Fish and chip shops: a flawed HMRC investigation to cash-only businesses

In the more than 30 years that I have been in the tax profession, certain types of business have been seen as likely candidates for an old-fashioned ‘back duty’ investigation. At or very near the top of that list will be a local takeaway catering establishment. Indeed, even training exercises provided for tax advisers outside HMRC, looking at how to respond to such investigations, are likely to focus on these kinds of business.

The case of Georgiou v HMRC [2022] UKFTT 455 (TC) concerns such a case.

The facts of the case

Mr Georgiou was a director of Georgiou & Co Ltd (the company). He owned 25% of the shares of the company.

As at 2013, the company had owned four fish and chip shops, which operated on a cash-only basis. During the 2014/15 tax year, the company ceased trading at two shops but retained the premises. It let them to other businesses and received rental income for the premises. The two retained premises were known as Dove House and Harvey’s. The company ceased trading altogether on 18 November 2017, again with the premises rented out to unconnected companies. Mr Georgiou continued working at Dove House as an employee of the new owner. By March 2018, the company was insolvent and entered a creditors’ voluntary liquidation.

In the meantime, the company had been assessed for the following:

• a VAT assessment of £141,238 for undeclared output tax for nine quarters up to March 2016;
• a deliberate behaviour penalty amounting to £84,037;
• corporation tax discovery assessments for the three years ended 31 March 2016 totalling £230,647.40;
• a corresponding penalty assessment of £137,235.55 (later reduced to £121,090 to reflect co-operation not previously recognised); and
• a notice of determination of corporation tax for the year ended 31 March 2017 charging £68,833.

The corporation tax assessments included sums due under the loans to participator rules (as well as in relation to under-assessed profits).

The company’s appeal was brought by the liquidator. In addition, Mr Georgiou faced a £64,885.32 personal liability notice in relation to the VAT penalty. This was later reduced to £47,857.76 in partial recognition of the fact that Mr Georgiou had not taken over control of the business until April 2015 following his father’s death.

Dove House and Harvey’s were the company’s original sites acquired in 2011 and 2012. Dove House was in a more affluent area, as well as being well positioned on a main road. Harvey’s was not so well sited, nor was its local clientele as affluent.

Until his death, Mr Georgiou’s father was the driving force behind the business and the one who dealt with the company’s finances. Mr Georgiou’s English was better than his father’s and he often dealt with suppliers to the company. Mr Georgiou also did the actual cooking. Mr Georgiou’s wife and mother (the other two shareholders) assisted with the cooking and cleaning. There were no external staff.

The investigation

The investigation started in February 2016. The officer requested documents showing the shops’ sales and cash purchases for the year ended 31 March 2015. The exercise was to ascertain how much cash would have passed through the business in the year. Twelve of the sheets were missing.

From the sheets made available, the total came to £490,000. By taking an average, given the missing sheets, the officer estimated the total for the year to be £552,450. The officer noted that the declared sales, however, came to only £507,000.

A similar exercise for the December 2015 quarter showed total cash of £117,030 compared with gross sales in the VAT return of £88,025. The officer noted a similar shortfall in the purchases, amounting to £33,000. Although he recognised that a possible explanation would have been that purchases were being accounted for in the wrong period, the officer reached the conclusion that this was further evidence of suppression of trading activity.

Similar shortfalls of £3,000 (sales) and £17,000 (purchases) were identified for the September 2016 trading quarter.

HMRC also raised concerns about the till data known as Z-readings. Z-readings show the daily totals and will, ordinarily, contain a date or other marker to allow the readings to be kept in sequence. However, the company’s tills appeared to be reset each day making such sequencing impossible. HMRC had also asked for the records of the individual sales but the company claimed that these were unavailable due to an apparent fault in the tills – the only records of the sales were within the Z-readings themselves.

Following a change of HMRC officer leading the investigation, the new officer made an unannounced visit to both shops in July 2017 but was refused entry by Mr Georgiou’s wife. This was on the basis that Mr Georgiou was not present.

In the meantime, the company purchased new tills (which were said to be HMRC-compliant) but bought those at the bottom of the range. HMRC asked the company’s accountant to provide HMRC with the electronic journal for each shop. Instead, the accountant sent only the Z-readings, saying that he considered those to be the same thing.

A further visit was arranged at which a representative of the tills’ supplier would also be present. HMRC’s own till specialists explained what information was wanted from the tills and this was taken by the supplier’s representative and saved onto a brand new USB stick. However, it transpired that the records taken by HMRC were not complete as there was nothing in the folders for the individual records. HMRC concluded that these had been deleted or somehow not recorded by the new tills. On the company’s behalf, it was suggested that the new tills were faulty but it was unable to afford any further replacement.

Covert observations were carried out at both shops on Friday 17 and Wednesday 22 November 2017. On the second occasion, officers also made test purchases. During the observation, it was noted that at one stage, a large payment was taken whilst the till was already open, with no keying into the till at that time. There were other instances when the till was kept open but, on those occasions, individual sales were keyed in.

The assessments

Having been told of the cessation of the business, HMRC then proceeded to make assessments on the basis of the information that it had. The Z-readings for Dove House in the September 2016 quarter showed the number of transactions in the quarter, allowing an average transaction of £7.31 to be calculated. The officer derived an average number of transactions for Fridays and Wednesdays which were then taken as representative for weekend and weekday trading days. As no Z-readings were taken for Harvey’s, the sales information for Dove House was used instead.

These average transactions were lower than those actually observed in November 2017. For example, the recorded data showed 104 transactions on average in the September 2016 quarter, whereas 163 were observed on Wednesday 22 November 2017. HMRC thus calculated a 36% suppression rate. In the company’s favour, HMRC ignored the fact that Dove House was also open at lunchtimes.

By adopting a similar method, the suppression rate for weekends was calculated as 49%. For Harvey’s, suppression rates of 26% (weekends) and 19% (weekdays) were calculated.

It was noted by HMRC, however, that declared sales went up significantly once the investigation started. Therefore, in later periods, the suppression rates were taken to be lower.

Subject to those adjustments and an estimate as to the split of the turnover between the different shops, this enabled HMRC to estimate the under-assessments and to quantify the penalties.

The company and Mr Georgiou appealed against the assessments and penalties. A point was taken by the company’s representative that the VAT assessments were late because they were made more than a year after the original meeting at which point HMRC had obtained the cash figures.

The First-tier Tribunal’s decision

The case was heard by Judge Marilyn McKeever and Member Susan Stott.

As for the initial question about the timing of the assessments, the tribunal dismissed the company’s argument. Although the cash figures, as HMRC put it, suggested that there had been suppression, HMRC was not able to quantify the assessments until after the covert observations in 2017. The VAT assessments were made well within a year of those observations. They were therefore made in time.

But for that single point, the taxpayers were completely successful in their appeals. The company was able to demonstrate that HMRC’s calculations were riddled with errors, thereby undermining their credibility.

HMRC had identified anomalous trading figures on the VAT returns comparing the March 2014 and March 2015 figures. However, the company was able to show that the March 2014 return reflected the period from the commencement of trading (seven and a half months) compared with the 12 month trading period for the following return. Once the period lengths were properly taken account of, the trading pattern was smoother.

Secondly, HMRC had failed to reflect the fact that the company was operating four shops at the beginning of the period under investigation, and only two by the end. Once one also factored in the fact that the shops effectively closed following Mr Georgiou’s father’s death, the reported figures looked considerably more realistic.

Thirdly, HMRC had not taken account of the business problems that arose around the time of and subsequent to the father’s death. Not only did the business close during the busiest period of the year because of a flood (in the run up to Christmas) but the company employed casual staff when Mr Georgiou was away in Cyprus in the aftermath of his father’s death. At that stage, portion sizes and the quality of the output varied, leading to a loss of clientele, something that Mr Georgiou attempted to stabilise following his return to the helm.

Fourthly, HMRC’s totting up of all bankings failed to exclude the rental income received from two of the premises.

Fifthly, although the analysis of purchases correctly disregarded expenditure that was clearly personal in nature, there were still a number of outgoings marked ‘cheque’ or ‘transfer’ that did not necessarily relate to purchases of stock.

Sixthly, it was only when the ‘missing’ cash sheets were factored in that HMRC’s methodology gave rise to any possible under-assessment.

Seventhly, the observations, which lay at the heart of HMRC’s case, were not sufficiently reliable. First, they took place too closely together. Secondly, the second one was carried out when the business was no longer being carried out by the company.

Eighthly, HMRC seemed to work on the basis of there being 15 Wednesdays in any particular quarter (when usually there are only 13, occasionally 14).

Ninthly, the absence of evidence from any of the officers conducting the observations undermined the reliability of the evidence derived from those occasions. Although hearsay evidence is admissible in the tribunal, it is often weak, particularly when there are uncertainties as to what was observed and recorded. In particular, the record that there were 56 customers arriving between 5pm and 5.30pm was considered unlikely by the tribunal, given that photographs show it looking full with only 20 individuals present. Indeed, a later visit several months later showed Dove House to be ‘almost empty’ at 5.50pm.

Tenthly, HMRC had failed to factor in (and refused to acknowledge) the difference in the two remaining shops’ locations and the fact that a local factory (which provides a number of customers) closes for two weeks during the summer.

Finally, HMRC had failed to exercise any credibility check to the numbers calculated. The tribunal noted that Mr Georgiou was cooking and not handling the cash. Indeed, the observations showed that all transactions were being keyed in – the exception observed related to a telephoned order which was keyed in when the order was made and therefore not duplicated when the cash was received a little later that evening.

As the tribunal noted, HMRC’s figures suggested that £850,000 of profits had been taken from the business over a four year period. HMRC had not checked whether these alleged takings had impacted upon Mr Georgiou’s lifestyle. The HMRC officer had failed to recognise that this was simply loss-making business (which was eventually required to close).

The tribunal concluded that the VAT assessments were not made to best judgment. The penalties and other matters were set aside for similar (and in some cases additional) reasons.

Commentary

Reverting to my opening comments, I do not doubt that cash businesses such as this represent a risk for HMRC. Cash is inherently hard to trace, trading activity can be hectic and it is rare for such businesses to have a dedicated bookkeeper. But it is, in my view, fundamentally wrong to proceed on the assumption that the proprietors of such businesses are routinely suppressing sales so as to evade their tax liabilities.

However, this case suggests that HMRC officers proceeded from that assumption. Indeed, the tribunal itself noted that ‘following the cash reconciliations which “suggested” the suppression of sales and purchases, [the officer] concluded that was the case and interpreted all the information provided and other evidence in a way which supported that conclusion’.

What to do next

Fortunately, Mr Georgiou appears to have engaged an experienced adviser who not only identified flaws in HMRC’s methodology but also looked at the background facts that underlay the figures and was able to set out his case with clarity to the tribunal. In cases such as this, it can often be impossible to show what the correct numbers should be. However, if HMRC’s assessments are so fundamentally flawed (for lack of objectivity), it will be unnecessary for the taxpayer to prove to a tribunal what the correct figures should be.

© Getty images/iStockphoto