Discovery Assessment Staleness: Case Law Analysis

The Supreme Court’s decision in HMRC v Tooth [2021] UKSC 17 is the leading modern authority on the threshold for “deliberate inaccuracy” in a tax return and on the requirement that any discovery must be genuine and not merely a reappraisal of previously-known facts. This guide analyses the staleness doctrine and its interaction with the s29 TMA 1970 discovery assessment regime, equipping advisers to identify and run these arguments before the First-tier Tax Tribunal.

Introduction: Scope and Practitioner Relevance

Discovery assessments are the principal mechanism by which HMRC raises additional tax once the statutory enquiry window under s9A TMA 1970 has closed. For the adviser, they are simultaneously the most common vehicle for out-of-time HMRC challenges and the most technically vulnerable: the conditions that must be satisfied before a valid s29 assessment can be raised are strict and two of the most powerful challenges, staleness and the deliberate inaccuracy threshold, remain underutilised in practice.

The staleness doctrine holds that HMRC cannot manufacture a “discovery” by reviewing material it already held and simply drawing a different legal conclusion. The deliberate inaccuracy doctrine, as restated by the Supreme Court in Tooth, holds that the 20-year extended time limit requires an intention to mislead HMRC, not merely a technical error in the return, however significant. Together, these two principles represent a substantial body of law that can defeat discovery assessments entirely.

This guide is directed at accountants, solicitors and tax advisers acting for individuals and businesses facing discovery assessments. It covers the statutory framework, the key authorities, the procedural steps for mounting a challenge and a worked example drawn from Common Reporting Standard (CRS) data scenarios, which are increasingly the context in which HMRC seeks to deploy the 20-year extended window.

The s29 TMA 1970 Framework

Section 29 of the Taxes Management Act 1970 provides the gateway for discovery assessments once the s9A enquiry window has closed. The structure is tripartite.

The Three Conditions

A valid s29 assessment requires satisfaction of all three of the following conditions:

1. Discovery: An officer of HMRC discovers that an amount of income tax or capital gains tax which ought to have been assessed has not been assessed or that an assessment has become insufficient, or that relief has been given which is excessive (s29(1) TMA).
2. The “staleness” gateway: An assessment may not be made where the insufficiency was attributable to information provided to HMRC which was sufficient for an officer of HMRC, on the basis of the information available, to have been aware of the insufficiency (s29(5) TMA, the “reasonable officer” gateway that protects the taxpayer who has made full disclosure).
3. The extended time limit conditions: Where the ordinary four-year window has passed, HMRC must also demonstrate that the insufficiency was attributable to “careless” or “deliberate” conduct by the taxpayer (or their agent) in order to access either the six-year or twenty-year limit respectively.

Time Limits at a Glance

• 4 years: The ordinary window. HMRC may make a discovery assessment within four years of the end of the relevant chargeable period without needing to prove any fault (s34 TMA).
• 6 years: Where the loss of tax is attributable to careless behaviour by the taxpayer or their agent. “Careless” means failure to take reasonable care (s118(5) TMA).
• 20 years: Where the loss of tax is attributable to a “deliberate inaccuracy” on the part of the taxpayer (s29(4) TMA, as qualified by s118(7) TMA). This is the extended window that Tooth directly addresses.

The Hypothetical Officer Test

Section 29(5) TMA is one of the taxpayer’s strongest shields. It provides that a discovery assessment cannot be made if, at the time the assessment could first have been made, an officer of HMRC could reasonably have been expected, on the basis of the information available to HMRC, to be aware of the insufficiency. The standard is objective: the question is what a hypothetical diligent officer would have been aware of, not whether the actual officer in fact made the discovery at an earlier point.

Critically, the “information available” for purposes of s29(5) is confined to what was provided to HMRC by the taxpayer in the return and in accompanying documents. It does not extend to information HMRC could have obtained by enquiry or from third parties. This statutory limitation on the “pool” of attributed information was central to the reasoning in Langham v Veltema.

The Staleness Doctrine: Origins in Cenlon Finance

The concept that a “discovery” must involve a genuinely new finding of fact, rather than a fresh legal analysis of facts already known to HMRC, originates in the House of Lords in Cenlon Finance Co Ltd v Ellwood [1962] AC 782.

The Principle in Cenlon

In Cenlon, HMRC (then the Inland Revenue) had made assessments based on facts that were, at the time of the assessments, already fully known to the Revenue. The question was whether this constituted a valid “discovery.” The House of Lords held that it did not. A “discovery” for the purposes of what is now s29 TMA requires that the officer forms a view, a new finding of fact or a newly-acquired appreciation of a material fact, that he or she did not previously hold. A reassessment of the legal consequences of a state of affairs already known in full is not, of itself, a discovery.

The Cenlon principle gave birth to what practitioners now call the staleness doctrine: where HMRC has, at an earlier point in time, formed a view that there may be an insufficiency in the taxpayer’s return and has then done nothing (or done something legally ineffective), a later officer reviewing the same file cannot rely on that earlier awareness as a “fresh” discovery.

How Staleness Arises in Practice

Staleness most commonly arises in the following fact patterns:

• HMRC received CRS or Automatic Exchange of Information (AEoI) data relating to offshore accounts several years before the assessment, but delayed action;
• HMRC opened an enquiry using the wrong statutory vehicle (as in Tooth) and then closed or abandoned it, subsequently attempting to “rediscover” the same issue via a discovery assessment;
• HMRC received a risk-referral or internal risk-flag in year one but an assessment officer did not act until a much later date, relying on the same underlying information;
• An officer exchanged correspondence with the taxpayer acknowledging the issue, then took no further steps before the relevant assessment window began to close.

Langham v Veltema [2004]: The Reasonable Officer Test Unpacked

The Court of Appeal decision in Langham v Veltema [2004] EWCA Civ 193; [2004] STC 544 is the primary authority on the s29(5) “hypothetical officer” test and the limits of the information pool attributed to HMRC.

The Facts

Mr Veltema was a shareholder-director whose company paid him a sum that was classified in the company’s accounts as a loan. His personal tax return did not include any income referable to that transaction. HMRC sought to issue a discovery assessment on the basis that there had been a beneficial loan or a distribution. The question was whether, on the information provided by the taxpayer in his return and accompanying documents, an officer of HMRC could reasonably have been expected to be aware of the potential insufficiency. If yes, the assessment was barred by s29(5).

The Court of Appeal’s Holding

The Court of Appeal held that the taxpayer’s own return and only that return and documents submitted with it, constituted the information pool for purposes of s29(5). HMRC was not taken to know what it could have found out had it investigated further. The taxpayer’s return had not disclosed the transaction in a way that would have made a diligent officer aware that there might be under-assessed tax. Accordingly, s29(5) did not bar the discovery assessment.

The ratio of Veltema for practitioners runs in both directions. It is a reminder that the s29(5) bar operates only on the basis of what the taxpayer actually disclosed, not what HMRC could theoretically have obtained. However, it equally confirms that where the return does disclose the material facts, even if by way of a white space explanation, a note or a set of accompanying computations, the hypothetical officer test may well bar the discovery assessment.

The Information Pool: What Is and Is Not Attributed

Following Veltema, the settled position is that the information pool attributed to HMRC for s29(5) purposes includes:

• The content of the taxpayer’s self-assessment return in the relevant year;
• Any documents and computations submitted with the return;
• Any claims, elections or amendments submitted in connection with that return.

The following are not included in the attributed pool:

• Information that HMRC held about the taxpayer from other sources (third-party data, CRS reports, other taxpayers’ returns);
• Information HMRC could have obtained had it opened a timely enquiry;
• Intelligence held by a different HMRC team or compliance unit.

HMRC v Tooth [2021] UKSC 17: Facts, Procedural History and Supreme Court Ratio

The Facts

Mr Tooth participated in a tax avoidance scheme designed to generate an employment-related loss in the tax year 2008–09. The intention was to carry that loss back to the 2007–08 return and set it against income in that earlier year. Due to a software deficiency in the tax return preparation tool used by his advisers, the loss was entered on the partnership pages of the 2007–08 return rather than on the employment pages. Recognising the discrepancy, the advisers included a “white space” narrative explanation in the return, making clear that the taxpayer was seeking to carry back an employment loss from 2008–09 under the relevant provisions.

In August 2009, HMRC wrote to Mr Tooth stating that it intended to open an enquiry under Schedule 1A TMA 1970 into the claim made in the 2007–08 return. Unfortunately, Schedule 1A was the wrong statutory vehicle for the enquiry: the correct route was an enquiry under s9A TMA 1970 into the return itself. The Schedule 1A notice was therefore legally ineffective. The 2007–08 return became final by the end of 2010. In 2014, HMRC issued a discovery assessment under s29 TMA 1970, asserting that the taxpayer had “deliberately” brought about an insufficiency, thereby invoking the 20-year extended window.

Procedural History

• First-tier Tax Tribunal: Held that HMRC had made a discovery in October 2014, but found that there was no deliberate inaccuracy, the return, read as a whole, did not contain an inaccuracy at all. Allowed the taxpayer’s appeal.
• Upper Tribunal: Held that there was no discovery in 2014 because HMRC had formed its view in 2009, and any discovery made at that point had become “stale” by 2014. Additionally, the UT confirmed that there was no inaccuracy in the return read as a whole.
• Court of Appeal: Agreed there was no qualifying discovery, but the majority found that there was a deliberate inaccuracy in the partnership-pages entry, notwithstanding the white space explanation.
• Supreme Court (Lords Reed, Briggs, Sales, Leggatt and Burrows, judgment delivered May 2021): Dismissed HMRC’s appeal.

The Supreme Court’s Ratio: Deliberate Inaccuracy

The central holding of the Supreme Court in Tooth concerns the meaning of “deliberate inaccuracy” within s118(7) TMA 1970. The Supreme Court held, restoring the UT’s approach, that the phrase means a statement that was deliberately inaccurate , that is, a statement the maker intended to be misleading or as to the accuracy of which the maker was reckless. It does not mean merely a deliberate act that happens to produce an inaccurate result.

The Court adopted a grammatical and purposive analysis: the word “deliberate” as an adjective qualifies “inaccuracy” as a whole. The natural meaning is that the person making the statement intended the inaccuracy, in the sense of intending to mislead HMRC or being indifferent as to whether they did so. A technical error, however significant in its effect, that arises without any intention on the taxpayer’s part to misrepresent their position to HMRC does not satisfy the test.

Reading the Return as a Whole

The Supreme Court also affirmed that a tax return must be read in its entirety and in context, not through artificial tunnel vision applied to individual boxes. HMRC provides white space in returns specifically to permit taxpayers to supplement and explain what would otherwise be incomplete box-level entries. Having designed the return to accommodate explanatory narrative, HMRC cannot then assert a purely mechanical interpretation that ignores that narrative when determining whether there is an “inaccuracy.”

In Mr Tooth’s case, any reasonable reader of the complete 2007–08 return, including the white space explanation, would have understood exactly what the taxpayer was claiming: a carry-back of an employment loss from 2008–09. The fact that the software had placed the figures on the partnership pages rather than the employment pages did not make the return inaccurate when read as a whole. There was accordingly no inaccuracy in the return and therefore no deliberate inaccuracy of any kind.

Staleness: Left Open at Supreme Court Level

The Upper Tribunal’s staleness finding, that any discovery made by HMRC in 2009 had become stale by 2014, was not the basis of the Supreme Court’s decision and the Supreme Court expressly left the staleness doctrine as an open question at that level. However, the Court did not disapprove the doctrine and its discussion of the issue confirms that staleness remains a live and viable argument before the FTT and Upper Tribunal. The UT’s analysis in Tooth is accordingly the leading direct authority on the staleness point.

What Tooth Confirms and What It Changes

What the Decision Settles

The Supreme Court’s decision in Tooth definitively resolves three points that had previously been uncertain or contested:

• The deliberate inaccuracy threshold is subjective. HMRC must show that the taxpayer intended to mislead or was reckless. Technical or structural errors in a return, even errors that substantially misstate the taxpayer’s position, do not satisfy this test without evidence of dishonest intent.
• Returns must be read holistically. White space explanations, covering letters and accompanying computations are part of the return for the purpose of determining whether it contains an inaccuracy. HMRC cannot cherry-pick individual boxes and declare them inaccurate in isolation from the explanatory context provided.
• The 20-year window is reserved for genuine wrongdoing. The extended time limit is not a general residual power: it is an exceptional measure calibrated to fraud or deliberate misrepresentation. HMRC cannot reach for it whenever a return turns out to have understated a liability.

What Remains Open

The Supreme Court did not finally resolve the staleness doctrine, meaning that the principle, while well-established in the authorities below, has not yet received Supreme Court endorsement or rejection. This leaves a degree of uncertainty for practitioners in framing how robustly to advance the point. However, the lack of Supreme Court disapproval, combined with the consistent acceptance of the doctrine at FTT, Upper Tribunal and Court of Appeal level, means it remains a strong ground of challenge in appropriate cases.

The decision also does not resolve the precise moment at which a discovery becomes “stale,” nor the temporal gap required between the earlier awareness and the assessment for staleness to apply. These are questions of fact and degree, to be assessed case by case on the available evidence.

Running a Staleness Argument at the FTT

A staleness challenge is inherently a fact-intensive exercise. The following steps represent best practice for building and presenting such an argument.

Timeline Evidence

The starting point is a comprehensive chronology of everything that passed between HMRC and the taxpayer from the date the relevant return was filed to the date the assessment was issued. Gather:

• All correspondence from HMRC, including opening letters, information requests and any notices (whether valid or not) issued before the assessment;
• Any internal HMRC documents disclosed during the appeal process or obtained via FoIA or Subject Access Request;
• The date on which CRS, AEoI or other third-party data was received by HMRC (this can sometimes be inferred from published HMRC statistics or from the assessment letter itself);
• Evidence of any risk-flag or referral within HMRC, sometimes visible in the narrative of the assessment letter or in documents produced under disclosure.

Freedom of Information Act and Subject Access Requests

A Freedom of Information Act 2000 (FoIA) request to HMRC can yield internal case history notes, risk-assessment records and HMRC’s own timeline of its engagement with the taxpayer’s file. A Subject Access Request under Article 15 UK GDPR is separately available and is directed specifically at personal data held about the taxpayer. Together, these routes can reveal:

• The date on which HMRC first received third-party data about the taxpayer’s offshore assets;
• The date on which an HMRC officer first noted or documented a view about a potential under-assessment;
• Whether any prior enquiry, informal approach or internal referral preceded the formal assessment.

HMRC is not obliged to disclose all internal deliberative documents under FoIA and certain exemptions apply. However, the factual record of what HMRC held and when, as opposed to internal legal advice, is generally disclosable and in complex cases can be highly probative of staleness.

HMRC’s Disclosure Obligations in Appeal Proceedings

Under the Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009 (SI 2009/273), HMRC is required to provide a statement of case and to disclose the documents on which it relies. The FTT also has broad case-management powers under r5 to order further disclosure. An adviser running a staleness argument should consider making a targeted disclosure application seeking the production of all documents recording HMRC’s awareness of the alleged insufficiency prior to the date of the assessment.

What the FTT Looks For

In practice, the FTT approaches a staleness challenge by asking two questions:

1. At what point did an officer of HMRC form the relevant “discovery”, that is, at what point did an officer actually reach the conclusion that there was an insufficiency?
2. Is the formal assessment based on a genuinely new finding of fact or appreciation of the evidence or is it based on the same material that the earlier officer had already processed?

The answer to question (1) requires factual evidence, typically from HMRC’s own records. The answer to question (2) requires analysis of the assessment letter and any officer’s witness statement against the earlier materials. Where an assessment letter is substantively identical in its reasoning to earlier correspondence, that is powerful evidence that no new discovery was made.

Worked Example: CRS Data and an Offshore Account

The following worked example illustrates how the staleness analysis operates in a typical fact pattern arising from HMRC’s use of Common Reporting Standard data.

The Facts

Consider a UK-resident taxpayer (“Client A”) who held a Swiss bank account throughout the period 2015–2021. The account earned interest and dividend income. Client A failed to declare this income in any of his UK self-assessment returns for those years. In 2021, Switzerland reported Client A’s account details to HMRC under the Common Reporting Standard. In 2025, HMRC issues a discovery assessment for 2015–16, invoking the 20-year extended window on the basis that Client A “deliberately” omitted the income.

The Staleness Analysis

The adviser’s first task is to determine when HMRC made the alleged discovery.

• Step 1, Date of CRS data receipt: HMRC received CRS data from Switzerland in late 2021. This information disclosed the existence of Client A’s account, its balance and its income. It is arguable that an officer of HMRC, on receipt of this data in 2021, formed a view that there was an insufficiency in the 2015–16 return. That would place the discovery in 2021.
• Step 2, What happened between 2021 and 2025: If no enquiry was opened, no informal approach made and no protective assessment raised during 2021–2024, the adviser should ask: was the 2025 assessment based on the same CRS data as HMRC held in 2021 or on genuinely new information? If the answer is the former, there is a plausible staleness argument that any discovery was made in 2021 and became stale before the 2025 assessment.
• Step 3, The hypothetical officer test: Does the CRS data received in 2021, viewed objectively, constitute sufficient information for a hypothetical officer to have been aware that there was an insufficiency? Almost certainly yes, CRS data discloses account income directly. If so, HMRC’s discovery should have occurred in 2021.
• Step 4, Time limit for 2015–16: For the 2015–16 tax year (ended 5 April 2016), the 20-year limit runs to 5 April 2036. The assessment in 2025 is therefore within the 20-year window. However, if HMRC can only access that window by proving a “deliberate inaccuracy,” the question of whether Client A acted deliberately must be assessed on Tooth principles.
• Step 5, Deliberate inaccuracy under Tooth: Was the omission of the Swiss account income a deliberate inaccuracy within s118(7) TMA, as interpreted in Tooth? The question is whether Client A intended to mislead HMRC or was reckless as to whether he did so. A taxpayer who was unaware of the CRS reporting obligation or who took a genuinely mistaken view of their domicile status, is not a deliberate inaccuracy case. HMRC must adduce specific evidence of intent, not merely the fact of omission.

Practitioner Checklist: Challenging a Discovery Assessment on Staleness and Deliberate Inaccuracy Grounds

1. Identify the asserted basis for the extended time limit. Does HMRC rely on “careless” (6-year) or “deliberate” (20-year) conduct? The Tooth ratio applies directly only to the deliberate inaccuracy gateway. Identify which provision is in play before structuring the challenge.
2. Establish the date of the alleged discovery. Request the assessment file and all accompanying officer notes. Identify the first point at which any HMRC officer documented a view that there was an insufficiency in the relevant return.
3. Serve a Freedom of Information Act request. Request all documents recording HMRC’s receipt of third-party data (CRS, AEoI, Connect flags, employer reports) about the taxpayer for the relevant year, with the date of receipt. This is foundational for any staleness argument.
4. Serve a Subject Access Request under UK GDPR. In parallel with the FoIA request, serve a SAR. This route is specifically targeted at personal data held about the taxpayer and may yield the case history notes and contact logs that confirm the timeline.
5. Analyse the assessment letter against earlier HMRC correspondence. If the reasoning in the 2025 assessment letter is substantively identical to a risk-referral or informal enquiry letter sent in 2021, that is powerful staleness evidence.
6. Assess the deliberate inaccuracy point under Tooth. Does HMRC have direct evidence of intent to mislead or is it simply asserting that the omission of income demonstrates deliberateness? A bare allegation of deliberateness is not sufficient under Tooth. Require HMRC to particularise the conduct it relies on.
7. Consider the “return read as a whole” argument. If the return contained any white-space explanation, accompanying note or separate claim that put HMRC on notice of the relevant transaction or position, that context must be taken into account. A technically incorrect box entry explained by a white-space narrative may not constitute an inaccuracy at all.
8. Review the s29(5) hypothetical officer gateway. Even if there is a valid discovery, could a diligent officer, reading only the information disclosed in the return and accompanying documents, have been aware of the potential insufficiency? If yes, the assessment is barred regardless of any culpability question.
9. Consider making a targeted disclosure application at the FTT. Apply under r5 of the Tribunal Procedure Rules for production of all internal HMRC documents recording awareness of the alleged insufficiency prior to the assessment date. This is often the most effective way to build the evidentiary record for a staleness challenge.
10. Assess the limitation interaction. If the staleness or Tooth argument succeeds and the assessment is quashed, calculate whether HMRC has any remaining time to issue a fresh assessment. In many CRS-driven cases, if the staleness argument places the operative discovery in 2021, the 6-year careless window for years prior to 2019–20 will already have expired, giving the taxpayer a full limitation benefit from winning the staleness point.

Frequently Asked Questions