WM Morrison Supermarkets PLC v The Commissioners for HMRC

The appellant, the well-known supermarket, Morrisons, (“Morrisons”) appeals against a decision of the First-tier Tribunal (“FTT Decision”) published as WM Morrison Supermarkets PLC v HMRC [2021] UKFTT 106 (TC). The FTT rejected Morrisons’ argument that certain products it sold (Organix Bars and Nakd Bars) were zero-rated as “food of a kind used for human consumption” (under Group 1 Schedule 8 Value Added Tax Act 1994 (“VATA 1994”). Instead, the FTT agreed with HMRC that the products fell within the exception the legislation provided to the zero-rating of such food for “confectionery in Item 2 of Group 1 of Schedule 8 of VATA 1994.

Although Morrisons’ position was that the Organix and Nakd bars were zero-rated, it had, when it sold the products on to customers, accounted to HMRC for standard rate VAT at 20%. It thus sought repayment of that output VAT (just over £1 million in relation to the Nakd bars in the period October 2014 to July 2018 and £97,000 in relation to Organix bars in the period October 2013 to July 2017).

The FTT upheld HMRC’s decisions that the products bore standard rate VAT and also HMRC’s refusal to repay the amounts in respect of output VAT Morrisons sought. With the permission of the Upper Tribunal, Morrisons now appeals to the Upper Tribunal against the FTT’s Decision on the grounds the FTT, in analysing whether the products were “confectionery” wrongly treated certain factors as irrelevant, namely: (1) (i) the actual or perceived healthiness of the products and/or (ii) the products’ marketing as such (2) the absence of cane sugar, butter and flour (being ingredients associated with traditional confectionery).

Section 30(2) VATA 1994 provides for zero-rating of goods described within Schedule 8, Group 1 of which includes “Food of a kind used for human consumption”, along with a list of excepted items and interpretative notes. Item 2 is the relevant exception for the purposes of this appeal. This excepts from zero-rating:

“Confectionery, not including cakes or biscuits other than biscuits wholly or partly covered with chocolate or some product similar in taste and appearance”.

Note 5 to the list of excepted items provides:

“for the purposes of item 2 of the excepted items ‘confectionery’ includes chocolates, sweets and biscuits; drained, glacé or crystallised fruits; and any item of sweetened prepared food which is normally eaten with the fingers”.

Unless otherwise stated paragraph references are to those in the FTT decision. The relevant products Morrisons sold were two different types of bars manufactured by Organix (Organix Carrot Cake Soft Oaty bar, Organix Banana Soft Oaty bar) and 18 types of bars manufactured by Nakd (listed in full at [12] to [15]) which could be broken down in the following three categories: (1) “Fruit and Nut bars” (such as “Nakd Cashew Cookie Wholefood Bar” and “Nakd Cocoa Orange Wholefood Bar”, (2) “Oaties” (such as “Nakd Apple Pie Wholefood Bar” and “Nakd Banana Bread Wholefood Bar”) and (3) “Crunchies” (“Nakd Banana Crunch Wholefood Bar” and “Nakd Strawberry Crunch Wholefood Bar” ).

The FTT concluded all were “confectionery”. They therefore fell within the exception to zero-rating and thus were standard-rated for VAT purposes. The FTT went on to consider, and reject, the appellants’ argument, in respect of both the Organix and Nakd bar products, that even if they were confectionery, they were zero-rated because they were cakes. The appellants do not pursue any such alternative argument before us.

There is no dispute that the general approach the FTT adopted, as to whether the products were confectionery, was correct. That was to conduct a multi-factorial evaluation of the various attributes of the products, and to come to a view on whether the ordinary person on the street would regard a given product as confectionery. The FTT did not seek to define the term “confectionery” but on the basis of the wording in Note 5, and on the discussion of the meaning of “confectionery” in the case-law, it held that all confectionery had the characteristics of sweetness, and that it was “subject to a process” (any process of mixing or compounding would in principle be sufficient” ([171]). It held the ordinary person would also consider that confectionery had the characteristics that it was (1) normally eaten with the fingers and (2) held out to be eaten as a treat or snack ([172]). Again, neither party suggests the FTT’s analysis was wrong in these respects.

There is also no challenge to the FTT’s underlying findings of fact. The FTT made these findings from a variety of evidence before it: documentary evidence (including lists of ingredients, website marketing for certain products), the witness statements of the appellant’s witnesses (a Food Developer at Organix, the appellant’s buying manager for “Free from” products, and a customer manager for the appellant’s “Customer Insights” team) ([78]). The tribunal was also provided with samples of some of the products and tasted them.

As mentioned in the introduction, the appellant’s appeal is on the basis that the FTT wrongly failed to consider two particular factors ((1) healthiness, and the products’ marketing with regard to healthiness (2) comparison with ingredients of cane sugar flour and butter in traditional confectionery) which were, in the appellant’s submission relevant. As we look at in more detail below when discussing the grounds of appeal, the FTT considered the factors irrelevant given its view of the case-law principles.

Given that limited scope of the appeal, we can summarise the FTT Decision relatively briefly. Its overall structure was to make its findings of fact and then conduct a multi-factorial evaluation in relation to, first the two Organix products, and then to carry out the same exercise with the Nakd bar products. The FTT did not distinguish between the different Organix and Nakd Bar products. Neither party criticises that approach but the appellant points out that in principle it was open (and remains open should the matter be re-determined) for the products to be considered individually (so for example that a different outcome could be reached in respect of a particular individual type of Organix or Nakd bars within those product ranges).

The FTT made extensive findings of fact (at [110] to [131]) relating to the bars. These included the bars’ ingredients (principally wholegrain oats and raisins), and the process by which the bars were made (mixing together, “sheeting out” and cutting into oblong bars” which were then baked and cooled). In relation to the target market and marketing, the target market was parents of toddlers and young children, and the bars were held out for sale with other baby snacks. Morrisons’ website described the bars as “ideal toddler snack bars”. The FTT set out, in detail, the wording and colouring of the packaging. It also inferred from the packaging that the purchasers were “health-conscious”. Both bars’ texture was described as soft. Regarding the banana bar’s taste and texture, this was “sweet and crumbly with some banana flavour”. The carrot cake bar was also sweet.

Having considered the parties’ legal and factual submissions, the FTT went on to conduct its multi-factorial evaluation (at [177]) taking account of the following: as to sugar content although this was half of the sugar content of comparative confectionery, it was more than 25% by weight. Both bars were sweet to taste, were subjected to a process, were normally eaten with fingers, and were held out for sale as snacks, when consumed. The FTT noted the bars were normally eaten between meals, the same as traditional confectionery. It noted the packaging which was bright and colourful, indicated the bars were held for sale as treats in the same way as confectionery. Regarding the ingredients, it rejected the appellant’s submission that the absence of cane sugar, flour and/or butter was a factor pointing to the bars not being confectionery.

As for the physical positioning of the product when sold, the FTT was unable to make findings on whether the bars were positioned near traditional confectionery, but noted that even if they were not, that was insufficient to outweigh the other factors. It concluded “…Instead the multi-factorial test gives the clear answer the Organix Bars were confectionery ([178]).

The FTT made detailed findings of fact (at [185] to [202]) in relation to the Nakd bars under similar headings to those above in relation to the Organix bars.

In terms of the ingredients, it found the main ingredient in all was dates (between 37% to 58%), that the Fruit and Nut bars contained significant percentages of nuts and/or dried fruit. For the Oaties, the main ingredient, after dates, was oats. For the Crunch bars, the next main ingredient was “soya protein crunchies”. The processing consisted of mixing the ingredients and then cutting the mix into rectangular shape bars.

The FTT made extensive findings on the wording on the packaging and colouring and set out extracts from the product descriptions on Nakd’s website, customer comments on that website together with extracts from comments which appeared on Morrisons’ website. The FTT accepted, on the basis of the packaging and marketing, that the purchasers of Nakd bars were health conscious. Regarding the physical positioning in store, this moved over time from the “free from” aisle, to healthy biscuits and cereals, to healthy snacks ending up by the time of the FTT hearing between protein flapjacks and Morrisons’ own brand fibre bars. The bars were sold on the website under “Biscuits and crackers – cereal bars and breakfast biscuits – healthier cereal bars” and under “crips, snacks and nuts – healthier options”. Regarding taste and texture, the FTT noted all the samples provided tasted sweet and it inferred the same of products in respect of which no sample had been provide. It noted all had a texture “similar to fudge with finally chopped nuts and/or fruit”. The Cocoa Orange “tasted like the well-known product by another manufacturer known as chocolate orange” and the Cocoa Delight tasted “like liquid chocolate”.

After setting out the parties’ submissions, the FTT set out its multi- factorial evaluation (at [206]) by reference to the various factors it had taken into account: As to sugar content - this was between one third and one half of each bar, the product had more sugar than Green & Black’s organic dark chocolate – this was consistent with confectionery. The FTT noted the product was sweet to the taste, subjected to a process (mixing together and pressing), and normally eaten with fingers. They were held out as snacks, normally eaten between meals, positioned as treats (despite being sold as healthy food), and held out as filling the same role as confectionery (despite being positioned as more healthy). Regarding the product names, the FTT rejected the appellant’s argument that references to natural products pointed towards the general food classification. In relation to the packaging, it noted the cocoa-based products (which were brown – the colour of chocolate, others were brightly coloured) were similar to products which were clearly confectionery. As for the ingredients, the absence of sugar, flour, butter was not considered, as with the Organix bars, to be a relevant factor. The physical positioning of the product, which had changed over time, was at best a neutral factor.

The FTT concluded “having considered and balanced all the above factors” that it was clear all the Nakd Bars were confectionery.

The UT (Judge Richards) set out the scope of the permission to appeal as follows:

As will become apparent when we come on to discuss these grounds in more detail below, both errors are said to derive from the FTT’s mis-reading of the case-law (Kalron Foods Ltd v HMRC [2007] EWHC 695 (Ch) in the case of Ground 1, and HMRC v Premier Foods Ltd [2007] EWHC 3134 (Ch) in the case of Ground 2).

HMRC argue that even if the appellant were successful on the grounds above, the FTT’s decision should be upheld on an additional ground that the products fell within Note 5. The issue here is whether the reference to “sweetened” includes products that are already sweet. HMRC argue the FTT wrongly rejected HMRC’s argument that case-law (Premier Foods) held that “sweetened” must be read in this way.

The parties agree here that the FTT was to give the word “confectionery” its ordinary meaning and to ask what the view of the informed ordinary person was as to the nature of the product and whether or not it was one which fell within the relevant category.

As both parties acknowledge, this calls for a multi-factorial evaluation and accordingly appellate caution in interfering with that evaluation. They agree matters of weight are for the first-instance tribunal. Both parties also acknowledge the statutory jurisdiction on appeals to the Upper Tribunal (s12 Tribunal Courts and Enforcement Act (“TCEA 2007”) requires that there is an error of law, but that even if an error of law is found, the decision as to whether the FTT decision should be set aside is a matter of Upper Tribunal discretion.

The parties disagree however on what constitutes an error of law where the particular allegation is that the FTT has omitted to consider a relevant factor. Central to HMRC’s defence of the FTT’s decision, is the argument that it is not enough simply to show the FTT omitted consideration of a relevant factor but that there is an additional threshold of perversity. HMRC thus argue that the omission only constitutes an error of law where a tribunal failed to take into account a matter which no tribunal properly instructed would have left out of account.

Furthermore, in assessing the materiality of the error of law (for the purpose of deciding whether the UT should set aside the FTT decision), HMRC argue it is not enough that the error of law might have made a difference (as the appellant argues). The appellant (HMRC submit) had to show the error would have made a difference.

Before addressing the above disputed points of legal approach to appeals from the FTT to the UT, it is helpful to remind ourselves of the rationale and scope of the principles surrounding the need for appellate caution when considering appeals against so-called multi-factorial evaluations. That an appeal court should accord due deference to the FTT’s role in carrying out an evaluative multi-factorial exercise is not disputed. Nor is it disputed that the issue before the FTT of determining whether the products were “confectionery” involved precisely that kind of evaluation.

Proctor and Gamble v HMRC [2009] EWCA Civ 407 (“Pringles”) concerned an appeal from a fact-finding tribunal on a similar type of classification question (whether Pringles fell within the exception to general zero-rating for food products for “potato crisps…and similar products…”). The High Court had overturned the fact-finding tribunal’s conclusion, which had agreed with HMRC, that the product fell with the exception. The Court of Appeal, in allowing HMRC’s appeal and finding the tribunal had not made any error of law, addressed the principles regarding the approach on appeal to “value judgments” of the primary decision maker. This was explained by Jacobs LJ as follows at [9]:

“Often a statutory test will require a multi-factorial assessment based on a number of primary facts. Where that is so, an appeal court (whether first or second) should be slow to interfere with that overall assessment—what is commonly called a value-judgment.”

Jacobs LJ went on at ([10]) to refer to an earlier judgment in which he had gathered together various authorities on this topic from different subject areas including Biogen v Medeva [1997] RPC 1 where Lord Hoffman had explained the rationale for caution as follows:

“…specific findings of fact, even by the most meticulous judge, are inherently an incomplete statement of the impression which was made upon him by the primary evidence. His expressed findings are always surrounded by a penumbra of imprecision as to emphasis, relative weight, minor qualification and nuance (as Renan said, la verité est dans une nuance), of which time and language do not permit exact expression, but which may play an important part in the judge’s overall evaluation….Where the application of a legal standard such as negligence or obviousness involves no question of principle but is simply a matter of degree, an appellate court should be very cautious in differing from the judge’s evaluation.”

At [11] Jacobs LJ also noted the House of Lords authority cited in Toulson LJ’s judgment in Pringles (Baroness Hale in AH (Sudan) v Secretary of State for the Home Department [2007] UKHL (at [30])) concerning the deference to be given to specialist tribunals as the primary decision maker given their special expertise.

Toulson LJ explained at ([60]):

“Where a Tribunal has taken into account all relevant factors, and has not been influenced by impermissible factors, a court will only exceptionally entertain a challenge based on the Tribunal’s evaluation of those factors for the reasons given by Baroness Hale [in AH (Sudan)]. The challenger would have to show the decision was perverse and in this case there is simply no foundation for such a challenge. The tribunal was not obliged to accord a separate grading for each factor. It was entitled, as it did, to look at the matter in the round.”

Another reason, as Lewison LJ pointed out in Fage UK Ltd & Anor v Chobani UK Ltd & Anor [2014] EWCA Civ 5 (at [114]) for appellate caution in the evaluation of primary facts is the expertise of the trial judge in determining what facts are relevant to the legal issues to be decided.

As regards the scope of the principle that appellate caution should be exercised, returning to Pringles, Jacobs LJ explained the kind of issue to which the principle applied by reference to what Lord Hoffman said in another case Designers Guild namely where “the decision involves the application of a not altogether precise legal standard to a combination of features of varying importance”.

Taking account of the above, we note the focus on appellate caution is directed towards to analysis of weight or matters of degree and, in the context of a multi-factorial evaluation the first-instance court or tribunal’s overall evaluation. (The excerpt from Pringles at [9] refers for instance to the tribunal’s “overall assessment”). None of that is controversial. However, the appellant points out there is nothing in these principles which suggests an appellate tribunal should defer where the fact-finding tribunal has taken account of an irrelevant factor, or as they say happened in the instant case, disregarded a relevant factor.

HMRC on the other hand referred to [22] of Jacob LJ’s judgment. There he said:

“So one can put the test for an appeal court considering this sort of classification exercise as simply this: has the fact finding and evaluating Tribunal reached a conclusion which is so unreasonable that no reasonable Tribunal, properly construing the statute, could reach?”

It is important however to recognise the context in which the statement by Jacob LJ in the preceding paragraph was made. This makes clear the classification exercise Jacob LJ had mind concerned the issue: “what is the reasonable view on the basis of all the facts” (see [21]). In other words, the high threshold of perversity related to challenges to the tribunal’s overall evaluation of the facts.

In support of the proposition that appellate caution does not apply where the criticism of the tribunal is that it took account of irrelevant factors, or disregarded relevant factors, the appellant emphasises the extract from Toulson LJ’s judgment in Pringles (at [60] – see ([31] above). There, Toulson LJ specifically assumed (before addressing the need to show perversity) that the decision was not one where the tribunal had failed to take account of relevant considerations.

HMRC submit this passage does not however deal with the required intensity of scrutiny (in other words whether perversity is required to be shown where the challenge concerns disregard of relevant factors). That argument must be rejected simply in terms of the way the relevant paragraph of Toulson LJ’s judgment ([60]) is expressed. The paragraph takes as given that the tribunal has considered the right factors and makes clear perversity needs to be shown where “evaluation of factors” i.e. the weighing up is challenged. The plain implication is that in cases where the criticism is that a relevant factor had not been considered, that would not require deference. The point is made clearer by the preceding paragraph which explains the scope of arguments the Court of Appeal was addressing:

“59.

On the question of similarity, Mr Cordara [counsel for the appellant] was critical of the Tribunal for saying that it did not regard the shape of Regular Pringles or the size of the packaging as particularly important factors. But those are classic examples of matters which were for the judgment of the Tribunal, and their view cannot be said to have been perverse. Otherwise, Mr Cordara did not suggest that the Tribunal either took into account any matters which it ought not to have taken into account or failed to take into account any matters which it ought to have taken into account. His primary criticisms are that the Tribunal failed properly to evaluate the various factors, because it failed to weight or grade them appropriately, and that it failed to give adequate reasons for its conclusion (emphasis added)”

HMRC relied on a number of authorities at Court of Appeal and Upper Tribunal level in support of their submission that it is only where a tribunal has failed to take into account a matter, which no tribunal properly instructed would have left out of account, is there an error of law (i.e. perversity). We consider however that, for the reasons we explain below, none of those authorities on closer analysis support that proposition. Rather the consistent message, in line with the Court of Appeal’s approach in Pringles, is that, while a requirement to show the decision is perverse applies in relation to matters of weight /evaluation, failing to take account of a relevant factor or taking an irrelevant factor into account will constitute an error of law (albeit there will be subsequent issue of whether any such error is material to the decision in question).

We find it convenient to start with the Court of Appeal authorities as this will help to properly understand the subsequent Upper Tribunal decisions HMRC rely on.

Davis &Dann Ltd v HMRC [2016] EWCA Civ 142 concerned an MTIC appeal requiring the fact-finding tribunal to make a determination on whether the trader ought to have known its transactions were connected to fraud; a classic situation where a number of factors fell to be considered in order to reach an overall conclusion. The particular context for the appeal was that the tribunal had to consider whether certain factors were firstly relevant and secondly whether they were probative (in that case to the “no other reasonable explanation” standard). At [77] Arden LJ, as she then was, dealt with various submissions (grouped under the heading dealing with the allegation that the tribunal had failing to take account of countervailing factors):

“In my judgment, this submission goes to the weight to be attached by the primary decision-maker to certain matters in relation to others. While the categorisation of a fact as probative of a particular issue is a question of law, the question whether it is so probative is a question of fact. On an appeal on a question of law, it is well established that an appellate tribunal whose function is restricted to questions of law cannot revisit questions of fact unless no reasonable judge could have come to that conclusion. Mr Scorey [counsel for the taxpayer] has not contended that that test applied in relation to any of these countervailing factors: there is no cross appeal in this Court. Therefore, the existence of these countervailing factors cannot determine the outcome of this appeal.” (emphasis added)

Arden LJ referred back to this paragraph (at [101]): (This passage, as will be seen, was referred to in a UT decision (Northside Fleet Ltd v HMRC [2022] UKUT 256 (TCC)) which HMRC rely on and which we address further below) :

“I have no doubt that the categorisation of fact in the present case constitutes a question of law which founds this Court's jurisdiction and that of the UT (see paragraph 77 above). (Moreover, if the UT found an error of law, it had jurisdiction to substitute its own decision: see Pendragon). There are, however, as appears from paragraph 77 above, limits to this: how the tribunal applies any categorisation of the fact to the circumstances of a particular case is likely to be a question of fact and not of law. For the reasons given in this paragraph, this Court is entitled to ask whether the UT was correct to evaluate the facts in the way that it did, namely as indicating normal market transactions, or whether, as HMRC contend, that conclusion was itself in error.”

While this decision confirms a perversity hurdle (“no reasonable judge…”) must be surmounted where the challenge concerns the factual issue of whether a factor was probative on the facts of the case, it does not suggest the question of whether, a factor was relevant, as a matter of principle, similarly faced such hurdle. Rather, Arden LJ confirmed the question of relevance was a question of law.

The next case relied on by HMRC is Teinaz v London Borough of Wandsworth [2002] EWCA Civ 1040, which concerned a discretionary decision by an employment tribunal of whether an adjournment to a hearing should be granted. HMRC rely on the following passages in the Court of Appeal’s discussion regarding the limited bases on which an appellate court will intervene:

The extracts do not, in our view, support (and indeed contradict) HMRC’s argument that there is an additional hurdle of perversity where the allegation is that the tribunal omitted to consider a relevant factor. Rather, the Court of Appeal confirms that, even in the slightly different context of challenges to a discretionary decision of a first instance tribunal, consistent with the principles we have already discussed, determination of what is relevant or irrelevant is a matter for an appellate tribunal, determination of what weight a factor should bear is not, unless that evaluation or balancing of such factors reaches the threshold of perversity. (The proviso at the end of [36] concerns the materiality of the error, not whether there is an error of law in the first place).

HMRC also seek to link the additional perversity hurdle with the approach to relevance of factors in public law decision making where there is “no statutory lexicon of factors” per Laws LJ in London Borough of Newham v Khatun & Ors [2004] EWCA Civ 55. We did not however find the cases HMRC relied on to be on point to the kind of multi-factorial evaluation at issue here. The passages relied on in the Supreme Court’s decision Samuel Smith Old Brewery (Tadcaster) & Ors, R (on the application of) v North Yorkshire County Council [2020] UKSC 3, at [29] – 32] per Lord Carnwath JSC) concerned the meaning of what was a “material consideration” in a planning statute and related case law – including Wednesbury - on whether a factor had to be taken into account “as a matter of legal obligation” (i.e. as matter of statutory construction), or whether on the facts of case were “so obviously material”. The public law context where Parliament has granted a discretion to a non-legal decision maker of deciding between competing policy factors is not necessarily comparable to the interpretative judicial task of determining whether something falls within a statute or not. There was, as far as we can see, no consideration in these public law cases of the case-law principles (from cases such as Pringles) as one might expect if it was envisaged a hurdle was being imposed where in case such as Pringles (for the reasons we have discussed) none was required.

Mr Watkinson relied, in addition, on a number of UT decisions. Aria Technology Ltd v HMRC [2018] UKUT 363 (TC) was another MTIC case concerning whether the trader knew or should have known its transactions were connected to fraud. The UT explained at [12] that the core of the appeal involved a challenge to “most of the FTT’s findings of fact” which included [12(2)] that the FTT took into account irrelevant considerations in making its findings or failed to take into account relevant considerations.

HMRC rely in particular on the following passages:

Northside Fleet was another MTIC related appeal to the UT. Amongst the grounds of appeal were that the FTT wrongly ignored certain factors (submission 2 (at [32(2)]). The UT recorded the parties’ agreement that the UT had power to interfere if it identified fundamental flaws in the fact-finding process (for example if the FTT failed to take account relevant considerations); it continued:

In the end, as the UT indicated, it did not need to address the question of relevance/ irrelevance and the above comments were, therefore, obiter. The parties were agreed the factors were relevant; the issue was whether the FTT took those relevant factors into account or not.

In our judgment, neither of the extracts from Aria or Northside Fleet, properly analysed, supports HMRC’s case that there is an additional perversity hurdle. Both passages in Northside Fleet [28] and Aria [34] relied on by HMRC must be read in the light of the contrast the Court of Appeal drew in Davis & Dann , that the question of what facts are relevant is a question of law (i.e. at level of principle) with the question of whether the fact is actually probative on the facts (which is question of fact). (We have some doubt therefore whether the statement in [29] of Aria “…The decision of what primary facts are relevant for the purpose of reaching the overall factual conclusion is essentially one for the fact-finding tribunal: it is part of the fact-finding process. …if an appellate body were simply to substitute its own assessment of what primary facts should be taken into account, it would itself be engaged in fact-finding and not restricting the appeal to a question of law” is correct as it seems to be at odds with the principles explained in Davis & Dann and also Teinaz at [35] and [36]).

Both decisions, read in the above context, were making the point that perversity needed to shown as regards a challenge to an evaluation of factors i.e. the weight to be given to factors given the particular facts of the case. Thus, in Aria it was noted the overall conclusion of fact was based “…on an evaluation of an assemblage of findings of primary fact” ([29]). In Northside Fleet [34] when the UT was referring to the categorisation adopted by the fact-finding tribunal having to be perverse, it was referring to the question of whether in the particular circumstances of the case, the fact was actually probative of the issue in question (that question being a question of fact). That question was to be contrasted with the question of legal relevance, namely whether the tribunal could be said to have been properly instructed as to what to take into account or leave out of account at level of principle.

The appellant’s grounds in this case are firmly in the latter category. This is evident from the fact that in respect of both grounds, the conclusion the FTT came to that the particular matters were irrelevant and should not be taken into account was derived from examining the case-law (with the appellant arguing that the FTT erred in so doing).

Mr Watkinson also referred us to [50] of Aria where the UT noted the application of the proper approach to an appeal was critical:

“…If each paragraph of a long judgment, involving the analysis of a wide range of facts following substantial evidence, is examined with a fine toothcomb, it may often be possible to find some factual error or lack of clarity, or to identify particular facts which were arguably relevant but are not referred to or assessed. If such flaws meant that the decision must be quashed as erroneous in law, little would be left of the cautionary words of Evans LJ in Georgiou. In our judgment, having regard to the wording of s12(2) of TCEA which we have emphasised and to the authorities we have discussed, even if we find some errors regarding any factual matter considered, or which should have been considered, in the decision below, such that any tribunal properly instructed would have taken that matter into account, or left it out of account, only if that matter is material to the overall factual conclusion is that a basis for setting the decision aside. The same approach applies if the reasoning expressed in support of a particular factual finding is unclear or deficient.”

However, those observations are clearly directed to the separate question, which we come on to shortly, of whether once error of law is identified, the error is considered material. The UT was acknowledging there may be errors in the fact-finding process which do not justify setting aside decision.

Mr Watkinson also took as to passages in Hargreaves v HMRC [2022] UKUT 34 (TCC) at [17(2)]) and Vital Nut Co Ltd v HMRC [2017] UKUT 192 (TCC) at [43] making a link between challenges to the fact-finding process based on ignoring relevant considerations with having to meet the Edwards v Bairstow test. It is correct the UT in Hargreaves UT 17(2) –referred to Edwards v Bairstow applying but that included the situation where the fact-finding process was flawed because relevant considerations were omitted. Even if it is assumed the aspect of Edwards v Bairstow said to apply was that “no reasonable tribunal, properly directed in law, could have reached that finding” that would not necessarily mean imposing an extra perversity hurdle: a tribunal omitting to consider a relevant consideration would not be said to be a reasonable tribunal, properly directed in law. The UT in Vital Nut (at [43]) similarly confirmed the exclusion of legally relevant and probative evidence is “Edwards v Bairstow territory” – but there is nothing to suggest a requirement to show additionally that the exclusion of relevant and probative evidence was such that no reasonable tribunal properly instructed would have excluded it.

Mr Watkinson also relied on Anna Cook [2021] UKUT 15 (TCC) (at [19]) and HMRC v Netbusters (UK) Ltd [2022] UKUT 175 (TCC) ) (at [23]) to say there is a high threshold / the bar is particularly high where the FTT has been called upon to make a multi-factorial assessment (at [23]). We agree, but it is important to recognise the areas in relation to which that high threshold applies. Both cases confirm what we have already discussed, that perversity needs to be shown in relation to challenges to findings of fact based on the weight of evidence, or regarding the particular weighting of factors, or matters of degree in making an overall assessment.

In conclusion, we agree with Ms Sloane’s submissions that HMRC’s case, that there is an additional perversity hurdle to surmount, must be rejected. If it is demonstrated that the FTT ignored a relevant factor that will be sufficient to constitute an error of law. We do not need therefore to deal with the other bases on which the appellant put its case regarding there being an error of law (which for instance included an alternative argument that the alleged errors were ex facie errors of law, which Mr Watkinson objected to as being out of scope of the permission to appeal).

Under s12 TCEA 2007 where the Upper Tribunal finds that the making of the FTT decision “involved the making of an error on a point of law” the Upper Tribunal “may (but need not) set aside” the FTT decision.

As mentioned above, as regards the Upper Tribunal’s discretion to set aside the decision, HMRC argue for in essence a more stringent test: that the FTT decision would have been different rather than that it might have been different.

The most recent higher authority in relation to the exercise of the Upper Tribunal’s discretion to set aside is Degorce v HMRC [2017] EWCA Civ 1427. The issue was whether the taxpayer, Mr Degorce, who had participated in a film scheme, was carrying on a trade. It was admitted the FTT made an error (which the court described as an error of approach) in not taking account of the taxpayer’s other film-related activities because no findings regarding them had been made by a court or tribunal because they were subject to an enquiry by HMRC ([96]). The UT nevertheless declined to exercise its discretion to set aside the decision. In the Court of Appeal the taxpayer argued the UT had misstated the test as requiring that if the evidence had been approached correctly “[The FTT’s] doing so would, or at least might have affected the outcome”:

Henderson LJ considered the test of materiality will have a (([95]):

He continued:

“At least in case of the present type, I find it difficult to envisage circumstances in which the Upper Tribunal could properly leave the decision of the FTT to stand, once it is satisfied that the error of law might (not would) have made a difference to that decision. As a taxpayer, Mr Degorce is entitled to be taxed according to the law, and if an error of law is detected in the FTT’s decision, which is material in the sense I have mentioned, justice will normally require nothing less than that the decision is set aside.”

Applying the test to the facts, Henderson LJ went on to conclude the admitted error of approach was immaterial ([100]).

Mr Watkinson draws a distinction between the materiality test where the error of law is an error of approach. There, he accepts it is enough that the error might have produced a different conclusion. In contrast, where the error concerns a multi-factorial assessment, he submits, it needs to be shown the FTT’s decision would have been different. That, he argues, would be extremely difficult as the factor wrongly ruled as irrelevant would have to be of such great weight that it outweighed the cumulative weight of all the relevant characteristics that were taken into account.

In our view, the facts of this case are such that we do not need to decide the scope of the materiality test and whether it varies, as HMRC submit, according to the type of legal error. That is because the error in issue in Degorce is directly analogous to the error here. The issue of whether Mr Degorce was carrying on a trade also required a multi-factorial analysis and the relevant error concerned the FTT’s disregard of a relevant factor, which the FTT had mistakenly ruled out of account. Similarly, if the appellant’s case here is made out, it rests on the FTT wrongly ruling out a factor as irrelevant. Moreover, both errors bear the same character of being said to stem from a misconception of the law. Thus, if the appellant successfully makes out its case, that the FTT made error(s) of law by disregarding one or more relevant factors, there is no reason to adopt a different test of materiality from that adopted in Degorce.

We in any case are sceptical of the distinction HMRC advance. As Ms Sloane pointed out, the underlying rationale – that justice will require the person is taxed according to the law – does not suggest the principle is limited depending on the type of error of law. (Against that we recognise however that by prefacing his views with “at least in the case of the present type” Henderson LJ was leaving scope for the argument a different test might apply).

Mr Watkinson also relied on the UT’s decision in Aria at [50] (see [54] above) – but that simply highlights the significance of applying a materiality test – the UT was not referred to and therefore did not mention Degorce and it did not say anything about the particular test to be deployed.

Mr Watkinson referred us to the Supreme Court’s decision HMRC v BPP Holdings Ltd [2017] UKSC 55 at [21]) in relation to a discretionary decision (whether to bar the Respondents from proceedings). The Supreme Court suggested an appellate court could interfere with the tribunal’s decision if it could be shown that irrelevant material was taken into account or relevant material ignored “unless the appellate court was quite satisfied that the error made no difference to the decision”. However that test is, we consider, entirely consistent with the “might have made a difference” test in Degorce for deciding whether to set aside.

That leaves the principal higher authority Mr Watkinson relies on which was the Court of Appeal’s decision in R v Chief Registrar of Friendly Societies, Ex parte New Cross Building Society [1984] QB 227, in particular an extract from the judgment of Griffiths LJ at 260. The case concerned a judicial review of the building society registrar’s decision which the applicant sought to quash – Griffiths LJ explained (42b All ER pg76 Auth) the applicant had to either satisfy the court that the registrar misconstrued his powers or that he had abused them (which Griffiths LJ explained involved considering the reasons for the decision to see if they could be challenged on Wednesbury principles). He considered the court should not allow those principles “to be erected into immutable propositions of law”:

“Take as an example the proposition that the decision must take into account relevant considerations and leave out irrelevant considerations. In a decision involving the weighing of many complex factors it will always be possible to point to some factors which should arguably have been taken into account or left out of account; even if they should have been, the court should not intervene unless it is convinced that this would have resulted in the decision going the other way.”

We do not consider this case supports HMRC’s submission. The context was weighing up of factors, and this decision pre-dates, or does not consider, the authorities which make clear that disputes over the weight a tribunal has given to a factor will not be regarded as an error of law, so it can be seen why a high threshold of the court being “convinced” the decision would have gone the other way was being set. The public law context of the decision is also different (see [46] above). The case was not referred to in Degorce which is more on point given it relates to the particular statutory discretion under s12 TCEA and which suggests a markedly lower hurdle for a decision to set-aside an FTT decision justified by a tax-based rationale.

Finally, Mr Watkinson took us to the Upper Tribunal’s decision in Ingenious Games LLP v HMRC [2019] UKUT 0226 (TCC) at [66] – [74]. There the UT was dealing with the similar question of whether the taxpayers were carrying on a trade and whether the taxpayers were carrying on business with a view to profit.

At [74] the UT summarising the legal principles in relation to appellate interference said extreme caution was required before setting aside conclusions based on careful evaluative findings of fact made on the basis of extensive evidence in contrast to the position where an ex facie error of law is identified. Applying that to the facts the UT considered that none of the challenged errors was so significant in itself that the FTT could not have reasonably reached the conclusion it did on the issue in question. The UT was aware of Degorce (it referred to a different point on it at [70]).

We do not read the UT’s approach to materiality in Ingenious Games as being different from that of the Court of Appeal in Degorce and it would be surprising if it was given it was dealing with very similar kind issue. Ingenious Games, in this respect, is best understood as a case where the Upper Tribunal had to decide whether the impugned factual findings might have made a difference to the ultimate decision on the issues in question and decided that they did not. In Degorce, when it came to applying the materiality principle to the facts, the Court of Appeal similarly considered that the errors of approach in that case were immaterial.

In conclusion we consider the relevant principles in deciding whether or not to set aside the decision if Grounds 1 and/or 2 disclose an error of law are those set out in Degorce.