Customs and Excise v Pegasus Birds Ltd.

This is an appeal by the Commissioners of Customs and Excise against a decision of the VAT and Duties Tribunal (Mr Theodore Wallace and Mr Cyril Shaw FCA) released on 7^th June 2002. The issue before the Tribunal was whether an assessment to VAT notified to the Respondent Company, Pegasus Birds Limited ("Pegasus"), on 16^th April 1997, and contained in a formal notice of assessment dated 14^th May 1997, was made to best judgment. The Tribunal decided that it was not and allowed the Respondent’s appeal. The Commissioners now appeal to this Court on the basis that the Tribunal misdirected itself or at least misapplied the law as to what constitutes a failure to exercise best judgment. In the alternative they contend that if the Tribunal did correctly direct itself as to the law, there was no evidence to support the critical findings of fact upon which they based their own judgment.

The scale of the proceedings before the Tribunal is alarming. As recorded in the Tribunal’s decision, this was the second longest VAT appeal ever. It lasted some 42 days, spread over a period of eight months between May 2001 and February 2002. During the course of the hearing Mr Bakewell, the Customs officer who was responsible for the investigation and the subsequent assessment, was cross-examined for seven days. At various points in the proceedings challenges were made to the honesty and good faith of the investigation and assessment, although Mr Bakewell was ultimately accused of no more than intellectual dishonesty and a failure properly to consider material available to him which was said to be at odds with, and critical of, the assessment. The Commissioners were represented before the Tribunal by Mr Mc Nicholas of Counsel, and Pegasus by Mr Freeman of Messrs Horwath Clark Whitehill, who is a forensic accountant and produced a number of reports in support of the Respondent’s appeal. His costs alone are estimated to be in the region of £1m. The assessments in question covered the ten accounting periods from 1^st January 1993 to 30^th June 1995 and totalled £658,388. That assumes undeclared tax-inclusive sales of some £4.05m, compared with declared sales of a little under £0.9m. On these figures some 82% of sales were suppressed and off-record. Although not required to decide the matter in the light of their findings about best judgment, the Tribunal indicated that its own view on quantum was that there was a liability of no more than £27,000. Mr Freeman on behalf of Pegasus had sought to justify liability to VAT of between £3,000 and £6,000.

On 8^th October 1996 Mr Brett Hammond, the Director of Pegasus, pleaded guilty to various offences under ss.72(1) and/or 72(3) of the Value Added Tax Act 1994 ("the 1994 Act"). The particulars of the offences charged were that between 1^st January 1993 and 24^th May 1995 Mr Hammond and a Mr Eric Foster purchased live birds in Belgium and Holland and transported them to England, where they were sold for cash. The two Defendants were alleged to have failed to keep proper or complete records of the purchases and subsequent sales and to have failed to account for the VAT that became due.

The trial was set to commence on 8^th October 1996. Mr Foster had pleaded guilty on 2^nd October and Mr Hammond entered a guilty plea on the first day of the trial, on the basis of a document headed "Basis of Plea" which indicated an acceptance that there was evidence of a joint enterprise between Hammond and Foster, but gave notice that there would be a later "Newton Hearing" to determine the scale of Mr Hammond’s evasion by reference to 18 purchasing trips referred to in the opening note of the Crown. That note refers to a minimum of 18 known trips on which it was said that Mr Hammond must have purchased birds, none of which appears in the accounting records of Pegasus. The Basis of Plea document went on to indicate that at the Newton Hearing Mr Hammond intended to adduce evidence (including expert evidence) about the number and value of the birds involved.

In preparation for the trial the prosecution had obtained four reports (referred to as A-D) from Mr Colin Booker ACMA, who was to give expert evidence. Booker C (dated 30^th April 1996) contained a calculation of the VAT evaded between January 1993 and May 1995 by both E & J Birds (the firm run by Mr Foster) and by Pegasus. Booker D (dated 6^th June 1996) replaced Booker A and B, and contained an analysis of the contemporaneous trading records of Pegasus, including the so-called Money Taken sheets, till rolls, paying-in slips, VAT returns and worksheets, purchase invoices, cash books and bank statements. One of the points urged on the Tribunal by Mr Freeman and repeated in submissions before me was that Pegasus operated an established business which did record purchases and sales and did provide VAT returns. The assessments to VAT which were the subject of the appeal to the Tribunal were, as already indicated, based on an estimate of off-record sales, calculated by reference to trips made to the continent to purchase birds in which Mr Hammond is said to have participated, or at least been supplied through. Mr Booker did not, however, make factual assumptions of his own as to the number of those trips and the level of purchases. He relied upon various Trip Schedules prepared by Mr Bakewell and, on the basis of these, calculated what he considered was the likely measure of evasion. At the request of the prosecution he produced Booker C, containing his assessment of the amount of VAT avoided by both Pegasus and E & J Birds. It is calculated on an assumption of at least 165 trips between January 1993 and May 1995, with an average value of the purchases per trip of some £15,305.61. This produced total purchase costs for the period of £2,525,096, which he adjusted to £2,272,586 to take account of pre-sale mortalities. He then applied mark-ups of 159.61% for retail sales and 104.43% for trade sales. If all sales of the birds were trade sales, this would result in a sales figure for both businesses of £4,645,847 and a VAT liability of £691,935. If all sales were retail, the corresponding figures would be £5,907,679 and £879,718.

Mr Bakewell produced a statement dated 1^st November 1996 in which he updated his Trip Schedule for purposes of the Newton Hearing. The statement was discussed with Prosecuting Counsel (Mr Seddon Cripps: now HH Judge Cripps) and approved by him. It contains a number of exhibits, including invoices obtained from foreign traders and invoices for the hire of vans and trailers. The Crown’s case for the trial (supported by the Trip Schedule in its original form) was that Mr Foster and Mr Hammond had been involved in 165 trips, resulting in VAT evasion of between £692,000 and £880,000, and although Mr Hammond had entered a guilty plea by reference to 18 specific trips, it is clear that this in no way bound the Crown or altered its own view of the scale of the evasion involved. Mr Bakewell therefore proceeded to calculate the amount of VAT avoided on the basis of the 165 trips, in sums ranging from £681,248 to £2,200,723, depending upon the average value of the consignments (£13,699 to £20,690) and the mark-ups applied (80-285%). He assessed Mr Hammond’s share of each consignment as 57%, based upon some invoices from a Dutch dealer named Kloeg, which broke down a particular consignment between "O" and "X". O was said to be a reference to Mr Hammond and X to be a reference to Mr Foster. Prior to the Newton Hearing, Mr Foster through his accountant agreed to a mark-up for calculation purposes of 150%.

The Newton Hearing was fixed for 17^th February 1997. The trial judge (HH Judge Pontius) had indicated at a procedural hearing that for purposes of sentencing he needed a bracket rather than a precise assessment of the tax evaded and that the first bracket would be up to £50,000. Judge Cripps gave evidence to the Tribunal and described what happened at the Newton Hearing. Some of the Commissioners’ foreign witnesses who were to give evidence at the trial had been stood down following the guilty pleas. He was concerned about the evidence of foreign purchases available to him at the Newton Hearing but also wished to secure a custodial sentence. His view as a prosecutor was that, once a level of evasion of £50,000 was reached, any increase above that was unlikely to have anything but a limited effect on sentence. After discussion with Counsel for Mr Hammond, it was agreed that, for sentencing purposes only, Mr Hammond would accept that he was responsible for the evasion of VAT in a figure not exceeding £50,000. On this basis Mr Hammond was sentenced to 18 months’ imprisonment, which was later reduced to 12 months on appeal.

In these circumstances it was unnecessary for any evidence about the scale of the evasion to be called by the prosecution. Equally, no evidence was given by Mr Hammond or any admissions made by him beyond what was necessary for sentencing purposes. The Commissioners therefore obtained no further information from the Newton Hearing for the purpose of making their assessment of Pegasus to VAT, other than the agreement reached about mark-ups and the material contained in the two reports by Mr Freeman. The first of these is dated 30^th September 1996 and was prepared for the main criminal trial. The second is dated 20^th January 1997 and was intended for use at the Newton Hearing. In his first report Mr Freeman concluded that there was no real evidence of any off-record sales and that the Money Taken sheets (which Mr Booker considered to be indicative of a fraudulent system) were consistent with the annual accounts audited to 31^st March 1994. These conclusions were of course negated by Mr Hammond’s plea of guilty. In his second report Mr Freeman calculated the level of VAT evaded to be between £13,000 and £16,000. These reports contain a number of specific criticisms of Mr Booker’s evidence, including the following:

As already indicated, the assessments were raised and notified to Pegasus in April 1997 by Mrs Skam of the Basildon VAT Office. Her letter of 16^th April indicated that the assessment figure of £658,388 was based on a total of 166 trips between March 1993 and June 1995. Each trip was given an average value of £18,688, coupled with the agreed mark-up of 150%. The consignments were split 57/43 between Pegasus and E & J Birds on the basis of the Kloeg invoices. A Notice of Appeal against the assessments was lodged by Mr Freeman on behalf of Pegasus, dated 6^th June 1997. In his accompanying letter Mr Freeman stated that the grounds of appeal were that the assessment was "excessive and not in accordance with information provided". Only £14,500 of unpaid tax was admitted. There was no express challenge to best judgment. But on 31^st July Mr Freeman notified the Tribunal that he wished to rely on two additional grounds: (i) that the assessments were not made to best judgment in accordance with s.73(1) of the 1994 Act; and (ii) that the tax periods up to and including the quarter to 31^st March 1995 were out of time.

The second issue about time limits was heard and rejected by the Tribunal in its decision released on 22^nd October 1997. There were then unsuccessful appeals to the High Court ([1999] STC 95) and to the Court of Appeal ([2000] STC 91). A petition for leave was presented to the House of Lords and was refused on 29^th June 2000. Pegasus (through Mr Freeman) then applied for the issue of best judgment to be decided as a preliminary issue, but withdrew this application at a directions hearing held on 16^th October 2000. On 27^th November 2000 Mr Freeman served what was described as an "Extended Notice of Appeal". In relation to the issue of best judgment, it sets out a series of allegations against Mr Bakewell, the relevant ones of which can be summarised as follows:

The directions given on 16^th October 2000 were made on the basis that both issues on the appeal (i.e. best judgment and quantum) were to be dealt with at the same hearing. The evidence was accordingly directed to both issues. Mr Bakewell (whose conduct is central to the issue of best judgment) produced a composite statement of some 39 pages which was served on 30^th May 2001. It exhibited his earlier statement of 1^st November 1996 produced for the Newton Hearing, together with the Trip Schedule. I should mention at this stage that there were a number of procedural wrangles before the Tribunal concerning the late production of witness statements and disclosure. Mr Bakewell’s composite statement was not produced until the thirteenth day of the hearing (pursuant to a direction given on 11^th May 2001) and there was late disclosure by the Commissioners of the unused material seized from Mr Foster and Mr Hammond during a raid in May 1995. There were some 40 boxes of this material, which were only produced during the hearing. There was no agreed timetable for witnesses prior to the hearing and no agreed bundle of documents. This clearly caused the Tribunal a considerable amount of justifiable irritation, which is evident in their decision. Equally, the Tribunal and Mr Mc Nicholas had to deal with a number of new issues raised by Mr Freeman, which undoubtedly lengthened the proceedings. I mention these points only to put on record that neither party has relied upon the procedural difficulties as relevant to the issues on this appeal. It is not therefore necessary to say anything further about the detail of them. I should, however, refer to one matter which resulted from the directions that both best judgment and quantum were to be dealt with at the same hearing. Included in the evidence were statements by a number of individuals (including at least ten overseas witnesses) which had been prepared for use in the criminal proceedings. This material was relevant to best judgment in the sense that it was available to Mr Bakewell at the time of the making of the assessment and was considered by him. But the Tribunal seems to have raised the question whether, because the statements had not been served on Pegasus in the appeal proceedings under Rule 21 of the Tribunal Rules, they should be admitted to prove the truth of their contents without the makers of the statements being called. The Tribunal ruled that the depositions of the overseas suppliers and the oral hearsay evidence of Mr Bakewell as to what they had said should be excluded on the issue of quantum. They were, however, admissible in relation to best judgment. Inevitably this had the consequence that when witnesses such as Mr Bakwell were examined about their assessment of matters such as the likely number of trips and the size of the consignments, they were able to give evidence of how they relied upon information provided by the overseas suppliers, but the Tribunal would not receive that evidence in relation to the issue of quantum to which it was also highly relevant. The failure to separate quantum and best judgment also had the consequence that the Tribunal embarked upon a fact-finding exercise (which it was bound to do in order to determine quantum) and had formed its own view of the correctness of the assessment when it came to consider the question of best judgment. This is not inherently objectionable, provided that the Tribunal does not allow its own determination of the facts to dictate its conclusions on best judgment. There are obvious difficulties here, and in paragraph 63 of its decision the Tribunal itself acknowledged that the difference between the evidence on best judgment and that on quantum had caused considerable problems during the hearing. Although (as I shall come to later) the parties to this appeal argue for different tests of what constitutes best judgment, Miss Sharpston QC (for Pegasus) rightly rejected any suggestion that the Tribunal could infer from the mere fact that it took a different view of the evidence, and therefore the correctness of the assessment, that the Assessing Officer had not been entitled to take a different view on the same material and had on that basis acted in good faith. The essential criticism of the Tribunal made by the Commissioners on this appeal is that the Tribunal fell into that very trap and failed to disentangle its own view of the likely scale of evasion from the question whether the Commissioners had exercised best judgment in coming to a different conclusion.

Although the issues of best judgment and quantum were both before the Tribunal at the hearing, it is clear that most of the argument and the cross-examination of Mr Bakewell was directed to the question of best judgment. That is evident from the Extended Notice of Appeal lodged by Pegasus in November 2000, which elaborates the case on best judgment over 11 out of its 12 paragraphs and restricts the appeal on quantum to a single sentence to the effect that the amount of VAT due was no more than £8,000. At the start of the hearing in May 2001 Mr Freeman also provided Mr Mc Nicholas with a document headed "Draft Legal Submissions", which was intended to put the Commissioners on notice of the case they had to answer. It takes issue with the assessments broadly speaking on the same grounds as are set out in the Extended Notice of Appeal: i.e. the number of trips, the 43/57% split, the size and value per consignment and the inability of Pegasus to have conducted off-record sales on the scale alleged. It also relies on a submission that Mr Hammond was not a co-venturer with Mr Foster, but merely became involved in the evasion of VAT as a necessary consequence of importing or purchasing birds which had been smuggled into the UK to avoid MAFF quarantine regulations. The document was prepared without sight of the Commissioners’ authorities and, despite its title, does not in fact contain any detailed submissions on the law. But in relation to the issue of best judgment it contains the following paragraphs:

Mr Bakewell began his evidence in cross-examination on 30^th May 2001 by re-asserting that he believed that he had made the assessments to best judgment having considered all the material in the case and that he stood by them. He was asked about his life and career before joining Customs and Excise and said that he had been interested in birds since childhood and had worked in Paris as the manager of the bird and reptile section of a large pet store. There then followed several days of cross-examination, during which he was asked about various matters, including the book-keeping at Pegasus, a control visit made by Customs and Excise in 1994, his observations of the Foster trips and the presence on them of an additional vehicle, his own visits to Pegasus’s shop and the relative sizes of Pegasus and E & J Birds. There was then a break in the evidence from June until September.

Prior to the resumption of the hearing there was an exchange of correspondence between Mr Mc Nicholas and Mr Freeman in which Mr Mc Nicholas raised the concerns of the Commissioners about certain investigations which had been or were being carried out by Pegasus into the background of Mr Bakewell. In a fourth supplementary statement Mr Freeman said that these investigations revealed that Mr Bakewell and Mr Hammond had lived within a radius of two miles of each other as boys and were both interested in birds. There was a possibility that they had known or at least met each other through a local bird club, but Mr Hammond in a statement could not recall whether this was in fact the case. On 2^nd September 2001 Mr Freeman wrote to Mr Mc Nicholas in these terms:

It is clear from this correspondence that the case being advanced by Mr Freeman at this stage was that the assessments made by the Commissioners through Mr Bakewell were essentially dishonest and were motivated either by the wasted costs issue referred to in paragraph 1 of the letter or by some longstanding animosity between Mr Bakewell and Mr Hammond which had existed since adolescence. When the hearing resumed on 3^rd September, Mr Mc Nicholas raised before the Tribunal the issue of dishonesty referred to in his recent correspondence with Mr Freeman. Mr Freeman was asked by the Chairman whether he was now alleging mala fides. He said that he was alleging bad faith in the form of the CITES allegation, but disowned any suggestion that Mr Bakewell in some way profited from his conduct. He was also saying that the Commissioners had "deliberately and carefully" over-emphasised the importance of some evidence and ignored evidence to the contrary.

Following this exchange Mr Mc Nicholas produced overnight some written submissions setting out the authorities relied upon for the proposition that an allegation of dishonesty must be clearly pleaded prior to it being put to a witness in cross-examination, so that the witness and the party against whom the allegation is made should have a proper opportunity of knowing what the allegation is and of preparing to deal with it. Paragraphs 2.13 and 2.14 of the submissions read as follows:

On 4^th September Mr Freeman repeated that Pegasus did have a strong suspicion that there was mala fides on the part of Mr Bakewell and the Commissioners. There was then a discussion as to whether, under the Tribunal Rules (which make no provision for pleadings as such), the Tribunal could direct an applicant to particularise allegations of dishonesty in the way that Mr Mc Nicholas was asking. In the event, after an exchange with the Chairman about what had to be proved in order to challenge best judgment, Mr Freeman indicated that he did not intend to pursue the CITES allegation and that point dropped away. But when Mr Bakewell resumed his evidence in cross-examination on 6^th September, Mr Freeman began by asking him about his school career and whether he had met Mr Hammond at that time. Mr Bakewell said that he had not come across Mr Hammond until about 1991. The only possible purpose of this line of cross-examination was to establish an early link between the two men, relevant to the allegation of a dishonest motive in relation to the assessments. The evidence of Mr Bakewell resumed on 12^th December, following the interposition of a number of witnesses including Mr Hammond. By then Mr Freeman had produced a further document entitled "Appellant’s Focus Regarding Best Judgment" dated 9^th December 2001, the purpose of which was to show that the 1997 assessments did not satisfy either of the possible tests on best judgment canvassed on the appeal. The relevant parts read as follows:

Mr Bakewell was cross-examined on 12^th, 13^th, 14^th and 17^th December. The cross-examination covered most of the points raised by Mr Freeman in his various statements and in the Extended Notice of Appeal in support of Pegasus’s case that the assessments were excessive in amount. I shall come to the detail of some of this later in this judgment. Consistently with the "Focus" document, most of the cross-examination was directed to putting to Mr Bakewell the reasons why he should have taken a different view of the evidence he had to consider. Dishonesty or a lack of good faith was not put to him in terms, except that at the very end of his cross-examination it was put to him that his evidence was not reliable and, to use Mr Freeman’s words, "could possibly be a lie". In his written closing submissions following the conclusion of the evidence, Mr Freeman analysed the key points in Mr Bakewell’s assessment of the scale of Hammond’s evasion, with references to extracts from the transcript. There are references in that document to evidence being ignored, but the submissions for the most part concentrate on the alleged weaknesses in the evidence and methodology relied on by Mr Bakewell and are directed to persuading the Tribunal to take a different view of the facts. Nowhere in this document is there any analysis of whether it was possible for Mr Bakewell to have taken a different view, or of the reasons he gave in cross-examination for doing so. In particular no express allegation is made of any conscious failure to pay due regard to any of the relevant material or any bad faith.

In his closing submissions to the Tribunal Mr Freeman said that, in relation to best judgment, the test of dishonesty was intellectual dishonesty, not financial dishonesty. During the course of this appeal Miss Sharpston was at pains to emphasise that whatever the position may have been at the outset of her client’s appeal before the Tribunal, it was made clear by the end that no allegations of dishonesty or conscious bad faith were being made. Mr Freeman confined himself to submitting that the assessment was capricious. Mr Bakewell had rejected contrary evidence out of hand and had taken no account of Mr Freeman’s reports.

Article 22(2) of the Sixth Directive requires every taxable person to keep accounts in sufficient detail for VAT to be applied and inspected by the tax authority. This is given statutory effect by paragraph 6 of Schedule 11 to the 1994 Act and by the 1995 VAT Regulations (SI 1995/2518: regs 31 and 33). The Sixth Directive does not itself prescribe machinery for the collection of VAT, but instead Article 22(8) of the Directive provides that:

The 1994 Act deals with the evasion or non-payment of VAT by imposing penalties for mis-declarations of the tax and failures to keep proper records (see ss.63 and 69) in addition to the sanctions for the dishonest evasion of VAT contained in ss.60 and 72. When a taxable person fails to make returns or to keep the necessary or complete records, then the Commissioners are given power to assess the amount of VAT payable. Section 73(1) of the 1994 Act provides as follows:

On an appeal under s.83(p) the Tribunal has the power to reduce the amount of the assessment. Section 84(8) of the 1994 Act provides that:

Before considering the authorities on best judgment it is necessary to make two preliminary points which are obvious from the terms of s.73(1) and the circumstances in which the power to assess falls to be exercised. The s.73(1) power only arises and can only be exercised when the person chargeable to VAT has failed to make the necessary returns or to keep the records required under the Act, or where those returns appear to be incomplete or incorrect. The degree of non-compliance may therefore vary from cases of deliberate and outright evasion to cases where returns are made but on an incorrect or inadequate basis. The material and information available to the Commissioners in making their own assessment will therefore vary according to the circumstances, but in every case, to a greater or lesser degree, they will have to deal with a trader whose records or returns are incomplete. This necessarily requires the Commissioners, as s.73(1) makes clear, to exercise judgment in making their own assessment of the amount of VAT actually due. The second point is that it is the Commissioners who are required to make the assessment to the best of their judgment. This is not intended to be, and cannot by its very nature be, a guarantee of accuracy. The Commissioners can do no more than bring to bear their own best assessment of what tax is due, on the basis of the information which exists and is available to them. If the trader is dissatisfied with the assessment, he can appeal to the independent Tribunal and put before it any arguments in favour of a reduction in the assessment which he has. The right to appeal on quantum is therefore a recognition by Parliament that the assessment may, by its very nature, be inaccurate or flawed, and the Tribunal is given power to adjust it accordingly.

The possibility of error in the making of a s.73 assessment has also been recognised in the authorities which consider what is required of the Commissioners of Customs and Excise in order to exercise best judgment. In Spillane v. Customs and Excise Commissioners [1990] STC 212, the Appellant had carried on the business of manufacturing and selling counterfeit sound-recording tapes. He had made no VAT returns at all nor kept any records. He also refused to tell the Customs officers whether he was selling the tapes wholesale or retail. An assessment was made, based on a selling price of £2 per tape, which was the maximum sale price justifiable on the information available. Subsequently additional information was obtained and the Commissioners reduced the assessment by some £40,000. The Tribunal rejected the contention that the original assessment was not made to best judgment, finding that the officer had done his best on the information then available and had acted in good faith. Simon Brown J dismissed the trader’s appeal. At page 218 he said this:

The authorities referred to by the learned judge begin with the judgment of Woolf J in Van Boeckel v. Customs and Excise Commissioners [1981] STC 290. That case concerned an assessment of the takings of a public house. There were no visits made to the premises and the manager was not interviewed. Instead the Commissioners based the assessment on the takings during a test period of five weeks. The Tribunal reduced the assessment to take account of pilferage alleged by the taxpayer, but held that the assessment was nonetheless made to best judgment. Woolf J dismissed the appeal. On the issue of best judgment he said this (at page 292):

The judgment of Woolf J was considered by the Privy Council in the case of Bi-Flex Caribbean Limited v. The Board of Inland Revenue (1990) 63 TC 515. This was an appeal from Trinidad and Tobago concerning assessments to income tax following a failure by the taxpayer to make returns. The relevant legislation required the assessments to be made by the Board "to the best of its judgment". At page 522 Lord Lowry said this:

These authorities have most recently been considered by the Court of Appeal in the case of Rahman v. Customs and Excise Commissioners (No. 2) [2002] EWCA Civ 1881; [2003] STC 150. Following an investigation by Customs officers it was calculated that there was an under-declaration of takings by the trader, amounting to a shortfall in the VAT paid of £9,897 in the relevant accounting period. The officer notified the Appellant that she would treat VAT for earlier accounting periods as under-declared by the same proportion. An assessment was therefore made in the sum of £17,249 and during the appeal the Commissioners accepted that this should be reduced to £15,117. One of the members of the Tribunal dissented on the issue of best judgment and did not participate in the determination on quantum. Carnwath J set the decision aside and ordered a re-hearing before a differently constituted Tribunal. By the time of the re-hearing the Commissioners had accepted that there were errors in the calculation of the earlier assessment, including the double counting of purchase invoices and delivery notes, and agreed to reduce the assessment to £7,683. The Tribunal held that the Commissioners had nonetheless exercised best judgment in relation to the original assessment and confirmed it in the reduced amount of £7,683. The taxpayer appealed and contended that the concessions made by the Commissioners that there had been a miscalculation, including double counting, showed that the original assessment could not have been made to best judgment. Lawrence Collins J dismissed the appeal. He rejected the argument that the Tribunal should adopt the objective test of what a reasonably competent officer would have done when preparing the assessment on the basis of the material available. He said ([2002] STC 73 at [20]):

The Court of Appeal dismissed the taxpayer’s appeal from that decision. At paragraph [32] Chadwick LJ dealt with the taxpayer’s principal argument as follows:

During the course of the hearing before the Tribunal Mr Freeman did, as I have described, seek to impugn the honesty of Mr Bakewell and the good faith of the investigation and the subsequent assessments. In his opening submissions he put his case in two ways. He said that the assessments failed to satisfy the test laid down in Wednesbury (Associated Provincial Picture Houses Limited v. Wednesbury Corporation [1948] 1 KB 223) in that Mr Bakewell had disregarded relevant material and had reached an estimate of undeclared takings which was wholly unreasonable. He also submitted that Mr Bakewell had acted "vindictively, capriciously and made a spurious estimate or guess in which all elements of honest judgment are missing". 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. These terms were not further explained to the Tribunal and were, in the circumstances, misused. Capriciousness means a tendency to sudden or unpredictable changes of attitude, but it is used in the authorities I have referred to as denoting a decision made on a whim: in other words, a conscious lack of judgment. It is not an objective criticism. In the light of the abandonment of the case based on dishonesty, however, it is clear that Mr Freeman was not using the term in that sense. He linked capriciousness to intellectual dishonesty. 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.

Ultimately, however, what matters is the test which the Tribunal chose to adopt. In its decision the Tribunal set out what it considered to be the test of best judgment as follows:

This decision predates the judgment of the Court of Appeal in Rahman (No 2). For the Commissioners Miss Foster QC submits that it misapplies the law on best judgment. The reference in the authorities to an assessment being "wholly unreasonable" is not intended to be, she says, a distinct test of an objective kind, nor a matter unconnected with the officer’s state of mind. It is simply a shorthand for saying that the officer has failed to do his honest best. The question to be answered in all cases is not simply whether the assessment is apparently excessive or wrong, but whether it represents (even if excessive or wrong) a genuine and honest attempt to assess the taxpayer’s liability on the material available. That question cannot be answered by reference to an objective test of the Wednesbury kind and is inconsistent with such a test. When making his assessment, the officer is of course required to base it on the material available to him and to fairly consider that material. It is difficult to see how his assessment could be a genuine and honest one, were he to do otherwise. But it is not enough merely to show that he took a mistaken view of the facts or made some other error of calculation or methodology. The Tribunal did not apply the test laid down by the Court of Appeal in Rahman (No. 2) and instead carried out a detailed critique of Mr Bakewell’s methodology and conclusions, based upon their own assessment of the evidence. Because they reached a different conclusion, they inferred from this that Mr Bakewell had not exercised best judgment. This, she says, is evident from two passages in paragraphs 232 and 250 of the Tribunal’s decision, where it sets out its conclusions on best judgment as follows:

I shall come later in this judgment to consider whether (applying the correct test) there was any evidence before the Tribunal from which it could properly have concluded that Mr Bakewell did not exercise best judgment. But the first issue to determine is whether the Tribunal did in fact apply the right test. It seems to me that I am bound to apply the law as set out by the Court of Appeal in Rahman (No. 2), which explains the earlier authorities and lays down as the standard for the exercise of best judgment the test of whether the Commissioners, through their officers, made a genuine and honest attempt to calculate the tax due. 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.

Miss Sharpston seeks to meet these points in two ways. First, she submitted that for an assessment to be made to best judgment, it was not enough merely to show that the officer, however mistaken, had acted in good faith. That would create, she said, an impossible burden for the taxpayer to overcome and would be inconsistent with the references in the judgments of Woolf J in Van Boeckel and Carnwath J in Rahman v. Commissioners of Customs and Excise [1998] STC 826 to an assessment not being made to best judgment if it is wholly unreasonable. The Tribunal, she submits, adopted this test. To decide whether an assessment is wholly unreasonable, the Tribunal must review the evidence in detail to see where the parameters of what would be a reasonable conclusion lie. Having done this, they are then able to compare the officer’s conclusions with that assessment and to determine whether it lies outside the boundaries of what is justifiable. The Tribunal followed this course and concluded that Mr Bakewell’s assessment lay outside the parameters of the reasonable: see paragraph 250 of the decision.

In relation to the Court of Appeal’s decision in Rahman (No. 2) Miss Sharpston referred to paragraph 32 of Chadwick LJ’s judgment (quoted above) and in particular to the passage in which he says that 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". This extract, she submits, indicates that the "wholly unreasonable" test promulgated by Woolf J in Von Boeckel has been preserved.

Alternatively, as her second line of argument, Miss Sharpston submitted that Rahman (No. 2) was decided per incuriam because the Court of Appeal failed to take into account the impact of Article 1 of the First Protocol to the Convention for the Protection of Human Rights and Fundamental Freedoms 1950 ("Article 1") and the principles established under European Community Law governing respect for human rights, good administration, the right to a fair hearing, legitimate expectation and proportionality. At the start of the appeal I gave leave for the Respondent’s Notice to be amended to include these arguments as different or additional grounds for upholding the Tribunal’s decision.

I am not persuaded that on a proper reading of Chadwick LJ’s judgment a finding that the assessment was wholly unreasonable is sufficient to vitiate the assessment on grounds of best judgment, absent a finding that errors in the assessment were due to a failure to make a genuine and honest attempt to make a reasoned assessment. In the passage relied upon by Miss Sharpston Chadwick LJ was merely stating the alternative possible explanation for an erroneous assessment: i.e. not an honest and genuine attempt. He was not seeking to lay down an alternative test. The assessment under consideration in Rahman was based on double counting, which could justifiably be characterised as wholly unreasonable, but that did not amount in itself to a lack of best judgment. It follows that unless the decision in Rahman (No. 2) can be disregarded in favour of the principles of European law which Miss Sharpston relies upon, then there was a misdirection and misapplication by the Tribunal of the law in relation to best judgment. I turn, then, to those arguments.

Article 1 of the First Protocol of the ECHR provides as follows:

In Georgiou and another (trading as Mario’s Chippery) v. United Kingdom [2001] STC 80 the taxpayer challenged a penalty assessment to VAT and contended, inter alia, that the way in which the assessment was reviewed by the Commissioners and the Tribunal constituted a violation of his rights under Article 1. The ECHR rejected these arguments and (at page 90j) said this:

Because the principles involved are not disputed, it is enough, I think, for me to summarise the main principles of Community law relied upon by Miss Sharpston for this purpose. I can do so under six main headings.

VAT is an EC tax derived from the Sixth Directive. In relation to the implementation of the VAT Directives by a Member State the fundamental principles of EC law apply. In Ali and Begum v. Commissioners of Customs and Excise [2002] V & DR 71 (another case on VAT penalties) the Chairman of the London Tribunal, Stephen Oliver QC, summarised the position as follows:

National legislation implementing an EC directive is, whenever possible, to be construed in a way that is compatible with EC law: see Marleasing SA v. La Comercial Internacional de Alimentacion SA [1990] ECR 1-4135.

The protection of fundamental human rights is an integral part of the principles of EC law, and the ECJ has endorsed a liberal interpretation of legislation so as to accommodate this: see Nold v. Commission [1974] ECR 491; Marks and Spencer Plc v. Commissioners of Customs and Excise [2002] STC 1036 at paragraphs 73-77.

The Commission as a public authority under s.6 of the Human Rights Act 1998 must give effect to primary legislation (in this case s.73 of the 1994 Act) in a way that is compatible with Article 1. The meaning of best judgment should be modified to take account of the EC principles of law relied upon.

The taxpayer is entitled to rely upon procedural guarantees of good administration under EC law, including a duty of care in public administration and the right to be heard. An important component of the right to be heard is the right to have one’s argument taken into account. The right to be heard is a fundamental principle of EC law: see Hoechst v. Commission [1989] ECR 2859 at page 2923. In Belgium v. Commission [1986] ECR 2321 at page 2348 (a case on State aids) the ECJ said that:

Apart from Article 1 this is also an established principle of EC law. To decide whether a provision complies with the principle of proportionality:

In Marks and Spencer (supra) the ECJ at paragraph 44 ([2002] STC at page 1059) said that:

This summary of the principles of EC law which are relied upon does less than justice to the quality and comprehensive nature of Miss Sharpston’s presentation of the authorities. But she accepted that, though capable of exposition as separate rules, all the principles I have referred to could be summarised and bracketed as part of a general obligation on the part of the Commissioners as the revenue-collecting authority to act reasonably and fairly in carrying out their assessment of the tax due. A duty to act fairly involves, she submits, respecting the procedural safeguards such as the right to be heard. Looked at more widely, the machinery of assessment, coupled with the right of appeal, has to be judged in terms of proportionality, having regard to the consequences for the taxpayer of a wrong and excessive assessment which will remain binding on him unless he is able to set it aside or reduce it by challenging it on grounds of best judgment or quantum. This, she says, places a considerable burden on the taxpayer which can only be justified if the test as to what constitutes best judgment imports an objective criterion of reasonableness and is not merely dependent on the assessing officer acting honestly and in good faith.

As I indicated earlier in this judgment, the formula adopted by the Court of Appeal in Rahman (No. 2) as the test of best judgment accommodates an obligation on the part of the officer concerned fairly to consider all the material made available to him. A genuine and honest attempt to reach a correct assessment requires the officer to take those matters into account as part of his calculations. But I do not accept that honest mistakes and differences of view between the assessing officer and the taxpayer are enough to contravene the procedural safeguards inherent in the right to be heard and to have the tax fairly administered, nor that they render the assessment process as a whole disproportionate and unconscionable.

It is clear from the authorities cited to me that the application of the principles of EC law relied upon is essentially a pragmatic one, and that in order to determine whether the particular governmental process under consideration complies with the principles of good administration and proportionality, it is necessary to consider that process as a whole and to pay proper regard to the procedural safeguards (if any) which are built into the system. Miss Sharpston’s principal criticism of s.73, as interpreted by the Court of Appeal in Rahman, is that it places too high a burden on the taxpayer of proving the lack of a genuine and honest attempt to assess. If the standard of best judgment is an objective one based on proportionality, then the question will be whether the Commissioners acted reasonably in making the assessment they did on the material available. Questions of motive and good faith (or the lack of it) will be avoided in favour of an objective analysis of the merits of the assessment and whether reasonable steps were taken fully to inform that process on the basis of the material available.

The difficulty, however, about this argument is that it gives too little weight to the taxpayer’s right of appeal and to the role of the Tribunal. The right of appeal under s.83 of the 1994 Act has, in my judgment, to be taken into account when assessing the compatibility of s.73 with the principles of EC law under consideration. It provides an internal mechanism for correcting the assessment procedure. An appeal lies both on the issue of best judgment and on quantum. The Tribunal on an appeal against quantum is not in any way bound by the officer’s own assessment and is free to form its own view on the evidence presented to it. In cases of hardship, payment of the tax is postponed until after the appeal. If the taxpayer is able to establish that the assessment is unreasonable and wrong, then the appeal on quantum will succeed. If Miss Sharpston’s argument is correct, that might also be sufficient to show that the assessment was not made to best judgment. But the consequence of that would be to discharge the assessment and not merely to reduce it to the correct amount. Regardless of the correct test of best judgment, a taxpayer who appeals against the assessment under s.73 will be able to have it reduced merely by showing that it was wrong. Given that the purpose of s.73 is to enable the recovery of undeclared and unpaid VAT, I can see nothing obviously wrong or disproportionate about a system which allows the excessive assessments to be corrected in that way, but leaves the validity of the assessment intact. Miss Sharpston’s argument is not in truth directed to the taxpayer’s ability to reduce the assessment. That can be achieved under s.73 on an objective basis by an appeal on quantum. Her argument is that the taxpayer should be afforded the opportunity of discharging the assessment in toto merely by being able to show that it was unreasonably or mistakenly calculated, and that ss.73 and 83 contravene the principles of good administration and proportionality if an attack on the validity (rather than the amount) of the assessment requires proof that the assessment was arbitrary in the sense that it did not represent a genuine and honest attempt to calculate the tax due.

I have no hesitation in rejecting that argument. It seems to me that it ignores the realities of what s.73 is designed to deal with: i.e. cases where there has been an under-declaration or evasion of the tax and in which the Commissioners have to work from material that is often limited or even non-existent. I can see nothing unfair to that class of taxable person in a system which gives them a right to appeal against excessive assessments to an independent tribunal, but does not allow the validity of the assessments to be impugned except upon proof of a lack of good faith on the part of the Commissioners. The decision of the Court of Appeal in Rahman is fully in line with the sentiments expressed by Woolf J in Van Boeckel and Simon Brown J in Spillane and cannot be said to be disproportionate. It is a valuable and welcome re-affirmation that challenges to the validity of s.73 assessments on grounds of lack of best judgment should be rare. As it is, they are only too often pursued at great length and expense solely for the tactical reason that, due to the passage of time, the Commissioners will be unable to make new assessments for the same period if the assessments under challenge are set aside. At the end of his judgment in Rahman Chadwick LJ expressed the view that Tribunals should in most cases decide the issue of quantum and then consider whether to decide the issue of best judgment. He went on to say that even if the Tribunal found that the Commissioners had failed to exercise best judgment, it might nonetheless exercise the power contained s.84(5) of the 1994 Act to preserve the assessment in a different sum. Attractive as that latter solution is, it will only be available when the amount of the tax found to be due exceeds the assessment made. It will not therefore be of assistance in most cases where the assessment itself is found to be excessive and the Tribunal is then faced with having to decide on the taxpayer’s appeal whether it should also be set aside on the ground that it was not made to best judgment. I share the view of Chadwick LJ that the Tribunal should in all such cases assess quantum, but that it should only allow an attack on best judgment to be developed if it becomes apparent from the evidence on quantum that there is a real ground for such a challenge, based on a lack of good faith. The Tribunal should not hesitate to adopt a pro-active stance towards the management of the appeal on this issue. The adoption of the two-stage approach in this and other appeals leads to an artificial distinction between best judgment and quantum and to the unnecessary lengthening of the hearing. It also has the unfortunate and undesirable consequence (evident from the decision in this case) that no effective decision is made on quantum when that is, or ought to be, the essence of the exercise.

If the officer does his honest best to take account of the evidence and any representations that are put to him, then the procedural safeguard of the right to be heard is satisfied. I do not accept Miss Sharpston’s submission that, having considered the material provided by the taxpayer, the officer is required to disclose his proposed assessment to him before it is made and give him an opportunity of commenting on it. If that were right, the process would become endless. Nor can I see on what basis Mr Freeman can be said to have had a legitimate expectation that he would be consulted in that way. In these circumstances I am not persuaded that the test of best judgment, in order to render the process of assessment proportionate either under EC law or in accordance with Article 1, requires the importation of an objective test of some kind in the way which Miss Sharpston suggests.

It follows that I do not accept that the test of best judgment propounded by the Court of Appeal in Rahman (No. 2) is in any way inconsistent with Article 1 of the Convention or with the other principles of EC law which I have summarised. Nor do I believe that the Court of Appeal would have been persuaded to take a different view, had the full range of authorities been cited to it. A point of particular significance, as Miss Sharpston accepted, is that the ECHR in Georgiou must have accepted as proportionate the test of best judgment set out by Sir Louis Blom Cooper QC (sitting as a Deputy High Court Judge) [1995] STC 1101, who rejected the taxpayer’s appeal at first instance and whose formulation of the test was accepted with only minor amendments by the Court of Appeal. At page 1116 he set out the following propositions of law in relation to s.73:

The remaining question is whether, on the Tribunal’s findings of fact and on the evidence before it, Mr Bakewell did exercise best judgment in the Rahman sense: i.e. did he make a genuine and honest attempt, on the material available to him, to assess the tax due. Pegasus submits that even if the Tribunal misdirected itself or misapplied the law on best judgment, it would, on its findings of fact, have reached the same conclusion on a correct application of the law. The Commissioners dispute that, but contend in the alternative that if the Tribunal’s findings disclose an absence of best judgment in the correct sense, there was no evidence on which it could properly have reached those conclusions.

I should mention at this stage that in the Respondent’s Notice Pegasus relies upon a number of additional factors such as the value placed on each trip and the mark-up of 150% as being wholly unreasonable and not therefore the subject of best judgment. These criticisms were rejected by the Tribunal and, for the reasons explained earlier in this judgment, the allegedly unreasonable nature of these matters is not in itself sufficient to make out a failure to exercise best judgment. I do not therefore propose to go behind the Tribunal’s findings on these issues.

As indicated earlier, the assessments were calculated upon four basic assumptions:

In its conclusions on best judgment the Tribunal (at paragraph 243) described this method of calculation as having an initial appearance of logic, but also as being highly speculative. The figure of £18,508 is said in the same paragraph to be inflated and the 43/57 split based on the Kloeg invoices dubious. The assumption that Pegasus was involved in 166 trips is said to be unwarranted. Paragraph 243 sums up these findings as follows:

The figure of £18,508 is based on Mr Bakewell’s estimate of the cost of birds purchased on trip 165 (22^nd May 1995) which was the last trip before the customs raid on 24^th May. On 24^th May 1995 (as described in paragraphs 92-99 of the decision) a driver was arrested and a clipboard was found with some instructions. These were to collect purchase items from some six foreign dealers. The driver (a Mr Daniel) had made a trip to Belgium on 22^nd May and had been making weekly trips since 1993. Some of the invoices found at his home were for bird-food and cages. The £18,508 figure was made up of invoices from all six suppliers, and the Tribunal sets out in some detail how it was arrived at and what assumptions were made. In the case of the invoices expressed in Dutch florins, Mr Bakewell adopted an exchange rate of 2.3 florins to the pound, compared to a bank rate of 2.4823 florins. Mr Daniel was observed on 22^nd May delivering boxes to Pegasus from the van which had been taken to the continent. Mr Bakewell also calculated the value of trip 159 (8^th April 1995) as £18,613 and used it as a cross-check. Mr Foster’s accountants, Grant Thornton, prepared a schedule for the Newton Hearing. This showed 165 trips by Mr Foster between 1993 and 1995, at a range of values up to £15,525. Many of the trips were given a purchase value of less than £1,000.

The Tribunal (at paragraph 233) criticised the exchange rate used by Mr Bakewell and described the figure of £18,508 as inflated, given that one of the dealers (Van der Feesten) was paid by cheque and the purchase was not, therefore, off-record. It was accepted by Mr Mc Nicholas before the Tribunal to be the most robust figure available, i.e. one based on a trip for which the most evidence was available. The Tribunal concluded that it was not wholly unreasonable as a figure for trip 165. Mr Bakewell was cross-examined on this issue on 31^st May 2001 and again on 17^th December 2001. He explained that it provided, in his view, the best evidence of the value of a vanload and that, prior to the raid, he had no way of knowing the value of the earlier trips. On 31^st December Mr Freeman questioned him about the correctness of using the £18,508 figure, based on one trip, as a multiplier for all 165 trips:

The Tribunal had to consider how many trips Mr Foster or Mr Hammond made and whether Pegasus was involved in all of those trips. On the first issue the Tribunal accepted that with one exception (trip 114) there was evidence to support a total of 165 trips and that there "was nothing unreasonable in assuming that either or both Mr Foster and Pegasus were involved in some 166 trips apart from some duplication". The only real issue was whether there was sufficient evidence to link Pegasus as well as Mr Foster to all 166 trips, as opposed to the 18 admitted by Mr Hammond at his criminal trial. As the Tribunal records, the main reason for linking Pegasus to all 166 trips was what has become known as the "six out of six theory". On the six occasions in 1994 and 1995 (1^st September and 28^th November 1994; 21^st January, 21^st March, 8^th April and 22^nd May 1995) when Mr Foster’s vans were followed, they delivered purchases to Pegasus. These were trips 107, 133, 139, 155, 159 and 165 on the Trip Schedule. Mr Bakewell’s evidence, contained in his composite statement prepared for the Tribunal hearing, was that:

Mr Freeman in his second report of 20^th January 1997 said that there was no evidence linking Pegasus with most of the trips and that some support for this was provided by a diary kept by Mr Foster. This records sums owed by and to Mr Foster during 1995. The document was seized in the May 1995 raid. Mr Freeman submitted to the Tribunal that the diary showed that there was no joint enterprise between Mr Foster and Mr Hammond. It was put to Mr Bakewell during his cross-examination on 1^st June 2001, but he said that he did not consider it to be a complete record. The Commissioners rely inter alia on the fact that trips 145 (7^th February 1995) and 165 (22^nd May 1995) are not referred to in the diary. The response to this is that trip 145 was an entirely on-record trip by Mr Hammond, when he bought some £15,525 worth of parrots from a dealer called Orni-Mondo. In relation to trip 165 it is said that Mr Foster did not have time to write his diary up before the raid on 24^th May.

Mr Bakewell was cross-examined about trip numbers on various occasions during the hearing. He accepted that he had largely discounted evidence produced by Mr Ost (one of the continental suppliers) that Mr Hammond was invariably present. Mr Ost had described Mr Hammond (wrongly) as having long blonde hair. He also accepted that there was no direct evidence linking Pegasus to 54 of the trips. But he said that on every one of the six occasions (bar one) when the trips were observed, Mr Hammond was present, even though not referred to on the ships’ manifests, and on every one of these six occasions the vans unloaded at Pegasus. He said that it was his belief that on most trips other vehicles were involved, but that the Customs followed Mr Foster’s vans and not known Pegasus vans. The selection was therefore random. When it was put to him by Mr Freeman that he had what was described as an idée fixe that Mr Hammond was involved throughout, he said this:

The Tribunal concluded that it was unreasonable on the evidence before Mr Bakewell for him to have assumed that Pegasus was involved in 166 trips: see Decision paragraph 242. In reaching this conclusion it relied on the fact that Mr Ost’s evidence linking Mr Hammond to the Foster trips was unreliable and that the "six out of six theory" was flawed because on four occasions Customs officers knew in advance that Pegasus was involved and followed the Foster vans on these occasions for that reason. I do not accept that that necessarily robs the observed trips of their evidential value as an indication of the regularity of the involvement of Pegasus in the trips, but I am not concerned to resolve issues of that kind on this appeal. They are questions of fact for the Tribunal. What is important is that the Tribunal accepted Mr Freeman’s submissions on this point. These concentrated in detail on the various trips and on the evidence about Pegasus’s involvement. No reference was made by Mr Freeman in his submissions to the allegation that Mr Bakewell had a closed mind on the point. His attack was confined to challenging Mr Bakewell’s assumption that it was right to infer from the six observed trips that Mr Hammond and Pegasus had a regular involvement. The only part of Mr Freeman’s cross-examination specifically referred to was that on 17^th December 2001, when the following exchange took place:

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.

Mr Freeman accepted the 150% mark-up as reasonable (Decision paragraph 235) and I need say no more about it. That leaves the 57/43 split as the last of the four components to be considered. This was based on three pairs of Kloeg invoices in July and August 1994. Mr Kloeg said that the "O" invoices were Pegasus and the "X" markings denoted Mr Foster. As the Tribunal found, seven Kloeg invoices for six earlier and seven later trips contained no "X" or "O" markings. The Tribunal concluded (paragraph 114) that although "O" denoted Pegasus, it did not follow that all the purchases on the "O" invoices were for Pegasus alone. It was not, they said a consistent system.

Again I have no difficulty with these findings as such. It was for the Tribunal to decide whether the apportionment based on the invoices that were marked "O" and "X" was justified. But the Tribunal’s rejection of this aspect of Mr Bakewell’s methodology is not sufficient to challenge best judgment. Nor, to be fair, does the Tribunal appear to have based its decision on that point. There is no finding in paragraph 234 of the decision about a lack of good faith or even a lack of reasonableness. The Tribunal is content to record that it felt bound to give much less weight the reliability of the Kloeg invoices as a guide to the split between Pegasus and Mr Foster. This is simply a judgment on the issue of apportionment. It seems to me that Mr Bakewell was entitled to take a different view and that there was some evidence, in the form of the Kloeg invoices, to support that. During cross-examination Mr Bakewell said that the instructions to pack "X" and "O boxes were found at Pegasus. In these circumstances there is no basis on which to challenge his assessment of this evidence as not being to best judgment.

Although the Tribunal accepted many of Mr Freeman’s criticisms of the four basic components in Mr Bakewell’s assessment, it is clear that it also attached a considerable amount of weight to other material which was relevant to a consideration of whether the assessments were reasonable. Again it is not necessary for me to deal extensively with the detail of these matters so much as the way in which the Tribunal approached the task of assessing Mr Bakewell’s own response to this material.

In summary the Tribunal took the view that evasion on the scale contemplated by the Commissioners’ assessment was beyond the ability of Mr Hammond and Pegasus. On Mr Bakewell’s calculations there were suppressed off-record sales of 80%, generating net sale proceeds of £2.6m. This would, the Tribunal said, have required a level of sophistication on the part of Pegasus which it did not believe it had. The Tribunal based these conclusions on a number of factors which were stressed by Miss Sharpston during the course of this appeal. Most significant were the control visit by Customs and Excise in January 1994, the extent of on-record sales by Pegasus and the level of documentation, including audited accounts; the size of its premises and the inability of the authorities to detect the large sums of cash which would have been generated from the sales, either in specie or in the form of the level of Mr Hammond’s personal expenditure.

The control visit took place on 5^th January 1994 and was conducted by two officers. They spent an hour at the premises with Mr Hammond and then visited Pegasus’s accountants, Messrs Goldberg Chandler. The officers reported that most of the birds sold had been purchased in the UK. There were five staff at the shop and takings were reconciled and banked daily. The premises used by Pegasus comprised a small office, the pet shop and a café area. At Goldberg Chandler the officers inspected records of takings, sales and purchase day-books, sales and purchase invoices, till rolls and bank statements. They also traced export invoices. They concluded that there was no apparent reason to doubt the credibility of the business. If Mr Bakewell is correct, then for the three periods 03/93 to 09/93 covered by the control visit the output tax should have been £158,649 more than the £36,214 declared. This would have involved additional off-record sales of £906,560, compared with the £228,390 disclosed on the company’s returns.

Goldberg Chandler checked the documents inspected at their premises as part of their annual audit, with the exception of the till rolls. They also attended an annual year-end stock-take at the premises. They had certified that the accounts for the year to 31^st March 1994 gave a true and fair view as at that date of the year’s profit.

The raid on 24^th May 1995 also revealed sheets headed "Money Taken" for the period from 5^th to 16^th May 1995, when Mr Hammond was away on holiday. They recorded till takings, trade cheques and credit card receipts, together with monies paid out in cash or by cheque. For the twelve days they totalled a maximum of £23,520.

The Tribunal calculated that in order to generate sales on the scale contemplated by Mr Bakewell’s assessment, it would be necessary to turn over the entire stock held at the premises 31.3 times each year or every eleven to twelve days. The Tribunal accepted that turnover was higher and faster than shown in the audited accounts, but that it was not likely to have been five times higher and faster, as required. Put shortly, the sales which Mr Bakewell considered to have taken place would have involved a level of activity at the shop not witnessed by Mr Bakewell or supported by other evidence. Although some sales took place at fairs and exhibitions, the stock was initially delivered to Pegasus and there was no evidence of it being stored elsewhere.

As to Mr Hammond’s lifestyle, there was evidence that he owned a sports car and enjoyed foreign holidays, including a safari to Africa, but from the 40 boxes of papers seized during the 24^th May raid, no evidence was revealed of hidden assets.

In his composite statement Mr Bakewell did not shrink from asserting his belief that there were substantial off-record sales by Pegasus. He said that the trade in wild birds imported into Europe from Africa was significantly off-record and that MAFF statistics suggested that only 10% of trade was documented. He accepted, however, that off-record birds were likely to have been sold as part of Pegasus’s general stock and not merely at off-site events such as exhibitions. He also accepted during cross-examination that some of his estimates of stock at Pegasus’s premises during his visits may have been inaccurate. But it was not put to him that his assessment (mistaken or not) was not honestly made or that he had deliberately or consciously failed to take into account the points about stock, turnover and the level of suppression which subsequently impressed the Tribunal. Indeed it is clear that the main thrust of Mr Freeman’s cross-examination of Mr Bakewell was to challenge the four components of his assessment and, in particular, the assumption that Pegasus was involved in all the 166 trips. The points about the on-record sales and the ability of Pegasus to suppress sales at the level required were largely presented to the Tribunal as a matter of submission by Mr Freeman as to why Mr Bakewell’s assessment could not be right.

This explains why the Tribunal in paragraph 250 of its decision was not able to make, and did not make, a finding based on the examination of Mr Bakewell, that he did in fact deliberately fail to consider relevant material before him. Instead it based its view of the way he conducted his assessment on an inference from the Tribunal’s own view of the evidence which was available to him at the relevant time. This is confirmed by its reference to the assessments having no logical basis. The points about turnover and the ability of Pegasus to suppress the required level of sales are undoubtedly powerful ones. But to conclude that Mr Bakewell closed his mind to any material which did not fit his case, merely from the fact that the assessments were, in the Tribunal’s own view, excessive or wrong, excludes the possibility of an honest mistake. This was not the subject of any consideration by the Tribunal.

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. The Tribunal’s determination that the assessment was not made to best judgment must therefore be set aside and the Commissioners’ appeal on that issue allowed. The challenge by Pegasus to best judgment fails. I shall remit the remaining undecided issue of quantum to a Tribunal to be determined. In the absence of agreement, I shall hear Counsel as to whether the issue of quantum should be decided by the same or a differently constituted Tribunal.