First-tier Tribunal (Tax)

 

TC06940

 

Appeal number:  TC/2016/00920          

 

CORPORATION TAX - Appellant's large-scale cash purchases of scrap gold and bullion in Birmingham's 'jewellery quarter'

Were there understated profits chargeable to corporation tax in accounting period ending ('APE') 31.1.11? - Yes - What were they? - Tribunal has made findings which will permit re-calculation

Having so found, can 'the presumption of continuity' be applied to earlier and/or later years so as to support assessments for those earlier and later years? - Yes, but, on the facts, only to APE 10 - Presumption of continuity displaced for years other than APE 10 and APE 11

Assessment for APE 10 was a discovery assessment - Were the conditions in FA 1998 Schedule 18 satisfied in relation to APE 10? - Yes - Was the discovery stale by the time of the assessment? - On the facts, No

Can the understated profits for APE 10 or APE 11 be treated as a loan or advance by the Appellant - No

ALTERNATIVE DISPUTE RESOLUTION - remarks on its potential utility in this dispute

 

FIRST-TIER TRIBUNAL

TAX CHAMBER

 

	STIRLING JEWELLERS (DUDLEY) LTD		Appellant
	- and -
	THE COMMISSIONERS FOR HER MAJESTY'S REVENUE AND CUSTOMS		Respondents
TRIBUNAL:		JUDGE CHRISTOPHER MCNALL
		MR NICHOLAS DEE

 

Sitting in public at City Centre Tower, 5-7 Hill Street, Birmingham B5 4UU on 15-23 January 2018, and on 2 February 2018 at Taylor House, Rosebery Avenue, London EC1R 4QU, with further written submissions (from the Appellant) on 9 February 2018

 

Thomas Chacko, Counsel, instructed by Avery West Ltd, Chartered Accountants, for the Appellant

 

Sadiya Choudhury, Counsel, instructed by the General Counsel and Solicitor to HM Revenue and Customs, for the Respondents

DECISION

 

1.             This is our decision in relation to an appeal (made by way of a Notice of Appeal dated 3 February 2016) against a series of decisions (made on various dates, as set out below), which were upheld at departmental review on 20 January 2016, in respect of:

(1)          The Appellant limited company's liability (the assessed amounts amounting to about £10.986m) to corporation tax on profits; and

(2)          The Appellant's liability for corporation tax under section 419 Income and Corporation Taxes Act 1988 ('ICTA 1988') (for accounting periods ending before 1 April 2010);

(3)           The Appellant's liability for corporation tax under section 455 of the Corporation Tax Act 2010 ('CTA 2010') (for accounting periods ending after 1 April 2010) (the assessed amounts under s 419 and s455 together amounting to about £8.26m).

2.             HMRC's decisions relate to eight successive years, from the APE ('Accounting Period Ending') 31.1.07 up to and including APE 31.1.14.

3.             HMRC seeks to maintain all the amendments and assessments, but in revised amounts.

4.             The figures below are HMRC's (re)-revised computations, presented on the last day of the hearing, appearing at page 52 of the attachments to HMRC's closing written submissions, and headed 'Revised tax computations including the additional purchase invoices provided as exhibits BM1 and BM4 to Ms Barbara Mullinder's witness statement dated 27 February 2017'. Hence, these figures revise both the figures in HMRC's Statement of Case (19 August 2016) and HMRC's letter of 5 October 2017.

5.             At the outset, we observe that the very fact that both parties found it necessary to repeatedly revise their competing figures is a clear indicator of the situation (which could fairly be described, even putting it most favourably to the Appellant, as close to chaotic) which, for a period, pertained in this taxpayer's affairs, and which, in consequence, continued to pertain between the taxpayer and the HMRC, both before and during this appeal.

6.             Since this is a lengthy decision, we consider it appropriate to set out, at the very outset, the outcome of the Appeal.

7.             We have allowed this appeal against all the assessments and amendments except for the corporation tax assessments relating to APE 11 and APE 10.

8.             The corporation tax assessments for those two years still stand, but in amounts which must be re-calculated and revised on the footing of our findings of fact below.

HMRC's Decisions

 

9.             The following outlines HMRC's decisions.

APE 31.1.07 ('APE 07')

 

10.         Inquiry opened 30 December 2008.

11.         Closure notice with revenue amendment: issued on 22.7.15, amendment varied on 21.1.16, further varied on 5.10.17.

12.         Additional corporation tax due on profit = £212,135.34

13.         Corporation tax due under s 419 ICTA 1988 = £30,601.75 

APE 31.1.08 ('APE 08')

 

14.         A protective inquiry opened on 25.1.10.

15.         Closure notice with revenue amendment: issued on 22.7.15

16.         Additional corporation tax due on profit (as amended) = £93,302.79

17.         Corporation tax due under s 419 ICTA 1988 = £14,891.00

APE 31.1.09 ('APE 09')

 

18.         'Protective' discovery assessment issued on 20.1.15 (just within the extended 6 year time limit). That was done on the basis that the Appellant's lax approach to record-keeping made it impossible to argue that no additional tax was due.

19.         Additional corporation tax due on profit = £88,264.16

20.         Corporation tax due under s 419 ICTA 1988 = £4,537.00

APE 31.1.10 ('APE 10')

 

21.         Revenue amendment issued on 22.7.15, but cancelled on 20.1.16 because no enquiry had been opened.

22.         Discovery assessment issued on 22.1.16

23.         Additional corporation tax due on profit = £817,299.79

24.         Corporation tax due under s 419 ICTA 1988  = £714,520.00

APE 31.1.11 ('APE 11')

 

25.         Full cross-tax inquiry opened on 13.6.12.

26.         Revenue amendment issued on 22.7.15, but cancelled on 20.1.16 because no closure notice had been given.

27.         Closure notice issued on 29.1.16 and revenue amendment pursuant to that closure notice issued on 4.2.16.

28.         Additional corporation tax due on profit = £2,370,021.92

29.         Corporation tax due under s455 CTA 2010 = £2,151,189.50

30.         APE 31.1.11 has particular significance for the purposes of this appeal because (i) it was subject to a full cross-tax enquiry; (ii) it was in any event subject to particular scrutiny by HMRC and the Appellant's advisers for a long period of time; (iii) it has been used by HMRC as an index year; (iv) the reasoning and outcome of HMRC's analysis for that year has been carried across (both backwards and forwards) to other years which are in dispute, making use of the presumption of continuity.  

APR 31.1.12 ('APE 12')

 

31.         Revenue amendment issued on 22.7.15, but cancelled on 20.1.16 because (on HMRC's case) no enquiry had been opened.

32.         Discovery assessment made on 22.1.16.

33.         Additional corporation tax due on profit = £3,164,659.98

34.         Corporation tax due under s455 CTA 2010 = £2,983,421.20

APE 31.1.13 ('APE 13')

 

35.         Revenue amendment issued on 22.7.15, but cancelled on 20.1.16 because no enquiry had been opened.

36.         Discovery assessment made on 22.1.16

37.         Additional corporation tax due on profit = £2,435,449.80

38.         Corporation tax due under s455 CTA 2010 = £2,575,948.50

APE 31.1.14 ('APE 14')

 

39.         Closure notice and amendment made on 4.8.15 and issued on 5.8.15

40.         Additional corporation tax due on profit = £913,491.30

41.         Corporation tax due under s455 CTA 2010 = £1,019,098.50

Burden and Standard

42.         The burden is on the Appellant to show that it has been overcharged by the assessments and amendments.

43.         However, in relation to the discovery assessments (APEs 09, 10, 12 and 13) the burden is on HMRC to show that the statutory conditions for the making of such an assessment have been satisfied.

44.         The applicable standard of proof in any event is the balance of probabilities.

The evidence

 

45.         The hearing bundle originally prepared by the Appellant comprised 29 files, which included 19 files of the Appellant's existing purchase invoices for APE 11. As we understand it (from Mr Bosley, who had looked at all the available invoices for that year) there were 15,000 - 20,000 invoices. However, on 3 January 2018 (that is to say, shortly before the hearing) the Appellant informed HMRC that it no longer intended to rely on those files and neither party produced them, or relied on them, at the hearing.

46.         For the Appellant, we read witness statements and heard oral evidence from:

(1)          Mr Alexander (also known as 'Ali') Ragonesi, a 50% shareholder and director of the Appellant;

(2)          Mr Peter Bosley, a chartered accountant;

(3)          Ms Barbara Mullinder, the Appellant's book-keeper;

(4)          Ms Christine Noble, a current employee of the Appellant;

(5)          Mr Trevor Bennet, a former employee of the Appellant

(6)          Ms Valerie Heighway, a former employee of the Appellant, and Mr Ragonesi's sister-in-law;

(7)          Mr Andrew Scadeng, a director of Samson Gold Ltd, who sold scrap gold to the Appellant;

(8)          Mr Jonathan Haag, the owner of E Haag Jewellers, likewise;

(9)          Mr Nigel Blackburn, a director of Lois Jewellery Ltd, one of the Appellant's competitors;

(10)      Mr Scott Moore, a director of Mercia Jewellery Ltd;

(11)      Mr Tom Reeder, the owner of Style Line MFG jewellery manufacturers.

47.         We also read witness statements, but did not hear oral evidence, from

(1)          Mrs Carole Ragonesi, Mr Ragonesi's wife, a 50% shareholder and the Appellant's Company Secretary;

(2)          Ms Sharna Ragonesi-Browne, a present director of the Appellant, and Mr Ragonesi's daughter.

48.         For the Respondent, we read a witness statement, and heard oral evidence from Mrs Caroline Spaughton, a caseworker in HMRC's Wealthy and Mid-Sized Business Compliance Team. She had taken over the case in March 2015 from Mr Martyn Cope.

HMRC's case

 

49.         On 30 December 2008, HMRC opened an inquiry into the Appellant's APE 07.

50.         Its review of the Appellant's underlying business records led it to express a number of concerns in relation to the Appellant's record keeping systems, including a lack of records of cash transactions, missing purchase invoices, and a 'negative cash' position based on the Appellant's SAGE records.

51.         As to the latter: the 'negative cash' position, put simply, is that the Appellant, in HMRC's view, did not seem to have had sufficient recorded cash-at-hand to make the purchases which it did. HMRC took the view that, therefore, there must have been undeclared cash in the business.

52.         On 13 June 2012 HMRC opened a full cross-tax inquiry for APE 11, but it did not open any enquiry for the intervening three years (i.e., APEs 08, 09, 10).

53.         At a meeting on 12 February 2013, the Appellant's accountants at the time, Grant Thornton, acknowledged that there were failings in the Appellant's record keeping, and made available the records for APE 10 and APE 12, but did not make available the records for APE 11.

54.         Following that meeting, HMRC reviewed the underlying records for APEs 10 and 12. HMRC considered these to be incomplete, including incomplete sales invoices, no prime records of cash purchases, and missing bank statements. HMRC's initial comparison of cash purchases, cash receipts, and cash drawn from the bank suggested that there was insufficient cash to meet cash purchases: i.e., the Appellant was appearing to spend more money than it actually had, i.e., that there was similarly a 'negative cash' position.

55.         In April 2013, HMRC reviewed the business records for APE 11 (i.e., the intervening year). Its concerns were not allayed. It found that the day-to-day recording of sales was not accurate, with the Appellant issuing invoices retrospectively to match banked amounts not previously invoiced; the purchase day book included large entries for cash purchases covering weekly/fortnightly/monthly periods; the Appellant's book-keeper Ms Mullinder had made a series of journal adjustments amounting to £4.23m, and a further year end journal adjustment of £250,000 to reflect unrecorded cash purchases.

56.         HMRC relies on the fact that its reviews of the Appellant's records for APE 07, 10, and 11 are consistent as between themselves and - HMRC contends - establish a pattern whereby the Appellant's records are inadequate in the following respects:

(1)          Sales invoices were not always issued contemporaneously to customers and were only issued subsequently to match banked accounts;

(2)          Purchase records were incomplete and significant monthly adjustments were made for cash purchases which had later been established to be incorrect.

57.         In relation to APE 11, HMRC added:

(1)          No documentary evidence had been provided of how the journal adjustments made in July 2010 by Ms Mullinder (which increased cash purchases by £4.23m) had been calculated;

(2)          If those adjustments were correct, then there was insufficient cash to meet cash purchases in the later part of the year;

(3)          The purchase records for 24 whole days of the year when the Appellant was open for business were missing, and significant amounts of cash could not be accounted for.

58.         In relation to APE 11, HMRC's calculation was that £98,035,574 was stated as paid out in cash, but HMRC could only find records for £88,956,885. On that footing, HMRC formed the view that the Appellant's costs of sale were overstated and refused to allow the difference (= £9,078,689) of the cost of sales.

59.         £9,078,689 was 6.42% of the declared cost of sales. £9,078,689 reduced the cost of sales to 91.95% of turnover.

60.         Having arrived at that point for APE 11, HMRC rely on the presumption of continuity to apply or carry across a similar disallowance to all the other years in issue (although for APE 09, its methodology is slightly different).

61.         In its Statement of Case, HMRC puts it in this way: 'the amendments for the APEs 07, 08, 10, 12, 14 were made by adjusting the purchases figure by adding back the same percentage of the overall cost of sales as in the APR 11 of 6.42%'. That statement was not completely accurate because, as HMRC's letter of 21 January 2016 made clear, the corporation tax adjustment which it made for APE 13 was made on the basis of the 6.42% as well (£169,630,767 x 6.42% = £10,890,295): see Paragraph 5 of that letter.

62.         In relation to APE 09, HMRC used a different metric. Since HMRC did not hold the relevant information regarding cost of sales for that year, HMRC estimated the cost of sales as 91.95% of turnover: see HMRC's letter dated 21 January 2016 at Paragraph 6.

63.         When it comes to the alleged continuity, HMRC's position is that the problems that occurred in APE 11 'were likely to have been present in all the other accounting periods at issue, so that a loss of tax had occurred in those accounting periods. By estimating the overstated purchases with reference to the cost of sales, HMRC had adopted a reasonable and appropriate methodology to quantify that over-statement'. It then applied 6.42% to that figure.

64.         Having considered there to be an under-declaration in the cost of sales for APE 11, HMRC then went on to contend that 'in the absence of any evidence to the contrary, HMRC consider that the cash on hand as at 31 January 2011 of £9,072,251 should be treated as having been drawn by Mr Ragonesi' and hence liable to charge.

65.         HMRC went still further than this, and also assessed corporation tax on the basis that any sums so calculated, in all years of assessment, and not just APE 11, had been lent or advanced by the Appellant to Mr Ragonesi. Its letter of 21 January 2016 said (at Paragraph 7): 'Where expenditure has been disallowed because there is no records" (sic) "HMRC's view is that these amounts should be treated as having been drawn by the director" (i.e., Mr Ragonesi).

66.         In a nutshell, the outcome of HMRC's reasoning is that the Appellant actually made profits of about £47m rather than (the declared) £4.9m, and that Mr Ragonesi was to be taken as having taken about £38m - whether in undeclared cash or equivalent value of gold, or a combination of the two - from the Appellant.

67.         These are the core principles of HMRC's formal case, as advanced before the Tribunal in response to the Appeal. Those core principles have remained the same, albeit subject to some adjustment of the figures.

68.         At this point, one thing which we must note is that HMRC's position (as opposed to its formal case) can be seen, from the extensive suite of documents placed before us, to have evolved significantly over the course of time. Part of this seems to flow from the fact that there was a change of officer at HMRC. The officer dealing with the Appellant at HMRC was, until at least 24 February 2015, Martyn Cope, when he was replaced by Mrs Spaughton. One important feature of this case is that, throughout 2014 and into early 2015, Mr Cope undertook a determined 'push', which went through a number of iterations, to arrive at feasible, substantiable, figures for the business in APE 11.

The Appellant's case

 

69.         In broad outline, what follows summarises the Appellant's case, taken both from the Grounds of Appeal (and the letters dated 2 February 2016 which accompanied them) and also from the Appellant's Skeleton Argument.

70.         The Appellant accepts that some of the primary records that could demonstrate the Appellant's cash expenditure were and are not available. The Appellant's position is that they have been lost, although it accepts that there is no hard-and-fast evidence as to how they came to be lost (in the sense that they existed and then went missing, as opposed to never having existed in the first place) and when and in what circumstances this happened.

71.         The Appellant contends that it would not have been possible for the Appellant to maintain the sort of profit margin that HMRC's amendments indicate, which is in the region of 8-9%. The Appellant argues that its true margin was in the region of 2%, which, it argues, are in line with those of its competitors (of which there are only three or four) and cannot so far exceed those of other businesses.

72.         The Appellant argues that there are daily records of its gold sales to the smelter (Engelhard, in Cinderford), which was the Appellant's predominant customer, for the relevant period, which show the minimum amount the Appellant must have spent on gold to a high degree of accuracy. The Appellant argues that a 'reconciliation' or comparison of physical gold sold by it (to the smelter) with physical gold purchased by it supports the cash purchase figures in the accounts.

73.         The Appellant argues that it can establish that it did make cash purchases on a number of days for which prime records are missing in APE 11 and can, to a tolerable degree of accuracy, reconstruct those figures.

74.         The Appellant also argues that it cannot be presumed that any errors made in APE 11 were also made in the entire period 2006-2014, since there are reasons why the accounting system for 2010-11 was less likely to be reliable than in other years.

75.         The Appellant argues that there is no evidence to indicate that Mr Ragonesi has either stolen, taken 'or otherwise extracted' excess cash or made 'unrecorded withdrawals' from the Appellant. The Appellant relies on a 6 year private side review by Grant Thornton which did not find any irregularities in the private financial affairs of Mr or Mrs Ragonesi, as well as on the absence of corroborative evidence, and inherent likelihoods.

The Appellant's 'new' Ground of Appeal

 

76.         In his closing submissions, Mr Chacko, on behalf of the Appellant, raised what he considered to be a new argument.

77.         This related only to APE 12. HMRC made a Revenue amendment on 22 July 2015. This was cancelled by HMRC on 20 January 2016 on the ostensible footing (on HMRC's case) that no inquiry had ever been opened. On HMRC's reasoning, if there were no open inquiry, then HMRC could not amend the return: see Schedule 18 Paragraph 34 of the Finance Act 1998. 

78.         On 22 January 2016 (that is to say, two days after withdrawing the amendments) HMRC issued the discovery assessment which is now in issue in this dispute.

79.         The Appellant (relying on Schedule 18 Paragraph 24 and the decision of the Upper Tribunal in Portland Gas Storage v HMRC [2014] UKUT 0270 (TCC)) argues that there was in fact a valid inquiry into the return for APE 12, which inquiry was validly closed by way of the amendments - meaning that HMRC could not, once it had withdrawn its amendments to the return, issue a discovery assessment instead unless the conditions in Schedule 18 Paragraphs 43 or 44 of the Finance Act 1998 were met.

80.         We required this new argument to be formally framed as a new ground of appeal. That was done.

81.         HMRC opposed the introduction of a new ground of appeal. Particular reference was made to the decision of the Court of Appeal (as affirmed by the Supreme Court: [2017] UKSC 55) in BPP Holdings v HMRC [2016] EWCA Civ 121.

82.         At this stage, and without needing to engage in detail with the merits, we simply need to consider whether it was open to the Appellant to raise this new point.

83.         We consider that the Appellant can.

84.         The overriding objective, set out in Rule 2 of the Tribunal's Rules, expressly directs us to deal with cases 'fairly and justly': Rule 2(1). Rule 5 confers extensive case-management powers on us, including the ability to extend time for compliance.

85.         We consider it fair and just to afford the Appellant the opportunity to advance this argument, notwithstanding the late stage at which it was advanced.

86.         There is no doubt that the guidance of the Court of Appeal and the Senior President of Tribunals given in BPP Holdings v HMRC binds us. But that guidance does not prescribe any particular course of action in any particular situation; nor in this one.

87.         Lateness, in and of itself, is only part of the picture. The real point is one of prejudice - for example, whether something is raised at such a late stage that there is no opportunity for any meaningful response to it.

88.         That is not the case here. Firstly, the point was initially made, in the Appellant's representatives' letter to the Tribunal dated 8 February 2016, which said "The assessments and amendments are something of a dog's breakfast in that HMRC incorrectly made assessments and amendments for years 2010, 2012, and 2013 when they should have made discovery assessments". Although put a little colloquially, and without the level of detail later supplied by Mr Chacko, it makes the point.

89.         It is true that the argument seems to have been washed away in the tide of adversarial and largely non-progressive correspondence which ensued, and was not then raised again until the Appellant's Supplementary Skeleton Argument dated 31 January 2018, filed in advance of a hearing on 2 February 2018. But, although time was tight, it was raised with sufficient time for experienced Counsel appearing for HMRC to consider the point.

90.         Moreover, the Ground of Appeal does not require the hearing of new oral evidence, or the consideration of any new documents. We were not asked to recall any witnesses. The new Ground of Appeal is essentially one of law. Insofar as the new Ground of Appeal does deal with matters of fact, those are facts already at large between the parties in this appeal.   

The facts

 

91.         The Appellant limited company operates from commercial premises in the 'Jewellery Quarter' in Birmingham. At the material times, its sole director and governing mind was Mr Alexander (also known as 'Ali') Ragonesi. He and his wife were equal shareholders. The Ragonesis bought the Appellant in 1998, but they had been in the jewellery business for about 20 years before then, trading together in a partnership known as 'The Golden Purse'.

92.         The Appellant's business was originally a retail jeweller - that is, buying and selling pieces of jewellery to the public. But, the global financial crisis and the 'credit crunch' led to a significant increase in the retail price of gold. The price increased from about £435 an ounce in August 2008 to a peak of around £1,165 an ounce in September 2011.

93.         These factors encouraged people the length and breadth of the United Kingdom to rummage around to find any unwanted gold so as to sell it and turn it into cash.

94.         Many sales were to local agents and shops. Such was the demand, these were not always jewellers. Even businesses such as butchers were buying gold. The small amounts being sold by individuals were compounded into increasingly larger batches, passing through the hands of 'middle-men', and up the transaction chain.

95.         The Appellant (as well as a small number of other firms in Birmingham) was at the top of the supply chain - immediately below the smelter.

96.         The main focus of the Appellant's business became buying and selling scrap gold. Much was jewellery such as ear-rings, rings, and necklaces. Much of this was 9 carat gold. It also bought also coins, such as sovereigns and 'Krugerrands', and gold bars. Those were of a higher grade of gold. Occasionally, the Appellant bought more exotic objets d'art (sic), such as a gold heraldic bedstead said to have weighed about 100kg.

97.         The Appellant sold most of the gold to the smelter, where it was refined and processed. The overwhelming majority (at least 95%) of the Appellant's sales were to the firm of Englehard in Cinderford. The Appellant would pay for the gold to be transported from its premises to the smelter. The Appellant kept gold coins, which it tended to sell to coin dealers rather than send to be melted down. There were two reasons for this. The smelter paid the same for gold whatever its form (that is, irrespective of whether it was scrap, worked into fine jewellery, or coins) and the Appellant could make slightly more profit - perhaps 1% - by selling coins of numismatic interest to dealers.

98.         It was not in dispute that the Appellant, and other firms in its position, were suddenly handling extremely large quantities of gold.

99.         Nor was it in dispute that much of this business was being done for cash. These two factors, in combination, are what have given rise to this appeal.

The increased declared turnover

 

100.     It was not in dispute that the Appellant's turnover - in response to the changing economic conditions - increased both rapidly and tremendously.

101.     The figures are striking.

102.     In the year ending 31 January 2006, the Appellant's declared turnover was £2.961m.

103.     In the year ending 31 January 2007, which is the first year in dispute in this appeal, and which is the first year in which the Appellant became involved in larger scale buying and selling of gold, its declared turnover leapt to £12.856m - an increase of just under £10m (£9.895m) or about 334%.

104.     These are the other declared turnover figures (with APE 11 highlighted):

(1)          y.e. 31.1.08                   £5,285,013

(2)          y.e. 31.1.09                   £5,547,708

(3)          y.e. 31.1.10                   £50,730,246

(4)          y.e. 31.1.11                   £143,802,136

(5)          y.e. 31.1.12                   £206,081,126

(6)          y.e. 31.1.13                   £170,282,961

(7)          y.e. 31.1.14                   £72,224,418

 

How the business bought and sold

 

105.     The Appellant arrived at the price which it paid to its sellers on the basis of three factors:

(1)          The 'spot price' (fixed twice daily) for 'fine gold' (i.e. gold of 99.999% purity);

(2)          Weight, and

(3)          Purity.

106.     The spot price for fine gold pertaining at the time of the sale was always available online. The Appellant's business was not a gamble on price movements. The Appellant usually used the morning fix, and carried on using it all day, but occasionally, if the price was moving rapidly, or if it had been asked, it would use the afternoon fix.

107.     Since the Appellant was buying and selling gold much of which was of a lesser purity than fine gold (for example 9 carat) then the price per gramme of the thing actually being bought was adjusted by the Appellant.

108.     The Appellant aimed to pay its sellers, but did not always pay, the equivalent of 97.5% of the spot price. However, the Appellant did adjust the price paid - albeit on an unknown (and now, unknowable) number of occasions - to take account of the identity of its customer. Whilst many of its sellers got 97.5% of the spot price, some of its bigger customers sometimes got up to 1%, or even 1.5%, more. Thus, the Appellant's margin was 2% at most: but sometimes, especially for larger sellers, significantly less.

109.     To add further complexity, neither weight nor purity are completely reliable measures.

110.     As to weight, things are sometimes heavier than their gold content (of whatever purity). For example, hooped earrings are hollow and sometimes have brass wire looped through them, or are heavier than they should be simply because the owner has worn them whilst washing and they get soap on them.

111.     As to purity, an expensive 'gun' (a 'thermo analyser', costing about £20,000) is used to ascertain the purity of a piece of gold. The Appellant had one. We were given an interesting demonstration of this 'gun' by Mr Ragonesi on the afternoon of the first day of the appeal. But although scientific, the gun is not completely reliable where (for example) a piece is coated with gold, or a gold bar has a non-gold core, which is only discovered at the smelter when the bar fails to melt.

112.     The smelter would pay the Appellant 99.5% of the spot price. But the smelter did not simply accept everything which the Appellant sent to it, as and when it was sent. The smelter itself, buying at 99.5% of fix (i.e. a margin of 0.5% at most) was operating a different business model to the Appellant.

113.     Mr Ragonesi described a pattern where he would estimate, on Day 1, the amount of gold he was likely to buy. He would do that on the basis of phone calls to him from trade customers who intended to come into the shop. On the morning of Day 2, before 10am, he would then speak to the smelter and tell them how much fine gold he would be selling to the smelter. But, since he was not buying fine gold, but was buying gold of lesser purity, this would involve him estimating, on the basis of his knowledge and experience, how much gold of lesser purity he would be receiving, and then 'back-calculating' to arrive at the figure for the smelter.

114.     This mass was called his 'cover' and was recorded by the smelter on the Appellant's 'metal account'. If the Appellant had agreed to sell 10kg of fine gold to the smelter, then it was obliged to sell 10kg. If it sold less, then the shortfall would be deducted from its metal account. There were sometimes discrepancies in the amount actually bought. For example, the Appellant would buy more gold than expected. This would exceed the 'cover' but would be 'forward sold' (i.e., added to Day 3's cover).

115.     This is not an exact science because the smelter would take the gold and melt it down. Impurities would be burned off. But there is also a phenomenon known as 'melt loss'. The overall volume of fine gold produced, and paid for, would be somewhat less than the overall volume of gold sold to the smelter. Moreover, some impurities - such as silver, platinum and palladium - were themselves recouped in the refining process, weighed, and paid for. For example, the Appellant's sale of about 88kg of gold to Englehard on 1 February 2010 produced about 40kg of fine gold, 8kg of silver, and smaller amounts of platinum and palladium.

The role of cash

 

116.     It is not in dispute that a significant proportion of the Appellant's purchases of gold were made in cash.

117.     The bulk of the Appellant's business was buying from 'trade' dealers who themselves had bought from the public. But the Appellant did also buy from members of the public who walked through the doors or posted the gold to him, although this was not the main focus of his business. The Appellant still buys gold through the post, and Mr Ragonesi brought in a large bag of gold scrap to show us at the hearing, including items which had come to him through the post. There were a mixture of scraps, and shavings, and gold dust, and scrap jewellery. Each of these would have to be tested with the gun, sorted into carat value, individually weight by carat value, and valued. This was such a time-consuming business that Mr Ragonesi thought it worth the Appellant's while to buy gold guns for regular sellers, so that they could do some of the testing and sorting, and he and his staff would not have to.

118.     We heard from some of the Appellant's trade customers. Mr Scadeng was one step below the Appellant in the chain. His evidence was that he was selling £40,000 worth of gold per week (approximately £2m per year) to the Appellant. Mr Moore was in a similar position. He was regularly selling anything between £5,000 and £40,000 worth of scrap gold to the Appellant. Mr Haag was based in Chesterfield, and his evidence was that he was selling about £20,000 worth of gold per month.

119.     These people were paid in cash because they themselves were paying their (sub)sellers in cash.

120.     It is therefore unsurprising that what Mr Ragonesi described as his 'biggest stress' was 'the constant pressure for cash'. On the above turnover figures, this pressure must always have been acute. Taking the figure for APE 11, £143m divided (say) by 300 gives a daily figure of £476,000. If made up (say) of £20 notes (which Mr Ragonesi said it was) that would be about 23,000 banknotes each working day, albeit that these would often be in sealed blocks of £10,000 (500 notes).

121.     We heard startling evidence from Mr Ragonesi (which we accept as truthful) as to the lengths to which he had to go in order to secure sufficient supplies of cash to feed the supply of gold coming through the doors, which was - both figuratively and literally - a gold rush. He was operating cash-drawing (not overdraft) facilities with at least two banks, with a daily combined facility of up to £1.25m, and, until late 2010, was spending 2-3 hours every day driving around between them, each and every working day, extracting as much cash - hundreds of thousands of pounds - as he could.

122.     The need for cash was so great that, without access to an overdraft facility, Mr Ragonesi would sometimes need to buy cash from competitors. For example, between 2 February 2010 and 8 February 2010, he bought £650,000 worth of cash from two of his competitors. Mr Ragonesi's evidence was that he would do this by selling them bullion.

123.     We heard evidence from other people in the industry which was consistent with Mr Ragonesi's evidence. Mr Blackburn is the director of another firm in the jewellery quarter. He was managing to arrange deliveries of £400,000 per day, but it still was not enough. He went so far as to buy two armoured cars which he would use to go to the bank to withdraw 'hundreds of thousands more' at this peak.

124.     Mr Ragonesi eventually arranged for the bank to make cash deliveries, but this was still a slow process, and was limited to £200,000 a day which later, part-way through 2011, increased to £600,000 a day.

125.     Mr Ragonesi would keep a cash float at the business. Ms Heighway's evidence was that this could be £100,000, although Mr Ragonesi told Ms Mullinder that he would expect £500,000-£600,000. But there are no contemporary or reliable records, before January 2013, as to the cash position at the end of each day.

How cash transactions took place

 

126.     Here, there was consistent and compelling evidence from Mr Ragonesi and his trading counter-parties. Their evidence spoke volumes as to the nature of this business, and the standards of commercial integrity applied by its participants.

127.     Whilst it is (to put it mildly) unusual to encounter a goods business, buying and selling from one modest commercial address, turning over up to £1/4 billion in a single year, a large part of which was in cash, we are satisfied that, in broad terms, this was genuinely the position here.

128.     The purchasing side of the Appellant's business was almost wholly conducted then-and-there, and face-to-face. Most importantly, it was conducted on the basis of an obvious and exceptionally high degree of trust. Mr Ragonesi was dealing with people who he knew (sometimes whom he had known for many years), who knew him (likewise, often for many years), and who knew where he was to be found if anything went wrong. Mr Ragonesi had a commercial reputation to maintain.

129.     One consequence of the exceptionally high degree of trust amongst the repeat participants in this particular market was a markedly relaxed attitude to paperwork and record-keeping - at least during the early years of the 'gold rush', which (for better or worse) are the years with which we are most acutely concerned in this appeal.

130.     Not only were Mr Scadeng, Mr Haag, Mr Blackburn and Mr Moore people who had done business with Mr Ragonesi, but they were prepared to come to the Tribunal to give evidence as to his commercial integrity, and the integrity of the business which he ran. None of them complained that they had ever been short-changed by Mr Ragonesi, or subject to any form of sharp practice. The market seems to have operated with remarkable ease and lack of formality.

131.     There was credible evidence, which we accept, both from Mr Ragonesi and his counter-parties as to the actual way in which these cash transactions worked. It paints a very clear picture of a fast-moving, high-value, business, with large amounts of gold and money changing hands, day-in, day-out.

132.     The evidence of Mr Haag presents a telling picture, and touches on some of the points which are driving this dispute:

"I would often not be paid in full for the gold I bought. Sometimes I would not be paid at all when I dropped off the gold and have to return later for payment. Sometimes if I was there at the right time I would receive full payment. I would normally be given an IOU with the rest of the payment I was owed noted on it and I would come back later to collect payment. I would usually be issued an invoice when I was paid in full. Sometimes I would only have to wait a few days, but there were times I was waiting 3 months for payment..."

133.     Mr Reeder describes a similar situation. He was bringing in £6,000 - £18,000 worth of scrap gold to the Appellant every week. He would tend to be paid in full 'most of the time' but there were times when the Appellant did not have enough cash to pay in full for the gold. When that happened, Mr Reeder said that 'they would usually be able to pay me 50-60% of the price, and I would be given a note so that I could come back later to collect the rest. Sometimes I would only have to wait until that afternoon, sometimes I would need to wait around 4 days'.

134.     This is consistent with the Applicant's records wherever they showed no cash at hand, and instead showed a minus figure. Mr Ragonesi's explanation of this (given at a meeting on 21 July 2014, and in relation to APE 10) was that 'this would reflect purchases where payment has not been made. The Company did not keep a record of such transactions, but the invoice given to the customer would say that moneys were owed. The cash would be paid when it was available'.

135.     We are confident that the descriptions of how the trading worked which we were given were truthful. The witnesses were speaking from their own knowledge and experience, and without anything obvious to gain by giving evidence.  

136.     We have no doubt at all that if the Appellant (which, in reality, meant Mr Ragonesi) owed someone money for gold which it had bought from them, then they would know how much they were owed, and would be paid what they were owed, even if they were not paid straightaway. 

137.     The sector as a whole was a volume business, working on modest margins, driven by the 'spot' price of gold.

138.     This was a highly-competitive, 'word of mouth' business. Established commercial relationships were key. Nothing forced anyone to sell to the Appellant or Mr Ragonesi. A seller who was less than entirely satisfied, for whatever reason, with Mr Ragonesi or the Appellant, could go to one of several other shops in the very near vicinity to sell their gold. Moreover, the commercial consequences for the Appellant had it defaulted or fallen short on a single deal with a regular customer would have been immediate and severe. The personal consequences for Mr Ragonesi would have been likewise. His reputation - and hence the Appellant's reputation - would be gone, and its regular trade would have disappeared overnight.

139.     We accept Mr Ragonesi's evidence that only he and Ms Noble and Ms Heighway actually bought gold. Ms Heighway confirmed that she would give customers - but only people she knew - handwritten notes that stated how much was still owed - that is to say, 'IOUs'. She gave invoices '98% of the time' - that is to say, she did not give invoices 100% of the time. Ms Noble would give IOUs as well. Ms Heighway confirmed that the amounts of IOUs were tens of thousands. All three confirmed that they would honour IOUs, even though these sometimes did not even give the creditor's full details, simply because Mr Ragonesi, Ms Heighway, and Ms Noble recognised each others' handwriting. The only real precaution taken against fraud was that all would decline to pay against an IOU if they did not physically recognise the bearer, even if the handwriting was authentic.

140.     We accept Mr Ragonesi's evidence that he had 400-700 regular trade cash customers, of whom a significant number - he put it at around 100 - would be willing to sell to the Appellant regularly on credit. This led to a number of customers with whom the Appellant ended up having a rolling credit. Mr Ragonesi's evidence was that sometimes over £1m - that is to say, a day or two's turnover - would be owing to various customers. His evidence was that he was sometimes so busy that he did not know whether people had even been given a receipt. As Mr Ragonesi candidly acknowledged: "basically, it was mayhem for a short while".

The Appellant's book-keeping regime

 

141.     Taking all the above features into account, and even taking account of the inherent trust and confidence between the market's participants, there were several obvious inadequacies with the reliability and resilience of the Appellant's record-keeping and mercantile habits:

(1)          Absence of contemporary documentation from the Appellant's counter-party/ies;

(2)          Deferred payment, whether of whole or of part, and significant volume of rolling credit;

(3)          No nominal ledgers were kept;

(4)          Informal documents, such as 'IOUs', were used, rather than formal credit notes;

(5)          IOUs were destroyed once retrieved;

(6)          Invoices were not always issued at the time;

(7)          Failure to issue interim invoices recording part-payment.

142.     Schedule 18 Paragraph 21 of the Finance Act 1998 sets out the Appellant's legal obligations. It provides that a company which may be required to deliver a company tax return must (a) keep such records as may be needed to enable it to deliver a correct and complete return for the period, and (b) preserve those records until the end of the relevant day. The relevant day is, unless HMRC specify an earlier day, the sixth anniversary of the end of the period for which the company may be required to deliver a company tax return. The records required to be kept and preserved include records of (a) all receipts and expenses in the course of the company's activities, and the matters in respect of which the receipts and expenses arise, and (b) in the case of a trade involving dealing in goods, all sales and purchases made in the course of the trade.

143.     The Appellant's 'back-office' - book-keeping, invoicing, cash management and control - were (perhaps) appropriate for the type and size of business up to and including APE 09, where annual turnover was approximately £5m.

144.     But thereafter, beginning in APE 10 (when turnover increased to approximately £50m) the position changed dramatically. However, the Appellant (for the detailed reasons set out below) failed to change with sufficient agility to keep pace with the increased turnover and focus of the business.

145.     The Appellant's books were kept by Ms Barbara Mullinder. Ms Mullinder impressed us as an intelligent and meticulous person. Her evidence was given truthfully and with candour.

146.     She had been with the business from 1988. She was not a trained accountant, and not even a formally trained bookkeeper. She was a part-time sole-trader. She was self-taught and had learned book-keeping as she went along with various jobs. Her own evidence was that she did not necessarily work in a conventional way, but based her work on what she described as common sense. She reached an important birthday in December 2009 (i.e. in APE 10) and, in response, scaled-down the number of her clients from about 15 to 5. Of these, the Appellant was the only bullion dealer. Ms Mullinder was charging the business £15 an hour for keeping its books. Later on, and at the height of the gold rush, she was charging £700-£800 per month. On any view, the Appellant was appropriating a miniscule proportion of its resources towards keeping its books in proper order. This was a glaring false economy.

147.     The Appellant placed extremely heavy reliance on her. But, for the reasons which we explain more fully below, the Appellant's continued reliance on Ms Mullinder (i) when the turnover of the business increased from about £5m in APE 09 to about £50m in APE 10, and (ii) when the focus of the business (again, in APE 10) changed from retail to to buying scrap and bullion, was a significant management failure by Mr Ragonesi, and therefore a significant management failure by the Appellant.  

148.     We must be clear - the entire responsibility for managing and directing change rested on Mr Ragonesi, and not on Ms Mullinder. It was part of Mr Ragonesi's directorial responsibility to ensure that the Appellant's records were complete and correct.

149.     Even though Mr Ragonesi, on his own evidence, was a poor administrator, it would nonetheless have been completely obvious to any objective bystander and should have been completely obvious to him, at the time, that the Appellant's back-office was just inadequate to keep track of the gold and money surging through the business.

150.     Until February 2010 (that is, until APE 11), the Appellant did not have any electronic invoicing system at all at its premises to deal with gold-buying, although there may have been a computer which had been used to deal with the other, earlier, retail, aspects of the business.

151.     There were manual carbon-copy purchase invoice books which the staff would use to issue purchase invoices on a day-to-day basis. The carbon-copy would be kept and the top copy would be given to the seller. Mr Ragonesi's evidence was that he and the staff - mainly Ms Heighway and Ms Noble - between them would often have several books - two or three - 'on the go' at once.

152.     Even then, the system was open to muddle and lack of completeness. Ms Heighway accepted that invoices were not being given 100% of the time. Ms Mullinder made the telling point that it was more than possible that during the very busy periods some purchases might have been agreed without the invoice being properly produced. For example, she said that "it is entirely possible that Mr Ragonesi could have been distracted mid-way through a transaction and did not return to it". Having had the opportunity to hear Mr Ragonesi give evidence, we consider this likely. He is an extremely dynamic individual, and his focus was on making money, and not on keeping the paperwork in order.

153.     This all feeds into our overall view, set out in more detail below, about the inadequacy of the paper-based system.

154.     When each invoice book was finished with, it would be left for Mr Ragonesi to take away to deliver to Ms Mullinder. She did not work at the Appellant's premises but, during this period, worked entirely from home in Telford. Mr Ragonesi would take her the invoice books and she would post details into her personal SAGE system. She would also keep her own manual ledgers. She would tot up the figures in the invoice books using her calculator, and would then tear off the print off and staple it to the front of the books.

155.     Over the course of a year, there would be hundreds of books. There is no evidence that these books were sequentially numbered, as between successive books, or securely stored. Hence, there was no sufficiently sure way of knowing if books went missing or were mislaid. Indeed, part of the problem in this appeal was the re-discovery, at some point, by Ms Mullinder of three invoice books, which contain invoices ostensibly issued on six of the 24 'missing' days, during the period 1-8 February 2010: her exhibit BM1.

156.     BM1 illustrates this point quite well. Invoice book 1 contains 50 sequentially numbered invoices (2501 to 2550) dated from 29 January 2010 to 22 February 2010. Book 2 contains 50 (2101 to 2150) dated from 26 January 2010 to 15 February 2010). Book 3 contains 50 (2151 to 2200) dated from 26 January 2010 to 9 February 2010. The books therefore (i) are not in sequence (because 2200 to 2501, or 300 invoices, or 6 books, are not present), (ii) in any event, were 'on the go' simultaneously; (iii) were not ruled off at the end of the accounting year. Close examination of BM1 reveals other anomalies, such as dates being out of sequence (see, e.g., BM1/28 (9 February 2010) and BM1/29 (8 February 2010). Mistakes were obviously being made, on the hoof, at the time. It is simply not possible, almost 10 years down the line, to disentangle these mistakes.

157.     Ms Mullinder did not think that those three duplicate books were in fact all the books which were kept for that period, although she was not sure.

158.     This system was self-evidently inadequate. Firstly, it was entirely dependent on the books being kept fully and accurately, so as to record each and every transaction. But this was not happening. There were several books on the go at once, and a real possibility of transactions not being recorded at all or of being recorded incompletely or inaccurately. No-one could say for sure was happening. No-one was checking. We are quite sure that transactions were not being properly recorded.

159.      The integrity of the system (sic) was also dependent on the Appellant taking care to retain all the books, once they had been finished with, so as to be in a position where it could identify (for example, though sequential numbering) if any were missing. That was not done here. The Appellant tries to suggest that books and papers have gone, or must have gone (which is not the same thing) missing whilst in the hands of a succession of professional advisers, with the inference that this is not the Appellant's fault. This is simply not good enough. There is no evidence that the books and papers were ever complete. On the contrary, there is evidence that they were never complete. Ultimately, it is the Appellant's responsibility, and its responsibility alone, to keep the books in proper order.

160.     The 'rolling credit' and/or IOUs being used by the Appellant would inevitably carry the risk of conflict or confusion with the books. There is no evidence that anyone was cross-checking the books, one against the other. Ms Mullinder was not doing it. To compound the problem, no individually named or identified client ledgers or accounts were kept, so as to allow the picture for any individual trader to be easily and quickly ascertained. It remains far from clear to us, even now, how the 'closing' financial position in relation to any trading counter-party could have been settled except ultimately through reliance on memory, goodwill, and trust.   

161.     To his credit, Mr Ragonesi - albeit belatedly - recognised that this system needed to change. He recognised that things (in his words) "had got out of hand" When things got busy (again, in his own words) "paperwork was the last thing on my mind". But, and as he candidly volunteered, he was "having the time of his life", simply revelling in being busy and (as he put it) going out to work every day to "play shops", buying hundreds of thousands of pounds worth of gold every day.

162.     At some point towards the end of 2009 - we do not know exactly when - Mr Ragonesi commissioned a computer system. Although it was supposed to arrive at Christmas 2009, it did not arrive until February 2010.

163.     It came into operation on 9 February 2010.

164.     This system was designed for the Appellant by a Mr Mollison-Smith, from whom there was no witness statement, and from whom we did not hear any evidence. Although he may well have had material evidence to give, we draw no adverse inferences from his absence. The thrust of the evidence from those who used the system was that it was problematic. This was an under-statement.

165.     Mr Ragonesi described these problems as 'irritating design flaws'. There was a 'cash' default option, which led to mistaken posting of purchases as cash when they were made by cheque or telegraphic transfer. Another was the system's retention of the details of the preceding invoice, which sometimes carried across. That is to say, individual invoices were not being saved on the system. That is - put neutrally - a most surprising 'design flaw', since it undercuts most of the practical utility of having a computer-based system in the first place. As Ms Noble, who used the system, put it: "You had to be careful to change everything from the previous invoice or you would include something you didn't mean to".

166.     There was also evidence from Ms Mullinder and Ms Noble that the system sometimes broke down or "crashed". We do not know how often, or for how long. Ms Noble did not remember any full day when the system had been down, but said it "could be hours before the computer man would come".

167.     In the event of a breakdown, Ms Mullinder's view was that the duplicate books should have been used as a paper-based fail-safe, with the contents 'put through the system ... when it came back up'. However, she could not say whether or not that had been done each and every time. As she remarked, she was not there, physically, in the building, day in day out. We do note that BM1 books 1 and 2 each contain invoices for several dates on and after 10 February 2010, when the computer system should have been up and running. This is suggestive that the use of the invoice books, once the computer system had been introduced was, for whatever reason, somewhat more prevalent that the explanations given to us would suggest. 

168.     A further problem was that the system was not capable of producing daily totals. The daily totals could only be arrived at by someone gathering up the duplicates (during this period, up to 100 at a time) and adding them up, just as Ms Mullinder had been doing earlier. The copies would be put into an envelope and the total written on the outside of the envelope to give to Ms Mullinder. The adding-up and writing on the envelope was often done by Trevor Bennett (who mainly handled shop sales) or Christine Noble (who mainly handled postal sales). But that sum was not always double-checked. As Ms Noble said, it depended on how busy they were. Ms Mullinder confirmed that she had seen, on her visits from July 2010 onwards, the shop so busy that Mr Ragonesi and Ms Heighway "didn't have time to breathe."

169.     A further problem, identified by Ms Mullinder in her evidence, was that, for a period which she thought lasted from February to July 2010, quite a few suppliers came in with their own paperwork (i.e., self-billed invoices) and "it took quite a while for it to filter through that those invoices weren't actually getting onto the computer system (that is, they were not making it onto the computerised scrap system) "so, therefore, they weren't added in the totals". We accept Ms Mullinder's evidence that she personally could not have identified this particular problem at the time because she was not visiting the premises, and she did not see the scrap invoices.

170.     Ms Mullinder's evidence was that she was not given the envelopes with the invoices inside, and so was not in a position, after 9 February 2010, even to cross-check the figures on the outside of the envelopes to make sure that they tallied with the invoices inside the envelopes. However, she had done this for BM1 books 1 and 2.

171.     Here, further error creeps in. Mr Ragonesi's evidence was that latterly the Appellant's representatives had realised, at some point during the course of preparing for this appeal, that the totals on the outside of the envelopes could be wrong because of missing or transposed digits.

172.     So, whilst Ms Mullinder did write down daily figures for each and every day from March to November 2010, these figures are not reliable and cannot now be relied upon because, even if they came from the outside of the envelopes, it is accepted that the figures on the outside of the envelopes are unreliable. 

173.     To round things off, there were loose invoices - not in any envelope at all. There were at least four of these 'loose between the envelopes for the missing days' which Ms Mullinder exhibits as BM4.

174.     The original computer system was frankly amateurish. We would go so far as to say that it was not really a 'system' as such. It was simply a surrogate typewriter, and not a very good one at that. It did not seem to represent any significant step forward in improving the quality or reliability of the Appellant's record-keeping. Paradoxically, in some respects, it seemed to have been somewhat worse - less satisfactory and less immune from error - than the carbon-paper based system which it was aiming to replace. At least the carbon-paper based system had sequentially numbered invoices in each book, with the carbon flimsies bound together in books, with each invoice being written separately, from scratch each time, meaning that the same invoice could not be accidentally written and issued twice. Mr Ragonesi acknowledged that, when the computer system was put in, things got worse before they got better.

175.     The computer system did not even automatically add unique invoice numbers until 7 September 2010.

176.     Even then, there are still significant gaps in the records, because there are 259 missing invoices between numbers 1228 (29 October 2010) and 1487 (4 November 2010) which period includes 4 working days. It was only the adding of unique individual invoice numbers which permitted the Appellant to identify that a swathe of invoices were missing.

177.     From 'around September 2010' (but no-one could give the exact date) the system would give the daily total and Ms Noble or Mr Bennett would just copy this down onto the outside of the envelope, but without any cross-check. Of course, the daily total would only ever be as accurate and reliable as the inputted data, and we have already identified significant, and un-remedied, inadequacies with that system. It is common ground that one of these notes, for 19 May 2010, is almost half a million pounds less than the invoices held for that day.

178.     In our view, whilst the introduction of the computerised system in February 2010 was ostensibly progress, of a sort, it was still far from sufficient to get to grips, in a sufficiently careful, thorough, or organised way, with the need for the Appellant to keep records sufficient to allow its purchases to be fully, accurately, and contemporaneously recorded and intelligently scrutinised. That remained the same despite the 'tweaks' applied to the system throughout 2010.

179.     There remains doubt as to whether the first computer system was replaced by another system in late 2010 or early 2011, or whether the underlying system remained there, but with modifications.

180.     The evidence all points strongly to a failure in the Appellant's record-keeping system which continued after 9 February 2010, and which therefore affected APE 11.

181.     We have regard to other changes which were taking place.

182.     In mid-2010, Ms Mullinder started visiting the premises around once a month.

183.     PWC were engaged in mid-2010 in place of Bentley Jennison (who later became part of Tenon). Mr Ragonesi believed that PWC were 'top accountants' with experience of the bullion trade. His evidence was that he expected them to 'pick up if we had any shortfalls or flaws in our paperwork, or had not handled adjustments properly'.

184.     But, having looked at matters in the round, we do not regard the increased frequency of Ms Mullinder's attendance, or the involvement of PWC, as effecting any sufficient, demonstrable, improvement in the Appellant's record-keeping systems to the point where it could really be said that the Appellant's record-keeping was adequate.

185.     Indeed, paradoxically, the attendance of Ms Mullinder at the business in mid-2010 led directly to the creation of one of the difficulties which subsequently bedevilled this dispute - namely, the £4.23m adjustment to the Appellant's books.

186.     PWC were not keeping the Appellant's books on a daily basis. PWC did not conduct an audit. PWC may well also have been working to a threshold of materiality of £1m, which was, by that point, significantly less than 1% of turnover. No 'management' letters from PWC (of the sort which would ordinarily accompany annual accounts) were put into evidence before us, although Mr Bosley had seen them, as well as such letters from Grant Thornton, and so we cannot assess what advice the Appellant was (or was not) getting from PWC and what it was (or was not) doing in response to such advice.

187.     The back-office systems were simply inadequate for this business which had moved very rapidly from the retail of individual pieces of jewellery to the wholesale purchase and sale of hundreds of kilograms of gold scrap. The situation was (both figuratively and literally) a gold rush. As Ms Mullinder put it herself, both in her written and her oral evidence "The business was still growing so rapidly in 2010 that there were certainly problems and the staff used to dealing in staff purchases were not bookkeepers".

188.     Indeed, Mr Ragonesi accepted (in his meeting with HMRC on 21 July 2014) that he had been careless with his book-keeping. That was expressed in general terms, but it sensibly reflected the inevitable realisation - years too late - that, had the record keeping been done with greater care, the situation facing the Appellant would have been very much different.

The £4.23m adjustment

 

189.     In our view, the system was always perilously close to breakdown. That breakdown eventually came when Ms Mullinder took a few weeks off for an operation in March or April 2010 (i.e., during APE 11). After the operation, she was advised not to drive and so she did not visit the Appellant's premises at all during that period. She was out of action for about eight weeks.

190.     When she started work again, she found herself unable to balance the books, no matter how hard she tried. There was a discrepancy of about £4.23m in the cash position which was not supported by the prime records. She took this up with Mr Ragonesi in mid-August 2010, when she was preparing the VAT return for that quarter, asking him words to the effect of whether "he happened to have £5m in the safe". Ms Mullinder did not know how much cash-in-hand Mr Ragonesi normally kept because she had never been involved in counting the cash. The very fact that the one and only book-keeper did not know, even vaguely, the cash position in a largely cash business throws a harsh and revealing light as to the inadequacy of the Appellant's systems and record-keeping. According to Ms Mullinder, whose recollection and evidence we believe, Mr Ragonesi "more or less" laughed her off. But it should have sounded the alarm straight away that something was badly amiss. Unfortunately, it did not.

191.     According to Ms Mullinder, she told Mr Ragonesi that there was "a big hole somewhere and a big mess" and she was going to "just start again". This was a short conversation - no more than a few minutes - and Mr Ragonesi did not really engage with it although Mr Ragonesi did tell her that he kept between £500,000 and £600,000 in cash.

192.     In the absence of any meaningful guidance or support, Ms Mullinder was put in an impossible position. She did various long-hand exercises trying to work out where the figures had gone wrong. She simply could not get to the bottom of it. In August 2010 she decided - in good faith, and approaching the matter as a book-keeper, and with the need to 'balance the books' taking priority over all over features - to post this £4.23m discrepancy as a loss split across a number of months' trading. She explained this as she thought it 'prudent' to break it down into months "rather than put the big £4.2 through as a lump sum". When it came to the amount per month, she had Mr Ragonesi's £500,000 to £600,000 figure in her head. We wish to emphasise that there is no suggestion - and we do not find - that this was not done by her with any ill intent. It was certainly not done with any intent on her part to mislead or deceive. But it was not a sensible thing to have done since it effectively disguised the discrepancy in the books. Mr Ragonesi should not have allowed this to happen.

193.     The £4.23m discrepancy (roughly equivalent to a few days' turnover at the time) was both a symptom and a cause. It was a symptom of the haphazard nature of the record-keeping. It was a cause (although not the only cause) of a pervasive and uncorrected inaccuracy in the books.

194.     We wish to be clear that this should not be read as critical of Ms Mullinder. She is not to blame for this mess. In our view, she was doing the best that she could, in good faith, and in difficult circumstances. But she was simply not equipped for the task which faced her, and she was overwhelmed by it. The business had become too big, too fast. She described it - fairly - as "an explosion". She said - accurately - that sometimes she wondered if she had "put enough noughts in her calculator" because of the figures the Appellant was dealing with. Mr Bosley - a sensible and measured individual - said that he had imagined it must have been "pandemonium".

195.     No adequate systems were put in place to cope with this increase in turnover, or to give Ms Mullinder any appropriate guidance or support. The Appellant, directed by Mr Ragonesi "playing shop", was doing everything on a wing and a prayer. The failure to balance the books was inevitable. It was Mr Ragonesi's responsibility to make sure that the books were being kept correctly, and the failure to discharge this responsibility is his.

196.     The Appellant only began to maintain a cash reconciliation of a conventional kind in January 2013. Ms Mullinder had not done so before then: she had not been directed to do so. The cash reconciliation was introduced by Mr Bosley. It started with the day's opening cash figure, adding in cash sales, cash deliveries, or bank withdrawals, and deducting the cash paid out to record a closing figure. Ms Mullinder started attending the premises weekly in January 2013, and she would check the reconciliation. Ms Mullinder described the work she started to do in January 2013 (i.e., the end of APE 13) being to check the bank statements, tracing back any TT or cheque payments to the bank statement, identifying anything said to have been paid by TT or cheque which had not been, and adjusting the cash figure so as to produce a daily running cash total. At the end of every month, she would then post the cash items onto SAGE, when it normally "agrees within pennies". Her view was that, from January 2013, the system is tight, and leaves no room for error. We accept that evidence.

197.     The deficiencies in record-keeping particularly affected the years up to and including APE 11.

The involvement of Mr Bosley

 

198.     In late 2011 (that is, during the course of APE 12) Mr Ragonesi approached Mr Peter Bosley in connection with Mr Ragonesi's intended take-over of another firm. Mr Bosley had qualified as an accountant in 1966. After an accountancy career with various firms, he became self-employed, and was a Fellow of the Institute of Chartered Accountants.

199.     Mr Bosley impressed us as a witness. He was trustworthy and non-partisan. He made appropriate concessions in his evidence. He struck us as an individual who was professionally intrigued by the apparent discrepancies in the Appellant's financial affairs, and was determined to get to the bottom of it, whichever way the materials happened to lead.

200.     Between December 2011 and July 2012 Mr Bosley's involvement with the Appellant was on what he described as a "fairly casual" basis. During this period, he was trying to get to grips with the business and familiarising himself with its systems, so that he could get used to how the Appellant was actually supposed to perform. He spoke with Mr Ragonesi and Ms Mullinder and tried to (as he put it) 'bridge the gap' between Ms Mullinder as a book-keeper and Mr Ragonesi as the director. That was insightful on Mr Bosley's part because, as is clear from our discussion above about the £4.23m discrepancy, there was certainly a serious communications gap to be bridged.

201.     From July 2012 (i.e., during APE 13) Mr Bosley began visiting the Appellant's offices monthly, and began to work more closely with Ms Mullinder.

202.     From January 2013 (i.e., the very end of APE 13, and into APE 14) Mr Bosley's involvement became weekly. He did not record or upload any raw data into SAGE. His role primarily focussed on producing running management accounts, although he would look at primary documents if he needed to.

203.     In our view, and beginning in January 2013, Mr Bosley was providing a second, and objective, viewpoint as to whether the figures made sense.

204.     There are two aspects to his involvement which are important in this appeal:

(1)          Establishing the true financial position in APE 11; and

(2)          The effect of Mr Bosley's involvement upon the accuracy and adequacy of the Appellant's record-keeping in APE 12 and onwards.

205.     We can deal with (2) first.

206.     We are confident that Mr Bosley's involvement from January 2013 had the effect that the figures reported were accurate and reliable.

207.     Even though Mr Bosley only became involved about half-way through APE 12 (and then on a fairly casual basis) we accept his evidence that the records for APE 12 and APE 13 were intact when he considered them. It is consistent with Ms Mullinder's evidence. We accept his evidence the records for APE 12 and APE 13 had been properly recorded and retained. To avoid any doubt, we consider that his evidence (of which we have considered the transcript) although referring to '12/13' was being given with reference to accounting periods ending, rather than to calendar years. That was the whole basis upon which he was working.

208.     We also accept Ms Mullinder's evidence that there were no significant errors in book-keeping after July 2012.

Mr Bosley's reconciliation for APE 11

 

209.     We now turn to deal with (1) and Mr Bosley's involvement with the APE 11 position. Mr Bosley had no involvement with the Appellant during this period. His involvement came only later.

210.     In early 2013, Mr Bosley performed a full cash reconciliation for APE 11 although the records for APE 11 were not intact.

211.     Mr Bosley's first cash reconciliation for APE 11 was provided to HMRC at a meeting on 22 August 2013. Some confusion creeps in here, because Mr Bosley's work was being adopted and presented by the Appellant's accountants as if it was coming from them rather than from Mr Bosley. But Mr Bosley was the one doing the detailed legwork.

212.     Even Mr Bosley was not able to balance the books for APE 11. His efforts were hampered by pervasive and irresolvable discrepancies between the daily totals of purchase invoices held and purchases as recorded on the Appellant's SAGE records. For example, in July 2010, the cash purchases in SAGE were £7.837m, but there were only physical records to support purchases of £5.196m - a difference of approximately £2.641m or about 50%.

213.     But, despite this, we consider that his work is extremely valuable and the conclusions he arrived at are deserving of respect and close attention. He is appropriately qualified. His approach was methodical and attended with professional rigour.

214.     Whilst some infelicities had crept into his work, this was hardly surprising given the state and condition of the information and material he was given to go on for APE 11.

215.     Mr Bosley remarked - candidly - that he had reached a stage 'where I did as much as I could but I realised that you couldn't actually get to a finite figure because of the way things were paid very often in stages, didn't relate back to the invoices'. He amplified this, fairly, by saying 'Even as an accountant you become drunk with figures sometimes ... I'm quite sure you could get half a dozen accountants to come up with different figures within reason. As I say, half a million on 100 million is not necessarily acceptable, but it's not a great difference ....". Here, the 'half a million' refers to the eventual difference between himself and HMRC's Mr Cope as to the feasible level of the Appellant's cash purchases.

216.     Unfortunately, there was never a direct discussion between Mr Bosley and Mr Cope, but, as Mr Bosley said "one could do the exercise again, and come up with a different figure. It wasn't finite. You couldn't say exactly what the cash would be." His view was that "the gaps and lack of information on certain areas make it virtually impossible to arrive at the right figure."

The amount in dispute: APE 11

 

217.     HMRC's fundamental position is that any deduction claimed from profits for cash purchases should be supported by evidence of the expenditure and the nature of that expenditure, although HMRC is willing to accept that this evidence need not necessarily be the Appellant's primary records but could be 'other evidence' that the Appellant had incurred the expenditure sought.

218.     The readiness to accept 'other evidence' is both pragmatic and commendable. The ordinary position is that it is the responsibility of any taxpayer to be able to satisfy HMRC as to the accuracy of their figures. The conventional way of doing this is by keeping accurate and complete records. Where the records - as here - are inadequate, it was open to the Appellant to provide alternative evidence, albeit of satisfactory probative value.

219.     However, the application of these simple propositions to the realities of this Appellant's business, against the background of the facts and matters set out above, is very difficult. Indeed, even in relation to one of the years - APE 11, which is the factual and legal heart of this appeal - it is an exercise which has all but defeated HMRC, the Appellant, a series of professional advisers, and specialist Counsel. Between them, they have spent several years grappling and sparring, inconclusively, with the figures.

220.     The Appellant accepts that there is (at least, prima facie) a discrepancy, and has engaged in a series of attempts to establish whether the discrepancy comes from lost invoices or unrecorded purchases.

221.     The very fact that these attempts have still failed to satisfactorily address the discrepancy, and were still being made by the Appellant as late as the Appellant's written closing submissions, speaks volumes as to the quality of the Appellant's records. We bear firmly in mind that the Appellant bears the burden of displacing the assessment. However, if the Appellant discharges its burden, we have to move on to do our best to establish the correct figure.

222.     Efforts were still being made, by the ingenuity of Counsel, in early 2018 to fill the gap. Whilst we admire the ingenuity of Mr Chacko in seeking to drive the APE 11 discrepancy down to zero (in fact, to below zero) we are bound to say that his submissions are just that - submissions. He is not a witness, and his submissions are not evidence. We have to be careful to give full and proper regard to the admissible evidence which we have read and heard. 

223.     As a judicial body, we are required to identify and then analyse the best available evidence as to the true position. As a specialist Tribunal we are required to apply our specialist skills. In the absence of much (if any) common ground between the parties, we are driven to a re-constructive exercise to establish the true position, doing the best that we can, on the available evidence.

224.     HMRC helpfully relies on the cash reconciliation prepared by Mr Bosley, as subsequently amended by HMRC: see its closing submissions at Para. 16.

225.     In August 2013, HMRC was provided with Mr Bosley's cash reconciliation work for APE 11.

226.     HMRC remained concerned as to the figures for APE 11, and considered that the declared figure for purchases in the accounts of £98,035,574 could not be fully evidenced. We agree.

227.     Indeed, it is common ground that the Appellant did not have purchase records sufficient to demonstrate the £98,035,574 claimed for its cash purchases in APE 11. Mr Bosley had identified total cash purchases supported by physical records at £93,671,567 - a shortfall of about £4.5m.

228.     HMRC had also identified 24 days between February and November 2010 in which the Appellant was open for business but for which no purchase invoices had been provided.

229.     Mr Cope did not consider that there was anything sinister underlying the missing 24 days. We agree. We do not consider that there was any ill-intent or intent to deceive. Put bluntly, the missing 24 days were the result of the then-prevailing 'pandemonium' in the Appellant's business rather than any deliberate wrongdoing.

230.     The parties - through Mr Bosley and Mr Cope - engaged in what Mr Bosley described as 'a long tedious exercise', with about 'three weeks of hammering away at the figures'. Mr Bosley thought that both he and Mr Cope had done their best. Mr Bosley's assessment of the end product was that Mr Cope, as Mr Bosley, also came to realise "that there were many omissions really or lack of information so as to make a conclusive result." That is an entirely fair observation, by someone who had tried, long and hard, to get to the bottom of the figures.

231.     The purpose of the exercise done by Mr Bosley, 'with Mr Cope's check-up', was to produce a feasible figure for cash purchases in APE 11, no more and no less, and without that becoming confused or clouded by other issues for that year (such as the £4.23m).   

232.     HMRC's view of Mr Bosley's work was that the total purchases for which records existed equalled £88,956,885. That was the result of amendments to Mr Bosley's original figures made by Mr Cope, who had identified a number of infelicities in Mr Bosley's approach, including (a) certain invoices originally treated as cash invoices but which had been paid through the bank and (b) certain invoices which had been double-counted.

233.     Mr Cope's closing position for APE 11, based as set out on Mr Bosley's work, was that the Appellant's purchases were £139,905,945, made up of cash purchases of £97,544,062 and invoice and cheques of £42,361,883, against recorded sales of £143,655,262 (a profit of £3,749,317). 

234.     The £97,544,062 includes an add-back by Mr Cope of a figure for unrecorded cash expenditure for the missing days at £368,000 per day. He had arrived at the figure of £368,000 on the basis of his calculation of the average daily scrap cash purchases (on the footing of days when there were records) for the 6 months ended 31 July 2010. Although the Appellant advanced a competing calculation for this average, we consider Mr Cope's figure to be the most feasible and so we adopt it.

235.     On that basis, the resultant gross profit was £3,749,317.

236.     However, that is not the end position. The arithmetic position changes to take account of a development during the course of the case. This was the belated discovery of records for 6 of the missing days, being 1-5 and 8 February 2010 (exhibited by Ms Mullinder as BM1) as well as some loose invoices (exhibited by Ms Mullinder at BM4)

237.     HMRC - by this point acting through Mrs Spaughton and not Mr Cope - wound the figures back to £88,956,885 (that is to say, stripping out the aggregate of the missing days' average) but allowed £614,325 (which was the sum represented by BM1 and BM4). That took HMRC's figure for cash purchases from £88,956,885 to £89,571,210. But that figure had no regard at all to the other missing days.

238.     Dispute emerged as to whether the invoices in BM1 and BM4 do represent all the daily purchases for each of those days, or just some of them. It seems to us that these invoices are the best evidence available to us, and that the figures on these invoices have to be taken at face value and as fully recording the Appellant's purchases for those six days. We reject the Appellant's suggestion that there were or must have been other purchases made on those days not recorded on those invoices but in relation to which the records remain lost. We are invited to draw various inferences from movements such as peaks and troughs in daily cash figures. That exercise simply cannot be performed, at this distance of time. It would amount to little more than guesswork. We remind ourselves that the Appellant bears the burden of discharging the assessment for APE 11. In relation to the figures for the 1-5 and 8 February 2010, it has failed to do so.

239.     HMRC (Officer Cope's) figure of £97,544,062 included figures for the missing days.

240.     HMRC (Mrs Spaughton's) revised figure of £89,571,210 did not include any figures for any (remaining) missing days.

241.     In apparent contradiction to its earlier position, HMRC now argues that it is impossible to reconstruct the amount of cash purchases for any of the (remaining) missing days.

242.     We reject that argument. It was certainly not seen as an impediment by Mr Cope, who proposed that the figures involved could, in the absence of primary records, be calculated with reference to an average, which he put at £368,000 per day.

243.     There is an argument that this was a proposal advanced only in the course of settlement discussions between the Appellant and HMRC and for the purpose of those discussions. That argument would perhaps enjoy greater traction if the parties - both legally advised and represented - had chosen not to place a substantial tranche of the relevant correspondence before us.

244.     As it stands, Mr Cope's approach and his reasoning - whether done for the purposes of settlement or not - are squarely before the Tribunal. Mr Bosley had considered his approach and reasoning and gave evidence about it.

245.     In our view, Mr Cope's approach to the missing days was pragmatic and progressive. His approach acknowledges that there are missing days, acknowledges an absence of malicious intent, and takes what we regard as an empirically comprehensible position for the sums for the missing days. The very existence of BM1 shows that the Appellant was trading on missing days. This is consistent with the Appellant's evidence, which we accept, that there was trading on all the missing days. Insofar as there is a dispute on the point, we accept the Appellant's evidence that it was not open for business on Tuesday 6 April 2010 (which was the Tuesday immediately following Easter Monday).

246.     We see no reason to depart from Mr Cope's adoption of a figure for the missing days, or the figure which he arrived at.

247.     However, given that there are now, at large, figures for six missing days, which were not before Officer Cope, but which are before us, we consider it appropriate for Mrs Spaughton to have taken account of them - as evidence - instead of ignoring them and applying the £368,000 average regardless. Although the average of these daily figures is less than £368,000, we reject Mr Chacko's invitation to treat these figures as if they were £368,000 so as to 'bring them up to the average'. That is to impermissibly 'mix and match' evidence, and gives his client unwarranted credit notwithstanding the absence of records.

248.     That still leaves 17 missing days. For those, we consider the £368,000 per day appropriate. The business was open on those days. There is no reason to suppose that the turnover of the business on those days was zero. Officer Cope considered - rightly, in our view - that, doing the best he could, on the basis of the information and materials before him - those days should be treated, in a rational, empirical, and defensible way - as each attracting an unrecorded figure of £368,000.

249.     17 x £368,000 = £6,256,000.

250.     £89,571,210 plus £6,256,000 = £95,827,210.

251.     That is £2,208,364 less than the declared figure of £98,035,574.

252.     Mr Cope had applied adjustments of (minus £1,791,881) for 2010 sales and (plus) of £1,627,650 (being Englehard's closing balance). It seems to us that the adjustments applied by Mr Cope were rational and pragmatic and are the best that could be done with the available information and materials.

253.     That takes us to the position where, for APE 11, applying the above figures and methodology, the actual cash purchase figure is £95,827,210, to which Mr Cope's adjustments should be applied in recalculating the Appellant's corporation tax liability for APE 11.

254.     £95,827,210 minus £1,791,881 = £94,035,329 plus £1,627,650 = £95,662,979.

255.     We have adopted the above methodology to arrive at what we consider to be the most feasible position in preference to others placed before us.

256.     But, lest our conclusions should fall to be re-examined, we consider it appropriate to set out why we have rejected the other approaches presented to us.

257.     The first main approach is a 'gold reconciliation' approach. This was a business which dealt in large quantities of scrap gold. HMRC does not dispute the proposition that the Appellant must have bought the gold which it sold to the smelter. We have found that most (but not all, e.g. excluding coins) of the gold would have been sent to the smelter. We agree that the Appellant's gold reconciliation shows that there was a large amount of unrecorded purchases, because the Appellant must have been sending gold which it had purchased to the smelter. The smelter's records are, to this extent, reliable, since the Appellant did not invoice the smelter. The smelter sent the Appellant self-generated invoices telling the Appellant what the smelter had bought from the Appellant.

258.     We agree that it would have been unlikely for Englehard to have produced invoices showing the purchase of more gold than they had received.

259.     However, ultimately, we are not persuaded that the 'gold reconciliation' is the right approach for this case. Whilst it is superficially attractive - because the Appellant could not physically sell gold to the smelter which it did not have, and the smelter would not have paid for gold which it did not receive - the working-out becomes too complex and ends up running into non-trivial doubt as to the actual quantities of gold.

260.     There are several difficulties here which cannot be solved on the basis of the information and materials before us:

(1)          This is not an approach which the parties themselves, in their iterative discussions, chose to adopt as their primary means of analysis;

(2)          The Appellant did not sell all the gold to the smelter, but kept bullion coins for sale to dealers and numismatists;

(3)          The actual gold composition of the material sent to the smelter cannot be established with sufficient certainty. The Appellant was paid by the smelter for 'fine gold' (99.999% pure) but the Appellant did not supply the smelter with fine gold, but with a mixture of fine gold, and gold of lesser or different purity - a lot of it 9 carat - all of which was smelted;

(4)          Mr Ragonesi sold to the smelter at 99.5% of fix, but this - as he accepts - was subject to deductions for unanticipated impurities and general melt loss. There would also have been variations (as additions) for impurities (sic) such as silver, platinum and palladium;

(5)          Some sales to the smelter were not at 99.5%, but (according to Mr Ragonesi, 'in the earlier days', but he could not say when those ended) were at 99.75%;

(6)          The smelter often paid 'around 97%' of what the gold should be worth, with a subsequent payment of around 2.5% when the gold was fully weighed and assayed;

(7)          The gold spot value changes twice a day (at 10.30 and 3.30, but only Monday-Friday), and no-one before us in a position, using spot-prices, to pin down with sufficient precision how much gold was actually passing through the Appellant's hands to the smelter;

(8)          The parties resorted to averages of spot prices of various kinds, but the averages serve simply to disguise the actual price fluctuations;

(9)          Even small adjustments to the spot price or purity of the gold can have large effects on value, and are themselves generative of swings in income.

261.     The Appellant revised its calculations to reflect the updated gold reconciliation exhibited to Mr Bosley's witness statement. It is said that the recorded gold shortfall should be 339,166.03g (that is, about 339 kg, or 757 lbs, or 1/3 imperial ton of 24 carat gold). Using the annual average price, that gives additional unrecorded sales of £8,811,533.

262.     We have also had regard to arguments about the anticipated profit margin and whether there can be a back-calculation. The Appellant's point is that, if HMRC's figures for APE 11 are right, then his profit margin would mean that he would only have been paying his customers 91-92% of the fix price (and thereby making a profit of up to 8.5%).

263.     This point is a fair one, but it only goes so far. We agree that it is indicative that something has gone non-trivially wrong with HMRC's figures in arriving at the assessment for APE 11 which is under appeal.

264.     In this regard, we accept the thrust of the evidence - not only from Mr Ragonesi, but also from fellow-traders such as Mr Blackburn, owing and operating similar businesses in the same period - that this was always a margin business, and that the achievable margins were slender. We accept Mr Blackburn's evidence was that he did not believe anyone would sell to the Appellant if it was offering less than 97% of the fix. We share his scepticism. But we cannot say with any more precision what the profit margin actually was.

265.     Mr Bosley undertook a 'gold reconciliation', and found that the Appellant had sold the smelter "significantly more gold than [the Appellant] was able to locate purchase invoices for". This was relied upon by Mr Bosley in his opinion that he was confident that the Appellant could not have achieved the margin that HMRC has applied in its assessments.

266.     We agree. Market forces would have destroyed the Appellant's business overnight if it had only been offering 91-92% of the fix. Sellers would have gone elsewhere to sell their gold. Trade sellers would only have had to take a short walk through the jewellery quarter to find someone else willing to give a better percentage. As Mr Blackburn said: "..trade sellers will always try to squeeze you. If I am offering £10.60 for 9ct gold, I will have trade sellers walking in and saying that they could get £10.62 up the road, will I match it?" 2p is a 1/530th (or about 0.2%) difference.

267.     People in the position of Mr Scadeng - one step below the Appellant in the chain, but themselves still buying from trade customers - were themselves (and self-evidently) buying for less than the 'approximately 97-98%'. Perhaps someone, somewhere, at the bottom of the chain was paying clients with very small quantities, selling gold though the post, without knowledge of the spot price, 91-92% of the fix (or, in the case of Mr Ragonesi's sisters-in-law in Telford, even less). But we are completely satisfied that this Appellant was paying more than that.

268.     Whilst the profit margins of the Appellant's competitors, being in the region of 1-2% have some relevance in illustrating the overall contours of the market, they are not determinative for the purposes of this appeal since (i) different considerations may well pertain to those other businesses; (ii) even those businesses show a significant variation in profit margin (from 0.6% to 1.8%); (iii) none of those businesses show a profit margin of dead on 2%.

269.     Ultimately, neither the argument, nor the evidence, allow us to go beyond our recognition that the Appellant's point, to a certain degree, is a sound one. It is a useful and illuminating exercise. However, it does not produce any sufficiently resilient outcome to establish the Appellant's actual figures in APE 11.

270.     Nor does the available evidence support the application of a notional profit margin (i) so as to back-calculate the amount of gold bought, and sold, and, from that (ii) to calculate the figures.

271.     There are several reasons for this. 97.5% may well have been 'the default price', but this is not a reliable or immutable measure applicable to all purchases:

(1)          The price could and did change according to the identity of the seller. For example, Mr Scadeng would 'usually' (but not always) receive (according to Mr Scadeng) 'approximately 97-98%', but (according to Mr Ragonesi) 98%, so the evidence in relation to at least one of the largest counterparties is insufficiently reliable;

(2)          Ms Heighway, as Mr Ragonesi's 'second-in-command' or 'right-hand-woman' was also authorised to give different prices. She generally bought at 98% of the fix, but a few of the larger customers got a little bit more;

(3)          Coins were bought at 99% - i.e., at a premium over scrap;

(4)          On some occasions during APE 11, the percentage had not been re-set on the computer;

(5)          The Appellant almost always bought at the morning spot price, but sometimes changed its price to the afternoon spot price if there was a significant decrease - more than 1% - in the price over the course of the day.

(6)          The Appellant's profit margins in any event had to absorb the costs of transporting gold to the smelter, and the transaction cost of withdrawing large quantities of bank-notes from the bank every day.

272.     Hence, we end up with 'around 97%', and 'around 2.5%' set out above. That is all too vague to translate across to transactions totalling in excess of £650million in the years APE 08-14 inclusive, and in excess of £100m per year for three of those years. 

The presumption of continuity

 

273.     The index year is APE 11, but HMRC seek to rely on it and apply it for the calculation of profits and further cash amounts alleged to have been taken in all the remaining years.  

274.     Officer Spaughton puts it in this way:

"Both Mr Cope and I took the view that the Appellant's record keeping was inadequate - not only for 2011, but also for earlier and later years. Where this is the case, HMRC presume that similar errors/omissions would have taken place in those years ('presumption of continuity')"

 

275.     HMRC relies on Jonas v Bamford (Inspector of Taxes) [1973] STC 519, where Walton J (in dismissing the taxpayer's appeal from the Special Commissioners) said (at p 540a-b):

"But, so far as the discovery point is concerned, once the inspector comes to the conclusion that, on the facts which he has discovered, Mr Jonas has additional income beyond that which he has so far declared to the Inspector, then the usual presumption of continuity will apply. The situation will be presumed to go on until there is some change in the situation, the onus of proof of which is clearly on the taxpayer' (emphasis added by us).

276.     As already explained, we have found there to be a corporation tax liability for APE 11.

277.     We consider that the presumption of continuity can, in principle, and subject to the discussion below about discovery, be relied upon for APE 10. That was the year in which the turnover increased from about £5m to about £50m. In our view, the detailed discussion and criticism of the Appellant's record-keeping, above, applies equally to APE 10. The effect of that is that HMRC's original calculation of the discrepancy falls to be re-adjusted to take account of our figures for APE 11, and then applied to APE 10.

278.     In relation to the earlier APEs, up to and including APE 09, we reject HMRC's case that the presumption of continuity operates so as to permit it to raise assessments by retrospectively applying the discrepancy percentage for APE 11:

(1)          The presumption is just a presumption, and can fall to be displaced by a taxpayer;

(2)          There is a clear difference - both quantitatively and qualitatively - between APE 09 and APE 10. In APE 09, the business was much smaller, and did not deal mainly in bullion, but was retail and the recycling of scrap gold jewellery to sell as jewellery.

(3)          The book-keeping procedure and Ms Mullinder personally were under much less pressure;

(4)          The only year which HMRC looked at in tremendous and forensic detail was APE 11;

(5)          HMRC had not performed the same exercise which it had undertaken for APE 11 for any of the earlier years, although it could have done. HMRC said that it had 'effectively absorbed' the inquiries for APE 07 and APE 08 into the APE 11;

(6)          APE 11 was the principal focus of HMRC's questions and the Appellant's answers.

279.     We are persuaded that the Appellant has shown us enough to displace the presumption for the years before APE 10.

280.     In relation to APE 12 and later years, we do not consider that the presumption of continuity could properly be relied upon as to impose liability on the Appellant.

281.     The presumption is just that - a presumption. It is a starting point, and not the end point. It can be displaced by sufficient evidence of material change in the Appellant's situation.

282.     We are satisfied that there was such material change. This is the result of a combination of different factors. Given the discussion and our findings of fact above, we can set these out relatively briefly. The first is the involvement of Mr Bosley. The second is the eventual ironing-out of the difficulties with the new computer system, and the adoption from January 2013 of a proper, orderly, cash reconciliation system. The third is the confirmation by Mr Bosley and Ms Mullinder that the records for APE 12 and 13 were intact, and that no significant adjustments were needed for either of those years.

The new Ground of Appeal: APE 12

 

283.     Given our finding that the presumption of continuity does not apply in relation to APE 12 so as to have permitted HMRC to have raised an assessment, it is not strictly necessary for us to express a concluded view on the new Ground of Appeal. However, and in case our conclusion on APE 12 should fall for later consideration, we should set out our reasoning, and explain why we would not allow the appeal on the new ground.

284.     We do not agree that the questions raised by HMRC regarding the return for APE 12 mean that an enquiry was opened and notified to the Appellant within the meaning of Schedule 18 Paragraph 24 of the Finance Act 1998.

285.     This turns on whether an officer of HMRC gave 'notice of enquiry' to the Appellant.

286.     There is no prescribed form of Notice of Enquiry.

287.     In Portland Gas Storage Ltd v HMRC [2014] UKUT 270 (TCC) the Upper Tribunal (Judges Herrington and Powell) were called upon to consider whether HMRC had opened an enquiry. The relevant provision in that case was contained in Schedule 10 of the Finance Act 2003, which dealt with stamp duty, but, like the applicable provision here, imposed no requirements of form. In the course of correspondence, an HMRC officer wrote that he was "seeking advice from our policy team regarding the time limit", and the Tribunal decided that was sufficient to open an enquiry. At Paragraph [48] the Tribunal said that a communication should be regarded as giving notice of an intention to enquire provided that the intended effect is reasonably ascertainable by the person to whom it is directed.

288.     Consistently with this, and in the context of an inquiry under section 9A of the Taxes Management Act 1970, the Upper Tribunal (Judges Bishopp and Brannan) considered this in HMRC v Mabbutt [2017] UKUT 289 (TCC). It was common ground that the notice did not have to observe any particular formality (see Para 44) and all that was required was a document in writing informing the taxpayer of HMRC's intention to open an enquiry into a particular tax return. The question whether the disputed notice sufficiently made the taxpayer aware of HMRC's intention to open an enquiry into a particular return was an objective one. The test was whether a reasonable taxpayer, in the circumstances of the taxpayer in question, would have understood that HMRC had intended to open an enquiry into a particular tax return: it was not a matter of the parties' intentions or actual knowledge: see Para [45].

289.      Here, the two particular documents relied upon by the Appellant are a letter from Mr Cope to Corrina Watson at Grant Thornton, dated 10 May 2013, and attaching an 'action plan' following a records review undertaken by Mr Cope at Grant Thornton's office on 25 February - 1 March 2013 and 3-5 April 2013. He identified 'lots of questions', and said that 'what now stands in isolation will no doubt 'fit together' once the further information is available. The landscape format 'Action Plan' at file 27 pages 505-512.

290.     Following the guidance in the cases, we do not read this document as a notice of enquiry. Although not determinative, it is not without significance that neither the word 'notice' or 'enquiry' is used. Nor was this the first letter. This letter was part of an iterative and collaborative process whereby information was being considered and exchanged by the parties.

291.     The second document is the document at file 27 page 557, which is part of a meeting note of a meeting held on 31 October 2013 at the offices of Grant Thornton. Again, applying the guidance above, we do not regard this as a notice of enquiry either, whether taken on its own, or in conjunction with the other document referred to.

292.     As such, in our view, HMRC's position in January 2016 was right. In the absence of an enquiry, it could not make Revenue amendments, and hence could only, in principle, absent any enquiry, issue a discovery assessment.

Discovery

 

293.     There are four discovery assessments: APE 09 (issued 20 January 2015); APE 10 (issued 22 January 2016); APE 12 (issued 22 January 2016); and APE 13 (issued 22 January 2016).

294.     The only basis upon which HMRC seek to assess for APE 09, 10, 12, and 13 is by virtue of the application of the presumption of continuity.

295.     For the reasons set out above, we have decided that the presumption of continuity is displaced for years before APE 10 and after APE 11. Therefore, it does not apply so as to permit HMRC to have assessed for APE 09, APE 12, or APE 13 and the assessments for those years must be set aside.

296.     However, HMRC is, in principle, entitled to assess using the presumption of continuity, for APE 10.

297.     But, it has done so on the basis of a discovery, and so is bound to satisfy us that it meets the statutory conditions to do so.

298.     The discovery assessment in this case is made under Paragraph 41 Schedule 18 of the Finance Act 1998. Where an officer discovers, as regards an accounting period of a company, that (a) an amount which ought to have been assessed to tax has not been assessed, or an assessment to tax is or has become insufficient, the officer may make an assessment in the amount or further amount which ought in their opinion to be charged in order to make good to the Crown the loss of tax. The power to make a discovery assessment under Paragraph 41 is only exercisable in the circumstances set out in Paragraphs 43 or 44 and subject to Paragraph 45.

299.     Paragraph 43 provides that a discovery assessment for an accounting period for which the company has delivered a company tax return - as the Appellant has done here - may only be made if the situation mentioned in Paragraph 41 was bought about carelessly or deliberately by (a) the company, or (b) a person acting on behalf of the company, or (c) a person who was a partner of the company at the relevant time.

300.     HMRC does not argue, in relation to any of the discovery assessments in issue in this appeal, that the situation mentioned in Paragraph 41 was brought about deliberately. Hence, we fall to consider carelessness.

301.     An assessment in a case involving a loss of tax brought about carelessly by the company (or a related person) may be made at any time not more than 6 years after the end of the accounting period to which it relates: Paragraph 46(2).

302.     HMRC bears the burden of showing that there was a discovery authorising it to assess for APE 10.

303.     For the reasons already explained, we are entirely satisfied that the assessment to tax in APE 10 was insufficient, and that HMRC was right in treating it as such.

304.     We are entirely satisfied that position was brought about by carelessness, being the Appellant's failure to take reasonable care in its record-keeping. The record-keeping was demonstrably inadequate for the business in APE 11, and was, for the same reasons, demonstrably inadequate for the business in APE 10. The inadequacy was completely obvious, at the time, and Mr Ragonesi either knew, or ought to have known, but was closing his eyes to the situation. Either situation suffices. In our view, the degree of care and skill applied by Mr Ragonesi to the discharge of his directorial and managerial responsibilities fell well below the standard to be reasonably expected. This was not malice; but it was certainly muddle.

305.     We have had regard to the argument that the Appellant was entitled to rely on the apparent competence of its accountants. As the First-tier Tribunal (Judge Cannan) remarked in Hanson v HMRC [2012] UKFTT 314 (TC) at [21]-[24]:

"21.   What is reasonable care in any particular case will depend on all the circumstances. In my view this will include the nature of the matters being dealt with in the return, the identity and experience of the agent, the experience of the taxpayer and the nature of the professional relationship between the taxpayer and the agent. In my view, if a taxpayer reasonably relies on a reputable accountant for advice in relation to the content of his tax return then he will not be liable to a penalty under Schedule 24.

 

22.    I am fortified in these conclusions in relation to paragraph 18 by the content of the HMRC Compliance Handbook at CH84540 which states in relation to paragraph 18 as follows:

 

         "A person cannot simply appoint an agent and deny responsibility for their tax affairs. The person still has a duty to take reasonable care, within their ability and competence, to make sure that what they are signing for is correct. The person has to show that they took reasonable care, within their ability and competence, to avoid default by their agent. This will include

 

·         making sure that they give the agent all relevant information        with which to work ...

·        implementing the professional advice received, and not       neglecting   some vital step

·        checking the agent's work to the extent that the person is able      to do           so.

 

For example, an ordinary person cannot be expected to challenge specialist professional advice on a complex legal point. But they ought to be able to recognise the complete absence of a major transaction.

 

A person saying and meaning ‘I leave it all to my agent' is hardly taking care, let alone reasonable care, over their obligations or the work of their agent.... The person has an obligation to choose an adviser who is trained and competent for the task in hand ...The benchmark is a person who goes to an apparently competent professional adviser

          ·          gives the adviser a full and accurate set of facts

          ·         checks the adviser's work or advice to the best of their ability               and competence and

          ·         adopts it.

The person will then have taken reasonable care to avoid inaccuracy on the part of themselves and their agent."

 

23.    At one extreme is an error of omission, for example failing to declare a source of income. In those circumstances it seems to me that a taxpayer will almost always be expected to identify the error. At the other extreme an error might involve wrongly construing a complex piece of legislation. In those circumstances the possibility of a penalty may still arise because of the carelessness of the agent, but the taxpayer's liability to a penalty might well be excluded on the basis that he took reasonable care but did not identify the error.

 

24.     I agree with the general thrust of the guidance given in the HMRC Compliance Handbook. In particular that a taxpayer cannot simply leave everything to his agent. A taxpayer must certainly satisfy himself that the agent has not made any obvious error. That might involve the taxpayer seeking to understand the basis upon which an entry on his return has been made by the agent. However in matters that would not be straightforward to a reasonable taxpayer and where advice from an agent has been sought which is ostensibly within the agent's area of competence, the taxpayer is entitled to rely upon that advice.  At the heart of this issue is the extent to which a taxpayer is required to satisfy himself that the advice he has received from a professional adviser is correct. The answer to that will depend on the particular circumstances of the case."

 

306.     Whilst that guidance is helpful, we do not consider that it really assists the Appellant in relation to APE 10. As Judge Cannan makes clear at the very beginning of that passage, the outcome ultimately depends on the circumstances of the case. Here, the circumstance which leaps out is that the Appellant - through Mr Ragonesi - should have been taking more care than it (he) was.

307.     Nothing which the Tribunal says in Hanson exempts a tax-payer from keeping full and accurate records, or from complying with the record-keeping requirements of Schedule 18.

308.     The instruction of accountants is not a cure-all or panacea. Tax returns prepared by an accountant can only ever be as good, full, and accurate, as the underlying prime records provided to her. The Appellant was wholly in charge of its own record-keeping. It was the Appellant's responsibility to keep full and accurate records, which in turn would have allowed its accountants to complete its tax returns fully and accurately. We are in no doubt, for the reasons set out in detail above, and applying to APE 10 as well, that the Appellant was failing to keep full and accurate records in APE 10, as a consequence of the 'gold rush' which the turnover figures show the Appellant to have been undergoing.

309.     The Appellant's subsidiary argument is that, even if a discovery assessment could have been made on 22 January 2016, it nonetheless was stale.

310.     One strand of the argument here is that for APE 09, a so-called 'protective' discovery assessment was issued on 20 January 2015, on the basis that the Appellant's lax approach to record-keeping made it impossible to argue that no additional tax was due. There then ensues a gap of (just over) one year between the issue of the discovery assessment for APE 09, and those for APEs 10, 12 and 13. But the January 2015 assessment was for a different year than that which we are considering, and by a different officer.

311.     The second strand is that the assessment in January 2016 was too late because HMRC had issued a Revenue amendment for APE 10 in July 2015, and had then realised - correctly - that there was no open enquiry, and hence had realised (in the internal departmental review) that the Revenue amendment could not be maintained.

312.     In Clive Beagles v HMRC [2018] UKUT 380 (TCC) the Upper Tribunal (Birss J and Judge Ashley Greenbank) engaged in a careful and comprehensive review of the existing reported decisions on the concept of 'staleness' in the context of discovery. The Upper Tribunal approved and followed the comments of Lord Glennie, also sitting in the Upper Tribunal, in Pattullo v HMRC [2016] STC 2043, especially at Paragraph [52] where he said (albeit in the context of TMA s 29) that there is 'a temporal aspect' to discovery.

313.     There, Lord Glennie rejected as absurd the idea 'that, having made a discovery [of the sort specified in s 29(1)] HMRC could in effect just sit on it and do nothing for a number of years before making an assessment ..." He also expressed a view that a delay of 18 months would have rendered the relevant discovery stale, "on any view" although he expressed that view in the context of a delay arising while waiting a decision of a relevant court or tribunal.

314.     Lord Glennie recognised the concession made by Counsel for the taxpayer in that case that the decision in each case will be fact sensitive, and he declined to try to define all the possible circumstances in which a discovery would lose its freshness and be incapable of being used to justify making an assessment. But Lord Glennie went on to add that it would only be in 'the most exceptional of cases' that inaction on the part of HMRC would result in the discovery losing its required newness by the time that an assessment was made.

315.     We agree. The present case is not the 'most exceptional of cases'. But, even if (looking to the remarks in Beagles at Paragraph [60]) it should eventually be held by the Court of Appeal or Supreme Court that a discovery is capable of becoming stale with the effect that an assessment raised on that discovery cannot be sustained (being a matter upon which we do not need to express any concluded view), we nonetheless find - temporally - that there is no operative 'staleness' or delay here in relation to APE 10 which would have defeated a discovery assessment even in January 2016. 

316.     The timeline makes this clear. Officer Spaughton took over the case from Officer Cope on or about 12 March 2015: see Officer Spaughton's letter at file 28 page 428 (which is the best evidence available, as contemporary documentation). 

317.     The original decision made by her is set out in HMRC's decision letter (signed by Mrs Spaughton) on 22 July 2015, although she had written on 27 April 2015 (that is to say, about six weeks after taking over) indicating her intention to issue (inter alia) a discovery assessment for APE 2010, but inviting any further evidence or new representations before it was issued. Such further representations were made in a lengthy letter dated 18 May 2015. Overall, we do not consider that the time elapsed from 12 March 2015 to 22 July 2015 (just over four months), or the fact or substance of the continuing discussion between those two dates, deprived any discovery in July 2015 of any requisite newness.

318.     This was a large and complex dispute. We accept Mrs Spaughton's evidence that she had no personal knowledge of anything before March 2015. Coming onto the scene, Mrs Spaughton had to be afforded a reasonable length of time in which to read and absorb and consider the mass of information which was then extant.

319.     Her decision letter of 22 July 2015 is confusing because, in relation to APE 2010, and consistently with her letter of 27 April 2015, it says that HMRC intended to issue 'a discovery assessment'. In fact, this is not what HMRC did in July 2015 for APE 10. It is common ground that HMRC issued a Revenue amendment in July 2015, and that it did not have the power to do so. The latter was not identified with any degree of clarity until the internal departmental review conducted by Officer Kennedy, the result of which was set out in his letter of 20 January 2016. He made it clear that although HMRC had intended in July 2015 to issue a discovery assessment for APE 10, HMRC was not in a position to do so since it had not given a closure notice for APE 10.

320.     The Appellant's argument here is that nothing relevant to HMRC's "discovery" (the inverted commas are the Appellant's) happened between 22 July 2015 and 22 January 2016, 'save that HMRC had attempted to give effect to their decision in an ineffective way'. The Appellant argues that the discovery thereby became stale in the sense identified in Pattullo v HMRC [2016] STC 2043 at [52].

321.     We disagree. We do not consider that HMRC's mistaken making of the Revenue amendment for APE 10 in mid July-2015 operated so as to bar HMRC from subsequently issuing the discovery assessment for APE 10 in January 2016. HMRC had not made a discovery and then 'sat on it' (to adopt the language of Lord Glennie). It was not that HMRC had done nothing. What happened in January 2016 was akin to the correction of an oversight. Whilst the April and July 2015 letters had referred to a discovery assessment for APE 10, HMRC had done something, albeit something different.

322.     Ultimately, the very best that could perhaps be contended for by the Appellant is that 10 months from March 2015 to January 2016 was too long, or that January 2016 was simply too late for a discovery assessment to be issued. But the discovery was later than March 2015. Even the discovery can be aligned to the letter of 27 April 2015, 22 January 2016 still did not render it stale. The decisions in the Upper Tribunal all make clear that each case turns on its own particular facts. On the facts of this case, we do not consider that 10 months was too long. This is always a matter of fact and degree, but it is still significantly less than the 18 months which Lord Glennie would have considered, 'on any view', too long.

323.     HMRC, through a different officer, Officer Cope, had already spent a long time trying to understand the sprawling mess of the Appellant's records for APE 11. Officer Spaughton was a new set of eyes. It took Peter Bosley months on end to try to get to the bottom of things. We do not see why HMRC and Officer Spaughton should have been in any different, or less advantageous, position. Moreover, HMRC had acted in the meanwhile, and, when it was identified, on review, that the Revenue assessment could not lawfully have been made for APE 10, the situation was rectified with considerable expedition thereafter.

The charges to tax under ICTA s 419 / CTA s 455

 

324.     ICTA 1988 s 419 deals with 'Loans to Participators' up to 1 April 2010. Where a close company, otherwise than in the ordinary course of a business carried out by it which includes the lending of money, makes any loan or advances any money to an individual who is a participator in the company, or an associate of a participator, there shall be due from the company, as if it were an amount of corporation tax chargeable on the company for the accounting period in which the loan or advance is made, an amount equal to [a specified percentage] of the amount of the loan or advance.

325.     Section 455 of CTA 2010 is in materially identical terms to ICTA 1988 s 419. It deals with 'Charge to tax in case of loan to participator' on or after 1 April 2010.

326.     Mr Ragonesi considers that HMRC has, 'in effect', accused him 'of stealing tens of millions of pounds from his company over a prolonged period.

327.     Given our conclusions on the assessments, and our rejection of all those except for APE 10 and 11, then we simply need to consider the assessments made under this head for APE 10 and APE 11.

328.     We are not satisfied that the charges to tax made in this regard in these periods should be upheld.

329.     This is a multi-factorial decision.

330.     Firstly, it is a relevant part of the background and context that it was not always HMRC's unequivocal position that the cash discrepancies identified must have been money taken from the business by Mr Ragonesi.

331.     Mr Cope's note of 7 May 2014 records his view, expressed at the meeting with Mr Ragonesi, that 'Frankly, dependent on the reasons emerging, there was potentially a considerable difference between the problems relating to record-keeping rather than the director taking large sums of cash from the business': emphasis added by us.

332.     This is consistent with Mr Cope's note of his phone call to Mr Ragonesi's then representatives (Grant Thornton) on 29 April 2014.

333.     Mr Cope was the individual officer at HMRC most closely linked with the Appellant, at the time. We did not hear evidence from him. But his correspondence is thoughtful, and reflective of the difficulties which he - and we - have had to struggle with in seeking to ascertain, doing the best that we can, the true position for APE 11.

334.     His thinking cannot simply be brushed aside as of historic interest only. In mid 2014 he tended towards the fact that at least some of the serious cash discrepancies identified earlier were just down to 'inadequate record keeping'.

335.     Mr Cope's thinking at that time is intuitively attractive because the assessments, taken together, otherwise can only be read as Mr Ragonesi having taken the best part of £40m in cash and gold from the business over the years in question.

336.     This has to be tested against reality. Part of the function of this Tribunal is to hear evidence from the participants, and to find facts.

337.     We are in the unique position of having the opportunity of having heard Mr Ragonesi give evidence. He gave evidence for more than a day. We heard him robustly (but fairly) cross-examined. He was conspicuously anxious to deal with the matter of the alleged cash removals head-on. He asked us, with some asperity, why, if HMRC considered he had genuinely removed this gold and money, why the Police had not been involved. He also asked us (not wholly rhetorically) why, if this was HMRC's case, the authorities were not digging up his back garden. There was and is force in his points.

338.     Mr Ragonesi repeatedly and firmly denied in his oral evidence - consistently with what he is recorded as having said to HMRC for several years - that he did not take any money or gold from this business. His denials were unequivocal.

339.     We believed him.

340.     We are not simply taking Mr Ragonesi's denial and say-so as the 'be-all-and-end-all'. Looking at this realistically and objectively, as we must, then, unless Mr Ragonesi is a deceitful and manipulative individual of a high order (which, to avoid any doubt, we do not consider he is) then this huge amount of money and/or gold would, in all likelihood, have left some trace, somewhere. But there is none.

341.     It cannot be ignored that, if Mr Ragonesi was indeed stripping tens of millions of pounds of cash and gold out of this business, then he was doing so against a backdrop of the missing days for APE 11. The 'missing' days stand out like a sore thumb. The absence of records inevitably excited the suspicion of the Revenue. But, the missing days are plainly inconsistent with Mr Ragonesi executing a cunning plan to surreptitiously 'siphon' money off from the Appellant company. To have done so against the backdrop of an open enquiry (albeit into an earlier year) and intense Revenue scrutiny of the books would have been such a blatant, bare-faced, strategy that it just makes no sense.

342.     The tax affairs of Mr and Mrs Ragonesi were subject to an in-depth private side review by a team at a firm of reputable accountants who did not find anything wrong. Mrs Ragonesi's evidence dealt with this point. Her evidence was:

"I am not aware of and do not believe that my husband has taken any undeclared sums of money from Stirling. I confirm that I have not taken any undeclared sums of money from Stirling."

 

343.     This evidence was not challenged.

344.     Ms Sharna Ragonesi made a witness statement with materially identical content. Her evidence was not challenged.

345.     HMRC has extensive investigatory powers, but was not able to point to anything awry in terms of unexplained wealth, assets, or economic activity. The Ragonesis admittedly enjoy what many would regard as a very comfortable lifestyle, but this is consistent with the success of this business, and the hard work that Mr Ragonesi in particular has put into it.

346.     We also take account of the surrounding circumstances. These are corroborative of the position that Mr Ragonesi did not have access to a hoard of hidden wealth. Mr Ragonesi was interested in buying another jewellery business in late 2011 for £1m, but his evidence - which we accept - was that he did not have the cash to do so, and was struggling to raise finance. Mr Bosley was aware of this, and confirmed it. It is not in dispute that Mr Ragonesi was running around Birmingham, day-in and day-out, desperately trying to get money from banks. The stress and pressure which this placed on Mr Ragonesi (and those working with, and for him) simply cannot be intelligibly reconciled with Mr Ragonesi, at the same time, squirreling away cash and gold.

347.     There was no evidence at all that any of the employees or other witnesses had ever seen Mr Ragonesi wrongfully remove or conceal any money or gold - let alone tens of millions of pounds in cash and/or gold. Mr Ragonesi was often away from the business premises. When he was there, he was seldom on his own. Ms Heighway helped him with cashing-up. If Mr Ragonesi was taking money and gold, she would doubtless have known. However, we do not consider that she would have been prepared to jeopardise her liberty by coming to the Tribunal to lie about it. We do not consider that she was lying.

348.     Mr Bosley - who we considered to be a person of scrupulous integrity - was deeply engaged in the later years in the financial affairs of the business. He had not detected any cause for concern either, apart from the obvious disorder in the book-keeping arising from the system which we have described above. Mr Bosley is a former partner in the well-reputed and local firm of Eden Currie, who (since about July 2014) have been the Appellant's accountants.

349.     We have no doubt that had Mr Bosley suspected that anything even remotely underhand was going on, that he would himself have declined to continue to act, and certainly would not have allowed his former firm to have become involved with the Appellant.

350.     We also assessed Ms Mullinder as open and honest. We have no doubt that she would have retreated from her involvement with this business had she thought that the discrepancy which she had identified was down to Mr Ragonesi taking money.

351.     If Mr Ragonesi was indeed removing cash and gold, then he was either doing so under the very noses of his family, employees, Ms Mullinder, and Mr Bosley without anyone noticing, or someone knew about it, but was not only turning a blind eye, but was prepared to come to the Tribunal to lie about it.

352.     We simply do not consider either scenario as plausible, or consistent with the evidence which we heard.

353.     Various criticisms are advanced by Mr Chacko as to whether HMRC were even entitled to pursue these assessments in the absence of any square-on challenge to Mr Ragonesi that he had been taking enormous amounts of money and gold which he knew to be taxable, with the intent of not paying tax. Ultimately, and taking into account our findings, including as to the credibility of Mr Ragonesi's evidence, we do not need to address these in detail.

354.     The fact that Ms Choudhury did not expressly challenge Mr Ragonesi as a thief, or a fraudster, or even as dishonest is understandable to us. Perhaps unusually, both members of this panel of the Tribunal are also members of the Bar and understand the professional obligations imposed on Counsel in Ms Choudhury's position.

355.     Nonetheless, the omission to positively put such a case does serve to expose the ambiguity and weakness at the heart of HMRC's case in this regard. HMRC's argument is that the money (i.e., the under-declared income and profit) must - as a matter of logic - have gone somewhere.

356.     To avoid any doubt, we consider that argument to be misconceived. We do not consider the fact (as we have found for APE 11 and APE 10) that there are under-declared sums liable for corporation tax inexorably supports a conclusion that money was loaned or advanced by the Appellant to Mr Ragonesi so as to generate a charge to tax.

357.     The response to this is simple. The statute is not a useful 'fall-back' which engages where a close company is found to have made an under-declaration. The statutes are clear that an assessment can only be raised where a close company can be shown to have loaned or advanced any money to an individual.

358.     We are not satisfied that the Appellant company did loan or advance any money to Mr Ragonesi.

359.     On the contrary, we are satisfied that the Appellant has shown us, on the evidence, not only to the appropriate civil standard, but to the much higher standard of beyond reasonable doubt, that it did not loan or advance any money to Mr Ragonesi.

Outcome of the Appeal

 

360.     The assessment of additional corporation tax due on profit for APE 11 is confirmed in principle, but is to be re-calculated in accordance with our remarks above.

361.     The assessment of additional corporation tax due on profit for APE 10 is confirmed in principle, but is to be re-calculated.

362.     The Appeal against all the other assessments and amendments is allowed.

A postscript - Alternative Dispute Resolution

 

363.     The Appellant's cover letter to its Notice of Appeal says that an application for ADR had been made. That was very sensible. If ever there was a dispute which cried out for a determined attempt at Alternative Dispute Resolution, it was this one.

364.     This appeal has been allowed in part, and dismissed in part. Put bluntly, both parties' cases as advanced and pursued before the Tribunal were either factually or legally wrong in significant respects. The Appellant still faces a significant tax bill, albeit this will end up being somewhat less than the amounts which were assessed and subject to the appeal.

365.     If this appeal and our resolution of it - which has been laborious - has shown anything, the process of ADR may well have been challenging. Both parties would have had to be prepared - in a confidential and without prejudice environment - to reflect on the genuine strengths and weaknesses of their respective positions.

366.     At the very least - even if a negotiated settlement, giving appropriate regard to the strengths and weaknesses of the parties' cases, could not have been reached - there was nonetheless still ample scope for the parties to have engaged co-operatively so as to properly identify and narrow the issues genuinely in dispute.

367.     At the end of the day, we cannot adjudicate on the ADR process, although in this case we happen to know quite a lot about it because the parties - each legally advised and represented - chose to place a significant amount of correspondence which pertained to settlement discussions before us.

368.     Posturing and skirmishing in correspondence ultimately proved to be non-progressive, and there is an impression (we cannot put it higher than that) that the tone of the correspondence ended up driving the parties into polarised, and entrenched, positions.

369.     Finally, we wish to acknowledge and commend the assistance given to us by Mr Chacko and Ms Choudhury.

370.     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.

 

DR CHRISTOPHER MCNALL

TRIBUNAL JUDGE

 

RELEASE DATE: 22 JANUARY 2019