First-tier Tribunal (Tax)

[2015] UKFTT 92 (TC)

[image removed]

TC04297

 

Appeal number: TC/2011/01237

 

Value Added Tax - Disputed turnover of fast food business - Discrepancy between VAT returns and till figures on Z rolls

 

FIRST-TIER TRIBUNAL

TAX CHAMBER

TENNESSEE FRIED CHICKEN (a partnership) Appellant

-and-

 THE COMMISSIONERS FOR HER MAJESTY'S

REVENUE & CUSTOMS Respondents

 

Tribunal:  JUDGE  HOWARD M NOWLAN

DEREK SPELLER F.C.A.

 

Sitting at Eastgate House, Cardiff on 5 & 6 February 2015

 

Martin Arthur on behalf of the Appellant

Luke Connell of HMRC on behalf of the Respondents

 

 

©CROWN COPYRIGHT 2015

 

 

 

DECISION

Introduction

1. This was an Appeal in which the Respondents contended that the Appellant, a partnership operating a Tennessee Fried Chicken franchised outlet, had fraudulently concealed turnover in the amount of £481,356 in the period from 26 May 2005 to 23 September 2008. As a result, additional VAT of £76,483 was charged. Ignoring the interest, there was then a penalty charge of £58,960.

2. The Respondents’ entire case was founded almost entirely on evidence obtained from their inspection of the Appellant’s cash register on the occasion of two visits made in May and September 2008. 

3. Our decision has been made somewhat difficult by the fact that HMRC had made no other enquiries in relation to information that might support (or possibly undermine) their case. We naturally also accept that the burden of proof falls on the Appellant to sustain its calculation of turnover, and to challenge HMRC’s adjusted figure. In this regard, the particular partner who appeared to be the more active of the two partners in the business. and in the few discussions with HMRC in relation to the dispute, hardly assisted the Appellant’s case by having been far from forthcoming in those discussions and also by refusing to attend the hearing before us. These factors have made it difficult for us to reach a conclusion in which we can have absolute confidence.

4. Our conclusion is, nevertheless, that the Respondents’ case is fundamentally made out. Since, however, we accept that figures of turnover obtained from cash registers will never be absolutely accurate, and that there is a real possibility in this case that some of the figures obtained from the cash register might have been swelled by genuine errors and staff training in the use of the cash register, rather than actual turnover, our decision is to reduce the figure of additional VAT in respect of undeclared turnover by 10% from the figure claimed to be appropriate by HMRC. That decision automatically reduces the penalty by the same 10%, but we otherwise confirm that we did not consider it appropriate to make any other adjustment to the amount of the penalty.

The facts

5. The Appellant was a partnership comprised of two individuals, Mr. Dan Pathirana, who lived in Gibraltar and seemed to play little active role in the business or the dispute with HMRC, and Mr Charitha Palitha Wijeratne (“Mr. Wijeratne”), who lived in Northwood, West London. While Mr. Wijeratne, and possibly both partners, also operated another Tennessee Fried Chicken outlet, the dispute in this Appeal related to the one that they operated from leased premises in Wembley.

6. While the trade had been conducted in earlier years, there was a period between 2002 and May 2005 when there was no trading because there had been a major fire at the premises. Following the receipt of insurance money the premises were repaired and re-equipped and trading re-commenced on 26 May 2005. There had been no visit from any VAT officers in the first periods of trading but, in May 2008, A VAT officer who gave evidence to us, Mr. Raymond Harris (“Mr. Harris”) visited the premises. The principal purpose of this visit was to review the repayment claims that had been made in relation to all the spending on replacement items, following the fire. No dispute resulted from these checks. Mr.  Harris obviously then discussed other trading matters with Mr. Wijeratne, and the following extract from his Witness Statement recorded some of these discussions, as follows:

“I asked him how profitable the business was and he told me that business was very bad due to competition. He said he was not a Tamil and he was losing custom to Tamil businesses. I asked him if he had any other sources of income and he said he had another Tennessee Chicken shop in Edgware, London. He also said his wife works and has rental properties. He said he has one till on the premises and that it was new when bought for the first day of trading on 26 May 2005 which is the date of VAT registration. The serial number is 0405660. He said all sales are rung into the till and that it is not used for staff training. He told me only he and his partner have keys to the till. He said he takes a Z reading once a week and X readings every day. He said that the Z reading usually agrees with the cash in the drawer.”

7. Mr.  Harris then examined the Z readings from the till for the three-month VAT period 10/07, all of which appeared to have been taken weekly, such that there were 13 readings for the relevant period. He did note that the Z readings did not contain the sequential numbers that the till could (and should) have been programmed to record, but he initially concluded that all was in order since the turnover shown on the 13 recordings corresponded to the turnover on the VAT returns for the relevant period. 

8. Mr. Harris had apparently known in advance that the till used was a particular type of Casio till, and he had accordingly taken with him the HMRC instructions as to how to obtain information from this till, regardless of whether the information was included on the Z readings or already printed in other ways. Before describing the information that he obtained, and the analysis of that information derived by another officer, Amanda Sleight (“Ms. Sleight”) who examined the data obtained by Mr. Harris, we need to explain the categories of information that were available on this till.

The various available readings from the Casio till

9. We were given the following information by another VAT officer who gave evidence before us, namely Christopher Nowak (“Mr. Nowak”). He explained the function of 7 different readings that this till would record.

10. The most significant reading was the Z reading, so-called because when a Z reading was taken, this “zeroed” the calculations on the till, such that the next Z reading taken would record the takings on the till, taken since the previous Z reading. The trader could take Z readings whenever he chose, and often traders would have the routine of taking Z readings weekly or daily.  In the Appellant’s case, it initially appeared that Z readings had been taken weekly because this is first what Mr. Wijeratne had said, and secondly Mr. Harris was shown 13 readings for the 13-week period ending 10.07. The Z readings would then be printed by the trader.

11. When, as here, the trader was producing his VAT returns on the basis of the “point of sale” system, the trader was in fact required to ensure that the till was programmed to record the sequential numbers of each Z reading on the print-out of each Z reading. Accordingly, had there been only 13 Z readings taken in the VAT period 10/07, and had the immediately preceding Z reading taken before the commencement of this period been the reading numbered 100, the readings for the 10/07 period would have been numbered 101 to 113, had the readings included the sequential number. As we mentioned in paragraph 7 above, somebody (quite possibly the supplier of the till, or possibly the supplier at the request of the partners) had in fact deleted this function so that there were no sequential numbers recorded on the printed Z readings. Accordingly by looking at Z readings, it was impossible to know whether in fact more than 13 Z readings had been taken in the period 10/07, with the extra Z readings being concealed, such that the turnover recorded on the missing readings would not be evident to anyone just looking at the un-numbered 13 readings.

12.  We were given three other material facts in relation to the Z readings. 

13. Firstly, the till did reveal to someone who knew how to access this information how many Z readings had in fact been taken since the till had been supplied, or at least since the till had last been re-set (probably by the manufacturer or a knowledgeable service agent). On the occasion of Mr. Harris’s May 2008 visit, 725 Z readings had been taken on this till. This illustrated that since there had only been approximately 156 weeks since trading re-commenced, and the till was thought to have been supplied as a new till, plainly many more Z readings had been taken than the one a week seemingly suggested by the claim recorded in paragraph 6 above.

14. The second material fact in relation to the Z readings was that the till could be programmed to ignore takings rung up on the till if a particular button was pressed, this enabling the trader to cause the till to ignore readings if they simply reflected occasions when staff were being shown how to operate the till, such that they were operating the till when there was in fact no real turnover. The third possible function of this till was that there was another button that enabled the operator to delate an entry that had initially been made on the Z readings. The purpose of this function was of course to enable the operator to delete mistaken till entries, and this function represented a second way in which training operations on the till, whilst initially recorded, could be deleted from the Z readings. We were told that the HMRC officers were told that the selected settings on this particular till, as at the date of the May 2008 visit, did not enable either of these functions to operate. In other words, if this information was true, then the Z readings on the till would count as perfectly ordinary turnover, both mistaken entries on the till and entries made when staff were being trained. We might mention (as we will explain more fully below) that another reading that Mr. Harris took illustrated that at the very least the second facility just mentioned (i.e. to reverse out an error) did appear to have been activated, and also to have been used.

15. The next reading that could be taken at any time on this till was an X reading, so-called because this reading enabled the operator to “examine” the current Z reading without zeroing the Z reading. Doubtless it was convenient for the trader to be able to access the running total of turnover without zeroing the Z reading, but the X readings were of little significance in this Appeal.

16. There were then three readings, referred to as GT1, GT2 and GT3, “GT” standing for “Grand Total”. The GT1 reading recorded the total takings on the till, since either first operation or since the manufacturer, or possibly a service agent, had re-set the till. Accordingly if the Z readings enabled training entries to be ignored or reversed out, and they enabled errors to be reversed, the GT1 reading would record the total figures in a similar manner, save for the fundamental feature that the GT1 figures would accumulate all the Z readings, and be quite unaffected by the fact that numerous Z readings might have been taken. We were told that the only way to zero the GT readings, other than by the manufacturer or service agent re-setting them, was for there to be both a mains and a back-up battery failure. Since the back-up battery could not be removed without cutting wires, and the battery was likely to operate for a considerable period, it was said to be inconceivable that the GT readings could have been re-set. In any event, as we will see, the revealing feature about the GT readings obtained by Mr. Harris on his May 2008 visit was that they illustrated vastly more recorded turnover on the till than had been returned by the trader in its VAT returns between May 2005 and May 2008. Accordingly we could obviously ignore any issue about the GT1 counter having been re-set since May 2005, because had that occurred, that would simply have rendered the disparity between the GT1 count obtained by Mr. Harris in May 2008 and the recorded turnover on the VAT returns for the period since May 2005 to be greater still. 

17. The GT2 and GT3 readings were both, like the GT1 reading, cumulative calculations of total turnover but each illustrating a variation. In the event that the GT1 readings had been reduced by reversing out errors and reversing out training entries, the GT2 figures treated  the reversed entries as further positive entries. Thus since the GT2 figures first recorded the wrongly inserted entry (say an error), and it then added to the total the till operation when the wrong entry was reversed, it followed that the GT2 total would exceed the GT1 total by twice the amount of the “wrong or training entries” than had been reversed out on the GT1 calculations.

18. The GT3 total was irrelevant in this Appeal. It apparently disclosed some information in relation to discounts, presumably showing the undiscounted turnover whilst the GT1 calculations would have illustrated the total actual, i.e. discounted, turnover when discounts had been given.

19. While we can ignore the GT3 readings, there were two other available read-outs on the Casio till which were relevant in this Appeal, one being particularly relevant. The first of these produced Z readings for every hour of the 24-hour cycle of hours in a day, and the other produced Z readings for every day of the month, in other words 31 as opposed to 24 separate readings. These two readings were zeroed whenever one of the relevant Z readings was taken but since nobody had ever  sought either of these readings, they both revealed  the total turnover since the till had first been operated. It accordingly happened that when Mr. Harris accessed the readings on the GT1 memory and the Z readings on the “hourly” and “each day of the month” basis, the totals of all these readings were exactly the same.

20. Before turning to the actual figures that were derived from Mr. Harris’ operation of the till in May 2008, we should indicate the significance of the information revealed by the hourly Z count that we have just mentioned. Since this count had never been re-set, and so had presumably been running from May 2005 to May 2008, when zeroed by Mr. Harris for the first time, it follows that the count for till entries made between midnight, i.e. 00.00 and 01.00, showed us the total number of till activations made in that time slot on all the days in the three-year period (say approximately 1095 days), and it also showed us the cumulative turnover recorded throughout the three-year period in that one hour slot. Thus to give three examples, we learnt that in all the one hour periods between 00.00 and 01.01, the till had been operated on 2,529 occasions and the cumulative turnover recorded in all those hourly periods had been  £4,952.52. The second example illustrates that whilst it seems odd that there will have been any till activations between the hours of 4 a.m. and 5 a.m., we learnt that during this hourly slot the till had been activated in fact on 19 occasions, and that the aggregate turnover recorded in that hourly slot had been merely £50.22. Turning finally to the third example and taking an hourly period when we know that the business would in fact have been open to customers, we ascertain that between the hours of 6.00 p.m. and 7.00 p.m. the till had been operated on 8,951 occasions and that the aggregate turnover in the three-year period during that hourly slot had been £20,863.25. Taking those two figures, it can immediately be calculated that the average takings on each of the 8,951 activations of the till in that hourly time slot will have been £2.33. 

21. We will ignore the equivalent figures obtained from the Z counter that recorded takings on each day of the month. Naturally the first day of each month could fall on any day of the week when looking at a long period of three years, so that this particular counter would not have revealed relative levels of turnover on weekdays, Saturdays and Sundays. The only relevant fact to record is that the total turnover on all days of every month in the period did precisely match the aggregate entries on the GT1 calculation. So too, of course, in the hourly Z count, the aggregate totals for all 24 hours of the 1095 days also equalled the same figure.

22. One other point that we should record is that while it is always possible that the till’s time clock may not have been correctly set, Mr. Harris did ascertain that it was approximately 1 hour wrong during his May 2008 visit. This may have been occasioned by the one hour difference between summer time and standard time. The reasonable inference that we drew is that the time error on the till was unlikely to have been fluctuating wildly over the three-year period. In other words, it was credible that it had always been 1 hour wrong, or that at some times of year it might have been accurate and 1 hour wrong at other times.

The information obtained on the May 2008 visit

23. The significant information obtained by Mr. Harris on his May 2008 visit was that:

·         according to the GT1 figure, and the Z readings for “days during the month” and “each hour of the days”, the cumulative figures of turnover recorded  in all three memories since the till had been set were all shown as £698,503.20, contrasted with the figure returned on the various VAT returns of £308,001;

·         725  Z readings had in fact been taken since the till had initially been set, whereas  approximately 156 Z readings would have been taken if, as suggested by Mr. Wijeratne, one Z reading had been taken each week; and

·         as Ms. Sleight’s analysis of the hourly Z reading figures revealed, it appeared rather oddly that considerable turnover had been recorded in the very early hours of the morning, well before any normal staff would have been on the premises. We were told that the business hours normally commenced at 11.30 a.m. but the Z readings record that some hourly Z readings had in fact been taken at some time during every single hour slot in the 24-hour periods. Since this reading always gave us the number of readings taken in each hourly slot, and the aggregate turnover in the same slot, Ms. Sleight calculated that the average takings recorded onto the till during the following slots, on each activation of the till, had been the following figures:

 

o   06.00 to 07.00 £239.29 in each of 15 activations

o   07.00 to 08.00 £230.71 in each of 38 activations

o   08.00 to 09.00 £254.29 in each of 200 activations

o   09.00 to 10.00 £128.75 in each of 667 activations and

o   10.00 to 11.00 £89.58 in each of 1751 activations

 

·         Ms. Sleight’s analysis also illustrated that during normal business hours the till had been activated on roughly 10,000 occasions during each hourly slot, there of course being roughly 1095 of each of such hourly slots during the three-year period, and that the average amount rung onto the till during each of the activations in the hours when the premises were obviously open was roughly £2.30. That figure of £2.30, whilst perhaps being surprisingly low for the average amount rung onto the till on each activation, was also extremely stable. During the entire period of 11 hours between 13.00 to 00.00 (i.e. midnight) this figure was never less than £2.21 and never more (save on one occasion) than £2.42.

·         Mr. Harris’s reading of the GT2 figures also showed that the turnover on the GT2 memory was £124,000 higher than that on the GT1 memory, illustrating that some

 

 

·         entries on the GT1 memory must have been reversed at some point. Since the GT2 memory recorded both the wrong entries (either reflecting mistakes or “training” entries), and it reflected as a positive item the reversal of those wrong entries, it seems that £62,000 of turnover, once recorded on the GT1 memory, had at some point been deleted.

24. When Mr. Harris left the premises on the occasion of his May visit, he re-set some of the settings on the till such that they would be “HMRC-compliant”, and he asked Mr. Wijeratne to double check that the till was operating properly. Apparently it was.

25. The three critical conclusions drawn from the above information was firstly that the 725 Z readings that Mr. Harris’s examination of the till revealed suggested that Z readings had been taken much more frequently than once a week; secondly that the cumulative turnover recorded onto the GT1 calculations and the “hours during the day” and “days during the month” memories were all at least twice the figure returned on the VAT returns, and thirdly that a considerable number of error entries or training entries had in fact been deleted from the Z readings and the GT1 figure that reproduced all the Z readings on the cumulative basis.

The second and third visits

26. Mr. Harris made two further visits to the premises, one in August and one on 23 September 2008. Neither partner was on the premises on the occasion of the August visit and since the keys enabling the till to be operated were not on the premises, nothing was achieved. On the 23 September visit, Mr. Harris took further readings. He was told on this occasion that since the first visit, Z readings had been take daily. The information that he obtained on this third visit from his examination of the information on the various memories on the till was that £135,643 of turnover was shown on the GT1 memory, the “hourly” and the “day of the month” calculations, compared with additional turnover shown on the VAT returns (duly adjusted for parts of months and VAT periods) of £44,789, and that 924 Z readings had now been taken. Ignoring the one Z reading taken by Mr. Harris in May, this meant that in the period of approximately 131 days between the first and the third visits, an additional 198 Z readings had in fact been taken. There was also the same feature of considerable turnover being recorded, in large blocks of turnover on each activation of the till, generally in the period from 7 a.m. until 11 a.m.

27. We should mention that Ms. Sleight had asked Mr. Harris to seek an explanation for the feature that some turnover had been booked in the very early hours of the morning, and that considerable amounts of seemingly “aggregated turnover” had been recorded onto the till in the hours 7 a.m. to 11 a.m. Mr. Harris failed to ask this question and therefore the Appellant was not asked to explain how the turnover had been generated in such an odd fashion when the premises were not even open.

Mr. Patel’s evidence

28. Since Mr. Harris had been transferred to another responsibility within HM Customs & Excise, Mr. Umesh Patel (“Mr. Patel”) took over the investigation in late 2008 or early 2009. Mr. Patel gave evidence to us, and we concluded that he was a very impressive officer.

29. Mr. Patel’s initial enquiries were based on the fact that it appeared at least possible that the Appellant had been fraudulently concealing turnover. Mr. Patel therefore proceeded with his enquiry under what is referred to as the PN 160 process, in which he explained that he was required not to feed possible suggestions to the Appellant that might explain the disparities in turnover. This was because the level of penalties would be reduced were the Appellant to make disclosures of wrong-doing and therefore the Appellant had to be given every opportunity to make unprompted disclosures, and of course to give genuine and credible explanations for the relevant disparities.

30. The investigation proceeded slowly in part because Mr. Patel found it difficult to arrange meetings with the Appellant. Several arranged meetings proved fruitless, either because the Appellant’s then representative was not present, or because the Appellant said that the meetings had been arranged at a time that proved unsuitable, or because he had not read the information in relation to the PN160 process that had been given to him. During these meetings, various possible explanations were offered to explain why the recorded turnover on the various memories on the tills exceeded the declared turnover on the VAT returns, though some were more in the nature of speculative suggestions, being unconfirmed by any evidence. We should also add that during the enquiries, the Appellant was represented by various different advisers. At the hearing, none of these advisers represented the Appellant, and the Appellant was represented by Mr. Martin Arthur, who appeared not to have been involved in any of the actual meetings between the Appellant and Mr. Patel.

The factual contentions

31. The various explanations offered for the disparities between the GT1 and other memory figures and the turnover recorded in the Appellant’s VAT returns between May 2005 and May and September 2008 were as follows.  In addition to summarising the suggestions, we will generally add our own comments in relation to each.

Whether the till retained entries and turnover referable to a previous owner

32. While Mr. Wijeratne had initially told Mr. Harris that the till had been new, when acquired in May 2005, it was subsequently suggested that since the till had not been in a box when it was delivered, it was possible (notwithstanding that it appeared to be “as new”) that it may have been used prior to being used by the Appellant. This might have explained why the turnover on the GT1 memory in May 2005 had been greatly higher than the figures returned for VAT purposes.

33. We found this claim fairly unconvincing for several reasons. Firstly the claim that the till had not been new, when supplied, only emerged when the disparities in turnover had been indicated to the Appellant. It also seemed improbable that even if the till had been second-hand, the supplier would not have re-set the counters on the various GT memories to zero (as the competent supplier would apparently have been able to do), and by taking Z readings on the “hourly” and “days in the month” counters, those memories would automatically have been set to zero in May 2005. We were also told by the HMRC officers who gave evidence that it would be very unusual for a supplier, installing and programming a second-hand till for a new user, not to re-set all the counters, including the GT counters to zero.

34 It is also significant that when the various readings were taken in September 2008, there again emerged to be, if anything, an even greater percentage disparity between the turnover recorded on the till between May and September 2008 and that returned for VAT purpose than there had been between May 2005 and May 2008. The Appellant had not appreciated at the end of the May meeting that Mr. Harris had accessed, and taken away with him, various figures from the till that were going to reveal the problematic disparities, and when it appears that the September 2008 figures revealed similar, if not greater, disparities, and also that more Z readings had been taken than (as claimed) one on each day, it was clear that these May to September disparities could have had nothing whatever to do with any prior operation of the till by some original owner before May 2005.  This of course also cast doubt on the suggestion that similar disparities in figures between May 2005 and May 2008 had been accounted for by any prior use of the till.

Whether the business was trading at a loss, such that the figures claimed by HMRC were improbable

35. It was next suggested that HMRC’s figures of turnover, based on the till readings, could not have been correct because the business was in fact trading at a loss, and the losses had had to be financed by the provision of various loans by the partners. We were, however, never shown any accounts, or income tax returns, and no evidence was supplied to confirm the claim in relation to loans. Furthermore HMRC pointed out that when the partners sold the particular business at the Wembley premises in 2009 or 2010, the partners were said to have received £87,500 for the goodwill of the business. This seemed to undermine the claim that the business had been loss-making, though no further evidence or information, or indeed denials, were provided in relation to this information that HMRC appeared to have obtained in some way.

Whether staff had been selling their own chicken, and ringing the sales through the till

36. The next suggestion, ventured to explain the relevant disparities, was that prior to the fire which had resulted in the business being closed in 2002, the Appellant had found that staff were bringing in their own chicken, cooking it on the Appellant’s equipment and selling it and of course keeping the proceeds. It was suggested that possibly the same practice had been undertaken by the staff since 2005, Mr. Wijeratne suggesting that neither he nor his partner were generally on the premises during the bulk of the day, so that he would not have known whether this practice had re-commenced. We found this suggestion particularly improbable because it would have seemed odd for fraudulent staff to have recorded their own sales of their chicken on the Appellant’s till, when pocketing the proceeds. Furthermore Mr. Wijeratne had said at the initial meeting that the Z readings had generally corresponded with the cash in the till, and obviously that would not have been the case had the staff pocketed the cash and then, rather oddly, rung the fraudulent sales up on the till. We reject this possible explanation as being wholly unconvincing. 

Whether some repair or maintenance to the till could have occasioned the collection of wrong entries

37. This suggestion was made on behalf of the Appellant, but never amplified. When the readings taken by Mr. Harris in both May and September 2008 both revealed the same substantial disparity between the figures recorded on the GT1 calculation and the figures returned for VAT purposes, it seemed improbable that some adjustment to the till during some unspecified act of maintenance would have had the same effect in relation to both periods. Moreover, the more likely consequence of some act in relation to maintenance would seem to have been that counters would have been re-set, and when the actual count on the GT1 and the two Z calculations for “hourly” and “days in the month” calculations all consistently recorded high, rather than low (and possibly some lost) figures, it seemed improbable that any repair or maintenance activity could account for the readings on the till.

Battery failure

38. It was suggested that there might have been a battery failure, but when we were told that this was unlikely, and that if it was likely to have had any effect on the memory functions, it would more obviously have re-set the memories at zero, rather than consistently reported double turnover in contrast to that reported on the VAT returns, this suggestion seemed to be groundless.

Errors and staff training

39. We accept that it was possible that errors had been made in operating the till, and it is also possible that when staff were being trained, training entries that would not have reflected actual sales may have been made on the till. Two factors that slightly undermine these two possibilities are that the GT2 count did establish (contrary to the suggestion made to HMRC that the till had been programmed in such a way that staff training could not be disregarded by the GT1 counter, and for the purposes of all the Z reports, and furthermore that error entries could not be reversed out) because the excess of the GT2 count over the GT1 count clearly demonstrated that negative entries had in fact been made on the till. We are inclined to accept that some staff training would have been undertaken, though we should record that it was periodically said that the till was only operated by the partners, and the odd feature of blocks of turnover being entered between 7.00 a.m. and 11.00 a.m. might indeed be explained by staff not having had the key to operate the till when the business was busy in the evenings, such that they recorded the takings in a notebook and possibly Mr. Wijeratne may have entered the aggregate amounts on the till on the following morning. Whatever the position in relation to this, we still accept that staff training may have been undertaken, and the possibility cannot be ruled out that even if some such entries had obviously been reversed out on the GT1 counter, others may not have been.

40. Four statements were produced by the Appellant, signed by each of four employees (albeit not people reported to HMRC as being full- or part-time employees) and these statements recorded that the relevant “employees” had made various errors when operating the till, and when being trained. None of these employees had been available to be questioned by HMRC or by us during the hearing, and although the Appellant’s representative criticised HMRC for not having sought out the relevant employees, we again note that the burden of proof in relation to the basic calculation of turnover did fall on the Appellant. Furthermore all four statements were in identical terms, obviously prepared by the Appellant rather than by any of the so-called employees, and they all referred to a period of time at least two years after the period currently in dispute.

41. We do accept that errors in operating the till and staff training that had not been reversed out on the till could have resulted in some over-statement of turnover on the GT1 calculation. It is for this reason that we eventually conclude that we should reduce the level of undisclosed turnover claimed by HMRC by 10%. When, however, there is no firm indication or supporting evidence that these factors did account for the vast discrepancy in the figures, and when the GT2 counter undermined the claim that the selected settings on the till precluded error entries from being reversed out, we felt unable to attribute any greater adjustment to this factor. It is worth noting that if staff training had at times been conducted outside normal opening hours, and there had perhaps been substantial staff training at certain times, it might have been highly prudent for the Appellant to have made Z reports covering the particular training, and the date and time at which it was undertaken, and it would then have been particularly appropriate to retain those Z reports, and mark them to refer to staff training. It is also pertinent to note that numerous Z readings had been made in excess of the originally claimed weekly actioned Z readings and in excess of the claimed daily readings supposedly taken in the period from May to September 2008. There seemed, therefore, to be a considerable likelihood that some Z readings had been taken and destroyed, rather than that the relevant disparity resulted to any substantial degree from having had to make adjustments, in an entirely unrecorded manner, for staff training. The final obvious point is that the very substantial disparity must suggest that there had been an immense amount of staff training.

42. While we have decided to reduce the claimed figure of hidden turnover by 10%, and we account for this reduction by referring to the way in which a till can produce excess readings when mistakes are not corrected or training operations on the till are not reversed out, we are naturally not oblivious to the possibility in the other direction that if till readings greatly exceed the turnover returned for VAT purposes, such that a fraud seems highly likely, it is always possible that some turnover might never be recorded on the till at all, such that the till readings might in fact understate the level of the fraud, rather than exaggerate it. We are ignoring this feature, since there was no evidence or claim to support this possibility, but we mention it simply because it is always a possibility.

The contentions made during the hearing by the Appellant’s representative

43. The Appellant’s representative at the hearing appeared not to have been involved in earlier discussions with HMRC in relation to the dispute, and his contentions consisted principally in a general attack on HMRC, a fair statement that HMRC had relied in its case solely on the till information and they had not sought to verify their conclusions from the till evidence in other ways, and finally there was the claim posed as a question as to “How could the Appellant possibly disprove HMRC’s case?”

44. In relation to other possible checks that HMRC might have undertaken in order to support their primary case, the Appellant’s representative referred to:

·         undercover visits to assess likely turnover;

·         likely profit margins in the business;

·         analysis of the means of the Appellant;

·         evidence derived from bank statements;

·         evidence derived from purchase invoices, and

·         contacts with the Appellant’s suppliers.

 

45. We conclude that in this case, some of those tests might have sensibly been undertaken, while others might have been fairly pointless. We also repeat that the burden of proof was indeed on the Appellant, and had the Appellant provided evidence in relation to losses, newly injected loans, levels of purchases, this might have greatly assisted the Appellant’s case. If the Appellant had wished to dispute the claim that £87,500 had been received for goodwill on the sale of the business, this might have been highly relevant. In relation to the Appellant’s representative’s question of “how could the Appellant disprove HMRC’s case?”, we volunteered the response that had Mr. Wijeratne been more co-operative in early discussions with HMRC and, more significantly, had he actually chosen to attend the hearing, he would at least have had an opportunity to convince us of some of the suggestions advanced on his behalf. Furthermore, it is for the Appellant to keep records of daily turnover, and in its own interests to keep sufficient information and documentation to sustain its VAT returns, and there has been little actual evidence, as distinct from speculative suggestions, that might support the Appellant’s case, and the original figures produced in the VAT returns.

46. There are, however, two specific respects in which we criticise HMRC’s approach to the presentation of the case.

47. While most of the evidence suggested that the till had probably been a brand new till in May 2005, or that at least even if it was second-hand the likelihood is that the GT counters would have been re-set to zero before the till was supplied to the Appellant, a late claim by a till expert on behalf of the Appellant suggested that in May 2005 this particular till had been obsolete for some years. Nobody had bothered to ascertain from Casio whether this was the case or not. It would have been helpful, had HMRC wished to dispute this claim, had they simply asked Casio whether or not the till was still a current model in May 2005. Surely this would have been simple.

48. The only “sanity check” that HMRC claimed or admitted that they had run to verify their claims based on the till evidence was a statement by Mr. Nowak that he had made a separate estimate of turnover, which he suggested confirmed that HMRC’s figures were reasonably realistic. Since he had based his estimate on a supposed average price for every order of approximately three times the price for each order that we derived from the figures recorded in the fourth bullet point in paragraph 23 above, we found this suggested sanity check to be somewhat useless.

The blocks of turnover entered onto the till in the period between 7.00 and 11.00

49. We accept that we were troubled by the feature that, in a rather surprising manner, considerable turnover was entered on the till in the early morning period before the business was open, and that the figures suggested that turnover was then entered in “aggregated amounts”. Mr. Patel suggested that it was quite common for turnover not to be recorded on tills, as it ought to have been recorded, as sales were made, but instead to be recorded in a notebook, and then credited to the till on the following morning. This might be particularly common when the business might have been very busy, or when inexperienced staff were alone on duty, or when neither of the partners had given any staff member one of the two available keys to operate the till functions. We are inclined to accept that Mr. Patel’s explanation will have been the correct one, though it is certainly unfortunate that when Ms. Sleight asked Mr. Harris to ascertain why turnover had been credited to the till at these times and in these amounts, neither Mr. Harris, nor any other officer, had actually put this question to the Appellant. We accept that so far as Mr. Patel was concerned, he considered himself to be precluded from posing such questions by the HMRC guidelines as to how to conduct a PN 160 investigation.

Our conclusions and decision

50. While it was unfortunate in this case firstly that the Appellant declined to appear in person, secondly that little reliable evidence had been advanced at any time on behalf of the Appellant and thirdly that HMRC had not undertaken any other enquiries to back up their claims based on the till evidence, we do fundamentally accept HMRC’s basic claim. The evidence from the till entries, particularly the way that similar evidence emerged to that obtained in May 2008 when the third visit was made in September 2008, does strongly suggest that the Appellant fraudulently concealed turnover.  When it is for the Appellant to prove its case to the standard of the balance of probability, we have to conclude that the Appellant has fallen well short of achieving this. In an arbitrary fashion, however, and in order to give recognition to the feature that tills will often produce slightly inaccurate figures of turnover, and that it is possible that errors and staff training might have recorded entries on the till in excess of actual turnover, we reduce the claimed excess turnover by 10%.

The penalty

51. As we have indicated, HMRC claimed an 80% penalty on the basis that VAT of £76,483 had been fraudulently avoided.  By reducing the amount of the concealed turnover by 10% and therefore the VAT avoided by 10%, we automatically reduce the amount of penalty by the same 10%. Since however the Appellant had never disclosed what we must conclude was a fraudulent concealment of VAT and the Appellant had shown little co-operation and the avoidance was serious, in terms of both amount and dishonesty, we otherwise confirm the penalty calculation.

 Right of Appeal

52. This document contains full findings of fact and reasons for the decision. Any party dissatisfied with this decision has a right to apply for permission to appeal against it pursuant to Rule 39 of the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009. The application must be received by this Tribunal not later than 56 days after this decision is sent to that party. The parties are referred to “Guidance to accompany a Decision from the First-tier Tribunal (Tax Chamber)” which accompanies and forms part of this decision notice.

 

 

 

HOWARD M NOWLAN

TRIBUNAL JUDGE

RELEASED: 26 February 2015