Suppressed Sales Investigations: HMRC Methodology & Best Judgment

Cash-business suppressed-sales investigations are among the most common and most technically complex in HMRC’s compliance arsenal. Restaurants, takeaways, retailers and other cash-intensive businesses face scrutiny through covert observation visits, Z-reading analysis and purchase-invoice testing. The First-tier Tribunal’s June 2025 decision in 3KH Ltd v HMRC [2025] UKFTT 748 (TC) provides the most detailed recent account of how HMRC builds a suppression case and how the tribunal evaluates it. This guide analyses the methodology, the best judgment standard, the deliberateness question and how to challenge quantum.

Why Cash Businesses Are Targeted

HMRC’s Connect system (see our resource on how HMRC finds out) compares declared turnover against benchmarks for comparable businesses in the same sector and geography. A restaurant or takeaway declaring significantly lower sales per cover than comparable competitors or showing a purchase-to-sales ratio inconsistent with the sector, generates a risk flag. Once flagged, HMRC’s compliance teams have a well-established suppression-investigation methodology that begins well before the business is told it is under enquiry.

HMRC’s Investigative Toolkit

The primary tools HMRC deploys in a cash-business suppression investigation are:

• Covert observation visits: HMRC officers attend the premises as customers, recording covers served, bills paid and cash/card splits, without identifying themselves as HMRC.
• Z-reading analysis: Comparing the end-of-day till readings (Z-readings) against declared takings to identify discrepancies between gross till income and declared sales.
• Bank statement analysis: Identifying cash deposits that exceed declared cash takings or cash withdrawals inconsistent with the business model.
• Purchase invoice testing: Using the ratio of purchases to declared sales to identify whether the business could be generating sales at the declared level with those inputs (a classic suppression indicator is a high input-to-output ratio).
• Third-party data: Bookings systems, food delivery platforms, card terminal records, landlord turnover-linked rent reviews.

Covert Observation Visits: 3KH Methodology

The 3KH Ltd v HMRC [2025] UKFTT 748 (TC) decision provides an unusually detailed account of how HMRC conducts and documents covert visits. In that case, nine pairs of HMRC officers made covert visits to a restaurant (Tipu Sultan Moseley) on 7 November 2015, from noon to 22:30. The officers recorded:

• The number of covers and waiting staff (consistently over 10 front-of-house staff, with the restaurant becoming progressively busier);
• Table occupancy and waiting times (up to 40 minutes on the busiest sessions);
• Six pairs paid by cash, recording exact bill amounts (£43.75 to £60.95 per table) and cash tendered;
• Additional cash payments by other diners observed.

The Z-reading for that day recorded total sales of £10,446.68 (split £8,214.35 card and £2,232.53 cash). However, the VAT ledger recorded total gross sales for that period based on a handwritten amendment showing card sales of £7,740 and cash of only £728.08, i.e. the VAT was declared by reference to the manually amended figures, not the raw Z-reading totals. The discrepancy between the Z-reading and the declared figures was the core of HMRC’s suppression case.

Z-Reading Analysis

Z-readings are end-of-day reports produced by till systems showing gross sales for the trading period. In an honest business, declared VAT takings should match (or be reconcilable with) the Z-reading totals. Common suppression patterns HMRC looks for include:

• Handwritten amendments reducing Z-reading totals before the figures are transferred to the accounts, as occurred in 3KH;
• Deleted or voided transactions at till-close that are not accompanied by genuine refunds;
• No-sale key usage to open the till without registering a transaction;
• Missing Z-readings for periods where the business was trading;
• Implausibly low cash-to-card ratios in a predominantly cash-paying customer base.

In 3KH, the appellants had consistently recorded in the VAT ledger figures lower than those on the Z-readings, attributing the difference to alleged promotional discounts and employee meals. The Tribunal rejected this explanation as inadequately evidenced and found the suppression proved.

Purchase Invoice and Bank Testing

As an alternative or supplement to Z-reading analysis, HMRC uses purchase invoices to reconstruct turnover. The methodology works by:

1. Identifying total purchases of key inputs (food, drink, packaging) from purchase invoices and bank records;
2. Applying gross profit margins consistent with the industry to derive implied turnover;
3. Comparing implied turnover against declared sales.

In 3KH, HMRC also identified a second restaurant (TSL) in respect of which input tax had been claimed but no sales at all were declared, a more obvious indication of suppression or a ghost operation. Bank analysis confirmed cash deposits that the declared turnover could not explain.

The Best Judgment Standard: Van Boeckel

Where a trader has failed to keep adequate records or render accurate returns, section 73(1) VATA 1994 empowers HMRC to assess the amount of VAT due to “the best of its judgment.” The governing authority on what this requires is Van Boeckel v Customs and Excise Commissioners [1981] 2 All ER 505, in which Woolf J held that HMRC must:

1. Genuinely exercise its judgment;
2. Base the assessment on material from which a reasonable assessment can properly be made;
3. Not act in a way that no reasonable officer could act.

Best judgment does not require HMRC to achieve the right answer; it requires a genuine attempt at the right answer. As the Court of Appeal confirmed in Pegasus Birds Ltd v HMRC [1999] STC 95, the standard is not perfection, if it were, HMRC would never be able to assess a business that had destroyed its records. But a methodology that is arbitrary, unfair or based on demonstrably incorrect assumptions will not qualify. The tribunal can substitute its own figure if it finds the best judgment standard was not met or if the assessment was excessive on the evidence.

3KH Ltd v HMRC [2025] UKFTT 748: Assessment Upheld

The Tribunal in 3KH upheld HMRC’s VAT assessments and corporation tax discovery assessments in full (save for some quantum adjustments). Key findings relevant to practitioners:

• Oral evidence discredited. The two directors (MJ and MB) were excluded from each other’s testimony. Their evidence was internally inconsistent and inconsistent with each other, the Tribunal expressed the view that they were fabricating evidence as the hearing progressed.
• Best judgment satisfied. HMRC’s methodology, covert visits, Z-reading analysis, bank and purchase review, was systematic and based on adequate material. The assessments were made to best judgment.
• Suppression as deliberate conduct. The systematic amendment of Z-readings was deliberate. This justified both the extended 20-year discovery assessment window and the Schedule 24 deliberate-inaccuracy penalty.
• Corporation tax link via s455 CTA 2010. The Tribunal found that the suppressed sales had been extracted by the directors as participators, justifying additional corporation tax charges via section 455 CTA 2010 and NICs assessments. The s455 charge treats the extraction as a loan from the company to the participator, a significant further tax cost on top of the VAT and income tax assessments.

Deliberateness, Time Limits and Penalties

The characterisation of suppression as deliberate has very significant consequences. As analysed in our guide on the deliberate behaviour test:

• 20-year extended assessment window under s36(1A)(a) TMA 1970 (direct taxes) and s77(4A) VATA 1994 (VAT), rather than the 4-year ordinary window;
• Highest Schedule 24 penalty bands (up to 70% deliberate not concealed, 100% deliberate and concealed);
• Personal liability notice risk where the suppression is attributed to a director under paragraph 19 Schedule 24 FA 2007.

It is therefore critical to challenge the deliberateness characterisation where there is any basis to do so. A business with genuinely chaotic records, inexperienced bookkeeping and no positive evidence of intentional manipulation may be careless rather than deliberate, capping the window at six years and the penalty at 30%. The distinction may be the difference between an investigation reaching back 6 years or 20.

Challenging Quantum

Even where suppression is proven, the assessment may be excessive. Common challenges include:

• Sample validity: HMRC’s covert visits may not be representative of typical trading. A single observation on the busiest night of the year produces a distorted extrapolation.
• Z-reading reliability: Where Z-readings are missing or unreliable for other reasons (technical faults, genuine promotional discounts), their use as a benchmark may be challenged.
• Gross profit margin assumptions: Where HMRC applies a sector-standard gross profit margin, evidence that the specific business had higher costs (waste, spoilage, employee meals, promotional pricing) may reduce the implied turnover.
• Double-counting: Where HMRC assesses both VAT and corporation tax on suppressed sales, care is needed to ensure the same suppression is not assessed twice in incompatible ways.
• Card vs cash split: HMRC may assume that all suppressed takings were cash. Where the business had a higher card-payment proportion than assumed, the suppression quantum falls.

Practitioner Strategy

Obtain All the Evidence HMRC Has

Request all observation-visit notebooks, Z-readings reviewed, bank statements analysed and internal HMRC workings. HMRC’s case is almost entirely factual, and a challenge that is only possible once you have seen the full evidence base. Gaps in HMRC’s methodology are often only visible from its working papers.

Reconstruct the Business’s Own Figures

Commission an independent reconstruction of turnover using available records (purchase invoices, bank statements, surviving Z-readings, staff rota and wage records). An expert accountant’s independent figure is usually essential for the FTT, even if it results in some acknowledgement of under-declaration.

Challenge Deliberateness Specifically

If the record-keeping failures are more consistent with carelessness than systematic suppression, run that argument directly. The time-limit and penalty consequences are too significant to leave unchallenged. See our guide on the meaning of deliberate behaviour.

Prepare Directors for Oral Evidence

As 3KH vividly illustrates, credibility of oral evidence is often determinative. Inconsistent or fabricated evidence can convert a partial success on the figures into a complete loss. Witness preparation must be rigorous and honest.

Practitioner Checklist

1. Obtain HMRC’s full evidence base: observation notebooks, Z-reading analysis, bank workings, purchase testing methodology.
2. Identify the methodology used and test each assumption, are observation visits representative? Are gross profit margins appropriate for this business?
3. Commission an independent turnover reconstruction using the client’s own records.
4. Assess the best-judgment standard under Van Boeckel , was HMRC’s methodology a genuine, reasonable attempt?
5. Decide the deliberateness argument: systematic Z-reading manipulation = deliberate; chaotic bookkeeping = potentially careless.
6. Check the time limits , deliberate: 20 years; careless: 6 years and quantify the benefit of recharacterisation.
7. Address the s455 / participator extraction question if a corporation tax discovery is also raised.
8. Prepare witnesses thoroughly , 3KH shows that inconsistent oral evidence destroys credibility and with it the entire defence.
9. Check the appeal deadline and see our guide on late appeals if it has passed.

Frequently Asked Questions