SDLT discovery: HMRC's evidence bar for the 6 and 20-year windows

When HMRC issues an SDLT discovery assessment outside the standard four-year window, the entire assessment turns on whether HMRC can demonstrate that the lost tax was brought about carelessly (the six-year window) or deliberately (the twenty-year window). The thresholds sound impressionistic. In practice they are statutory tests with a defined evidence bar — and the burden falls on HMRC.

This article walks through the statutory framework in paragraphs 28–32 of Schedule 10 to the Finance Act 2003, the HMRC published guidance at SDLTM56020 and the Compliance Handbook (CH81120 on carelessness, CH53000 on deliberate behaviour), and the tribunal authorities that set what HMRC has to put in front of the First-tier Tribunal to discharge that burden. It is a case-law summary, not advice — speak to a qualified adviser before responding to any HMRC correspondence.

The three windows in Schedule 10

Schedule 10 to FA 2003 sets out three discovery-assessment time limits, stacked on the behaviour of the taxpayer or their agent.

ProvisionWindowTrigger
Para 30(1)4 years from filingStandard cap
Para 30(2)6 years from filingLoss of tax brought about carelessly by the taxpayer or a person acting on their behalf
Para 30(3)20 years from filingLoss of tax brought about deliberately by the taxpayer or a person acting on their behalf

Two structural points carry through everything that follows. First, the windows are gated by behaviour, not by HMRC's diligence in spotting the issue. Second, "a person acting on behalf of the taxpayer" expressly captures the conveyancer who prepared the SDLT return — the behaviour of the agent is attributed to the taxpayer for time-limit purposes.

For the upstream rules on how HMRC opens an SDLT enquiry within the in-window period, see the SDLT enquiry windows and HMRC compliance letters guide.

What HMRC has to evidence — the threshold question

Before the carelessness/deliberate axis is reached at all, HMRC must clear the discovery threshold itself. The leading authority is Hicks v HMRC [2020] UKUT 12 (TCC), which held that a "discovery" requires a HMRC officer to newly form the view that there is an insufficiency. Re-reading the same papers that were available at the time of the original enquiry does not count. Practically, HMRC has to be able to point to a contemporaneous file note recording when the discovery was made and what new information or analysis prompted it.

If HMRC clears the Hicks threshold, the time-limit then turns on behaviour.

The carelessness bar (6-year window)

Carelessness is defined in paragraph 31A of Schedule 10 by cross-reference to paragraph 3 of Schedule 24 to the Finance Act 2007: a person acts carelessly if they fail to take the reasonable care that a reasonable person would take in the circumstances. HMRC's interpretive guidance is in CH81120, and the SDLT-specific application is at SDLTM56020.

The reasonable-care test is fact-sensitive but the case-law and HMRC guidance converge on a small number of operational markers. HMRC routinely relies on:

  • The signed SDLT1 itself, including the declaration that the return is correct and complete.
  • Correspondence with the conveyancer at the time of the transaction — disclosure of the relevant facts (a second property, a non-resident partner, a linked transaction) by the taxpayer to the agent shifts where the carelessness sits but does not extinguish it for time-limit purposes.
  • The conveyancer's contemporaneous file — letters of engagement, fact-find questionnaires, completion statements and any specific SDLT advice memos. Where these show the agent was on notice of the relevant fact and either misapplied the rule or failed to ask further, the carelessness analysis is straightforward.
  • HMRC's own pre-completion guidance — gov.uk pages, the SDLT manual entry in force at the date of the transaction, and any flowcharts an ordinarily careful conveyancer would be expected to have consulted.

What HMRC cannot rely on, on its own, is a bare statement that the return was wrong. The tribunal in Anderson v HMRC [2018] UKFTT 200 (TC) and the cluster of FTT decisions following it underline that a carelessness finding requires HMRC to articulate what the taxpayer or agent ought to have done differently and why, by reference to a defined standard. A "you should have known" finding without that articulation will not stand on appeal.

The deliberate bar (20-year window)

Deliberate behaviour sets a markedly higher bar. The Supreme Court in Tooth v HMRC [2021] UKSC 17 held that a "deliberate inaccuracy" requires HMRC to demonstrate that the taxpayer (or agent) made a statement knowing it was inaccurate — an intentional, not merely an attributable, act. See the Tooth v HMRC discovery deliberate-test guide for the case in full.

After Tooth, the evidence HMRC needs to discharge the deliberate bar is qualitatively different from the carelessness bar. The kinds of material that have supported deliberate findings in published tribunal decisions include:

  • Contemporaneous internal correspondence showing the taxpayer or agent recognised the SDLT position and elected to file inconsistently with it.
  • Documents prepared for a separate purpose — for example a mortgage application or a tax-planning memorandum — that record the inconsistent position in writing.
  • A pattern of returns treating substantively identical transactions inconsistently in the same period, with no explanation.
  • A failure to disclose a known fact that was specifically asked about on the SDLT1 (for example the higher-rates question on a second-property purchase).

A negligent omission, a misunderstood ruling, or an aggressive but stateable interpretation will not, on the Tooth test, support a deliberate finding. The 20-year window is reserved for cases where the inaccuracy was put there on purpose.

The Auld test in the conveyancer's hands

Where the substantive return was prepared by a conveyancer, the behaviour of the conveyancer is attributed to the taxpayer under the "person acting on behalf" wording. The Auld test (after Auld v HMRC and its progeny in the discovery line) frames the question as: did the conveyancer take reasonable care, judged against the standard of a competent conveyancer holding themselves out to handle SDLT? It is not the standard of a tax specialist.

Two consequences follow. First, taxpayers who supplied the relevant facts to a conveyancer in writing typically defeat a carelessness analysis at the taxpayer-attribution level, even where the conveyancer's own behaviour falls inside the carelessness bar. Second, an inaccuracy that was the product of an aggressive interpretation by a specialist tax adviser (as distinct from a high-street conveyancer) is judged against a higher standard of care.

Penalty and discovery interact, but they are not the same test

A discovery assessment in the 6- or 20-year window can stand without any penalty assessment issuing — and a penalty assessment can issue without a discovery assessment, where the in-enquiry route has been used. The behaviour standards are the same words (paragraphs 30(2) and 30(3) of Schedule 10 mirror Schedule 24 FA 2007 for penalty purposes), but they are pleaded separately.

In practice, an HMRC notice that pleads carelessness will normally be accompanied by a Schedule 24 paragraph-3 penalty assessment in the 0–30% band (15% prompted floor for unprompted, 30% maximum). A deliberate-but-not-concealed finding lifts that to 20–70%, and deliberate-and-concealed to 30–100%. The mechanics of the prompted/unprompted disclosure distinction sit in the CC/FS7a prompted vs unprompted disclosure guide.

How HMRC gathers the evidence

The evidence cited above is not handed to HMRC voluntarily. It is gathered, in most discovery-assessment cases, through Schedule 36 (FA 2008) information notices to the taxpayer and to the conveyancer as a third party. The conveyancer's contemporaneous file is the single most important evidentiary asset for both sides. The penalty consequences of failing to respond to a Schedule 36 notice are covered in the Schedule 36 information notice penalties guide.

For the routes by which an in-window enquiry can be partially closed before progressing to a discovery assessment, see the SDLT partial closure notice guide.

Volume context

Across 2025, HM Land Registry recorded 848,775 residential property transactions in England and Wales (Land Registry pp-complete dataset, queried 1 June 2026). Of these, 488,720 transactions cleared the £250,000 SDLT-return threshold above which the slab structure begins to bite, and 148,392 cleared £500,000 — the band where multi-strand SDLT analyses (residency, additional-property, chattels apportionment, linked transactions, FTB relief at the lower end) most often produce the kinds of inaccuracies a discovery assessment would later allege.

A second-property SDLT charge at £500,000 is £40,000 at standard rates (HMRC published rates, 2026); the FTB-relief contestable amount at £425,000 is £11,250. The dispute economics make the discovery-window analysis material on a per-case basis even where the headline absolute is modest. Use the stamp duty calculator to model the contested amounts for a specific transaction.

Quick reference

QuestionWhere it sits
Has there been a "discovery" at all?Hicks v HMRC [2020] UKUT 12 (TCC); HMRC file note
Was the inaccuracy "careless"?Sch 10 FA 2003 para 30(2); CH81120; the Auld test
Was the inaccuracy "deliberate"?Sch 10 FA 2003 para 30(3); Tooth v HMRC [2021] UKSC 17; CH53000
Whose behaviour matters?Taxpayer plus any person acting on their behalf (Sch 10 para 30)
What evidence does HMRC need?Contemporaneous file (taxpayer + conveyancer), HMRC published guidance in force at the date of the transaction, the SDLT1 itself
Companion penalty regimeSchedule 24 FA 2007 (paragraphs 3, 4 and 5)

Wider context on the Homecost blog

Speak to a qualified adviser before responding to any HMRC correspondence or making any filing decision. This article is general information about the statutory and case-law framework, not advice.