Edwards v Bairstow: When You Can Challenge The Tribunal's Findings

You sat through the hearing. You told your side. The decision arrived and the First-tier Tribunal found against you. One thought keeps returning: they got the facts wrong. They believed HMRC's officer and not you. Surely you can appeal that.

This is the hardest truth in tribunal practice, so here it is plainly. You generally cannot appeal a tax tribunal decision because you disagree with how it found the facts. An appeal to the Upper Tribunal lies only on a point of law. "The tribunal believed the wrong person" is not, by itself, a point of law.

The question that decides whether your factual loss is appealable at all was settled seventy years ago, in a House of Lords case about two businessmen and a second-hand spinning plant. It is called Edwards v Bairstow, and every tax appeal from the First-tier Tribunal still runs through it. This article explains what it decided, when a factual loss really is a point of law, and—just as important—when it is not.

For the step-by-step mechanics of an Upper Tribunal appeal, see our Upper Tribunal appeal guide; this article is the companion that explains the substance of what "a point of law" actually means. For where this sits in the decision you face right now, see after your tax tribunal decision.

You Lost In The Tribunal. Can You Appeal?

Here is the pivot, and it comes early because everything else depends on it. The Upper Tribunal is not a second hearing. It will not re-read your bundle, re-hear your witnesses, or re-decide whether your account was more credible than HMRC's. A tribunal's findings of fact are, in the ordinary case, final: the judge who heard the evidence decides the facts.

So a great many genuine, deeply felt grievances are simply not appealable. "They didn't believe me" is not appealable. "I think the evidence pointed the other way" is not appealable. "A different judge might have decided differently" is not appealable. What is appealable is a narrower thing: an error of law. Edwards v Bairstow is the case that tells you when a factual loss crosses the line and becomes one.

The Bind: Appeal Lies Only On A Point Of Law

The right to appeal a First-tier Tribunal decision comes from section 11 of the Tribunals, Courts and Enforcement Act 2007. It is deliberately narrow. The right lies "on any point of law arising from a decision made by the First-tier Tribunal other than an excluded decision" (s.11(1)), and you can exercise it only with permission from the First-tier Tribunal or the Upper Tribunal (s.11(2)–(3)).

Two limits in one sentence. First, your appeal must raise a point of law—not a complaint about the facts. Second, you cannot just lodge it; you must persuade a tribunal to let you appeal, by granting permission.

Do not confuse this deadline with the one you have already met. When HMRC first sent you an appealable decision, you had 30 days to appeal it to the tribunal. That stage is over. This is a different stage with a different clock: to apply for permission to appeal a First-tier Tribunal decision to the Upper Tribunal, you generally have 56 days from the date the tribunal sends you its decision with full written reasons. You apply to the First-tier Tribunal for permission first; only if it refuses do you renew the application to the Upper Tribunal itself.

There is a trap inside that sentence. If the tribunal only gave you a summary decision, the 56-day clock has not started—but you cannot simply wait. You must request full written reasons, and that request has its own, much shorter deadline: since 2 March 2026 it is 14 days from the date the summary decision was sent, even though GOV.UK guidance may still say 28 days. Miss that short window and you can lose the right to the very reasons your appeal depends on. So the first thing to do after a loss is check what you are actually holding—a summary decision or full written reasons—and, if it is only a summary, request the full reasons immediately. The mechanics of the 56-day window, and the written-reasons request and its 14-day trap, are covered in our Upper Tribunal appeal guide and our guide to the right to written reasons.

The First-tier Tribunal hearing itself cost £0 to bring, and in most cases each side there bears its own costs whatever the result. The Upper Tribunal is different, and this is part of what "higher-stakes" means: it can order an unsuccessful party to pay the other side's costs. Appealing a loss is therefore not cost-free—if you appeal and lose, you may end up paying some of HMRC's costs as well as your own. That downside is worth weighing before you commit; how the two costs regimes differ is explained in our guide to tribunal tracks and costs. The gateway to the Upper Tribunal is the point-of-law requirement. So the practical question for every losing appellant is the one Edwards v Bairstow answers: when is a loss that feels factual actually a point of law?

The Spinning Plant: What Happened In Edwards v Bairstow

The facts are almost quaint, which is part of why the case has lasted. Two businessmen, Bairstow and Harrison—a company director and a manager, neither in the spinning trade—jointly bought a complete spinning plant for about £12,000. They were not buying it to use: they intended to sell it on at a profit, and had made enquiries about reselling it before they even bought it.

They could not sell the plant as a whole, so they sold it off in lots—one week after buying it, the main "Botany" section alone fetched £15,000, and the rest went in further lots over the following months. The accounts annexed to the Case showed a profit of £18,225 on the whole transaction. The tax question was simple to state: was this single, isolated transaction "an adventure in the nature of trade", so that the profit was taxable as trading income? Or was it a one-off capital realisation, outside the tax net?

The fact-finding body was the General Commissioners—the old lay tribunal that decided tax disputes before the modern system. They found it was not an adventure in the nature of trade and discharged the assessments. The dispute then went up by a procedure called a "Case Stated", in which the Commissioners set out their findings and the question of law for the higher court.

That machinery is gone, and you need to know it is gone. There is no "Case Stated" any more. The General Commissioners were abolished in the 2009 tribunal reforms. Today, the fact-finder is the First-tier Tribunal (Tax Chamber), and an appeal goes to the Upper Tribunal on a point of law under s.11 TCEA 2007. When you read Edwards v Bairstow, mentally translate "the Commissioners" into "the First-tier Tribunal" and "Case Stated" into "appeal on a point of law". The principle survived the machinery intact.

The High Court refused to interfere, the judge considering himself bound to treat the question as one of pure fact. The Court of Appeal also dismissed the Crown's appeal. Then, on 25 July 1955, the House of Lords reversed. It held that the Commissioners' finding was wrong in law: on the primary facts they themselves had found (the basic facts the tribunal accepted as proved—what was bought, when, for how much, what was done with it), the only reasonable conclusion was that this was an adventure in the nature of trade and the profit was taxable. A finding that had looked like an unappealable fact had become an appealable point of law. That conversion is the whole of Edwards v Bairstow.

Lord Radcliffe's Test

The enduring passage is Lord Radcliffe's. It explains exactly when a court (today, the Upper Tribunal) must step in even though the question looks factual:

"When the Case comes before the Court, it is its duty to examine the determination having regard to its knowledge of the relevant law. If the Case contains anything ex facie which is bad law and which bears upon the determination, it is, obviously, erroneous in point of law. But, without any such misconception appearing ex facie, it may be that the facts found are such that no person acting judicially and properly instructed as to the relevant law could have come to the determination under appeal. In those circumstances, too, the court must intervene. It has no option but to assume that there has been some misconception of the law and that this has been responsible for the determination. So there, too, there has been error in point of law. I do not think that it much matters whether this state of affairs is described as one in which there is no evidence to support the determination or as one in which the evidence is inconsistent with and contradictory of the determination, or as one in which the true and only reasonable conclusion contradicts the determination."

Read it slowly, because three things matter.

First, there are two distinct ways a determination can be wrong in law. One is obvious: the tribunal said something that is "bad law" on the face of its decision—it misstated the legal test. The other is subtler: the tribunal did not visibly misstate the law, but its conclusion is one that "no person acting judicially and properly instructed as to the relevant law could have come to". When that happens, the court must assume the tribunal misunderstood the law somewhere, even if you cannot point to the sentence where it did. That second route is how a factual-looking loss becomes a point of law.

Second, Lord Radcliffe offers three ways of describing the same situation: no evidence to support the determination; evidence inconsistent with and contradictory of it; or the true and only reasonable conclusion contradicting it. These are not three separate tests but three angles on one idea—that the conclusion was not rationally open on the facts found. The phrasing practitioners use most often is the last: "the true and only reasonable conclusion contradicts the determination."

Third, a point of vocabulary. Later courts call this the "perversity" or "irrationality" ground, and you will see Edwards v Bairstow described as its source. That is right—but be precise about who said what. Lord Radcliffe's own wording is "no person acting judicially and properly instructed", not "perverse". The word "perverse" entered the Edwards v Bairstow speeches through Viscount Simonds, who quoted the Lord President in the earlier Scottish case Fraser v Inland Revenue (1942 SC 493) to the effect that a court may interfere where the tribunal has made a perverse finding without evidence to support it. He reached the same result as Lord Radcliffe. The label "perversity" is convenient shorthand; the operative wording is Lord Radcliffe's.

When A Factual Loss IS A Point Of Law

The courts have distilled Edwards v Bairstow and the cases after it into a recognised list of errors of law. Each is a way a factual-looking loss can still be appealable, with a plain-English example.

1. No evidence to support a finding. The tribunal found a primary fact for which there was no evidence at all. Example: the decision states "the taxpayer was warned by his accountant in March", but no witness said it and no document shows it.

2. A conclusion no reasonable tribunal could reach. The core Edwards v Bairstow ground: "the true and only reasonable conclusion contradicts the determination." This is the irrationality (or "perversity") limb. Example: the tribunal accepts every primary fact pointing to a trade—bought to resell, advertised before purchase, sold at a profit within weeks—then concludes it was not trading, with no rational route from those facts to that conclusion.

3. The wrong legal test, or a misdirection of law. The tribunal applied the wrong legal standard, or misunderstood the statute, even if its fact-finding was otherwise careful. This is Lord Radcliffe's "anything ex facie which is bad law". Example: the tribunal asks "was the taxpayer careless?" when the statute requires "deliberate". In Brady v Group Lotus Car Companies plc [1987] STC 635, a factual discharge of assessments was set aside because the Commissioners had misdirected themselves on the burden of proof.

4. Taking irrelevant matters into account, or ignoring relevant ones. The tribunal weighed something the law says is irrelevant, or failed to weigh something it says matters. Example: a penalty decision turns on the taxpayer's demeanour in cross-examination when the statutory test is whether the inaccuracy was deliberate.

5. Failure to make a material finding of fact. The tribunal did not decide an issue it had to decide to reach its conclusion, leaving a gap in the reasoning. Example: the case turned on when the taxpayer discovered the error, and the decision never finds a date or resolves the conflicting evidence.

6. Inadequate or unintelligible reasons. A decision so poorly reasoned that you cannot tell why you lost can itself be an error of law—the losing party is entitled to know, in substance, why. This overlaps heavily with our guide on the right to written reasons, which sets out the adequacy-of-reasons authorities (English v Emery Reimbold & Strick and South Bucks v Porter) in full. If you think your decision is inadequately reasoned, read that article rather than re-deriving the test here.

Notice what every one of these has in common. None is "I disagree with the result." Each identifies something specifically wrong with how the tribunal reached it—a missing evidential basis, an irrational leap, a wrong test, an irrelevant factor, a gap, or an unintelligible explanation. That distinction is the entire game, and the next section is a practical tool for drawing it.

Is My Complaint An Error Of Law, Or Just Disagreement With The Facts?

This is the question that decides whether you have an appeal at all. Most losing appellants instinctively believe they have an error of law when what they actually have is disagreement with the facts. The two feel identical from the inside. They are completely different in law.

Work through your own decision with the table below. Be honest with yourself—the tribunal will be. For each complaint you have, ask which column it falls into.

A reliable self-check: if your best version of the complaint still reduces to "the tribunal should have weighed the evidence differently and reached a different result", you are on the left. Weight of evidence is the tribunal's job, not the appeal court's. If instead you can finish the sentence "the tribunal was not entitled to reach that finding because…" with something specific—no evidence, no rational route, wrong test, missing finding—you may be on the right.

Two cautions. First, being on the right of the table does not mean you will win; it means you have something arguable. The bar, as the next section explains, is high even then. Second, the line is genuinely difficult, and even the Upper Tribunal sometimes gets it wrong. If your complaint plausibly sits on the right, that is the point to get proper advice on whether it is worth running—not the point to assume you have won.

"Get advice" is easy to say when you are unrepresented and that is the reason you are reading this. In practice the realistic options are a Chartered Tax Adviser or accountant who handles disputes, or a barrister you can instruct directly through public access without going through a solicitor. If paying for advice is not possible, the tax charities TaxAid and Tax Help for Older People assist people on low incomes, and Advocate (the Bar's pro bono charity) can sometimes find a barrister to help free of charge. None of these is guaranteed or means a particular outcome, but a short, focused conversation about whether you have an arguable point of law is usually worth far more than the time it takes to find it.

Why The Bar Is So High

The reason so few "wrong on the facts" complaints succeed is not procedural meanness. It is a deliberate choice. The First-tier Tribunal heard the witnesses and is the body Parliament charged with finding the facts. An appeal court that re-weighed the evidence would be doing the First-tier Tribunal's job, not its own.

Nowhere is this clearer than where the tribunal has made a multifactorial value judgment—a conclusion reached by weighing many factors together, none decisive. In Procter & Gamble (UK) v HMRC [2009] EWCA Civ 407—the "Pringles" VAT case, on whether Pringles were similar to potato crisps—the Court of Appeal held that an appeal court will not interfere with an evaluative judgment of that kind unless there is a genuine legal error, a perverse finding, or a misconstruction of the statutory test. It is not for the appeal court to substitute its own judgment, and the tribunal need not spell out every factor it weighed. The Supreme Court put the same principle firmly in Pendragon plc v HMRC [2015] UKSC 37: evaluative judgments are treated as findings of fact, and an appeal succeeds only if the appellant identifies an error in the legal test or shows a conclusion no reasonable tribunal could have reached. (Pendragon also holds that once the Upper Tribunal is entitled to intervene, it may re-make the decision and find facts itself under s.12 TCEA 2007—but that bites only after an error of law is established. It does not lower the gateway.)

For a live example, see our companion analysis of PGMOL v HMRC. Employment status is decided by weighing many factors together—exactly the multifactorial value judgment Procter & Gamble protects—which is why employment-status appeals so rarely succeed on the facts.

The restraint cuts both ways. In A Taxpayer v HMRC [2025] EWCA Civ 106, the First-tier Tribunal found "exceptional circumstances" prevented the taxpayer leaving the UK; the Upper Tribunal overturned that and re-made the decision for HMRC; the Court of Appeal then restored the First-tier Tribunal, holding the Upper Tribunal had wrongly interfered with an evaluative finding properly open to it. The same high bar that makes it hard for you to overturn a finding also protects findings made in your favour—and even the Upper Tribunal can wrongly cross the line.

How To Actually Argue It: The Georgiou Discipline

Suppose you genuinely have a point of law. There is a right way and a wrong way to plead it, and the wrong way gets unrepresented appeals struck out.

The right way comes from Georgiou (t/a Marios Chippery) v Customs & Excise Commissioners [1996] STC 463, where at page 476 Evans LJ set out the four steps an appellant on an Edwards v Bairstow ground must satisfy:

1. Identify the finding that is challenged.
2. Show it is significant in relation to the tribunal's conclusion—that it actually mattered to the result.
3. Identify the evidence, if any, that was relevant to that finding.
4. Show that, on that evidence, the finding was one the tribunal was not entitled to make.

That is a disciplined, surgical exercise: you point at one finding, show why it mattered, show what evidence bore on it, and show the tribunal could not properly have made it on that evidence.

Evans LJ also identified the trap in the same passage. What is not permitted is "a roving selection of evidence coupled with a general assertion that the tribunal's conclusion was against the weight of the evidence and was therefore wrong." Later courts call this "island hopping"—jumping from one favourable piece of evidence to another across a sea of evidence, hoping the accumulation looks like an error of law. It does not. It looks like what it is: disagreement with the facts dressed up as a point of law. It is the single most common reason unrepresented Edwards v Bairstow appeals fail before they are properly heard.

So when you draft your grounds, resist the urge to re-argue the whole case. Pick the specific finding and apply the four Georgiou steps to it, one at a time, in writing. Our guide to writing grounds of appeal walks through how to structure an error-of-law ground in practice—use the Georgiou discipline as its backbone.

Don't Confuse The Remedies

Three different procedures get conflated constantly. They do completely different jobs, and using the wrong one wastes one of your few chances—and your time.

The single most important distinction: Rule 38 set-aside is not how you challenge the tribunal's reasoning. Set-aside is for when the process miscarried—you never got the bundle, you weren't told about the hearing, an administrative error denied you a fair process. It is not for "the tribunal weighed the evidence wrongly" or "the tribunal applied the wrong test". Those are errors of law, and they go to the Upper Tribunal under s.11—the Edwards v Bairstow route—not back to the First-tier Tribunal under Rule 38.

Note the numbers, because the confusion is common: set-aside is Rule 38, permission to appeal is Rule 39, review is Rule 41—different rules doing different jobs. If you file a Rule 38 set-aside application when what you actually have is an Edwards v Bairstow error of law, you can burn weeks on the wrong remedy and find your 56-day permission window has closed in the meantime.

The Real Lesson: Win It In The First-tier Tribunal

Here is the conclusion the cases drive towards, and it is not the one most people want to hear. The door to the Upper Tribunal is narrow by design. Edwards v Bairstow opens it only a crack—for genuine errors of law, not for factual disappointment. The realistic place to win your case is the First-tier Tribunal, the first time. That changes what you do before and during the hearing, not just after it.

Get your evidence in. The Upper Tribunal decides on the facts the First-tier Tribunal found; it will not hear evidence you could have brought but did not. If a document or a witness matters, it has to be in front of the First-tier Tribunal. Our guide to preparing for your tax tribunal hearing covers how to build that record.

Get the findings made. An Edwards v Bairstow point often depends on a finding the tribunal did—or did not—make. Put the issues that matter squarely, so the tribunal has to decide them. A missing finding can be an error of law, but a clear finding in your favour is far better than the argument that one is missing.

Ask for full written reasons. You cannot run an error-of-law appeal—least of all an inadequate-reasons one—without them, and the 56-day permission clock runs from when they are sent. Our guide to the right to written reasons explains how and when to request them, and why a summary decision is not enough.

None of this is a guarantee. It is the honest position. Most "they got the facts wrong" instincts do not survive contact with s.11 and Edwards v Bairstow—not because the system is rigged, but because Parliament gave the fact-finding job to the First-tier Tribunal and meant it. Knowing that before your hearing is worth more than discovering it after. For where this fits in the wider arc of an HMRC dispute, see our tax dispute timeline.

If you are reading this after a loss, the prevention advice has sailed—so here is the order to work in now:

1. Check what you are holding. Is it a summary decision or full written reasons? If only a summary, request full written reasons immediately—that request has its own short deadline (14 days from the summary decision since 2 March 2026).
2. Diagnose honestly. Run your grievance through the self-test table above. Be ruthless: "the judge believed them, not me" is not a point of law.
3. Get focused advice on the arguable points—from a Chartered Tax Adviser, a direct-access barrister, or one of the free routes above—before the deadline, not after.
4. Mind the clock. Permission to appeal is generally 56 days from the date full written reasons are sent; you apply to the First-tier Tribunal first.

Key Legislation And Resources

The Judgment

• Edwards (Inspector of Taxes) v Bairstow and Harrison [1955] UKHL 3, 36 TC 207, [1956] AC 14 — House of Lords, 25 July 1955 (BAILII). The "1956" is the report year, while BAILII dates by decision year, hence the 1955 neutral citation.

Legislation

• Section 11, Tribunals, Courts and Enforcement Act 2007 — right of appeal to the Upper Tribunal on a point of law, with permission
• Section 12, Tribunals, Courts and Enforcement Act 2007 — Upper Tribunal disposal powers (set aside, remit, or re-make)
• Rule 38, Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009 — setting aside a decision for procedural irregularity (28 days)
• Rule 39, Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009 — application for permission to appeal
• Rule 41, Tribunal Procedure (First-tier Tribunal) (Tax Chamber) Rules 2009 — review of a decision (narrow self-correction)

Key Cases

• Georgiou (t/a Marios Chippery) v Customs & Excise Commissioners [1996] STC 463 (at 476) — Evans LJ's four-step discipline and the "roving selection" / "island hopping" prohibition; reported at [1996] STC 463, not on BAILII
• Procter & Gamble (UK) v HMRC [2009] EWCA Civ 407 — the "Pringles" case; appellate restraint over multifactorial value judgments
• Pendragon plc v HMRC [2015] UKSC 37 — Supreme Court restatement of the orthodox Edwards v Bairstow standard
• A Taxpayer v HMRC [2025] EWCA Civ 106 — the Court of Appeal restored a First-tier Tribunal evaluative finding the Upper Tribunal had wrongly disturbed; restraint protects findings both ways
• Brady (Inspector of Taxes) v Group Lotus Car Companies plc [1987] STC 635 — misdirection on the burden of proof as an error of law; reported at [1987] STC 635, not on BAILII

GOV.UK Guidance

• Appeal to the Upper Tribunal (Tax and Chancery) — the official route and the point-of-law requirement
• Tax tribunal: appeal against a tax decision — the First-tier Tribunal starting point
• HMRC Appeals, Reviews and Tribunals Guidance (ARTG) — HMRC's internal manual on the Upper Tribunal appeal route

On This Site

• Upper Tribunal appeal — the procedural companion: permission, forms, and the 6-step process
• After your tax tribunal decision — the decision point: should you appeal at all?
• Writing grounds of appeal — how to plead an error of law using the Georgiou discipline
• Right to written reasons — you need full reasons to find the error; inadequate reasons is itself a ground
• Preparing for your tax tribunal hearing — win it in the First-tier Tribunal, because appeal rights are narrow
• Tax dispute timeline — where a tribunal appeal sits in the broader HMRC journey
• Tooth v HMRC: deliberate inaccuracy — companion case analysis (a substantive test)
• Perrin v HMRC: reasonable excuse — companion case analysis (a substantive test)
• Martland v HMRC: late appeals — companion case analysis (a substantive test)
• BPP Holdings v HMRC: tribunal directions — companion case analysis (a substantive test)
• PGMOL v HMRC: employment status — companion case analysis; a live example of a multifactorial evaluative judgment the higher courts will not re-open

This article is for informational purposes only and does not constitute legal or tax advice. For advice specific to your situation, consult a qualified tax adviser, accountant, or solicitor.