19761
VALUE ADDED TAX — restaurant — observation of numbers of customers — whether observations accurate — yes, with some adjustment — calculated meal value — whether realistic — whether calculated takings capable of reconciliation with declarations — conclusion that there was suppression of takings in early period of observation diminishing in later months — assessment adjusted by elimination of two periods — parties left to agree figures for remaining periods — appeal allowed in part
MANCHESTER TRIBUNAL CENTRE
AKBARS BALTI RESTAURANT LIMITED Appellant
- and -
THE COMMISSIONERS FOR
HER MAJESTY'S REVENUE AND CUSTOMS Respondents
Tribunal: Colin Bishopp (Chairman)
Arthur Brown FCA
Sitting in public in Manchester on 15 to 18 May 2006
Richard Barlow, counsel, instructed by Independent VAT Consultants for the Appellant
Nigel Poole, counsel, instructed by the Acting Solicitor for HM Revenue and Customs for the Respondents
© CROWN COPYRIGHT 2006
DECISION
- As its name indicates, the Appellant trades as a restaurant. It has premises in Bradford. In this appeal it challenges an assessment issued on 27 February 2002 in the sum of £70,681. The assessment covers each of the Appellant's consecutive prescribed accounting periods from 04/99 to 10/00, and includes an amount of tax in respect of each of those periods. The Respondents' case, shortly stated, is that throughout the period assessed the Appellant understated its takings, and consequently accounted for too little output tax. The Appellant's case is that its declarations were correct and that there is no underdeclared output tax to assess.
- Before us, the Appellant was represented by Richard Barlow and the Respondents by Nigel Poole, both of counsel. We heard the oral evidence of Shabir Hussain, one of the directors of the Appellant and from several officers of HM Revenue & Customs: Sylvia Jones, the assessing officer, Ronald Kerr, Trevor Graham, Heather Matthews, Christine Marshall, Dean Foster, Mark Gibbons, Julie Hardwick, Martin Mummery, Anthony Johnston, Ruth Ross, Helen Stanley, Peter Lagowski, Janet Lagowski, Patricia Stebbings, Gordon Jenkins, Adam Jones, Jane Wood, Richard Mumford, Neil Preistman, Timothy Marshall, Angela McCalman, Thomas Simmons, Stuart Highy, Alan Keighley, John Kinghorn, Andrew Lumb and Mary Sharpe.
- The officers' evidence related almost entirely to observations which they had carried out at the Appellant's restaurant on three trading days, Wednesday 19 January, Saturday 26 February and Friday 21 July 2000 and to an unannounced visit which was made at the close of business on Friday 1 December 2000. Although we spent a great deal of time hearing that evidence, we do not propose to set it out since our doing so would add considerably to the length of this decision, while the detail of the evidence is of little value. The officers had only a limited live recollection of the observations—unsurprisingly since they were carried out some six years before they gave evidence—and for the most part they did little more than tell us what we could read for ourselves in their notebooks and in the observation records they compiled at the time. All of the observations were carried out internally as the officers ate in teams in the restaurant (or, in only one or two cases, made a take-away purchase). They counted the numbers of customers whom they observed; because of the layout of the restaurant it was necessary for there to be two teams in place at all times except very early in the evening. The officers asked to be seated in a pre-determined position, in order that the whole of the restaurant could be observed, but the seating request was not always met and there were consequently some limitations on their ability to see and count customers.
- Mr Barlow, very properly, tested the evidence thoroughly and he was able to demonstrate to our satisfaction that the officers had made some mistakes, of double-counting customers or in recording what they had seen. It would be a matter for some surprise if so many officers, carrying out observations in teams over three evenings, had not made some mistakes. As in most cases of this kind, the officers were attempting to count customers, and to segregate diners from purchasers of take-aways and from people who called at the restaurant but did not make any purchases, while avoiding drawing attention to themselves. We are satisfied that they carried out their task conscientiously but that they did indeed make some errors.
- Nevertheless, the scale of the errors Mr Barlow was able to demonstrate was not great. On 21 July 2000, for example, the officers' records show that there were 323 customers who ate within the restaurant and that seven take-away sales were made. It seems that the number of diners may be overstated, but by, at the most, 13 and probably fewer – that is, four per cent or less. The possible errors on the two other days of observation were of a similar magnitude. On the other hand, the total number of customers recorded on each of the days of observation excludes the officers themselves, yet the amount spent by the officers has not been eliminated from the subsequent calculations. Since we have also concluded that it is as likely that the officers omitted some customers (because they could not see them, or were not sure that they were indeed customers) as that some were double-counted, we intend to take the recorded numbers of customers on each of the three days of observations as a fair indication of the total number of diners, including the officers, who paid for meals in the restaurant on the days of observation.
- No observations were carried out on 1 December 2000. At five minutes before midnight four officers, accompanied (in case there should be a breach of the peace) by two police officers, entered the restaurant, identified themselves and took steps to ascertain the day's takings and to secure the meal bills which had been used. It is, we understand accepted that they were eventually successful in so doing, although they initially encountered some difficulties; Mr Poole did not suggest that the difficulties were deliberately created in order to obstruct the officers. Thereafter Mrs Jones, the assessing officer, made calculations of the average meal value (determined by dividing the takings for the day by the number of covers recorded on the meal bills) and of the takings which could be expected on the three days of observations (taking account of an intervening price rise). She then compared the results of those calculations with the amounts recorded for the same three days on the Appellant's daily gross takings listings, which were obtained from the Appellant's accountants, to whom they were provided in order that the Appellant's VAT returns could be prepared. That comparison led her to the conclusion that there had been under-declarations. She calculated the average rate of under-declaration over those three days and applied it, uniformly, to the declared sales in each of the seven accounting periods included within the assessment.
- The takings on 1 December 2000 amounted to £2,724.60, of which £2,504 could be identified to have been paid by diners and £220.60 by customers purchasing take-away meals. The meal bills showed that 239 customers had eaten in the restaurant; the average cost of a meal eaten in the restaurant was therefore £10.48 (though Mrs Jones arrived at £10.49). She calculated by analysis of the menus used in December and its predecessor, taken up on an earlier occasion, that the assumed average cost of a meal would have been about 43p less (that is, on her calculations, £10.06) in January and February, when the first observations were carried out. Using those figures and the observation records, she calculated the takings which could be expected on the days of observation and compared them to those actually declared. She then added the expected cost of take-away meals. On 1 December 2000 take-away sales represented 8.1 per cent of the total takings (or 8.8 per cent of the total cost of meals eaten in) and she assumed a similar ratio in respect of each of the days of observations. The result of her calculations is represented in the following table:
Date |
19.01.00 |
26.02.00 |
21.07.00 |
Diners (A) |
210 |
343 |
316 |
Meal cost (B) |
£10.06 |
£10.06 |
£10.49 |
Dining total (C): A x B |
£2,112.60 |
£3,450.58 |
£3,314.84 |
Take-aways (D): C x 8.8% |
£185.91 |
£303.65 |
£291.71 |
Expected takings (E): C + D |
£2,298.51 |
£3,754.23 |
£3,606.55 |
Declared takings (F) |
£910.00 |
£2,190.00 |
£3,168.00 |
Under-declaration (G): E - F |
£1,388.51 |
£1,635.23 |
£438.55 |
% under-declaration: G ÷ F |
152.58% |
77.17% |
13.84% |
By taking the aggregate expected takings over the three days, and comparing them with the declarations, Mrs Jones determined that the average rate of suppression was 55.87 per cent. The assessment was made by the simple expedient of adjusting the declared takings for each period by that percentage.
- We have a number of misgivings about Mrs Jones' approach. First, she has rounded her figures upwards rather than, as is usual in cases of this kind, downwards. The difference, seemingly trivial, in fact makes a noticeable difference when the result is extrapolated over a long period. Secondly, and more importantly, the number of take-away purchases identified on 1 December was 19 whereas on the days of observation, 2, 10 and 7 purchases respectively were observed. Since the value of the restaurant's take-away sales, as a proportion of the total, on those days cannot be ascertained it is impossible to determine whether the value of take-away sales on 1 December is representative, but a simple head count of customers suggests otherwise. We think it would be more fair to take the average value of the take-away sales identified on 1 December, that is £11.61, discount it modestly for price increases and then multiply it by the number of take-away customers observed in order to arrive at the expected volume of take-away sales.
- As the following table shows, these changes and the correction of Mrs Jones' minor arithmetical mistake lead to significant differences in the overall totals:
Date |
19.01.00 |
26.02.00 |
21.07.00 |
Diners (A) |
210 |
343 |
316 |
Meal cost (B) |
£10.05 |
£10.05 |
£10.48 |
Dining total (C): A x B |
£2,110.50 |
£3,447.15 |
£3,311.68 |
Take-away cost (D) |
£11.18 |
£11.18 |
£11.61 |
Take-aways seen (E) |
2 |
10 |
7 |
Take-aways (F): D x E |
£22.36 |
£111.80 |
£81.27 |
Expected takings (G): C + F |
£2,132.86 |
£3,558.95 |
£3,392.95 |
Declared takings (H) |
£910.00 |
£2,190.00 |
£3,168.00 |
Under-declaration (I): G - H |
£1,222.86 |
£1,368.95 |
£224.95 |
% under-declaration: I ÷ H |
134.38% |
62.50% |
7.10% |
The average under-declaration, taking those revised figures, becomes 44.94 per cent.
- Rather more significant still, however, is the striking feature of Mrs Jones' calculation, repeated in our own, that the rate of suppression (if her deductions are correct) diminished over time, and very rapidly. Indeed, if our own revised calculations are correct, there was substantial suppression in January, less although still significant suppression in February but very little in July, so little indeed that the discrepancy might be said to fall within an acceptable margin of error: certainly it is impossible to be certain that the apparent under-declaration is not explained by a modest difference in the average cost of a meal. Those experienced in cases of this kind know that traders suppressing their takings usually do so either at a constant or near-constant rate, or they begin to suppress their takings at a fairly low rate which increases over time. It is, of course, by no means impossible that Mr Hussain and his fellow directors (there was a change in the directors during the period with which we are concerned) suppressed the Appellant's takings, and at a high rate, for a period and then began to declare the correct amount, but Mr Barlow did not suggest that this had occurred. Instead, he urged upon us the conclusion that there had been no suppression at all, and that the explanation for the apparent under-declaration was that the average cost of a meal had increased substantially over the relevant period.
- There is some evidence which supports that proposition. On 23 September 2000 the Yorkshire Post published a review of the restaurant which was in glowing terms. Mr Hussain told us that there was a noticeable increase in the volume of customers following the publication of the review. Secondly, the restaurant had been damaged by fire in April 2000, which had led to its partial closure for five weeks. As the existing restaurant was refurbished, the Appellant took over the adjacent premises and expanded what had been a two-room restaurant into three rooms, adding about 50 per cent to its seating capacity. We are willing to accept that the number of customers did increase, for both reasons. Mr Hussain said that, before the publication of the review, the restaurant had a higher proportion of Asian customers, who ate less (and correspondingly spent less) than non-Asian customers. The article had the effect of increasing the proportion of non-Asian customers, with the result that the average cost of a meal also increased.
- An increase in the number of customers, however, seems to us to be of little relevance. The assessment is based not on assumed or calculated numbers of diners, but on observed customers. Any change in the number of diners is, therefore already accounted for in the calculations. Moreover, the number of diners on 1 December 2000, at 236, is considerable lower than the number on 26 February—before the fire—at 343. The contention that the favourable publicity of the newspaper review led to an increase in the amount spent per head has more merit. Nevertheless, we are not willing to accept that the increase was more than modest, even making an allowance for the changing ethnic origin of the customers. We have already observed that the difference between the declared and calculated takings for 21 July 2000 was very small; the difference might be accounted for by an increase in the average amount spent, but if so it is itself an indication that the increase was modest. We certainly cannot accept Mr Barlow's assertion that the publication of a review, however favourable, could lead to an increase in the average cost of a meal from £4.33 in January 2000 (which is the result of dividing the declared takings of £910 by the 210 observed diners on that day) to the (corrected) £10.48 calculated by Mrs Jones from the information obtained on 1 December. An increase of that magnitude (necessary if the difference between observations and declarations is to be explained by this factor alone) defies belief. Nor do we find it credible, even allowing for inflation, that the average cost per head in February 2000 was as little as £4.33. Accepting, as we do, that the numbers of diners observed are reasonably accurate, the only rational conclusion is that there was suppression in January and, although to a somewhat lesser extent, in February. On the other hand, there was either very modest suppression in July, or none at all.
- That, however, was not how Mrs Jones viewed the matter: as we have indicated, she merely took the average she had calculated for the three days and extrapolated that average over the entire period assessed. In fact, she first raised an assessment for a longer period but, for reasons not material now, that assessment was withdrawn and the assessment now under appeal was made to replace it. Mr Barlow did not argue that Mrs Jones' approach offended the requirements of section 73(1) of the Value Added Tax Act 1994, that an assessment of this kind must be made to the Commissioners' best judgment. There is no reason to think that Mrs Jones was doing anything other than her best; but our task is to determine, so far as it is possible to do so, what is the correct amount of tax payable and we have come to the conclusion that a different approach is to be preferred.
- We can only speculate about the reason, since we had neither evidence nor explanation, but we are satisfied that while there was substantial suppression in January and February there was none (or, at least, insufficient evidence of it to justify an assessment) in July. We have also considered whether there is any reason to suppose that the days of observation in January and February were atypical, or whether the accounting period in which they fell—04/00—was in any way out of the ordinary. There was no evidence before us to suggest that there was anything unusual about either 19 January or 26 February (or, for that matter, 1 December 2000—unusual factors then might have distorted the average cost of a meal). Nor can we detect from the observations or the calculations any reason to think that any of those days was affected by an unusual circumstance. The Appellant's declarations do not suggest that period 04/00 was out of the ordinary; the amount declared was less than that in the preceding period, but that preceding period included Christmas and the New Year, and a fall in the following period (the figures were £18,400 and £16,500 output tax respectively) is to be expected. Period 04/00 also included part of the time when the restaurant was partially closed. The declared output tax liability rose substantially in 07/00 to £22,600, a rise consistent with the enlargement of the restaurant and the declaration of correct figures.
- We are therefore satisfied that, whatever the level of suppression in the earlier periods, the assessment, so far as it relates to periods 07/00 and 10/00, must be discharged. The difficulty we face is that if we adopt in respect of the other periods included the average of the suppression detected in January and February 2000, even when it is adjusted in accordance with our own calculation (it is 83.60 per cent), the amount of tax assessed for those periods increases quite substantially (although, even so, there would be an overall reduction in the total assessment). We are reluctant to adopt the course of simply adjusting the assessment by assuming uniform suppression in the first five of the periods included within it without first giving the parties the opportunity of themselves reaching agreement on the correct amount of tax, in the light of our findings or, if agreement cannot be reached, of making further representations to us. Accordingly we go no further than the finding of fact we have made and give the parties permission to continue the hearing, before the presently-constituted tribunal, if they so wish. Any application for continuation of the hearing is to be made within three months of the release of this decision.
- We heard some submissions about costs but those submissions did not take into account the conclusion to which we have come. Accordingly we also give permission to the parties to make an application in respect of costs should they so wish.
COLIN BISHOPP
CHAIRMAN
Release Date: 5 September 2006
MAN/02/0179