Peppermint Foods Limited v The Commissioners for HMRC

This is a VAT case. The appellant (or “the company”) benefits from a Subway franchise which it exploits via two outlets one of which is located in Lakeside Retail Park, in West Thurrock, Essex, and the other is located at Lakeside Shopping Centre, in Thurrock, Essex. The appellant sells hot toasted sandwiches (“hot food”) and cold food and drink (together “cold food”) from those outlets. Hot food and cold food which is eaten in an outlet is standard rated for VAT purposes, as, too, is hot takeaway food. Broadly speaking, cold takeaway food other than confectionary is zero rated. It is HMRC’s view that the appellant’s staff have, at the point of sale, incorrectly rung hot takeaway food into the till (which should have been standard rated) as cold takeaway food (which was thus zero rated).

HMRC have, accordingly, assessed the appellant to additional VAT for the periods 05/15 to 02/19, in an initial amount of £214,854. Following an internal review by the assessing officer, and a subsequent statutory review, this was reduced to £144,383, which is the amount at stake in this appeal.

There was no dispute between the parties concerning the relevant law which we summarise below.

By virtue of section 73(1) Value Added Tax Act 1994 (“VAT Act”), where it appears to HMRC that tax returns made by a taxpayer are incomplete or incorrect, HMRC may assess the amount of VAT due from him to the best of their judgment and notify it to him.

Under section 73(6) VAT Act an assessment must be made not later than either 2 years after the end of the prescribed accounting period or 1 year after evidence of facts, sufficient in the opinion of the Commissioners to justify the making of the assessment, comes to their knowledge (whichever is the later).

This is, however, subject to the general rule that the time limit for making an assessment is capped at 4 years after the end of the relevant accounting period. This is found in section 7(1) VAT Act 1994.

Section 83 VAT Act provides:

There is then set out a series of actions, decisions, and other matters arising under the Act listed under paragraphs (a) to (z). Paragraph (p) is as follows:

In Van Boeckel v Customs and Excise Commissioners [1981] AER 505 (“Van Boeckel”) the High Court (Woolf J as he then was) considered the application of best judgment.

In the Court of Appeal decision of Customs & Excise Commissioners v Pegasus Birds Ltd [2004] EWCA Civ 1015, that court approved the approach of Woolf J. It went on to add that the tribunal’s primary task is to find the correct amount of tax on the basis of the material before it and in all but very exceptional cases this should be the focus of the hearing; any mistake which I consider that HMRC has made in its assessment may still be to best judgment if it is consistent with an honest and genuine attempt to make a reasoned assessment of the VAT payable; and an assessment which appears to be unreasonable or wholly unreasonable may still be the result of an honest and genuine attempt to assess the VAT properly due.

As regards time limits, these were considered in the High Court decision of Pegasus Birds ([1999] BVC 56) (“Pegasus”) in which Mr Justice Dyson set out and approved the legal principles to be applied as follows:

Generally, the burden lies on the taxpayer to establish the correct amount of tax due:

Group 1 of Schedule 8 to VAT Act 1994 provides for zero rating of food of a kind used for human consumption but excludes from zero rating any supply of such food for consumption on the premises on which it is supplied, and any supply of hot food for consumption off those premises.

We were provided with two bundles of documents which included some authorities. The assessing officer, Officer Pradeep Vaghela (“Officer Vaghela” or the “assessing officer”) gave oral evidence for HMRC. Oral evidence for the appellant was given by its director, Mr Mohammed Adjmal (“Mr Adjmal”). From this evidence we find the following as facts:

The appellant benefits from a Subway franchise which it exploits via two outlets one of which is located in Lakeside Retail Park, in West Thurrock, Essex, and the other is located at Lakeside Shopping Centre, in Thurrock, Essex. The appellant sells hot food and cold food from those outlets.

The company first acquired and operated a Subway franchise on and from 1 March 2011. It was told at that stage that there was a Subway case before the courts in which Subway were challenging HMRC’s position regarding hot takeaway food. Some franchisees were, therefore, not paying VAT on hot takeaway food. Other franchisees, including the appellant, did pay VAT on hot takeaway food on the basis that if the case turned out favourably, they would recover the overpaid VAT.

In 2012 the company opened a new outlet, in the shopping centre, and refitted that outlet. Since it was, therefore, seeking to recover a great deal of input VAT incurred on the construction costs, it was subject to a VAT inspection. No VAT discrepancies were identified.

In 2016 the company opened a second outlet in the retail park. Again, it was subject to a VAT inspection given it was seeking to recover input tax on construction costs. HMRC inspected the company’s receipts and it was discovered that only one was incorrect.

It was Mr Adjmal’s evidence that there was a further investigation in October 2017, but we were told that HMRC had interrogated its database and could find no record of that. However, those records show that there was a visit in October 2016. It was Mr Adjmal’s evidence, and we find as a fact, that at that visit (he thought there were visits in both October 2016 and October 2017) he had provided a basket of Z readings to the investigating officers and told them to select the readings they wanted. His further evidence was that they did not take the Z readings away, but made notes of those readings, and went on to tell him that they would be in touch if there were any VAT abnormalities but he heard nothing. The appellant therefore continued to operate as it had done to that date. HMRC’s records then go on to show that “this should have been tested by carrying out an invigilation but this was not done due to pressure of work”.

Before taking over its first Subway franchise in 2011, Mr Adjmal had no experience in selling food. He is, by training, an electrical engineer who moved into facilities management. Apart from his first 6 months operating his first franchise, he is not located full-time in the outlets. He works from an office which is located at neither outlet. However, he is always available to the staff, and lives sufficiently closely to the outlets to be able to attend them at short notice if a problem arises which warrants such attendance.

Whilst there is dedicated seating for the appellant’s customers on the mezzanine floor of the the outlet in the shopping centre, there is no such dedicated seating at the outlet in the retail park.

Officer Vaghela has considerable experience in investigating the VAT position of Subway franchises. Having been assigned to the appellant’s case, he checked HMRC’s systems and discovered that the average standard rated sales for the preceding four years was 58% which, in his experience, seemed low. He therefore decided to carry out test purchases, with a colleague, which he carried out on 20 June 2018 and 29 June 2018. These test purchases were carried out on each day, at both outlets, and his evidence, therefore, was that they carried out eight test eats, four on 20 June 2018 and four on 29 June 2018. Whilst we had direct evidence of the test eats on 20 June 2018, as we were shown the till receipts, we had no such direct evidence of the eats on 29 June 2018. Officer Vaghela’s evidence, however, which was largely unchallenged, was that it was “probably a similar story”.

That story runs thus: at the shopping centre outlet, the till receipt for Officer Vaghela’s purchase was correct. VAT of £1.10 was paid on a purchase of £6.60. However, the till receipt for the purchase made by his colleague was incorrect in that it identifies the purchase as “(eat in)” whereas in fact it was takeaway. However, the VAT, charged was correct.

The officers then attended the outlet at the retail park. They purchased a 12 inch “tuna sub” hot takeaway which, as the sale price was £5.70, should have borne VAT of 95p, but only 43p of VAT was charged. This test purchase also included a bottled juice on which VAT should have been charged, and was, and also three cookies which should have been zero rated but on which VAT was charged at the standard rate. The 43p of VAT represents one sixth of the price charged for the bottled juice and the cookies (£2.60). The same errors were made in respect of the second test eat of a 12 inch “chicken strips sub”. This too was priced at £5.70 but VAT of only 29p was charged. There appears to be VAT on the bottled water and a cookie also purchased, and from which it appears that VAT was correctly charged on the water but incorrectly charged on the cookie. We were, incidentally, told by Officer Vaghela that, at that time, tills provided by Subway to its franchisees were incorrectly categorising the sale of takeaway cookies as standard rated when they should have been zero rated, and that it was a problem across the sector.

Officer Vaghela then contacted Mr Ajmal and visited the appellant’s office premises on 28 August 2018. He shared the receipts from the test purchases with Mr Ajmal who explained that there had been some issues with the till which he had sought to rectify and gave him Z reports for August 2018 for both outlets. There was some evidential confusion about this process, given that, in the bundle, there was an analysis of Z readings relating to an associated, but entirely separate franchisee. However, we are happy that on the balance of probabilities, the appellant provided Z readings for both outlets for the vast majority of August 2018, as asserted by the assessing officer.

Officer Vaghela then analysed the information from those Z readings for each outlet individually, and also on a combined basis. That analysis runs from Wednesday, 1 August 2018 through to Sunday, 26 August 2018 and so includes both weekday and weekend sales. The Z readings provided the gross daily sales and the total VAT charged thereon. From that information Officer Vaghela was able to work out the proportion of standard rated to zero rated sales, and the overall rate of VAT charged on the overall sales. His analysis suggests that the standard rated sales for the outlet in the retail park over that period varied from a low of 71% to a high of 94%, with an unweighted average of 84%; whilst that for the shopping centre varied between a low of 56% to a high of 87%, an unweighted average of 70% When averaged, this gave an unweighted average proportion of standard rated sales of 78%.

Officer Vaghela then conducted a further analysis to generate the potential VAT assessment for the periods 8/14 to 05/18. He took the standard rated sales percentage for each of those periods, applied the average rate of 78% to each of those periods, calculated the VAT payable at that rate, and then gave credit for the VAT paid for each of those periods. This resulted in a potential VAT assessment of £127,334, which he sent to the appellant’s agent with a covering email on 4 September 2018. In that email he went on to say that he would like to carry out invigilation exercises at both outlets on two dates. The appellant’s agent responded stating they disagreed with the figures but were happy for HMRC to carry out the invigilation exercises. These were undertaken on 30 October 2018 and 11 January 2019.

On the first of those dates, the invigilation exercises involved HMRC officers attending both outlets in shifts throughout opening hours and observing the way in which sales were made, and the way in which those sales were rung into the tills. At the retail park outlet, the entries started at 7.14 in the morning, and the last entry is at 20. 29 in the evening. In the shopping centre outlet, the first entry is at 9.57 in the morning and the last is at 21.45 in the evening. On the second of those dates, an invigilation exercise was carried out only at the outlet at the retail park, and only between 11.38 and 14.40, and then again between 17.45 and 20.24.

Officer Vaghela told the appellant’s agent, by email of 19 February 2019 that he had carried out these invigilation exercises and wished to discuss the results. An appointment was made for a meeting on 18 March 2019, and at the meeting, the assessing officer told the appellant’s agent that the invigilation exercises had shown that standard rated sales were in fact higher than those shown by the Z readings. On 2 May 2019, the officer sent to the appellant’s agent and to Mr Ajmal, documents showing the results of the invigilation exercises. In summary, these showed that for the invigilation on 30 October 2018 at the retail park outlet, the declared percentage of standard rated sales was 95.59%, but “corrected” (in other words, what should have been declared) was 94.14%. In other words, more standard rated sales, and thus more output VAT, had been declared than was in fact due. The same was true following the invigilation at the shopping centre outlet where the declared percentage of standard rated sales was 95.6% but corrected it was 94.08%, and that for the 11 January 2019 invigilation at the retail park outlet, where the declared percentage was 93.59% whilst the corrected percentage was 91.85%.

When averaged out, however, this was a standard rated percentage of 94%, which, rounded, resulted in a preliminary assessment amount of £214,854. The appellant’s agent, in an email of 16 May 2019, indicated that there were a large number of issues in relation to the findings and sought confirmation that assessments would be raised. On 21 May 2019 the appellant was issued with a pre-assessment letter and accompanying schedule which showed the VAT due in each period. Those periods were 05/15 to 02/19.

In an email dated 6 June 2019 to the assessing officer, the appellant’s agent identified a number of factors and conditions which needed to be taken into account when considering the results of the invigilation exercises. These included seasonal factors and the time of year on which the invigilation exercises took place; the susceptibility to staff errors which could not be foreseen or accurately managed; that significant construction work had taken place during that time; that there were significant and disruptive software and IT upgrades which had taken place at that time, and the respective locations of the outlets.

In an email of 18 June 2019, Officer Vaghela responded to these points. He could not think of any factors that would affect the appellant’s normal trading pattern and having analysed the appellant’s VAT returns had found no correlation to those trading patterns and the time of year. In his experience, dealing with many Subway businesses, that whilst there might be seasonal variations in turnover, the ratio of standard rated and zero-rated sales remains constant. Whilst it was understandable that staff errors were made, it is up to the manager and the business owners to ensure that their staff are trained on how to use the till correctly. The significant construction work might have an effect on turnover but would not affect the ratio of standard rated to zero rated supplies. Although he understood that there were software changes, he had spoken to Mr Ajmal about the situation and he had looked at things in early August. The Z reports had been provided for that period, but at the time of the invigilations, there were no software issues. His analysis of the cold takeaway transactions was that the average was 4.31% which would make the standard rated sales 95.69%, but the invigilation showed 94%. He thus saw no reason to change his view.

On 18 July 2019 the appellant was issued with a notice of VAT assessment and accompanying schedule for £214,854.

Following further correspondence with the appellant’s agent, the assessing officer reviewed the assessment and his evidence was that given that the Z reports were for August, whilst the invigilation exercises had been carried out in October and January, there was some justification for reducing the assessment and basing it on the difference between the invigilation percentage (94%) and the August Z readings percentage (78%). On 20 November 2019 he sent an email to the appellant’s agent stating this and indicating that he intended to use 86% and enclosed a revised schedule setting out the impact on the VAT due for the periods, which he calculated as being £158,280. The assessing officer’s evidence, which was not challenged and which we accept, was that the 86% figure was a rough and ready attempt to take into account the seasonality difference between August, October and January, and the impact that would have on hot food sales.

Following further correspondence between the parties, in which the appellant’s agent expressed disagreement with these revised figures, that agent sought a statutory review of HMRC’s position which review conclusion was set out in the review conclusion letter of 27 February 2020. The reviewing officer reduced the assessment to £144,383 based on the revised schedule of £158,289, further reduced to take into account that period 05/15 was out of time.

The appellant appealed that decision to the Tribunal, which, given that there were certain issues regarding hardship, was made slightly out of time. HMRC take no point on this and nor do we.

In his oral evidence, Mr Ajmal explained that: he had not manipulated the figures; the tills were supplied by Subway and he had thus assumed that they would correctly record standard and zero rated sales; there had been times during the periods under assessment when the ovens did not work and accordingly they were not able to sell hot food; he accounted for VAT in accordance with what the till records showed; he did not realise that he had to challenge, quantitatively, HMRC’s figures; had he realised that, he would have done so; he had not undertaken a detailed analysis of his standard rated sales based on his till records and information that was available from the till software.

The burden of establishing that a valid assessment has been visited on the appellant lies with HMRC and the standard of proof is the balance of probabilities. If HMRC satisfy us that the assessment is valid, then the burden of demonstrating that it is incorrect lies with the appellant. Again, the standard of proof is the balance of probabilities.

The appellant submits as follows:

The invigilation exercises which were carried out, effectively over two and a half days are wholly unrepresentative of the overall period assessed. It does not take into account the matters raised in correspondence including seasonal variations, construction workers, and store location.

HMRC should have undertaken a year of daily invigilations. This would take into account the change in the seasons and thus customer’s buying patterns, as well as store locations and seating arrangements which have not been considered by HMRC.

HMRC have not allowed for or given any consideration for staff errors nor for the times during which the ovens were unavailable. Nor have they given any consideration to IT errors.

It is not fair that having undertaken detailed investigations in 2016 and 2017, and having given, in effect, the appellant a clean bill of health by not saying that anything was wrong, they should now be allowed to go back before those periods and assess the appellant. If anomalies had been identified then, then they would have been corrected.

It did not provide any quantitative evidence to refute HMRC’s figures because it did not realise that it needed to do so.

HMRC submit as follows:

The assessment was made to best judgment. It was based on test eats which took place in June 2018 which showed till inputting errors where hot takeaway food was wrongly rung into the till as zero rated. It was also based on the August 2018 Z readings, and then the two invigilation exercises which took place in October 2018 and January 2019.

There was therefore ample information on which the assessing officer could base his assessment. He could have used just the Z readings but gave the appellant the opportunity to agree to an invigilation exercise. Following that exercise he shared the results with the appellant.

The August 2018 Z readings showed a standard rated percentage of sales of, on average, 78%. The invigilation exercise showed an average standard rated percentage of sales of 94%. The assessing officer recognised that this might be high given that the exercise had been carried out in the autumn and the winter and that this might reflect a higher proportion of hot food and had mitigated downwards to 86% to take into account the August 2018 Z readings. It might be expected in that month that there would be a higher proportion of cold takeaway food.

The appellant was given ample opportunity to support its qualitative criticisms of the invigilation results with quantitative alternative figures. But failed to do so. In these circumstances HMRC cannot alter the figures since they have no alternative numerical data on which to base such alteration.

Our first consideration is to establish whether the assessment has been made to best judgment. We have set out the principles against which we need to test the facts, above. The bar for making a best judgment assessment is a low one. The cases are clear that there must be some basis for the assessment which cannot be arbitrary or capricious or irrational. But provided HMRC fairly consider the material before them and come to a decision based on that material which is reasonable and not arbitrary, and that decision is a result of an honest and genuine attempt to assess the VAT properly due, then an assessment which reflects that decision has been made to best judgment. In our view Officer Vaghela has made his assessment in accordance with those principles.

Officer Vaghela took into account four things when making his assessment. The first, as an officer experienced in the ways of Subway outlets, that the four year average of 58% standard rated sales, was low. Secondly some of the test eats had shown that hot takeaway food had been incorrectly rung into the till as cold takeaway food and that as a consequence no VAT had been charged on those sales which should have been charged at standard rate. Thirdly the August 2018 Z readings which showed an unweighted average of standard rated sales across both outlets of 78% (albeit that those percentages fluctuated at each outlet over that month). Finally, the October 2018 and January 2019 invigilation exercises. His numerical analysis of the Z readings and information obtained from those invigilation exercises were not seriously challenged by the appellant.

Once he had carried out the test eats, he shared his concern with the appellant. This was best practice. He then obtained the Z readings for August 2018, analysed them with commendable alacrity and shared his analysis with the appellant. He was intending to assess on the basis of those readings, which was on the basis of a 78% average of standard rated sales. The appellant was not prepared to agree the figures, so the invigilation exercises were instigated. These took place in October 2018 and January 2019. We accept that these are colder months and therefore it is likely that in that period, there might have been a higher proportion of hot food, but the evidence from those exercises, that the proportion was 94%, was then shared with the appellant whose views were sought. The appellant’s agent expressed dissatisfaction with the figures and explained that there were a number of factors which the assessing officer seemed to have failed to take into account, including seasonality, staff errors, construction work and software and IT issues. These were all addressed by the assessing officer in his email of 18 June 2019. He therefore assessed on the basis of 94% and provided his computations to the appellant and its agent. Thereafter the assessing officer reviewed his assessment to take into account the seasonality impact on hot food sales, and thus the difference between the 78% average derived from the August 2018 Z readings and that 94% average, thus bringing the latter down to 86%.

In our view this is a very long way from an arbitrary, capricious, or irrational process. It is evidence-based, takes into account matters raised by the appellant and its agent, reflects an ongoing dialogue with the appellant and its agent, and evidences a justification for, initially, assessing on the basis of 94%, and subsequently reducing it to 86%. Whilst that might be a fairly broad-brush approach, it shows that the assessing officer was prepared to take into account the impact of what might have been higher sales of cold food during the hotter month of August.

It is our view that the assessment was made to best judgment and was a valid assessment.

The appellant’s position was that the assessment was excessive, even if it had been made to best judgment in the first place. But as can be seen from the legal principle set out above, once an assessment has been made to best judgment, it is up to that taxpayer to show, on the balance of probabilities, that it is wrong and what is the more likely amount to which it should be assessed. It is not, at that stage, for HMRC to conduct a year-long invigilation exercise. It is for the appellant, having been apprised of the amount which HMRC intend to, and then go on to, assess, to show that that amount is excessive. As Ms Donovan put to Mr Adjmal, it was clear from the figures that had been provided by HMRC that there was a considerable amount at stake and that it was open to him to put forward alternative figures which HMRC would have considered to justify the qualitative criticisms of the original assessments, namely seasonality, failing to take into account the situation of the outlets, IT issues etc. But the appellant did not do this. So, Ms Donovan asked rhetorically, on what basis could HMRC have amended their assessment? We are sympathetic with this position.

Mrs Neill who has considerable professional experience in this area, enquired of Mr Adjmal whether he was aware of the capability of the tills to provide relevant and detailed information of the nature of the supplies which passed through them. Mr Adjmal did not seriously contest that the tills were capable of providing such details. And so, in our view, it has been open to him to provide considerable data to HMRC to refute HMRC’s figures. Furthermore, it would have been possible to have procured data for several days in and around the dates of the invigilation exercises to bolster the criticism that the exercises, having taken place only over one or two days, were unrepresentative. However, the appellant did not do this, and that leaves us in the difficult position of having little evidence on which to base a decision that the assessment is excessive.

The appellant referred to the First-tier Tribunal decision of Golden Cube Ltd [2018] UKFTT 0488. That, too, was a case involving a Subway franchisee. HMRC based its assessment on three days of invigilation, all of which were weekdays, and all of which took place between 12 and 21 October 2015. The assessment was challenged. Judge Cannan reviewed the three reasons why the proportions of standard rated sales evidenced on the invigilation days were considerably higher (82%) than the proportion set out in previous VAT returns (44% – 65%). These reasons were deliberate manipulation of the records, mistakes either in the till entries or in the till programming, or that the three days of invigilation were not representative. There was no evidence of manipulation nor of mistakes of till inputting or programming, and the Judge was satisfied on the evidence that the most likely explanation for the different proportions of standard rated sales was that the invigilations were not representative of the VAT periods to which they were compared.

But that is very different from the position in this appeal. Here, the test eats clearly showed mistakes of till inputting. The invigilations took place in October and the following January, not all in the same month. And the assessment was based not just on the invigilation exercise but also on the August 2018 Z readings. The evidence shows that mistakes had been made. Quite rightly, there is no suggestion that the appellant has manipulated the records or suppressed its takings. The allegation is that hot takeaway food has been rung into the till as cold takeaway food and thus supplies which should have been charged VAT at standard rate have been subject to VAT at the zero rate.

HMRC’s evolving assessment, based finally on a proportion of standard rated sales of 86% cried out for contrary numerical evidence to be supplied by the appellant to demonstrate what, in its opinion, the correct proportion was. The appellant failed to supply this evidence to HMRC and failed to supply it to ourselves. This is understandable given that the appellant did not fully understand that the ball was in its court to do so. But it does mean that, on the balance of probabilities, we cannot say that there is a flaw in the assessment, nor what the more likely proportion of standard rated sales was during the period covered by the assessment.

It is our view, therefore that the assessment stands good.

Under section 73 (6) VAT Act 1994, an assessment must be made within the time limits provided for in section 77 VAT Act 1994 and shall not be made after the later of (a) 2 years after the end of the prescribed accounting period, or (b) one year after evidence of facts, sufficient in the opinion of the Commissioners to justify the making of the assessment, comes to their knowledge.

In this case, the assessment was issued to the appellant on 18 July 2019. The period in section 77 is 4 years, and this is the reason why, on review, the assessment was reduced to take out of account the period 05/15 which was out of time.

Furthermore, the assessment was made within the one-year time period that the evidence of facts sufficient to justify the making of the assessment was made within one year of both the receipt of the August Z readings and the invigilation exercises on 30 October 2018 and 11 January 2019. They were, however, issued outside the one-year period following the test eats.

As can be seen from the evidence, HMRC undertook a visit to the appellant in October 2016 and were provided with Z readings. They did not uplift these, but made notes of them, and, we surmise, took those notes away and analysed them. At least they did enough for them to be able to make a note on the HMRC file that “this should have been tested by carrying out an invigilation but this was not done due to pressure of work”. This strongly suggests to us that some analysis had been undertaken of the Z readings which resulted in a preliminary conclusion. And it is that preliminary conclusion which should have been tested by an invigilation exercise. This follows very closely the actual process undertaken by Officer Vaghela who, having been supplied with the August 2018 Z readings, undertook an analysis, shared that analysis with the appellant, explained to the appellant that an assessment based on that analysis would be £127,334, and indicated to the appellant in his email of 30 August 2018 that whilst this was slightly lower than he would expect (it was based on the 78% average for standard rated sales) he would accept those figures going back four years if they were accepted by the appellant, but if they were not, he would need to carry out an invigilation exercise at the outlets.

One of the appellants repeated submissions was that following HMRC’s enquiries in October 2016 (and the appellant’s recollection in October 2017 as well, although no such enquiry was evidenced by HMRCs files) no suggestion was made that the balance of standard rated and zero rated sales was incorrect, and so he had assumed his VAT affairs were in order.

So is there merit in the appellant’s submission that, essentially, HMRC, having sat on the evidence of the Z readings which were inspected in October 2016, should not now be allowed to assess either at all (i.e. the assessment of 18 July 2019 is invalid) or should have assessed within one year of October 2016 for those accounting periods predating October 2016, and so the July 2019 assessment cannot be valid to the extent that it assesses those accounting periods.

This situation was addressed in the 6 principles set out in Pegasus and which are recorded at [11] above. If the Commissioners had actual evidence which was sufficient to justify the making of the assessment in question, yet failed to do so, then that failure can only be challenged on the basis that it was perverse or wholly unreasonable. As Mr Justice Dyson said in that case “in some cases, the taxpayer may complain that the Commissioners have made an assessment on insufficient material. In other cases, the complaint of the taxpayer may be that, in the light of the evidence of which they were aware, it was wholly unreasonable for the Commissioners to delay in making the assessment. In both cases an appeal will succeed if it is shown that the Commissioners approach was wholly unreasonable……. I recognise that this is a high hurdle for the taxpayer to surmount, but Parliament has entrusted these matters to the judgment of the Commissioners, and it is right that challenges to the exercise of judgment should only succeed when something has gone seriously wrong.”

He went on to say that “the question before the Tribunal on an appeal, therefore, is whether the Commissioners’ failure to make an earlier assessment was perverse or wholly unreasonable….”

The burden of proving that the Commissioners had sufficient evidence in October 2016 to make an assessment at that date, rests with the appellant (as does any challenge to HMRC’s assessment based on time limits). For the appellant to succeed in his contention that there was sufficient evidence before HMRC in October 2016 to justify the making of an assessment at that date (and that by failing to assess previous years within one year from that date, any assessment relating to those periods is invalid) it must show that failure to assess based on the evidence of the Z readings which were inspected in October 2016, was perverse or wholly unreasonable. We cannot say that it was. Whilst we think that HMRC’s failure to follow up what we believe to be an analysis of the information provided by those Z readings, with an invigilation exercise, which they themselves accept should have been undertaken but was not due to pressure of work, is unreasonable (and given that this information has only just come to light, the appellant may wish to pursue other avenues of redress) failure to assess based on Z readings alone is not perverse or wholly unreasonable. Whilst it does provide information on which to make an assessment, HMRC’s subjective judgment that they might need further information (a process evidenced by Officer Vaghela in his investigation) cannot be impugned as being perverse or wholly unreasonable. Indeed the officer undertaking the review in October 2016 would have needed to made an assessment to best judgment, and doing so on the basis of Z readings alone might have been impugned as not complying with that criterion.

It is our view that that the last piece of evidence of facts of sufficient weight to justify making the assessment was that derived from the invigilation exercises undertaken in October 2018 and January 2019. It was not perverse or wholly unreasonable for HMRC not to have assessed following the disclosure of the Z readings in October 2016. The one year time period therefore runs from January 2019 and the assessment of July 2019 is thus within the one year period.

In conclusion, therefore, in our judgment the assessment of 18 July 2019 was a valid, in time, best of judgment assessment, which has not been displaced by the appellant. Accordingly, our decision is that its appeal is dismissed.

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.