (a) to (z). Paragraph (p) is as follows:

“An assessment -

(i)

under section 73(1) or (2) in respect of a period for which the appellant has made a return under this Act....

… or the amount of such an assessment.”

A condition of the exercise of this right of appeal is that the amount of the assessment should have been paid or deposited, but this requirement may be waived in appropriate cases (as in this case) if the Commissioners or the Tribunal are satisfied that the Appellant “would otherwise suffer hardship” (s.84(3)(3A)). Although that provision is not in issue before us, Miss Foster accepts that it must be interpreted in such a way as to avoid possible conflict with Article 6 of the Convention (right to a fair hearing).

Section 84 contains "further provisions relating to appeals". Subsection (5) provides that where, on an appeal under section 83(p), it is found that the amount in the assessment is too low, the Tribunal may give a direction specifying the correct amount. Subsection (8) provides for the repayment of any amount paid or deposited under subsection (3) if it is found on the appeal that “the whole or any part of any (such) amount… is not due”.

Generally, 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.” (Bi-Flex Caribbean Ltd v Board of Inland Revenue (1990) 63 TC 515, 522-3 PC, per Lord Lowry).

There is no express provision defining the Tribunal’s powers where an assessment is found not to be “to the best of their judgment”.

In Rahman v Customs & Excise Commissioners [1998] STC 826, I drew attention to phrases used by Woolf J in the leading case under this Act (Van Boeckel v Customs & Excise Commissioners [1981] STC 290) and in previous authorities in other tax contexts, to explain the effect of the “best of their judgment” requirement:

“The passages I have underlined show that the Tribunal should not treat an assessment as invalid merely because they disagree as to how the judgment should have been exercised. A much stronger finding is required: for example, that the assessment has been reached ‘dishonestly or vindictively or capriciously’; or is a ‘spurious estimate or guess in which all elements of judgment are missing’; or is ‘wholly unreasonable’. In substance those tests are indistinguishable from the familiar Wednesbury principles ([1948] 1 KB 223). Short of such a finding, there is no justification for setting aside the assessment.” (p 835)

In Rahman (1) it was common ground between the taxpayer and the Commissioners that the Tribunal should adopt a “two stage approach”, which I described thus:

“… the practice is to consider these cases in two stages: (1) consideration whether the assessment was made according to the "best judgment of the Commissioners"; if not, the assessment fails, and stage (2) does not arise; (2) if the assessment survives stage (1), consideration whether the amount of the assessment should be reduced by reference to further evidence or further argument available to the Tribunal…” (p 476)

Before me, Mr Barlow for the Commissioners had supported the practice as a discipline for officers, which was well understood and gave rise to little difficulty in practice. I expressed my concerns:

“I accept the importance of the discipline, and I also acknowledge the desirability of not upsetting established practice without good reason. In principle there is nothing wrong in the Tribunal considering the validity of the assessment as a separate and preliminary issue, when that is raised expressly or implicitly by the appeal, and, as part of that exercise, applying the Van Boeckel test. There is a risk, however, that the emphasis of the debate before the Tribunal will be distorted. If I am right in my interpretation of Van Boeckel, it is only in a very exceptional case that an assessment will be upset because of a failure by the Commissioners to exercise best judgment. In the normal case the important issue will be the amount of the assessment. The danger of the two-stage approach is that it reverses the emphasis…” (p 836, emphasis added)

In that case, the two-stage approach was applied in such a way that one of the Tribunal, having dissented from the chairman’s correct decision (as I found) on the best judgment issue, then wrongly regarded himself as having no further part to play in the consideration of the amount of the assessment. I held that the case had to be remitted to the Tribunal. I concluded:

“This case illustrates the dangers of an over-rigid adherence to the two-stage approach. I do not wish to diminish in any way from the importance of guidance given by Woolf J to inspectors as to how to exercise their best judgment when making assessments. However, when the matter comes to the Tribunal, it will be rare that the assessment can justifiably be rejected altogether on the ground of a failure to follow that guidance. The principal concern of the Tribunal should be to ensure that the amount of the assessment is fair, taking into account not only the Commissioners' judgment but any other points that are raised before them by the appellant.” (p 840)

The latter passage was in terms adopted by Chadwick LJ in this Court in Rahman (No 2) [2003] STC 150, para 42. I share his surprise (see para 17) that the Commissioners, having succeeded before me on the best judgment issue and failed only on the issue of amount, had apparently acquiesced in a form of order under which both aspects were relitigated before the Tribunal. However, it gave this court the opportunity to give further guidance on the correct approach.

Chadwick LJ (para 5) noted that the wording of section 83(p) reflected “the two distinct questions” which may arise where an assessment purports to be made under section 73(1):

“First, whether the assessment has been made under the power conferred under that section; and, second, whether the amount of the assessment is the correct amount for which the taxpayer is accountable.”

Having referred with approval (para 31) to my judgment in Rahman (1) and that of Dyson J to like effect in McNicholas Construction Co v Customs & Excise [2000] STC 553, he addressed the taxpayer’s submission that because the tax due had been found to be less than half the amount of the assessment, the assessment could not have been to “best judgment” (para 32). He regarded that as a “non-sequitur”:

“The explanation may be that the tribunal, applying its own judgment to the same underlying material at the second, or ‘quantum’, stage of the appeal, has made different assumptions - say, as to food/drink ratios, wastage or pilferage - from those made by the commissioners. As Woolf J pointed out in Van Boeckel ([1981] STC 290 at 297), that does not lead to the conclusion that the assumptions made by the commissioners were unreasonable; nor that they were outside the margin of discretion inherent in the exercise of judgment in these cases. Or the explanation may be that the tribunal is satisfied that the commissioners have made a mistake - that they have misunderstood or misinterpreted the material which was before them, adopted a wrong methodology or, more simply, made a miscalculation in computing the amount of VAT payable from their own figures. In such cases - of which the present is one - the relevant question is whether the mistake is consistent with an honest and genuine attempt to make a reasoned assessment of the VAT payable; or is of such a nature that it compels the conclusion that no officer seeking to exercise best judgment could have made it. Or there may be no explanation; in which case the proper inference may be that the assessment was indeed arbitrary.” (emphasis added)

That formulation of the “relevant question” was part of the ratio of the decision in that case; it is binding on us, and on the Tribunal in future cases.

In the light of that authoritative statement of the law, I would caution against attempts to refine or add to it, by reference to individual sentences or phrases from previous judgments. In Rahman (1), as already noted I listed a number of phrases used in earlier cases as “examples”, to illustrate that the test was higher than was being submitted by the taxpayer. I added that the tests were “indistinguishable from the familiar Wednesbury principles”. In retrospect, I think the reference to Wednesbury principles was unhelpful and a possible source of confusion, and may raise as many questions as it answers (see the comments of Neill LJ in John Dee Ltd v Customs & Excise [1995] STC 941, 952; and of the Tribunal in W H Smith Ltd v Customs & Excise [2000] V&DR 1 para 124). Another phrase (used by Woolf J in Van Boeckel) referred to the obligation of the commissioners “fairly (to) consider all material placed before them”. As a general proposition that is uncontroversial. However, it should not be seen as providing a separate and sufficient test of the invalidity of the assessment, nor as justifying lengthy cross-examination to establish whether the relevant officers have in fact looked at all the available material. Even the term “wholly unreasonable” (also used in Van Boeckel) may be misleading if it is treated as a separate test, rather than as simply an indication that there has been no “honest and genuine attempt” to make a reasoned assessment.

Even if it is established that there has been a breach of the “best of their judgment” requirement in relation to some element of the assessment, it does not follow in my view that the whole assessment should be set aside. We were not referred to any case where this issue has been considered in the higher courts, no doubt because in none of the reported cases in those courts was a finding of breach upheld.

This point was touched on, but not decided, in Rahman (2). Chadwick LJ (para 44-45) referred to the powers of the Tribunal following a finding that the assessment had not been made to best of their judgment:

“… The tribunal may take the view, in such cases, that the proper course is to discharge the assessment. But even in cases of that nature, as it seems to me, the tribunal could choose to give a direction specifying the correct amount – with the consequence that the assessment would have effect pursuant to section 84(5) of the 1994 Act. It could not be criticised for doing so. The underlying purpose of the legislative provisions is to ensure that the taxable person accounts for the correct amount of tax.” (para 45)

Before us, Miss Foster did not seek to rely on section 84(5) as giving such a power. With respect, I think she was right not to do so. Although the subsection applies in terms to an appeal in respect of “any” of the matters referred to in section 73(1) (thus apparently applying whether the appeal is against the assessment itself or against the amount), the trigger is a finding that the amount “specified in the assessment” was “less than it ought to have been”. Thus, it is of no direct relevance where the complaint is that there was no valid assessment at all, or that the amount in the assessment is excessive. On the other hand, the existence of the power to increase the assessment, whatever the nature of the appeal, is some indication of the width of the Tribunal’s powers, in pursuit of “the underlying purpose” to ensure that the correct tax is paid.

Mr Woolf, following on this issue, submitted that, where it is shown that any aspect of the assessment is vitiated under the Rahman (2) tests, the Tribunal has no discretion; it must set it aside. However, this approach takes no account of the development of modern principles of administrative law, under which the traditional distinctions between “void” and “voidable” have largely been eroded (see e.g De Smith, Woolf and Jowell: Judicial Review 5^th Ed para 5-048). There is no general rule that a decision arrived at in breach of administrative law principles is of no effect; the consequences of the breach must be looked at in the context of the particular statutory scheme (see e.g., in another context, R v Wicks [1998] AC 92). Mr Woolf’s submission is not helped by his reference to Pawloski (Inspector of Taxes) v Dunnington [1999] STC 550.) In that case, it was held that the validity of a direction making an employee liable to PAYE tax could be challenged in the defence to tax collection proceedings. However, there was no right of appeal against the direction, and the only issue was whether the challenge had to be by judicial review (see p 557f). The case does not assist in construing the scope of the Tribunal’s powers under the statutory appeal created by section 83(p).

As has been seen, the Act lays down certain preconditions for the making of an assessment; requires the assessment to be made to the best of their judgment; and provides a right of appeal to the Tribunal against either the assessment or the amount. Although the Tribunal’s powers are not spelt out, it is implicit that it has power either to set aside the assessment or to reduce it to the correct figure. There is no doubt that an appeal to the Tribunal, rather than judicial review, is the appropriate remedy if there are grounds for treating it as of no effect (Harley Development Inc. v CIR [1996] 1 WLR 727). Thus in Argosy (see above) the assessment was set aside, because, under the relevant statute, it was a precondition of making an assessment that the commissioner should “be of the opinion” that the taxpayer was liable to pay tax. The commissioner made no attempt to explain how he had formed that opinion, in the face of clear evidence that any assumed profits would have been “swamped” by previous trading losses ([1971] 1 WLR at 516).

Where, however, the complaint in substance is not against the assessment as such, but is that the amount has not been arrived at by “best of their judgment”, I see nothing in the statute or in principle which requires the whole assessment to be set aside. Clearly much will depend on the nature of the breach. We were told by Miss Foster that the Commissioners would not seek to defend an assessment which was arrived at dishonestly in any respect. That is understandable as a matter of public policy. However, the issue facing the Tribunal is unlikely to be so clear-cut. Fortunately in this country, sustainable allegations of actual fraud or corruption on the part of public officials are likely to be very rare indeed. What is much more likely is an allegation that, in “the heat of the chase” of an apparent wrongdoer, the officers concerned have, consciously or unconsciously, cut corners or closed their minds to relevant material. Defining the boundaries of “dishonesty” in such cases is notoriously difficult (cf Twinsectra Ltd v Yardley [2002] 2 AC 164, 170).

In my view, the Tribunal, faced with a “best of their judgment” challenge, should not automatically treat it as an appeal against the assessment as such, rather than against the amount. Even if the process of assessment is found defective in some respect applying the Rahman (2) test, the question remains whether the defect is so serious or fundamental that justice requires the whole assessment to be set aside, or whether justice can be done simply by correcting the amount to what the Tribunal finds to be a fair figure on the evidence before it. In the latter case, the Tribunal is not required to treat the assessment as a nullity, but should amend it accordingly.

Before leaving this topic, I should say something about the use of European authorities, having regard to the emphasis given to them in Miss Sharpston’s submissions to the Judge and to us. It is of course trite that VAT is a tax derived from a European Directive, and that accordingly relevant principles of European law must be taken into account in interpreting and applying the English code. Furthermore, under the Human Rights Act 1998, we must also take account of the European Convention of Human Rights, and any relevant Strasbourg jurisprudence.

However, European authorities, like domestic authorities, are of little assistance unless they are concerned with the same subject matter, or can be relied on to demonstrate some relevant legal principle of general application. In the latter case it is particularly important that the requirement of the Practice Direction is observed to state “the proposition of law that the authority demonstrates, and the parts of the judgment that support that proposition” ([2001] 1 WLR 1001 paras 8.1, 9.3). Unless that discipline is observed, much time and expense will be wasted on analysing complicated authorities which either have no relevance to the case in hand, or, if they do, merely illustrate principles which are uncontentious. Indeed, if a principle truly is one of general application, it may often be demonstrated more convincingly, succinctly and economically by reference to one of the standard textbooks on European law (cf my comments on the use of textbooks in common law cases: CEL Group Ltd v Nedlloyd Lines UK Ltd [2002] EWCA Civ 1716).

With respect to Miss Sharpston, the present case demonstrated the importance of this discipline. None of the many cases cited in her Skeleton seemed to me to help in any way in the resolution of the critical issue before us. She referred in particular to two ECJ decisions: “Eugen Nöelle” v Hauptzollamt Bremen-Freihafen [1991] ECR I-5163; and Case T-62/98 Volkswagen AG v Commission ECR II-2702. Both related to entirely different subject-matter. The former concerned the Commission’s choice of “reference countries” when determining “normal value” for the purpose of anti-dumping legislation. Advocate-General Van Gerven referred to the task of the Court when reviewing the observation by the decision-maker of the principles of good administration, including the duty of care and to provide reasons (p I-5186-7), and the Court reaffirmed that even a wide administrative discretion was subject to review by the Court to –

“… verify whether the relevant procedural rules have been complied with, whether the facts on which the choice is based have been accurately stated and whether there has been a manifest error of appraisal or misuse of powers…” (p I-5203)

None of that would come as a surprise to an English administrative lawyer. Miss Sharpston, as I noted her submissions, relied on the European cases as showing that European principles of good administration impose “a greater duty to look at the material than is discernible from the English authorities”. However, it was unclear to me what those cases added to the ordinary duty of a decision-maker to take reasonable steps to inform himself of relevant material (see e.g Secretary of State v Tameside BC [1977] AC 1014, 1065B per Lord Diplock). In any event, Nöelle throws no light on how and by what tribunal, as a matter of domestic procedure under the VAT legislation, the review is to be carried out, and with what practical consequences.

The second case (Volkswagen) touches on the question of the consequences of breach, although again in a quite different context, and as an incidental aspect of a complex report running to over 130 pages. The case concerned enforcement action taken by the Commission in respect of an alleged concerted practice contrary to Article 85. The Court held, that it had a duty “to examine whether the Commission incorrectly assessed the facts…”, and for that purpose to ascertain whether it had “gathered sufficiently precise and consistent evidence to give grounds for a firm conviction that the alleged infringement took place” (para 43). It was held that the decision contained “errors of assessment” in relation to certain aspects (paras 72, 169), but on others the Commission had rightly found an infringement of Article 85(1). The consequence was that the decision was “annulled” in part (para 198), and the amount of the fine “in the exercise of the Court’s unlimited jurisdiction” was reduced (para 346).

Miss Sharpston also relied on a section of the judgment in which the Court dealt with a plea that the Commission had wrongly leaked information about the investigation to the press (para 279ff). The Court observed:

“It is settled case-law that an irregularity of the type found above may lead to annulment of the decision in question if it is established that the content of that decision would have differed if that irregularity had not occurred.” (para 283)

The plea was rejected because the irregularity had made no difference to the result.

Miss Sharpston relied on this case as European support for the proposition that a finding that the amount of the assessment was “wholly unreasonable” in any respect must result in its annulment. Apart from the general point that the subject-matter was quite different, there seem to me to be at least four obvious answers. First, the passage says “may” not “must”. Secondly, the context was one in which the primary obligation was on the Commission to adduce the relevant facts, rather than (as here) on the taxpayer to provide the information to enable him to be taxed correctly. Thirdly, if relevant, the judgment would prove too much. It is not, and could not on the authorities, be seriously argued that any “irregularity” or “error of assessment” will lead to the complete annulment of a section 73 assessment. The argument is concerned solely with a failure described as “wholly unreasonable”, and the issue is what that means. Finally, the fact that the decision was flawed in certain respects in the Volkswagen case did not result in the proceedings failing altogether, but simply in the reduction of the fine to the level thought appropriate by the Court.

Miss Sharpston also referred us to the decision of the ECHR in Giorgiou v United Kingdom [2001] STC 80, which followed the Court of Appeal decision in Giorgiou v Customs & Excise to which I have already referred. In the context of its discussion of Article 1 of the First Protocol, the Court rejected a complaint that the UK system did not require the Tribunal to apply the “best judgment” test when reviewing the assessment. It commented shortly:

“… the tribunal found as a fact that the commissioners had reviewed the assessments to the best of their judgment. Accordingly, the applicants cannot claim to be a “victim” of a violation of the convention with the meaning of art 34 in this respect.” (p 90)

Miss Sharpston sought to persuade us that the ECHR must be taken as having impliedly adopted the summary of the law relating to “best judgment” given in that case by Sir Louis Blom-Cooper QC in the High Court ([1995] STC 1101, 1116; a summary “provisionally” accepted by this Court as “an accurate and working summary of the law”: [1996] STC 463, 475j). That submission echoed part of the present Tribunal’s reasoning; it said

“… The judgment must meet some minimum standard in order to satisfy the balance between the end and the means which is necessary to satisfy Article 1 of the First Protocol, see Sporrong and Lön v Sweden (1982) 5 EHRR 35 at para 61; it must not be arbitrary, cf. Aston Cantlow PCC. In reaching its decision in Georgiou the European Court specifically referred to the fact that the tribunal considered that the assessments were to the best of the commissioners’ judgment, see [2001] STC at page 90e.” (Decision para 86)”

With respect to both, I do not think the European Convention and cases under it add anything to the effect of the English cases. As to Georgiou, the Strasbourg Court made no reference to the High Court’s summary, and had no reason to do so, nor to concern itself with the detail of the test.

In the light of the above discussion, I would make four points by way of guidance to the Tribunal when faced with “best of their judgment” arguments in future cases:

I have already referred to the scope and result of the proceedings before the Tribunal. In order to understand the enormity of the task which faced the Tribunal, and to put its conclusions in context, it is necessary to refer in a little more detail to its comments on the course of the hearing.

The Tribunal noted that apart from its exceptional length, the case had exhibited a number of unusual features. One was the fact that the assessments assumed undeclared tax-inclusive sales of £4.05m compared with declared outputs of just under £0.9m, implying suppression of nearly 82% of sales. Another was that, unlike most cases in which the Commissioners have very limited material when making assessments, in this case they had a large volume of material obtained for the criminal trial, much of which was not adduced in evidence before the Tribunal.

The order for directions made in October 2000 had included provision for exchange of statements of case, and lists of witnesses and documents, and for the production by the Commissioners of a common bundle of documents for the hearing. Unfortunately, compliance with letter or spirit of that order was limited on both sides. Mr Freeman, as already noted, had served an “extended notice of appeal” in November 2000, including various allegations against Mr Bakewell, though not of bad faith.

No single statement in response was served by the Commissioners before the hearing. The evidence of their principal witness, Mr Bakewell, was in the form of eight statements produced for the purposes of the criminal trial. It was not until the thirteenth day of the hearing, following a specific direction, that a composite statement by Mr Bakewell was produced (Decision para 30).

The presentation of documents seems to have been chaotic. A particular problem was the failure of the Commissioners, in spite of a number of requests, to disclose written material which had been seized from Mr Hammond and Mr Foster in 1995. It was not produced until, following the intervention of the Tribunal during the hearing, “the respondents produced 40 boxes, which the appellant had to sort out during the hearing….” (para 35). The Tribunal observed that the retention of these documents had been unlawful, and “was a major contribution to the length of the hearing.” (para 37). The Tribunal noted also that the Commissioners had failed to produce a common bundle of documents for the hearing as directed:

“Since there were probably in excess of 10,000 pages of documents, many of which were duplicated under different references, this caused great difficulty at the hearing and also contributed to its length.” (para 38).

The notebooks of the relevant officers, including Mr Bakewell, were not produced until the hearing. The Tribunal commented:

“… they should have been listed since without them the officers would have quite been unable to give evidence of the detailed events up to eight years ago which were recorded in them.” (para 40)

The Tribunal’s view was that these failings by the Commissioners made a substantial contribution to the length of the proceedings:

“…On most of the initial 17 days in May and June the Tribunal sat at 10.00am and did not rise until 5.00pm in an attempt to keep to the timetable agreed. Given the ambit of the dispute and the difficulty both sides experienced in selecting relevant material, this proved impossible. Mr McNicholas complained that Mr Freeman kept raising new matters. To a considerable extent this was due to the Respondents withholding material until such a late stage.” (para 40)

One consequence of the defective preparation was serious confusion about the use to be made of evidence and other material from the criminal proceedings. Some 90 statements had been served by the Customs in the criminal proceedings, including seven statements by overseas suppliers. These had all been available to the Commissioners at the time of the assessments, and therefore were in principle relevant in considering their validity under the “best of their judgment” test. However, most had not been the subject of formal notices under the relevant rules, to enable them to be produced as evidence before the Tribunal.

During the course of the hearing, the Tribunal heard submissions as to whether it should admit in particular the depositions by overseas witnesses, or allow hearsay evidence as to what they had said. It ruled against the Commissioners, on the grounds that the evidence was “highly contentious”, and “would have added substantially to the length of the hearing which had already taken 16 days”. It commented:

“This difference between the evidence on best judgment and that on quantum caused considerable problems during the hearing and undoubtedly was a major factor in its length. The burden of proof is of course on the appellant. However neither member of the Tribunal can remember another case when the Commissioners have not produced in evidence important parts of the material on which the assessment was based.” (para 63)

It is important to re-emphasise that the taxpayer’s Extended Grounds of Appeal before the Tribunal contained an allegation of “perversity”, but none of dishonesty or bad faith. In spite of some vacillation on the part of Mr Freeman before the Tribunal, and even by Counsel in this Court, that remains the position. Patten J summarised his understanding of Mr Freeman’s position:

“During the course of the hearing before the Tribunal Mr Freeman did… seek to impugn the honesty of Mr Bakewell and the good faith of the investigation and the subsequent assessments…. But those allegations of dishonesty or bad faith were dropped and replaced by the argument that Mr Bakewell was guilty of what was variously described as intellectual dishonesty or capriciousness…. Intellectual dishonesty is not a term used in any case on best judgment that I know of and has no certain meaning. I assume that Mr Freeman’s contrast of that term with dishonesty in the sense of mala fides indicates that it was intended to denote a lack of reasoning on Mr Bakewell’s part. But that could have a number of possible causes.” (para 29)

Patten J clearly understood, on the basis of Miss Sharpston’s submissions, that no allegation of dishonesty or bad faith was being sustained before him.

Surprisingly, however, the Skeleton lodged in this Court on behalf of Pegasus contained an apparent attempt to revive an allegation of bad faith. It suggested that Patten J had been misled by “an inaccurate concession made by the Appellant’s counsel during exchanges on the final afternoon of the hearing…” However, following an exchange with the Court, this part of the Skeleton was deleted, and Miss Sharpston properly and unequivocally accepted that she could not pursue any allegation that the assessments were made in bad faith.

The first two grounds of appeal (see above) turn on the correct interpretation of the “best of their judgment” principle, and its application to this case. The third and fourth grounds are directed to the form of the order.

The Tribunal’s understanding of its task (formed before the decision of this Court in Rahman (2)) is apparent from the following passages. Having considered the relevant authorities, including my judgment in Rahman (1), it said:

“We are unable to accept that the test of best judgment is wholly subjective. Of course bad faith or vindictiveness are subjective being dependent on the state of mind of the assessing officer. However the question whether an assessment is capricious or a spurious estimate or guess or wholly unreasonable does not depend on the officer’s state of mind although it may or may not result from it…. Although the Commissioners are required to ‘fairly consider all material placed before them and, on that material, come to a decision which is reasonable’ (see Van Boeckel), the Rahman (1) approach is that (the) standard of what is reasonable is anything which is not wholly unreasonable.

That however is not the whole picture. If the only question apart from bad faith is whether the result is wholly unreasonable, the requirement fairly to consider all material before them would be nothing but empty words. Such an approach would be difficult to square with the proportionality necessary under Article 1 of the First Protocol to the European Convention on Human Rights. Having collected a vast amount of material in the present case and having been presented with more by Mr Freeman, Mr Bakewell was under a duty to consider it all fairly and with an open mind. He was not in any sense obliged to accept everything put to him. He was however obliged to consider whether in the light of all the material the assessment made sense. In deciding whether he came to a credible conclusion the Tribunal must be entitled to consider his judgment in the light of the material before him….” (para 84-6)

Having examined the evidence in detail, the Tribunal returned to this subject in a section headed “Conclusion as to best judgment”:

“232.

In considering whether the Appellant has shown that the assessments were not made to best judgment we consider first the method of calculation and then consider whether the result is credible in the light of the material before the Commissioners and in particular Mr Bakewell. We consider this objectively. It is not however enough for the Appellant to show that the assessments did not reach the standard required of a reasonably competent officer, it must satisfy the Tribunal that the assessments were wholly unreasonable. We take this to mean that the assessments were outside the parameters of what could have been reasonable if all the material before the Commissioners had been fairly considered.”

It then expressed its conclusions on the various elements of Mr Bakewell’s calculation. It had serious criticisms to make of a number of aspects, including the figure of £18,508, which it regarded as “clearly inflated” though not “wholly unreasonable” (para 233); and the 57/43 split, which it described as based on assumptions which were “wholly unclear” and in one respect “simply wrong” (para 234). The most serious criticism was made of the assumption as to trip numbers. The Tribunal accepted that it was not unreasonable on the evidence to attribute 166 trips to “either or both of Mr Foster or Pegasus”; the issue was whether it was reasonable to attribute all trips to both (para 238-9). The Tribunal understood the main basis for attributing all the trips to Pegasus to be the “assertion” that Mr Foster’s vans had been followed six times, and that each time they went to Pegasus. However, the Tribunal rejected this assertion, accepting Mr Freeman’s contention that –

“… there was no coincidence at all since on at least four occasions Customs officers knew that Pegasus was involved in advance which was the very reason why they followed…”

It mentioned other occasions when Mr Bakewell “kept observation outside Pegasus’ premises without success…”, commenting –

“We are driven to the view that Mr Bakewell was less than frank on this aspect.” (para 240)

It also referred to Mr Foster’s diary as “cogent evidence” that Mr Hammond was not involved in seven trips between January and March 1995, adding:

“Whatever Mr Bakewell did consider, we do not accept that he fairly considered Mr Foster’s diary.” (para 41)

The Tribunal concluded:

“250.

Our conclusion is that far from the other evidence supporting the level of the assessments, every single indication shows the assessments to be wholly excessive. We conclude that the assessments were wholly unreasonable, being outside the parameters of the reasonable. All of the matters to which we have referred in the last five paragraphs were there for Mr Bakewell to consider. He either failed to consider them at all or failed to consider them fairly; he therefore failed to consider fairly all the material before him. We are driven to the conclusion that far from making any assumptions in favour of the trader (paragraph 194), Mr Bakewell closed his mind to any material which did not fit his case. The Appellant has satisfied us that the assessments were not to best judgment; there is no logical basis to uphold the assessment for any individual period.”

Patten J held that these passages disclosed an error of law, tested by reference to Rahman(2):

“… For the reasons set out in the judgment of Chadwick LJ quoted above, the fact that the assessment is wrong and appears to the Tribunal to be objectively unreasonable is only the starting point. Conclusions to that effect do not justify finding that the assessment was not made to best judgment, unless the only explanation for the errors is that they were produced as part of something other than a genuine and honest attempt to calculate the amount of VAT. The principal difficulty about the Tribunal’s decision is that at no point did they ask themselves whether Mr Bakewell did his honest best. Instead they seem to have inferred that he (to use the words in paragraph 250) ‘closed his mind to any material which did not fit his case’ from the fact that he reached conclusions which they considered to be wholly unreasonable. Bearing in mind that Mr Freeman had by then abandoned his earlier allegations of dishonesty and bad faith, that seems to me to be something of a non sequitur and, on my reading of Rahman (No 2), a misdirection.” (judgment para 32)

He discussed the detailed criticisms made of Mr Bakewell’s reasoning. On the number of trips, he commented:

“In paragraph 240 of its decision the Tribunal said that it accepted Mr Freeman’s submissions and that it was driven to the view that Mr Bakewell was less than frank about the randomness of the observed visits. It is not clear what they meant by this, but it cannot be read as a comment on Mr Bakewell’s honesty in relying on the alleged randomness of the six trips, because that aspect of the cross-examination was not relied upon by Mr Freeman, and there is in any case no evidence which could support such a finding. It was never put to Mr Bakewell or suggested that the “six out of six theory” was some deliberate concoction which he knew to be deeply flawed. It was simply put to him that the evidence obtained from the six observed trips did not justify the inference which he had made. Similarly in paragraph 241 of its decision the Tribunal records that Mr Bakewell did not fairly consider Mr Foster’s diary. But again this cannot mean any more than that he gave it less weight than the Tribunal was itself prepared to do. No allegation of bad faith was put to him about his treatment of the diary, nor on the evidence could it be. He regarded it as unreliable and incomplete, whereas the Tribunal took a different view. The Tribunal was therefore right to confine its conclusion in paragraph 242 quoted above to saying that the assumption that Pegasus was involved in 166 trips was an unreasonable one on the part of Mr Bakewell. The Tribunal was entitled to take a different view of the evidence and to characterise Mr Bakewell’s contrary view as unreasonable. But that is not enough to establish a lack of best judgment and it is clear that the Tribunal could not have made such a finding, based on these conclusions, had it applied the Rahman test.” (para 64)

Overall he concluded:

“ I am satisfied, for the reasons given earlier in this judgment, that the Tribunal misdirected itself and misapplied the law on best judgment. I am also satisfied that its findings about the assessment, properly analysed, amount to no more than an acceptance of Mr Freeman’s submission that Mr Bakewell was wrong in the assumptions he made. That is not enough to establish that Mr Bakewell’s assessment was not made to best judgment, and insofar as paragraph 250 of the decision appears to suggest that there was some conscious failure by Mr Bakewell to take an honest and genuine look at the evidence, there was no material before the Tribunal, beyond the fact that it disagreed with his conclusions, on which that finding could be based. That is not a proper foundation for such a finding, and the Tribunal’s treatment of the evidence or the lack of it amounts to an error of law. There is no evidence that Mr Bakewell did anything but his honest and genuine best, however mistaken he may have been….” (para 76)

The task of this Court is made easier by the parties’ agreement that we should confine ourselves to the question whether Patten J was correct to find that the Tribunal erred in law in the test it applied. This makes it unnecessary for us to consider whether, even if the Tribunal’s understanding of the test was correct, its conclusions on the facts were sustainable. This issue was raised by the Commissioners in the appeal from the Tribunal, but Patten J found it unnecessary to deal with it. It was renewed without objection by an amended notice of appeal. It is agreed that, if we reverse Patten J’s decision on the legal test, the case should be remitted to him to deal with this aspect of the appeal on the facts.

In summary the question is whether it was sufficient, as the Tribunal thought, to find that the assessment, viewed objectively, was “wholly unreasonable”, or that there had been a failure fairly to consider all the relevant material; or whether, as the Judge held, it was necessary to find or infer that there had been no “honest and genuine attempt” to arrive at a reasoned assessment. In view of my earlier discussion of the principles, I hope I shall be forgiven for expressing my conclusions relatively briefly.

If the question were simply whether there was a breach of the requirements as explained in Rahman (2), I would agree with the Judge that there was not. The Tribunal, of course, was not to be blamed for seeking to apply criteria which it extracted from the previous judgments (including my own). However, the result must be judged by reference to the law as now stated in the Court of Appeal. It was unfortunate that an allegation of dishonesty, even if only “intellectual dishonesty”, was allowed to persist in the submissions to the Tribunal, and without clarification or adequate particulars. The charge in the extended notice of appeal was not of dishonesty but of “perversity”. I have been troubled, as was the Judge, by the Tribunal’s conclusion that Mr Bakewell “closed his mind to any material which did not fit his case…” A deliberate decision by an officer not to consider relevant evidence, in order to mask the truth, would be dishonest. However, the allegation was not put in that way, either in the notice of appeal, or in cross-examination. The Tribunal’s comment on Mr Bakewell’s lack of “frankness” on one aspect (which is in any event challenged by Mr Mc Nicholas) was not directed to the making of the assessment as such. Overall, it found that the assessment were “wholly unreasonable”, and that Mr Bakewell had failed “fairly or at all” to consider some of the relevant material. However, that is not the same as a finding that there was no “honest and genuine” attempt at the time of the assessments.

In any event, whatever the correct interpretation of the Tribunal’s findings, I am quite satisfied that this was not an appropriate case for the assessments to be set aside. The background was a serious fraud on the Customs, to which Mr Hammond had pleaded guilty. There was no doubt that an assessment to VAT was appropriate. The burden was on Pegasus to show what was the correct amount. The Commissioners were entitled to be highly sceptical of any information coming from a convicted fraudster. They took the view that the fraud was on a much larger scale than Mr Hammond had admitted. They had plenty of material on which they could reasonably do so, even if their view of the amount can now be seen as unrealistically exaggerated. In those circumstances, they were entitled for the purpose of the assessments to take a broad view of the evidence. They were not required, at the assessment stage, to conduct a meticulous examination of all the evidence in the criminal trial, such as was conducted over 42 days before the Tribunal.

Against that background, in my view, it was wrong for the Tribunal to allow the “best of their judgment” issue to dominate the proceedings. The warning I gave in Rahman (1) against reversing the emphasis of the proceedings seems to have been ignored. It created unusual problems in this case, where, through no fault of the Tribunal, there was such a confusing discrepancy between the evidence relevant to each issue. The hearing would have been much more manageable if attention had been directed to the admissible evidence relevant to fixing the correct amount of tax. It was particularly unfortunate that at the end of this process the Tribunal failed to make a conclusive finding as to the correct amount, so that the real issue still remains undetermined.

These conclusions are sufficient to dispose of the first two grounds of appeal. The Judge applied the correct test. Applying that test he was correct to hold that the assessments were properly made. He was also right to reject the application to remit the “best of their judgment” issue to the Tribunal to redetermine (a point renewed in this Court by an amended notice of appeal). If the object of this suggestion was to give the taxpayer a chance to pursue an allegation of bad faith which it failed to advance in the first hearing, it was wholly misconceived. An appeal on law cannot be used to enable the appellant to make a new case on the facts. In any event, for the reasons I have given, a favourable finding on the “best of their judgment” issue would not entitle the taxpayer to have the assessments set aside.

Ground (iii) raises the question whether the Tribunal should be treated as having made a definitive finding that no tax was due for the periods before September 1994, so that remitter is inappropriate.

The Tribunal noted that Mr Hammond denied any evasion before that date, and that, in spite of the scale of the Customs operation, “virtually no evidence was put before the Tribunal on which to infer evasion before September 1994” (para 251). It made no express finding on that point, although the remainder of its observations on quantum was directed only to the later periods. The Judge saw “the obvious attractions” to the taxpayer of “starting from the position at which the Tribunal left off”, but declined so to order. He said:

“… it seems to me that it would not be right, and I am not satisfied that I have jurisdiction, in a case where the Tribunal has not assessed or determined the appeal on quantum, either to determine myself or send it back to the Tribunal with some kind of direction which ties the hands of the Tribunal and requires it to abide by what are necessarily provisional findings of fact….” (Second judgment para 11)

He saw no alternative but to remit all the assessments for determination, adding:

“It would then be for the Tribunal to decide whether and to what extent it is necessary on the material before it to reopen each and every aspect of the quantum appeal.”

He rejected an application by the Commissioners for the appeals to be remitted to a different tribunal, and there is no appeal against that aspect of the order.

Although I have sympathy with the taxpayer’s position, I see no answer to the Judge’s reasoning. It would have been open, in my view, for the Tribunal to make alternative findings on quantum, in case it were found wrong on the “best of their judgment” issue; but it did not do so. The Tribunal’s comments on the period before September 1994 were not expressed as a determination. Nor were they part of the reasoning leading to the conclusion on the “best of their judgment” issue. They were clearly provisional, as the Judge said. Miss Sharpston was unable to provide any authority for the contention that in such circumstances the Judge could have treated them as definitive, or used them as the basis for deciding the matter himself.

Ground (iv) challenges the Judge’s decision on costs, both in the Tribunal and in the High Court. The award of costs in the High Court followed the event, and I can see no basis for questioning it. I have more difficulty with the Judge’s decision on the costs in the Tribunal. Clearly the Tribunal’s own award could not stand when its decision was reversed. Therefore, it is unnecessary for us to consider Miss Foster’s criticisms of the indemnity basis of the award. However, that does not mean that we should ignore the criticisms which led to that award, the most serious of which related to the withholding of substantial material until after the beginning of the hearing (Tribunal’s decision on costs para 47). Nor can we ignore the fact that, if the Tribunal’s provisional view on quantum is ultimately upheld, Pegasus will have achieved substantial success by reducing estimates from over £600,000 to £27,000.

The Judge was invited by Miss Sharpston to defer the issue of costs until the Tribunal’s final conclusion was known, or to remit it to the Tribunal until that event. He declined to take either course. He commented that it would be wrong to “delegate” to the Tribunal the determination of costs in the light of his judgment. He considered further that “in deciding to press for a determination” on the best of their judgment issue, Pegasus “took upon itself the risk” that if it won that issue before the Tribunal but lost on appeal, there might have to be a further hearing of quantum; and that the focus was “necessarily directed” to whether the propriety of the assessment rather than the extent of the tax evasion.

Although this Court is always reluctant to interfere with the exercise of the Judge’s discretion on costs, I am unable to accept this reasoning. As I understand it, Pegasus did not “press for” a decision on the “best of their judgment” issues in advance of a decision on quantum. It abandoned its request for a preliminary hearing. The hearing was concerned with both issues. It was the Tribunal which decided to make no formal decision on the issue of quantum. In any event, in remitting the matter to the same Tribunal, the Judge clearly recognised that much of the material and evidence used for the earlier hearing, and the Tribunal’s appreciation of it, would be relevant to the renewed hearing. The order appears to take no account of that fact. Nor can I understand why it was thought inappropriate to remit the issue of costs to the Tribunal. The fact that the Tribunal would have to take account of the judgment is no more an obstacle in relation to a decision on costs than on substance. On the contrary, the Tribunal would be in a much better position than the Judge to form a view of the appropriate order as to costs in the light of the whole course of the hearing before it, and the ultimate result.

In conclusion, I would allow the appeal against the Judge’s order in relation to the costs before the Tribunal. I would substitute an order remitting that matter to the Tribunal to be determined at the end of the renewed hearing. On all other points I would dismiss the appeal.