Mr Arif appeals against a decision of the VAT & Duties Tribunal (“the Tribunal”) given on 18 October 2005. The effect of the decision was to uphold two assessments to VAT, with relatively minor modifications. One was an estimated assessment to the best of the Commissioners’ judgment, made under section 73 of the Value Added Tax Act 1994. The other was an assessment to a penalty under section 60 of that Act. The latter assessment was dependent on a finding of dishonesty against Mr Arif.

Mr Arif runs a fish and chip shop in Huddersfield, called “Trinity Fisheries”. The case against him was that he had dishonestly underdeclared his takings with a view to evading liability to VAT. The Tribunal found the case against him proved.

An appeal from the Tribunal lies to the High Court under section 11 of the Tribunals and Inquiries Act 1992 if any party to the proceedings is “dissatisfied in point of law” with the decision of the tribunal.

The Commissioners began to take an interest in Mr Arif in the summer of 1998 when an inspection of his records revealed certain discrepancies. In June 1998 the Commissioners wrote to Mr Arif calling upon him to keep a record of all payments as they were received by him from cash customers. At the relevant time Mr Arif had a till. The till could be used to provide readings of takings. There are two main types of reading. One is a “Z reading”. The other is an “X reading”. The Tribunal’s finding was that “Z readings” are cumulative since the last “Z reading was taken”; whereas an “X reading” can be taken without interrupting the “Z reading” sequence. If correctly set the till will give the date and time of each reading. In fact Mr Arif’s till was incorrectly set; but the displayed settings were adjusted before the Tribunal so as to correspond with real time.

On Thursday 25 February 1999 two Customs officers made an unannounced visit to Mr Arif’s shop at 23.30 hours, with a view to watching Mr Arif cash up his takings for the day. Mr Arif was not himself present, but was contacted by telephone and agreed to a meeting on the following day. One of the officers (Miss Norris) took a till reading at 23.50. It was an “X reading”. It showed £563.81. The cash in the till amounted to £438, but Mr Arif had been in that afternoon and taken cash from the till.

Customs officers carried out a full-day invigilation on the following day, Friday 26 February 1999. A till reading taken at the end of the day, at 23.00, showed takings of £946.50. The cash counted was £945.10, which the Tribunal regarded as an insignificant discrepancy. Mr Arif was accustomed to record his daily takings in an exercise book. This showed that Fridays were the best trading day of the week; that on occasions more than £800 had been taken on a Friday, but never more than £900.

Customs took the view that Mr Arif had dishonestly underdeclared his takings. The case against him was made up of a number of strands.

The results of the invigilation were that the observed takings for Friday 26 February 1999 were the highest takings recorded for any Friday. The Commissioners’ case was that this suggested an underdeclaration of takings.

The Commissioners’ case was that on the day of the invigilation Mr Arif had adopted “counter-measures”; that is to say that he had artificially kept his takings low on the day that they were observed by Customs officers. This would have had the effect of minimising any discrepancy between his declared takings and the takings that Customs officers were able to observe. The two matters on which particular reliance was placed was that fish were not being sold at full price, and customers were not served when the business ran out of food.

It is common experience that if a business is underdeclaring its takings, it will also underdeclare its purchases, otherwise it will be apparently running at a loss or at an abnormally low profit margin. So Customs investigated the two principal sources of supply for a fish and chip shop: potatoes and fish.

The records relating to the supply of potatoes were examined for the period 1 January 1998 to 2 March 1999. The examination showed that after the first three or four months of that period, supplies of potatoes were reconciled with purchases. The records relating to the supply of fish were also examined. This examination covered only one quarter. It revealed that cash purchases of fish had not been declared.

The Commissioners therefore contended that the evidence showed a partial non-declaration of potato purchases, and a partial non-declaration of fish purchases.

Following the invigilation, recorded takings at Trinity Fisheries rose by about £1,000 a week. The Commissioners relied on this as indicating that takings before the invigilation had been underdeclared.

On 26 February 1999 customs officers took away some till rolls for examination. The audit roll from 10 February 1999 appeared to have been through the till twice; so it was overprinted and not easy to read. Miss Norris extracted information from it. The Tribunal themselves examined the till rolls; and concluded that they did contain the information that Miss Norris extracted. Miss Norris focussed on three dates and extracted the following information:

She deduced that the differences between these figures appeared to represent the takings of the business. She excluded from the cumulative totals the takings of 25 and 26 February when customs officers had been present on the premises. Having done this, she deduced that the takings since 11 February 1999 were considerably in excess of the figures declared by Mr Arif on which his VAT returns were based.

Mr Arif was due to attend at Customs’ offices for interview on 1 April 1999. The date was changed to 6 May 1999, but in the meantime Mr Arif changed advisers and thereafter refused to attend for interview. There was, however, some correspondence between Customs and Mr Arif’s advisers. But the Commissioners contended that inadequate information had been provided.

The hearing before the Tribunal occupied ten days in 2005; two in February, three in May and the remaining seven in August. The Tribunal correctly directed themselves that the burden of proving dishonesty lay on the Commissioners; and that the burden of displacing quantum lay on the taxpayer.

Mr Arif was represented by Mr Nawaz, an accountant. The Tribunal thought that his cross-examination of the Commissioners’ witnesses was “impressively well-prepared and thorough”.

However, although Mr Nawaz succeeded in damaging the Commissioners’ case, the Tribunal nevertheless found that they had succeeded in proving dishonesty.

As I have said, an appeal lies from a decision of a tribunal only if a party is dissatisfied in point of law. A finding of dishonesty is essentially a question of fact. On what ground, then, can I interfere with this finding? The answer is provided by the well-known test in Edwards v Bairstow [1956] AC 14, 35 -36:

In Georgiou v Customs and Excise Commissioners [1996] STC 463, 466 Evans LJ amplified this as follows:

There is one other general comment that is appropriate at this stage. It relates to the evaluation of circumstantial evidence. Pollock CB famously likened circumstantial evidence to strands in a cord, one of which might be quite insufficient to sustain the weight, but three stranded together might be quite sufficient (R v Exall (1866) 4 F & F 922). Thus there can be no valid criticism of a tribunal which considers that one piece of evidence, while raising a suspicion, is not enough on its own to find dishonesty; but that several such pieces of evidence, taken cumulatively, lead to that conclusion.

I turn, then, to consider how the Tribunal evaluated the evidence before them.

The Tribunal found that the difference between the observed takings on the day of the invigilation and the declared takings for VAT were in part accounted for by the inclusion in gross takings of some purchases of zero-rated items. They took the figure for zero-rated sales at 5 per cent. However, even after making this allowance, the invigilation still suggested that the recorded figures for Mr Arif’s takings were low (para 37).

The Tribunal recorded that an innocent explanation had been given for the alleged counter-measures which could not properly be discounted (para 34). Later in their decision (para 68) they found that the alleged counter-measures had not been proved. Since the burden of proving this particular allegation was not discharged, it follows that for the purpose of the Tribunal’s decision, there were no counter-measures.

Having looked at the first two strands of the Commissioners’ case in isolation the Tribunal concluded (para 37):

Thus the Tribunal’s conclusion at this point was that the invigilation raised a suspicion, but not one of such strength as to enable a conclusion of dishonesty to be drawn from that alone.

The Tribunal considered that Mr Nawaz had succeeded in cross-examination in casting doubt on the reliability of parts of the Commissioners’ case. Their conclusion on the question of under-declaration of purchases was (para 48):

However, later in their decision (para 73) they said that although the supply side investigation was imperfect “it did raise questions”. Their overall conclusion on the supply side investigation was that (para 114):

The Tribunal came to a clear view on this question. They concluded (para 116):

The Tribunal considered that the evidence of the till roll, and the conclusions that could be drawn from it, were “the central issue in the case” (para 74). In evaluating that evidence they also attached importance to Mr Arif’s explanations of how it came about that there might be discrepancies between the till roll and the declared takings (para 70).

Mr Arif’s explanation was that he had used the till as a calculator, in helping him count his takings for his manual records, kept in the exercise book (para 88). The Tribunal summarised his evidence as follows:

The Tribunal rejected Mr Arif’s explanation. They did so for a number of reasons. First, having examined the till rolls, they could find no evidence of calculations or errors such as Mr Arif spoke about (para 152). Second, they noted that the explanation that Mr Arif had given to the Tribunal had not previously been given. If his explanation had been true he would have given it to his advisers, who would have raised it in correspondence with Customs (para 154). Third, they attached importance to the fact that Mr Arif must have known that till rolls were potentially a reliable record of his takings, and yet he was accustomed to throw them away (para 155). Fourth, they regarded the explanation that the till was used as a calculator as improbable because of the universality of pocket calculators (para 155). The Tribunal thus concluded that Mr Arif’s defence was “spurious” (para 157) and that Mr Arif had lied to the Tribunal (para 108).

The Tribunal considered this under the heading of mitigation. They gave more credit to Mr Arif for co-operation than the Commissioners had been willing to do.

Having come to their conclusion on the question of the till rolls, the Tribunal then re-visited the other evidence which, they had already concluded, raised questions that called for an answer. They expressed their overall conclusion as follows (para 158):

Quite rightly, they did not include in this list the alleged counter-measures which they had previously decided had not been proved. What they did was to take the various other strands of the Commissioners’ case cumulatively; and concluded from them that dishonesty had been established.

Mr Arif’s case on appeal was skilfully argued by Mr Conrad McDonnell. Had he made the self-same submissions to me as a fact finding tribunal, they might have been thoroughly persuasive. But as Evans LJ explained, the question for me is not whether the Commissioners had discharged the burden of proof, but whether there was evidence on which the Tribunal could have found that they had.

Mr McDonnell compared the till reading taken by Miss Norris at 23.50 on 25 February 1999 and that taken at 23.00 on 26 February 1999. The first of these two readings was taken at the end of the working day. 26 February was the day of the invigilation. The second of the two readings was taken at the end of that working day. So Mr McDonnell argues that the takings for that day are known and indisputable. If the two readings are compared, they overstate the known takings for 26 February by £565. Mr McDonnell submits that there are only three possible explanations for this discrepancy:

The sum of £565 was entered into the till early in its life and then the Z reading was reset;

One of the customs officers, unaccustomed to the way in which the till operated, entered £565 by mistake or

Mr Arif should have been believed when he said that he used the till as a calculator.

The first of these hypotheses depends on a more sophisticated understanding of the way in which the till operates than was apparently in evidence before the Tribunal. It depends on recognising that there are two different kinds of “Z readings” and two different kinds of “X readings”. These questions were not explored before the Tribunal; and cannot be opened up on appeal. The second of these hypotheses was not put to any of the officers who gave evidence before the Tribunal. The third of these hypotheses was that advanced by Mr Arif, which the Tribunal rejected.

Mr McDonnell acknowledged that this point was not argued below. If it had been the evidence might have been different. But in my judgment it is not open to an appellant, whose only recourse to this court is an appeal on a point of law, to re-run the case on the facts. I reject this criticism of the Tribunal as amounting to a legal error.

Based on the analysis of the till rolls the Commissioners attempted to calculate the extent to which Mr Arif had understated his takings (the “suppression rate”). The original suppression rate was 45.62 per cent. I have already described the outline of the methodology. The details of the calculation were attacked in evidence before the Tribunal; and Miss Norris accepted that some criticisms were well-founded. The revised suppression rate was 43.13 per cent, which the Tribunal accepted.

Mr McDonnell extrapolated the suppression rate to work out what Mr Arif’s takings should have been on the hypothesis that the suppression rate was correct. He did this in respect of three days; and tabulated the results as follows:

This showed, he said, that the calculation of the suppression rate was excessive. The takings for 25 and 26 February were known figures. They tallied with the till records. It is true that the takings for each of those days was higher than Mr Arif’s recorded average takings for that day of the week, but in the case of 25 February the difference was minimal; and in the case of 26 February the difference was not so great as to lead to a finding of dishonesty. The takings for 16 February recorded in the till were higher than the amount declared in Mr Arif’s books, and also higher than his declared average for that day of the week. But again the differences were not such as to lead to a finding of dishonesty.

Some attention was given to figures for individual days before the Tribunal. They said (para 161):

Concentrating for the moment on the takings for 16 February (which was the only one of these three days included in Customs’ calculations), there is an undoubted discrepancy between the receipts recorded by the till and those declared by Mr Arif (whether for that day or for an average day). The discrepancy was, in my judgment, evidence of dishonesty. Whether it was strong enough to satisfy the burden of proof (together with all the other material they considered) was a question of fact and degree for the Tribunal. In addition the exercise that Mr McDonnell undertook was not one that was placed before the Tribunal. It is not one which, in my judgment, the Tribunal were bound to undertake for themselves. The Tribunal are referees, not goalkeepers.

So far as the extent of the underdeclaration is concerned (i.e. the suppression rate) I have already quoted the Tribunal’s conclusion. No alternative rate was put forward by or on behalf of Mr Arif (not surprisingly, since he was contending that there had been no suppression at all).

Although the burden of proving dishonesty lies on the Commissioners, the burden of displacing the amount of the assessment lies on the taxpayer. Mr Arif failed to discharge that burden before the Tribunal. Whether he had done so or not was a question of fact for the Tribunal.

Mr McDonnell said that the unexplained £565 should in any event have been deducted from the overall takings established by the till rolls. If it had been then mathematically it would have reduced the suppression rate. However, the Tribunal found that the gross takings figures revealed by the till rolls did represent the takings of Mr Arif’s business over the period to which they related (para 150). Mr McDonnell’s point was not, so it appears, argued before the Tribunal and, in my judgment, is not open to him on an appeal on a point of law only.

In his skeleton argument Mr McDonnell raised other criticisms of the manner in which the hearing was conducted. However, these were not pressed in oral argument; and they did not appear to me to have been well-founded. So I say no more about them.

I find no error of law in the Tribunal’s decision. I will therefore dismiss the appeal.