In HMRC v Mr Raymond Tooth [2021] UKSC 17, the Supreme Court held that (1) a “deliberate inaccuracy” means an inaccuracy intended to mislead (not an inaccuracy known to be wrong with no intention to mislead) and (2) in interpreting a tax return, the whole of the tax return must be scrutinised, in the face of HMRCs’ submission that an HMRC computer would adopt what the Supreme Court called “tunnel vision” and look at each entry without any regard to an explanation of that entry elsewhere in the tax return. The Supreme Court also made important (obiter) observations on the notion of “staleness”, whereby an otherwise in-time discovery assessment would be made bad because HMRC had taken too long between making a discovery and issuing an assessment. The Supreme Court thought the notion of “staleness” to be bad, although the Court acknowledged that this issue would have to be dealt with finally in a case where it was actually in issue.
The taxpayer, Mr Tooth had entered an employment loss as a partnership loss because the software which applied to online tax returns did not permit the loss to be entered as an employment loss. Mr Tooth had explained the loss to be an employment loss in the “white space disclosure” in his return. HMRC had not sought to attack the eligibility loss for income tax relief within the normal time limits. Rather, HMRC issued a “discovery assessment” on Mr Tooth on the basis that HMRC had “discovered” an insufficiency of tax in Mr Toth’s return and because the loss was an employment loss, not a partnership loss (which Mr Tooth knew), he had entered a “deliberate inaccuracy” in his return, which attracted a 20 year time limit for issuing assessments (rather than the usual 4 year time limit).
The Supreme Court (reversing the Court of Appeal) considered that HMRC had, on the evidence, made a “discovery” that Mr Tooth’s self assessment was incorrect because the loss entered as a partnership loss was, in fact, an employment loss. The Court went onto observe that, had there been a long period of time between the discovery and the issue of a discovery assessment, “staleness” would not have made the assessment bad if the assessment was otherwise within the statutory time limits. However, the Court made it clear that its observations were obiter.
The Supreme Court also reversed the Court of Appeal in holding that the entry of an employment loss into a partnership loss box in a self assessment return could not be described as an “inaccuracy” at all (because the loss was clearly explained to be an employment loss in the “white space” disclosure box elsewhere in the return). And importantly, any alleged inaccuracy could not be described as a “deliberate” inaccuracy unless there was a n intention to mislead HMRC, which there was not in this case. The term “deliberate inaccuracy” meant an inaccuracy which was intended to mislead (or perhaps, without deciding the point, reckless).
And finally, in rejecting what the Court described as a “very unattractive” submission by HMRC, the Supreme Court held that, in common with other documents, a tax return must be interpreted as a whole. HMRC had submitted that there was an “inaccuracy” in the taxpayer’s return because the loss had been entered as a partnership loss. HMRC had submitted that because a computer had read the return, rather than “a sentient, literate human being”, the computer did not (and could not be expected to) read the document as a whole. The Supreme Court strongly rejected this HMRC submission and held that the tax return must be assumed to be read by a reader who was capable of reading the document as a whole and interpret all information within the document on that basis.
Julian Ghosh QC acted for Mr Raymond Tooth, instructed by Pinsent Masons LLP.