Every argument that starts “but HMRC’s guidance says…” is measured against this case. The representation must be clear, unambiguous and devoid of relevant qualification, read as a whole by an ordinarily sophisticated taxpayer, not assembled from a favourable sentence.
On this page
Full name: R (on the application of Davies and another) v HMRC; R (on the application of Gaines-Cooper) v HMRC
Citation: [2011] UKSC 47; [2011] 1 WLR 2625
Court: Supreme Court (Lord Wilson giving the majority judgment; Lord Mance dissenting)
Judgment: 19 October 2011
Subject: Legitimate expectation arising from HMRC published guidance; residence under the pre-2013 rules
Result: Both appeals dismissed. IR20 did not give rise to the expectation claimed, and in any event the taxpayers had not satisfied what it required.
Why This Case Still Matters
Residence itself is now governed by the Statutory Residence Test, which applies from 6 April 2013 and replaced the case law and guidance these appeals were about. The residence analysis in the case is therefore of historical interest for most purposes.
What survives, and what makes it one of the most cited authorities in HMRC disputes, is the standard for establishing a legitimate expectation from HMRC’s published guidance. Every argument that begins “but HMRC’s manual says…” or “the guidance told us…” is measured against this decision.
The Facts
Two sets of appeals were heard together.
Davies and James left the United Kingdom to work on a business venture in Belgium. They maintained connections with the United Kingdom and returned regularly.
Mr Gaines-Cooper had a home in the Seychelles and substantial connections there, but retained a property in England and a pattern of visits and involvement in English life which HMRC said meant he had never left.
All three relied on HMRC’s booklet IR20, the published guidance then in force on residence and ordinary residence. They contended that IR20 set out a day-counting approach which, if followed, would result in their being treated as non-resident, and that HMRC could not lawfully depart from it.
The claims were brought by judicial review rather than by statutory appeal, because a legitimate expectation argument of this kind lay outside the tribunal’s jurisdiction, the same forum problem addressed in Oxfam v HMRC.
The Ratio Decidendi
The three elements of the test
- Clear, unambiguous and devoid of relevant qualification. This formulation, drawn from the established public law authorities, is exacting. Guidance which is hedged, general, or expressly subject to the facts of the case will not do.
- Read as a whole. The court will not allow a taxpayer to isolate a helpful passage. The document is construed in its entirety, including its caveats and its statements of purpose.
- The ordinarily sophisticated taxpayer. The addressee is not a lawyer, but nor is the guidance read as a person might skim it. The standard assumes a reader who engages with the document seriously.
Lord Mance’s dissent
Lord Mance considered that IR20 was not sufficiently clear to convey the distinct break requirement to the reader the majority had in mind, and that taxpayers who followed what the document appeared to say should not be penalised for the department’s imprecision. The dissent is not binding, but it is the argument to deploy where guidance is genuinely opaque, and it is a reminder that the majority position was not inevitable.
Obiter and Related Points
- The observations on the drafting of guidance. Comments about the difficulty HMRC faces in producing guidance that is both accessible and accurate are commentary. They cut both ways in argument.
- Whether guidance can ever bind HMRC. The case confirms that it can in principle. The failure was evidential and interpretive, not conceptual. That is important: the answer to a legitimate expectation argument is not that guidance never binds.
- The residence analysis is superseded by the Statutory Residence Test for years from 2013–14 onwards. It remains relevant only to very old years and to the interpretation of the pre-2013 concepts where they still arise.
- The relationship with Preston. The Supreme Court did not disturb the principle that conduct amounting to an abuse of power can found relief. It held that the threshold had not been crossed.
Practitioner Application
Running a guidance-based argument
- Quote the whole of the relevant passage, and the surrounding text. If you have to extract a sentence to make the point, the argument will fail. Test it by reading the section as HMRC will read it back to you.
- Identify the qualification and deal with it. Most guidance contains a caveat. Explain why it does not bite, rather than hoping it is not noticed.
- Establish reliance with evidence. What did the client actually read, when, and what did they do differently as a result? Contemporaneous file notes and correspondence.
- Consider whether a specific assurance exists. A ruling, a clearance or an individual officer’s written confirmation is a far stronger foundation than a manual, because it is more likely to be clear, unambiguous and unqualified.
- Show the unfairness. Following Preston, the threshold is conduct amounting to an abuse of power, not mere inconsistency.
Getting the forum right
For residence disputes today
Residence is now a mechanical statutory test based on day counts, ties and automatic tests. The arguments are evidential rather than interpretive: where the taxpayer was, on how many days, and what connections they retained. Our resource on the statutory residence test covers the current framework. Gaines-Cooper matters now for what it says about guidance, not about residence.
Frequently Asked Questions
Can I hold HMRC to its published guidance?
Only if the representation is clear, unambiguous and devoid of relevant qualification, judged by how the document would be understood by an ordinarily sophisticated taxpayer reading it as a whole. Gaines-Cooper confirms guidance can bind HMRC in principle, but the standard is exacting and extracting a favourable sentence from a longer document will not do.
Why did the taxpayers lose?
Because IR20, read as a whole, required a taxpayer leaving the United Kingdom to make a distinct break in the pattern of their life here. It did not represent that day counting alone would secure non-resident status. The expectation they claimed was not one the document conveyed, and in any event they had not shown they satisfied what it did require.
Is the residence analysis still relevant?
Largely not. The Statutory Residence Test applies from 6 April 2013 and replaced the case law and guidance these appeals concerned. Residence is now a mechanical test of day counts, ties and automatic tests, so the arguments are evidential rather than interpretive. The case survives on the legitimate expectation point, not on residence.
What is stronger than a manual for founding an expectation?
A specific assurance directed at the taxpayer: a ruling, a clearance, or an individual officer’s written confirmation. Such a statement is far more likely to be clear, unambiguous and devoid of relevant qualification than a passage in general guidance, and it is much easier to evidence reliance on it.
Should I run this in the tribunal or by judicial review?
Generally judicial review for a direct tax dispute, which is why these appeals took that route. The tribunal usually cannot entertain a legitimate expectation argument in a direct tax appeal. Judicial review has a three-month promptness requirement, so advisers who try the tribunal first and lose on jurisdiction are frequently out of time. Decide at the outset and issue protectively if there is real doubt.