Cash-based businesses: a claim that profits were overstated

Cash-based businesses: a claim that profits were overstated
19 March 2024

We consider a case in which HMRC believed that a cash-based business had overstated its takings, and argued for a reduction of declared profits.

Key Points

What is the issue?

In April 2021, HMRC started to investigate a claim made by Café Jinnah under the Eat Out to Help Out scheme. It complained that the appellant had failed to provide evidence to substantiate the number of diners, and the total value of food and drinks sold under the scheme.

What does it mean to me?

The taxpayer did not even need to go to the stage of disproving HMRC’s assessment. The assessment was fundamentally flawed because of HMRC’s apparent predisposition to the idea that the taxpayer had made an excessive claim, and fell at the first hurdle.

What can I take away?

HMRC investigation into a cash-based, fast-moving business and its apparent refusal to accept any evidence other than electronic records is unrealistic. There is a difference between ‘evidence’ and ‘evidence which HMRC accept’ – and the tribunal will be looking for the former.

Having just passed the 20th anniversary of the first of my monthly case reports and seeing that the official publication date of this article is 1 April, I thought I would write about a case in which HMRC investigated a cash-based business and argued that the business’s profits had been overstated.

I wish to make it clear at this stage that I have not taken leave of my senses (or, if I have, that the previous paragraph is not proof). In the last week of February, the First-tier Tribunal published a decision in a case in which this is what actually happened.

The case is Café Jinnah LLP v HMRC [2024] UKFTT 159 (TC).

The facts of the case

In the summer of 2020, the country was in the middle of the Covid pandemic. The government decided to announce a scheme to generate business in one of the many sectors particularly blighted by the lockdown, the hospitality industry. The ‘Eat out to Help Out’ scheme offered discounts to ‘eat-in’ diners on Mondays, Tuesdays and Wednesdays in August 2020 (13 days in total) worth up to 50% of the food cost (including non-alcoholic drinks) or £10 (whichever is the lower). The discount would be reflected as a discount on the bill, with the restaurant (or similar establishment) then claiming that discount from the government.

The appellant operated a restaurant in Bradford. Unsurprisingly, the restaurant’s business had been badly affected by the Covid pandemic. The scheme offered what the tribunal described as a ‘lifeline’ to the business. So far as existing customers were concerned, many switched to eating on scheme days rather than weekends, so as to take advantage of the government-backed discounts on offer. The scheme also attracted new customers to the restaurant, many of whom came from the local area. This was a community where many individuals did not have bank accounts or credit cards and therefore many customers paid for their meals using cash.

According to the restaurant’s proprietor (the designated partner of the appellant), people were queuing outside the restaurant for hours in order to participate. He had to recruit family members to assist with the increased business – there were 12 individuals employed during the restaurant in August 2020. Furthermore, there were difficulties attending the local bank in order to deposit cash and physical dangers in queuing outside a bank with substantial amounts of money. As a result, the business retained the cash in a large safe it had on its premises and paid its staff and many key suppliers using this cash.

The proprietor advised that, even with social distancing restrictions, large family groups were able to sit together at a single table (strictly, tables being bunched together) so some groups were up to 30 in size. The restaurant was open for 12 hours on each day of the week, except Friday when shorter hours were worked.

The restaurant did not use a till. All orders are recorded on paper, with the top copy being left with the customer until it is presented to the cashier at the end of the meal. That top copy is then either retained by the restaurant or (if kept by the customer) the details are then added to an A4 spreadsheet which, at the end of the day, is bundled up with any of the remaining top copies of the order records.

In the weeks during which the scheme operated, average daily discounts totalling between £6,000 and £9,000 appear to have been claimed. For the first four weeks, the tribunal’s decision shows these in three-day batches representing each of the Monday to Wednesday periods covered by the scheme (thus, weekly totals of between £18,842 and £26,579). For the final week/day of the scheme, the Bank Holiday Monday 31 August, a discount of £9,461 was claimed. The total discount claimed under the scheme was £103,351.

In April 2021, HMRC started to investigate the appellant’s claim. This led to the appellant’s adviser sending in sample bills, records of daily takings (with table numbers and covers), table plans, etc. The adviser acknowledged that there had been an overclaim of £671.

Correspondence continued, during which the HMRC officer complained that the appellant had failed to provide evidence to substantiate the number of diners/covers who had used the scheme, and the total value of all eat-in food and non-alcoholic drinks sold. In particular, the officer was concerned about the lack of cash deposits into the business’s bank account so as to substantiate the appellant’s argument that there had been some cash sales.

As a result, the officer decided that the claim should be restricted to the amount of the credit card takings for the month, which were £28,984.71. The difference of £74,366.30 was the subject of an assessment made on 4 April 2022. In the course of an internal review, this was revised downwards by £10,600.

The appellant then notified the appeal to the tribunal.

The First-tier Tribunal’s decision

The appeal was heard by Judge Nigel Popplewell and Member Mohammed Farooq. It referred to the power to assess any ‘amount of a coronavirus support payment [received by a person] to which the person is not entitled’ in the Finance Act 2020 Sch 16 para 9(1). The test for the exercise of that power is that there has to be ‘an officer of Revenue and Customs [who] considers’ that there has been an overpayment of a coronavirus support payment. When that test is met, an assessment may be made ‘in the amount which ought in the officer’s opinion to be charged’.

The tribunal considered that, despite the different statutory formulations, this involved the same test as applicable to discovery assessments, as summarised by the Upper Tribunal in the case of Anderson v HMRC [2018] UKUT 159 (TCC). That has two elements. First, it means that the officer has to have the subjective belief that the assessment is justified. Secondly, that belief has to be objectively reasonable.

The tribunal considered that the officer, who was described as ‘a truthful and reliable witness’, did have the subjective belief that the coronavirus support payments had been overclaimed.

So far as the requirement that the belief be objectively reasonable, the tribunal noted the following aspects of the officer’s approach:

  • It necessitated every discount meal to have been paid for by credit card rather than by cash.
  • HMRC emphasised the fact that the sample bills provided did not contain a note of the date to which they related. The tribunal could not see how adding a date would have proven their authenticity.
  • The officer ignored the evidence as to the change of eating habits that the scheme had led to. (As the tribunal said, ‘the scheme was having precisely the desired effect’.)
  • The officer ignored the appellant’s real-time information (RTI) records showing the wages paid to staff in August 2020 and the fact that these were paid in cash (without any corresponding withdrawal from the appellant’s bank account).
  • The officer ignored the fact that the appellant had paid VAT and income tax on these takings that HMRC now argued had not been received (although the tribunal noted that these could have related to non-scheme days).
  • The officer ignored the fact that the appellant’s accounts showed an increase in cash in hand over the period.
  • The officer did not take into account the food purchases, which showed a higher supply of food than would be justified by merely the meals that could be supported by the credit card purchases.

In conclusion, the tribunal felt that the officer ‘started off from the position of suspecting that the appellant was not telling the truth and he was not prepared to accept the appellant’s position unless some form of empirical documentary evidence provided a smoking gun … He seems to have come to the decision that in the absence of dated and timed bills, there was nothing that the appellant could do to justify the additional cash takings … [and] did not seriously consider other matters which could have supported claim.’

For these reasons, the assessment was not objectively reasonable and the taxpayer’s appeal was allowed.


This is a further case where the taxpayer did not even need to go to the stage of disproving HMRC’s assessment. Because the assessment was so fundamentally flawed, it fell at the first hurdle. As a result, even the admitted £671 overclaim could not be recovered by HMRC.

What appears to have been fatal to HMRC’s approach was an apparent predisposition to the idea that the taxpayer had made an excessive claim. I doubt that that was necessarily the officer’s view at the beginning of the investigation (although HMRC’s selection of the case might not have been entirely random and could have had an influence on the officer’s thinking).

HMRC’s stated position to the tribunal was that it was not alleging any impropriety by the appellant and that there were no suggestions of falsification or fabrication of documents by the appellant or its members, staff or agents. However, as the tribunal pointed out, impropriety was the essence of HMRC’s case. It had refused to accept that the sample bills it had been provided with could be married to the sales records and this was because the officer did not believe that the daily takings records were accurate. As the tribunal continued to observe, that carried an implication that those daily records were compiled in a deliberate attempt to inflate the claim.

Although this is probably the first case to consider the Eat Out to Help Out scheme and possibly the first involving HMRC arguing that a taxpayer has over-declared its income, the underlying themes of the case are depressingly familiar. We have an HMRC investigation into a cash-based, fast-moving business and an apparent refusal to accept any evidence other than the gold-standard, date-stamped electronic records. That is unrealistic and means that taxpayers are forced to incur costs defending themselves against baseless assessments (and public funds are wasted as these cases are litigated).

There also appears to a complete disregard of the promise within HMRC’s charter which says (with my emphasis added): ‘We’ll assume you’re telling the truth, unless we’ve good reason to think you’re not. ’

It is inevitable that any organisation, which relies on individual officers to make decisions, is going to make mistakes from time to time. However, this was a case where the assessment had been subject to an internal review and was then taken over by a litigation team. There were plenty of opportunities for common sense to prevail. Is this evidence of a wider problem within HMRC involving a predisposition to disbelieving what taxpayers are saying or does it simply mean that one should not expect these separate stages to amount to a proper review of the strength of HMRC’s case?

Another point that might be worth reflecting is the fact that these businesses are typically (and possibly disproportionately compared with the wider population) run by members of ethnic minorities. Indeed, over the last few years, the published decisions of over-reach by zealous HMRC officers in relation to cash-based businesses predominantly involve such taxpayers (at least if one is to use the taxpayer’s name as an indicator of ethnicity). I do not suppose for a moment that there is any deliberate targeting of such groups by HMRC. However, if there are systemic problems with how HMRC investigates such businesses, that could amount to indirect (albeit unintentional) discrimination.

There is a small point, however, where I respectfully disagree with the tribunal’s approach – although I suspect the disagreement is possibly more one of form than of substance. When stating its (correct) conclusion that the approach it should take to para 9(1) was akin to that taken in relation to discovery assessments, the tribunal said that this conclusion is ‘supported’ by the fact that the procedural provisions governing discovery assessments and appeals against them are expressly imported into the para 9 rules for recovering excessive coronavirus support payments.

As noted, I fully agree with the tribunal’s interpretation of para 9(1) as it represents a sensible reading of the provisions there (i.e. the fact that it imports both subjective and objective elements). Furthermore, it is possible that the tribunal was saying no more than that it gained comfort for its conclusion by the fact that, in practical terms, a para 9(1) assessment would be treated in the same way as a discovery assessment and, therefore, it would not be unreasonable to assume that they would be subject to similar tests.

However, if the tribunal was actually saying that the invocation by para 9(3) of parts of the Taxes Management Act 1970 into the para 9 process is itself a reason contributing to its interpretation, then this is where I do depart from the tribunal’s approach. In my view, para 9(3) is not actually importing the rules in s 29(1) but other more generic aspects of the Taxes Management Act 1970 governing assessments more generally.

Secondly, and perhaps more importantly, para 9(1) contains its own test as to when and how an assessment may be made. Although para 9(1) does not use the word ‘discover’, ultimately it turns on the same key requirement of a discovery assessment being that an officer must form an opinion that there is something to assess. It is for that reason that I believe (in agreement with the First-tier Tribunal) that this requires there to be both subjective and objective aspects to the officer’s conclusion.

As I have said, this is a minor point. I consider the rest of the First-tier Tribunal’s decision unimpeachable.

What to do next

It is possible that the taxpayer’s arguments with HMRC in the course of the investigation were not helped by the paucity of conclusive records that could persuade HMRC that the discounts claimed were (on the whole) ones to which it was entitled. However, taxpayers should remember that there is a difference between ‘evidence’ and ‘evidence which HMRC accept’. It is the former which the tribunal will be looking for.

© Getty images/iStockphoto