Revenue And Customs v Honeywell Analytics Ltd

This appeal concerns a Binding Tariff Information (“BTI”) decision made on 19 December 2013 by the appellants, the Commissioners for Her Majesty’s Revenue and Customs (“HMRC”). It raises issues as to the correct approach that should be taken when deciding under which heading in the Combined Nomenclature (“CN”) of the European Union’s Common Customs Tariff (“CCT”) a product should be classified for the purposes of customs duty on imports.

The appeal is from the decision of the Tax and Chancery Chamber of the Upper Tribunal (Judge Timothy Herrington and Judge Nicholas Aleksander) (“the UT”) published on 10 March 2017, setting aside the decision of the Tax Chamber of the First-tier Tribunal (Judge Richard Thomas and Mr David Batten) (“the FTT”), which had classified the product in issue of the respondent under sub-Heading 8531 80 95 of the CN, and determining that the product should be classified under sub-Heading 90 26 80 20.

The Common Customs Tariff

Council Regulation (EEC) 2658/87 of 23 July 1987 established the CCT, the provisions of which govern how customs duties are levied on goods which are imported into the European Union. In particular, Article 1 of Regulation 2658/87 establishes the CN. The CN provides a taxonomy system of goods according to an eight digit numerical system, containing approximately 10,000 sub-headings under which goods can be classified. Each sub-heading defines the level of customs duties to which that product classification is subject.

Annex I to the Regulation is updated and published annually as a stand-alone Regulation, changing the content of the CN to reflect any changes to the text which have been made. Those changes take into account developments in the EU’s commercial policy as well as technological and statistical requirements. They also reflect policy changes that have been agreed by the EU at an international level, either within the World Customs Organization (“WCO”) with regard to the Harmonized System (“HS”) nomenclature or within the World Trade Organization’s (“WTO”) framework with regard to conventional duty rates.

The EU’s Customs Code has been through three major iterations to date.

The Community Customs Code (“CCC”) was originally laid down by Council Regulation (EEC) 2913/92 which was adopted on 12 October 1992.

This was superseded by the Modernised Community Customs Code (“the MCCC”), which was laid down by Regulation (EC) 450/2008 of the European Parliament and of the Council, adopted on 23 April 2008, and the relevant provisions of which came into force from 1 November 2013.

The MCCC was in turn was replaced by the Union Customs Code (“the UCC”), which was laid down by Regulation (EU) 952/2013 of the European Parliament and of the Council, adopted on 9 October 2013, and the relevant provisions of which came into force from 1 June 2016.

At the time the respondent, Honeywell Analytics Limited (“Honeywell”), requested a BTI from HMRC the CCC was the applicable Customs Code regime. Section 2 of Chapter 2 of Title I (General provisions) of the CCC provided for an importer to request and obtain from the national customs authorities a decision relating to the application of customs rules. Corresponding provisions were contained in Article 16 of the MCCC and are now contained in Article 22 of the UCC.

Such a decision could be annulled, revoked or amended in certain circumstances pursuant to Articles 8 and 9 of the CCC. Corresponding provisions are now contained in Articles 18 and 19 of the MCCC and Articles 23, 27 and 28 of the UCC.

Articles 11 and 12 of the CCC also enabled an importer to request and obtain a BTI specifying the tariff classification of the goods in question. Article 12 specified circumstances in which the BTI would cease to be valid. The corresponding provisions for BTIs can be found in Article 20 of the MCCC and Articles 33 and 34 of the UCC.

The classification provisions of the CN relevant to these proceedings are as follows.

Explanatory Notes known as Harmonised System Explanatory Notes (“HSENs”) published by the World Customs Organisation can also be an aid to the interpretation of the heading of the CN. It is not necessary, for the purposes of this appeal, to set out any of the HSENs applicable to Headings 8531, 9026 and 9027.

General rules for the interpretation of the CN (the “GIRs”) include the following.

“Classification of goods in the Combined Nomenclature shall be governed by the following principles:

1.

The titles of sections, chapters and sub-chapters are provided for ease of reference only; for legal purposes, classification shall be determined according to the terms of the headings and any relative section or chapter notes and, provided such headings or notes do not otherwise require, according to the following provisions.

…

3.

When, by application of rule 2(b) or for any other reason, goods are prima facie classifiable under two or more headings, classification shall be effected as follows:

(a)

the heading which provides the most specific description shall be preferred to headings providing a more general description. However, when two or more headings each refer to part only of the materials or substances contained in mixed or composite goods or to part only of the items in a set put up for retail sale, those headings are to be regarded as equally specific in relation to those goods, even if one of them gives a more complete or precise description of the goods;

(b)

mixtures, composite goods consisting of different materials or made up of different components, and goods put up in sets for retail sale, which cannot be classified by reference to 3(a), shall be classified as if they consisted of the material or component which gives them their essential character, in so far as this criterion is applicable;

(c)

when goods cannot be classified by reference to 3(a) or (b), they shall be classified under the heading which occurs last in numerical order among those which equally merit consideration.

4.

Goods which cannot be classified in accordance with the above rules shall be classified under the heading appropriate to the goods to which they are most akin.

…”

Honeywell designs, manufactures and sells gas detection solutions. In issue in these proceedings is one of Honeywell’s products, the Gas Alert Micro 5 (“the Device”).

Honeywell submitted a BTI application dated 2 September 2013 in respect of the Device to HMRC, claiming that, in its view, the product should be classified under sub-Heading 9026 8020 90 of the CN. On 16 October 2013 HMRC replied, issuing a BTI notification in respect of the Device under Heading 8531 8095 90.

On 1 November 2013 Honeywell wrote to HMRC requesting an appeal in the form of a formal departmental review of the BTI decision on the basis that sub-Heading 8531 8095 90 was an incorrect classification and that the Device should have been classified under sub-Heading 9026 8020 90. On 19 December 2013 HMRC wrote to Honeywell informing them that it had concluded that the Device was correctly classified under sub-Heading 8531 8095 90 and that it upheld the BTI decision.

On 13 January 2014 Honeywell issued a Notice of Appeal, which was heard by the FTT on 1 September 2015. By their decision of 20 November 2015 they dismissed Honeywell’s appeal against HMRC’s classification of the Device under sub-Heading 8531 80 95.

The FTT first noted (at para. [15]) that in its request for a review of HMRC’s BTI decision Honeywell described the Device as follows:

“The product is a gas monitoring device which is carried on the person (portable) and used by people who work in confined spaces and may have reason to come into contact with high levels of potentially dangerous toxic gases.

The product (as described in the technical specification) detects the following gases and provides a ‘parts per million’ (PPM) LCD readout of each of these gases in real time (i.e. on a continuing basis): H2S, CO2, SO2, PH3, NH3, HCN, CL2, ClO2, O3 and combustibles.

The units contains [sic] both audible, visible and a vibration alert mechanism.”

The FTT went on to consider (in para. [17]) the technical literature available online in relation to the Device, which stated under the heading “Protect yourself” that the Device could simultaneously monitor and display up to five atmospheric hazards.

Based on the factual evidence provided by Mr Christopher Townsend of Honeywell, which was accepted by the FTT, the FTT found (in para. [24]) that the Device detected whether there were gases at all in the relevant space entered into by the person wearing the Device and that, if there were, the level detected may or may not be dangerous. The alarm indications could be disabled at the discretion of the user, although Mr Townsend agreed that it was not best practice to disable the detection sensors in a gaseous environment.

The FTT also found (in paras. [35]-[36]) that levels of gas building up on the filters on the Device caused electrical currents to be generated which were proportionate to the level of the gas. Those currents were measured and displayed on the LCD screen on the Device. Optionally the user could lock the readings and the Device could store several months of continuous data on a removable memory card. When a calibrated level of gases was reached the Device showed the word “ALARM” on the LCD screen which became backlit and displayed the ambient gas readings. It flashed, made a sound and vibrated.

The FTT made the following finding (in para. [37]) as to the contents of the Device and how it works:

“From the facts that we have found as to the contents of the device and the way it works, and from the appellant’s written and Mr Townsend’s oral description of it, we find that the Gas Alert Micro 5 has the characteristics and properties of an alerting device. Those characteristics and properties include the ability of the device to detect pre-calibrated levels of dangerous gases and the three different alarms together with the LCD screen display when a predetermined level of gas is reached. The ability to disable one or more alarms does not alter that.”

The FTT made the following findings (in paras. [38]-[39]) as to the functions and use of the Device:

“38.

We also find that the intended use, and actual use, of the device is the alerting of its wearer to the presence of noxious levels of gas in a confined space and it does that by at least one and usually two or three different types of alarm signal, visual, audible and vibrating. Put another way, in answer to the question: “What is the device for?” we find that it is to do that alerting. And we find that alerting is the only thing the device is intended to be used for.

39.

We also find that one of the things the device does in order to be able to give its alerts is measuring (and as we have said HMRC do not dispute that measuring is one of the device’s “functions”). It measures the quantity of gas (in ppm) and it also measures by reference to time, so that it can, depending on how it is calibrated, give alerts when a selected gas is present at a given level or range of levels over a given period. But we find that the measurement is a means to an end, not an end in itself: measurement is not its intended use.”

The FTT said (in para. [77]) that:

“… the essential characteristics and properties and the only intended use (and in fact the only conceivable actual use) of the device is as an instrument for alerting its operator by visual, audible and vibrating signals to the presence of a dangerous build up or absolute level of particular hazardous gases and other noxious substances”.

Accordingly, the FTT held (in para. [78]) that Heading 8531 described accurately and clearly the essential characteristics and properties and the use (the only intended use) of the Device. Having referred to the wording of Heading 9206, the FTT then held (in para. [80]):

“In our view while the device is an instrument that does measure “the … level … of gases”, it is not a instrument for doing that. … [N]o employer would send an operator into a confined space with this device to measure the level of dangerous gases without the alerting functions being operative. We consider that the items listed in Heading 9026 are ones whose only function and use is to measure the level etc of gas etc.”

In support of that finding, the FTT drew support (in paras. [81]-[82]) from Case C-218/89 Shimadzu Europa Gmbh v Oberfinanzdirektion Berlin [1990] ECR I-4391 to hold, by analogy, that only apparatus whose very purpose is to measure levels of gas can be regarded as apparatus for measuring or checking such quantities within Heading 9026.

The FTT then addressed the distinction argued by Honeywell between a “gas detector” (“GD”) and a “gas measuring instrument” (“GMI”), namely that the difference between the two is that a GMI is able to do more than simply detect the presence of gas that is present and can provide measurement of the amount of the level of the gas that is present. Honeywell had described in its evidence how an example of a gas detector, a carbon monoxide (CO) detector which did not indicate the level of concentration of the gas to the user, would activate an alarm if a dangerous concentration of carbon monoxide was detected.

The FTT said (in paras. [91]-[92]) as follows:

91.

We agree though with HMRC on where the appellant draws the line between a GD (8531) and a GMI (9026). We think the appellant’s example of the CO detector makes HMRC’s case for it. The CO detector is clearly measuring, and in much the same way as the device in question. Mr Townsend says that the “current between the electrodes … is proportional to the CO concentration”. He also shows that the alarm is given over three different time frames covering three different levels or ranges of gas concentration. That must, we consider, involve a measuring of the differences in the current. But as with the device in question, the detector is not “for measuring”; the measuring is subordinate to the alarm function.

92.

We do not see either that it follows, as the appellant seems to argue, from the fact that a CO detector does not indicate the level of concentration of the gas to the user, and is a GD, that therefore a device that does indicate the level to the user is a GMI. We imagine that in any case an operator wearing the Gas Alert Micro 5 is not going to be scanning the small LCD screen to find out if the levels of gas are dangerous: they will use the alarm functions for that purpose. Any measurements the device makes that may be of interest to the operator’s employer will we presume be scrutinised when the operator has come out of the confined space or will be examined and analysed on a computer into which the device’s memory card has been transferred.”

Consequently, the FTT arrived (in para. [94]) at a definitive classification of the Device under sub-Heading 8535 80 95 without looking at the GIRs. They did, however, briefly consider how they would have applied GIR 3 had they found it necessary to do so. They said (in para. [95]):

“…. We find it difficult to say whether either heading is more accurate than the other (it must be remembered that we are now in the realms of deeming heading 9026 to apply) so we pass from GIR 3(a) to GIR 3(b). Here we consider that what gives the device its essential character is the alerting components. We would therefore find for 8531 on the basis of GIR 3(b). But this (on the hypothesis that 9026 applies) may be subject to the operation of Rule 1(m) of Chapter 85.”

Finally, the FTT dealt with the argument that, if the Device fell within both Headings, then Note 1 (m) of section 85 would apply with the result that it would have to be classified under Heading 9026. They said (in para. [100]):

“…. were we to have decided that the device could fall within both 9026 and 8531 then it seems that Note 1(m) of section 85 would apply.”

Accordingly, the FTT dismissed the appeal and upheld the BTI.

On 7 April 2016 Judge Bishopp granted Honeywell permission to appeal. The appeal was heard by the UT on 15 December 2016. By their decision of 10 March 2017 they allowed the appeal.

On the hearing of the appeal HMRC sought to argue for the first time that, if the Device could not be classified under Heading 8531, then it should be classified under Heading 9027 (which relates to gas analysis apparatus). The UT refused to permit HMRC to do so because (1) the argument had not been raised in HMRC’s response to the Notice of Appeal (HMRC stating at that time that it would not put in a response to the grounds of appeal and that it would seek to uphold the FTT’s decision for the reasons they gave) even though the notice of appeal was given a considerable time after HMRC said it had formed the view that Heading 9027 would be more appropriate than Heading 9026 should Heading 8531 be held not to apply; (2) the point had first been indicated by HMRC in an email less than one month before the hearing of the appeal; (3) the point was developed for the first time in the skeleton argument of HMRC’s counsel for the appeal dated 7 December 2016, which was after counsel for Honeywell had filed her own skeleton; and (4) had the point been raised by HMRC in its statement of case, it is likely that Honeywell would have expanded its evidence to deal with it since the point involved a significant variation to the case which Honeywell had to meet before the FTT.

Turning to Honeywell’s grounds of appeal, the UT first addressed the issue as to whether certain matters determined by the FTT were or were not pure findings of fact. The UT said the following (in paras. [49]-[50]):

“49.

In our view the FTT’s findings at [38] and [39] of the Decision are not pure findings of fact but are part of its reasoning as to ascertaining the objective characteristics and properties of the Device and its intended use. In any event the comments of Lord Carnwath in HMRC v Pendragon plc & Ors [2015] UKSC 37 at [49] to [51] are particularly apposite in a case such as this. In our view what the FTT was doing at [38] and [39] of the Decision was making an evaluation of the primary findings of fact it had previously made in the Decision. We should therefore adopt a more flexible approach as to whether the making of those findings involved questions of law.

50.

… On that basis, in our view, the findings in those paragraphs involved questions of law and we are entitled to consider whether the FTT erred in the approach it took in coming to the conclusion it did at [80] of the Decision, namely that the Device is not an instrument for measuring the level of gases and that the items listed in Heading 9026 are ones whose only function and use is to measure the level of gases.”

The UT went on to hold as follows (in para. [51]) that the FTT erred in the approach they had taken:

“51.

… It had made clear findings … that the Device took measurements of gases which were displayed on the LCD screen on the Device which were stored on a removable memory card. In those circumstances, prima facie the FTT had made a finding that the Device fell within the scope of Heading 9026 as well as Heading 8531. As Miss Sloane [counsel for Honeywell] observed, that finding differentiated the Device from a device such as a burglar, fire or smoke alarm, clearly falling within Heading 8531 and which did not display measurements. We therefore accept Miss Sloane’s further submission that based on the findings of fact made by the FTT, displaying measurements is an objective characteristic of the Device.”

The UT said (in paras. [52]-[53]) that the FTT, having found that the product fell prima facie within multiple headings, should have applied GIR 3(a) but had failed to do so. The UT continued (in paras. [54]-[55]) as follows:

“54.

In our view that finding [viz. that the measurement display was a means to an end, not an end in itself and that measurement was not an intended use, the only use being for alerting] was made on the basis of nothing more than an assumption as to the use to which the Device would be put. In our view in making that finding the FTT made a finding that went beyond an assessment of the objective characteristics and properties of the Device by reference to its inherent character. It did not confine its findings to the technical characteristics of the Device as it should have done …

55.

In making this error, the FTT may have been unduly influenced by the heading to the technical literature available in relation to the Device which … (“Protect yourself”) indicated that the primary use of the Device was as an alert to the presence of dangerous gases in the space in which the Device was being worn.”

The UT said (in paras. [52]-[62]) that the FTT was strongly influenced in its reasoning by the judgment in the Shimadzu case and that the FTT had incorrectly applied that case. The UT distinguished the Device in the Shimadzu case, where the measurements were not displayed, which made the Device more akin to a smoke alarm which measures smoke levels but only does so for the purpose of signalling its presence. By contrast, the Device had a display so that, in addition to signalling the presence of dangerous levels of gas, those levels could be measured.

The UT, therefore, rejected (in para. [63]) HMRC’s submissions that Heading 9026 is to be restricted to measuring apparatus whose principal purpose is to measure for the sake of measurement alone, on the basis that nothing in the wording of the Heading suggests that measurement has to be the “principal” purpose.

The UT concluded (in para. 64]) that the FTT “was wrong to have decided that the Device did not have measurement as one of its uses and was wrong to conclude that Heading 9026 was restricted to those devices whose only function was to measure the level of gas”.

The UT also rejected (in paras. [66]-[67]) an argument put forward by HMRC that Heading 9026 is only concerned with devices which measure or check “process variables” of liquids or gases, that is to say devices which are concerned with how gas occupies, or moves within, a space whereas, so HMRC argued, the Device was not measuring a “process variable” but was identifying the gases and measuring their concentration within a space. The UT said that the whole case had been argued before the FTT on the basis that the meaning of “level” included the composition and concentration of gases in a particular environment. The UT said, furthermore, there is no implication from the wording of Heading 9026 that it is confined to “process variables” and the examples given are not exhaustive of the kind of instrument falling within that heading.

Accordingly, the UT found (in para. [68]) that the Device prima facie falls within the scope of both Heading 8531 and Heading 9026, and that inevitably, by virtue of Note 1(m), the Device must be classified under Heading 9026.

The UT concluded (in para. [70]) that the FTT’s decision should be set aside and that, rather than remit the matter to the FTT, the UT would set aside the BTI and substitute a classification of sub-Heading 90 26 80 90.

By his order of 16 June 2017 Lewison LJ granted HMRC permission to appeal the UT’s decision on the following two grounds: (1) the UT erred in stating and applying the proper approach to customs classification, particularly in (a) adopting an approach to classification on the basis that a product’s intended or targeted use was not to be taken into account, and (b) treating the FTT’s finding as to intended use as a question of law and not a question of fact, and (2) the UT erred in applying the GIRs.

Honeywell has filed a respondent’s notice seeking to uphold the UT’s decision on the additional grounds that (1) insofar as they were findings of fact, the findings in paragraphs [38] and [39] of the FTT’s decision that “alerting is the only thing the device was intended to be used for” and that “the measurement is a means to an end, not an end in itself: measurement is not its intended use” were perverse; and (2) if and so far as there is a “principal function” test, the measurement function is the principal function or, alternatively, it is not possible to establish a prevalent function, and the FTT erred in finding otherwise.

By an application notice dated 18 January 2018 HMRC applied to add the following further ground of appeal:

“The Upper Tribunal erred in law in failing to consider whether the product subject to the contested BTI was properly classified under CN heading 9027 and/or in finding that it would be procedurally unfair to allow the Commissioners to advance such a case before the Upper Tribunal on the appeal. The Upper Tribunal should have considered whether or not the Product fell within CN Heading 9027 and should have concluded that it did, applying GIR Rules 3(a) and/or 3(b) and/or 3(c)”.

The background to that application was explained in a Supplemental Skeleton Argument also dated 18 January 2018. The explanation may be briefly summarised as follows.

It was said that, as a result of the restructuring of CN Heading 8531 at the international level, some of its sub-headings were amended with effect from 1 January 2017. This occurred after the FTT’s decision and after the hearing before the UT, but before the UT handed down their decision. As a result of this amendment, the BTIs that had been issued to Honeywell were automatically invalidated on 31 December 2016. Honeywell made fresh BTI applications which, in light of the UT’s decision, classified the Device under Heading 9026.

Subsequent to the UT’s decision, HMRC became aware that other EU Member States, in particular France, Germany and Ireland, had issued BTIs which classified similar products to the Device as gas analysis apparatus under Heading 9027. This produced a divergence in classification between the UK where, according to the UT’s decision, the correct classification was 9026, and a number of other Member States which applied 9027 as the correct classification.

Pursuant to Article 34(10)(a) of the UCC and Article 20(1) of the Commission Delegated Regulation 2015/2446 of 28 July 2015, the Commission advised Member States that the issue of BTIs in respect of Gas Level Measurement Apparatus was suspended for a period of 10 months.

The issue as to the correct classification of those products was discussed for the first time at the 183^rd Customs Code Committee meeting held between 23 and 26 October 2017. No Member State, including the UK, spoke against classifying such products under CN Heading 9027.

A draft Regulation was prepared by the Commission giving effect to that classification. It was discussed at the 184^th CCC meeting held from 18 to 20 December 2017. All Member States considered that products similar to the Device should be classified under Heading 9027 (and specifically under sub-Heading 9027 10).

The draft Regulation is currently subject to ‘Inter-Service Consultation’ within the Commission. That is to say, the draft is subject to scrutiny by other Directorates within the Commission and by their legal service.  HMRC says that it appears very likely that the draft Regulation will be adopted, thereby classifying goods equivalent to the Device under CN Heading 9027.

In those circumstances, HMRC considers that it can no longer contend that the Device should be classified under Heading 8531.

We heard the application to add the additional ground of appeal at the outset of the oral hearing of the appeal. We refused the application and said we would give our reasons in our substantive judgments on the appeal.

The reason for the refusal is straightforward. The decision of the UT to refuse the application was a case management decision. In making it, the UT did not make any error of principle and it is impossible to say that it acted outside the limits of a legitimate exercise of judicial discretion. The reasons the UT gave for the decision, which are set out above, were perfectly cogent reasons.

Moreover, if we were to have permitted the application, this Court, as much as the UT, would have placed Honeywell in a most unfair position since it would have been deprived of the opportunity to advance evidence, including expert evidence, on the differences between items within Heading 9026 and within Heading 9027 in the context of the Device and its functions. The Court itself would be placed in a most difficult position in attempting to resolve the issue without the benefit of full evidence.

Notwithstanding that permission to amend the notice of appeal was refused, Mr Kieron Beal QC, for HMRC, proceeded with the appeal against the UT’s decision to set aside the decision of the FTT which upheld the BTI classification of the Device to sub-Heading 8531 80 95.

This left the court with an appeal hearing tinged with a heavy air of unreality since the current stance of HMRC is, as I have said, that the Device is properly allocated to Heading 9027 but HMRC cannot challenge the decision of the UT on that ground. HMRC no longer contends that the FTT was correct to allocate the Device to Heading 8531, even though HMRC wishes to overturn the decision of the UT. It wishes to overturn the decision of the UT solely in order to set aside the UT’s classification of the Device under Heading 9026. The purpose of doing so is to be able to refute the claim by Honeywell for repayment of customs duty levied on the basis of the Device being classified under Heading 8531 80 95.

Mr Beal submitted that it is possible for HMRC to follow that course because the UT made errors of law in reaching their decision.

In order to assess whether HMRC is correct in that assertion, it is convenient to pare down the decision of the FTT to its critical elements.

They were as follows. Firstly (in para. [37]), the Device had the characteristics and properties of an alerting device. Secondly (in para. [38]), both the intended use, and the actual use, of the Device was the alerting of its wearer to the presence of noxious levels of gas in a confined space by different types of alarm signal; the Device was for doing that alerting, and that alerting was the only thing that the Device was intended to be used for. Thirdly (in para. [78]), Heading 8531 described accurately and clearly the essential characteristics and properties and the use (the only intended use) of the Device, and therefore the description in Heading 8531 applied to the Device. Fourthly (in para. 80), the items listed in Heading 9026 are ones whose only function and use is to measure the level of gas etc. Fifthly (in para. [82]), that was not true of the Device, and so Heading 9026 did not apply. Sixthly (in para. [100)], if, contrary to the decision of the FTT, the Device fell within the description in Heading 9026 as well as Heading 8531, then Note 1(m) in Section XVI would have applied, which would have excluded Heading 8531, and so the appropriate classification would have been under Heading 9026.

Turning to the decision of the UT, the critical elements of its decision were as follows. First (para. [51]), by virtue of its objective characteristics, the Device, on the face of it, fell within the scope of Heading 9026 as well as Heading 8531. Second (para. [63]), Heading 9026 is not restricted to measuring apparatus whose principal purpose is to measure for the sake of measurement alone. By implication, third (para. [63]), the items in Heading 9026 are not limited to ones whose only function and use are to measure the level of gas etc. Fourth (para. [68]), Note 1(m) in Section XVI applied, and consequently the Device had to be classified under Heading 9026.

I consider that the UT were both entitled and right to reach the conclusion they did for the reasons they gave.

It is convenient to begin with the UT’s decision on the legal scope of Heading 9026. As I have said, the FTT held (in para. [80]) that the items listed in Heading 9026 are ones whose only function and use are to measure the level of gas etc. The FTT reached that conclusion on the basis of the word “for” in Heading 9026. They said that “while the device is an instrument that does measure “the ... level … of gases”, it is not an instrument for doing that”. It appears that, in reaching that conclusion, the FTT were influenced by the reasoning and decision in the Shimadzu case. I can see no proper basis for such a restrictive interpretation of Heading 9026. As the UT observed (in paras. [56]-[61]), the Shimadzu case is clearly distinguishable on its facts. It is unclear whether the UT expressly ruled that the FTT’s interpretation was wrong but they plainly did so implicitly. It has not been argued on behalf of HMRC on the appeal that the FTT were correct in their interpretation. It is not a ground of appeal.

The UT was also correct to reject (in para. [63]) an interpretation of Heading 9026 limiting it to measuring apparatus whose principal purpose is to measure for the sake of measurement alone. Mr Beal advanced extensive submissions before us to the effect that there is a general principle that, where an item is capable of falling within the wording of more than one heading, the heading which is most consistent with the principal purpose of the item in question should be applied. I consider that it is clear that there is no such principle.

There is a principle that, where a heading specifies a product’s function according to a descriptive term which is undefined, an item will fall within that heading if that specified function is its principal or main function even though the item might also be used for some other purpose. In order to identify the principal function, as is the case whenever classifying an item, regard must be had to the objective characteristics and properties of the item in question. This is the decisive criterion in such an exercise in order to ensure legal certainly and ease of verification. The jurisprudence of the European Court of Justice (in its various incarnations) (“the ECJ”) provides that regard may be had to other evidence indicating the use to which the item is intended or likely to be put. Under its case law such intended use may constitute an objective criterion for classification if it is inherent in the product in question. That was the situation in, for example, C-395/93 Neckermann Versand AG v Hauptzollamt Frankfurt am Main-Os ECR I-4027 (concerning the undefined function of “pyjamas”), C-338/95 Wiener S.I. GmbH v Hauptzollamt Emmerich [1997] ECR I -6495 (concerning the undefined function of “nightdresses”) and T–243/01 Sony Computer Entertainment Europe Ltd v Commission of the European Communities [2003] ECR II-4089 (concerning the undefined function of “video games”).

There is also a principle that, where the objective characteristics and properties of an article are capable of falling within more than one heading, a use which is “theoretically conceivable but highly improbable” is to be ignored. That was the situation in C-459/93 Hauptzollamt Hamburge-St Annen v Thyssen Haniel Logistic GmbH [1995] ECR I-138 (especially at para. 16). It was also the situation, and the reason for the decision, in C-480/13 Sysmex Europe GmbH v Hauptzollamt Hamburg-Hafen 17 July 2014 ECLI:EU:C:2014:1097 (especially at para. 42).

There are many cases which are plainly incompatible with any principle that, where an item is capable of falling within the wording of more than one heading, that heading should apply which is most consistent with the principal purpose of the item in question. They are cases in which the ECJ has held that the item in question did fall within more than one heading but the Court never suggested that the conflict might be resolved by applying any such principle advanced on behalf of HMRC. The matter was resolved by reference to tie-break provisions in the Notes or the GIRs.

T-243/01 Sony Computer Entertainment Europe Ltd v Commission of the European Communities [2003] ECR II-4189 is a good example. In that case Sony applied for annulment of a Commission regulation which classified items corresponding to, among other things, Sony’s PlayStation®2 under Heading 9504. The regulation was intended to remove uncertainty about the classification of certain items to which a number of different headings could possibly apply. The reasons justifying the classification under Heading 9504 were stated in the regulation. They included GIRs 1, 3(b) and 6, a Note and the wording of various headings. The parties agreed, and the ECJ held, that at the time of the contested regulation Sony’s PlayStation®2 fell within Heading 8471 as an automatic data-processing machine. It was in dispute whether, as the Commission contended, it also came under Heading 9504 which included video games of a kind used with a television receiver. The ECJ held that it did since PlayStation®2 was intended mainly to be used to run video games. The ECJ, rejecting the submission of the Commission, held that neither GIR 3(b) nor any of the other reasons stated in the regulation as justifying the classification under Heading 9504 applied. The ECJ held that Note 1(p) in Section XVI (which stated that Section XVI, which included Heading 8471, did not cover articles within Chapter 95, which included Heading 9504) would have sufficed to classify PlayStation®2 under Heading 9504 but that provision had not been given as a reason in the regulation. The ECJ held that, accordingly, the contested regulation had to be annulled insofar as it classified items such as PlayStation®2 under Heading 9504.

That case shows clearly that, even though PlayStation®2 fell within Heading 9504 because it was intended mainly to be used to run video games, that did not disentitle it to classification under Heading 8471 as well. The final determination could only properly be done by reference to Note 1(p) in Section XVI. There was no hint of any general principle that where an item falls, on the face of it, within more than one heading, it is allocated only to that heading which is consistent with the item’s principal purpose, function or intended use.

Case C-14/05 Anagram International Ltd v Inspecteur van de Belastingdienst (Rotterdam) [2006] ECR I-6763 concerned a clarifying regulation which classified items similar to the gas-filled balloons in issue under Heading 9503 90 32 (other toys not mechanically operated). The ECJ held that the balloons in question fell under that heading whether or not they could also be used as festive articles within Heading 9505. There was no suggestion that, even without the clarifying regulation, the balloons could only fall within that heading which described the main use for which the balloons were intended.

Mr Beal referred to and relied on C-119/99 Hewlett BV v Directeur Général des Douanes et Droits Indirects [2001] ECR I-3981, C-288/09 and C-289/09 British Sky Broadcasting Group plc v Revenue and Customs Commissioners [2011] STC 1519 and C-58/14 Hauptzollamt Hanover v Amazon EU Sarl 11 June 2015 ECLI:EU:C:2015:385. In each of those cases the ECJ used a “principal function” test as a tie-breaker between different headings applicable to the item in question and took into account what consumers would consider to be the ancillary or principal functions. The ECJ did so, however, pursuant to the express provisions of Note 3 in Section XVI. As Ms Valentina Sloane, counsel for Honeywell, observed, the fact that the ECJ was required to resort to Note 3 undermines, rather than supports, HMRC’s argument for a general “principal function” tie-break test applicable to all headings. If such a general test existed, there would be no need for Note 3.

Similarly, in C-467/03 Ikegami Electronics (Europe) GmbH v Oberfinanzdirektion Nürnberg [2005] ECR I-2389 the ECJ resorted to Note 5(E) to Chapter 84 in Section XVI (machines performing a specific function other than data processing and incorporating or working in conjunction with an automatic data processing machine are to be classified in the heading appropriate to their respective functions) as a tie-breaker. There was no reference to a generally applicable principle of “principal function” which might have avoided the need to resort to Note 5(E).

It might equally be said that GIR 3(b) would also be unnecessary if there was a generally applicable tie-break test of “principal function”.

HMRC does not rely on either Note 3 in Section XVI or the GIRs on its appeal and did not do so before the FTT or the UT.

As I have said, it was not contended by HMRC before us, and it was not a ground of appeal, that the FTT were correct to hold, on the basis of the Shimadzu case, that Heading 9026 is limited to items whose only function and use are to measure the level of gas etc. Nor was it contended by HMRC that the UT were wrong to hold that the Shimadzu case was distinguishable on its facts. For completeness, however, I should record that, in his oral submissions in reply, Mr Beal made a reference to, but did not take us to, paragraphs [10] and [12] of the Shimadzu case. It is not clear to me for precisely what purpose he did so. Insofar, however, as he did so in support of an argument that the word “for” in the phrase “Instruments and apparatus for measuring the … level of … gases” applies only to items whose main or principal purpose is to measure, that was not an argument mentioned at all in the two skeleton arguments for HMRC on this appeal and it is not a ground of appeal. Unsurprisingly, we heard no submissions by Ms Sloane on the point.

In the absence of full argument on the point, I decline to express a firm view on it. Firstly, the language in paragraph [12] of the Shimadzu case was with reference to Heading 9030, and was that “only apparatus whose very purpose is to carry out checks on electrical quantities can be regarded as apparatus for checking such quantities”. It is not at all clear that a “very purpose” requirement is the same as a principal function requirement. Secondly, the language used by the ECJ was in the context of the very particular facts of that case, which are materially different from the facts of the present case. Thirdly, even if a “very purpose” requirement is the same as a principal function requirement” in the context of Heading 9030, it is not at all clear that the same is true in respect of all headings in which the word “for” appears. There was no discussion of the point in the context of Heading 9027, which is the heading now favoured by HMRC for the Device. We were not referred to any other heading by way of comparison or analogy on this point. Fourthly, the Sony case illustrates why caution must be exercised on this point. As mentioned above, in that case the ECJ held that Sony’s PlayStation®2 fell within both Heading 8471 and Heading 9504 and that the only legitimate tie break was Note 1(p) in Section XVI. The word “for” appears in both those headings. If that word imports a principal function test, it would not have been possible for Sony’s product to fall under both headings.

It is appropriate at this point to consider another argument advanced by Mr Beal which is not contained in the grounds of appeal and which was rejected by the UT. The argument is that Heading 9026 is, so far as relevant, only concerned with instruments or apparatus for measuring the level of gases whereas the Device is a gas and oxygen analyser, that is to say an instrument which identifies gases and measures their concentrations. Junior counsel for HMRC, Mr Brendan McGurk, first raised this argument in the course of his oral submissions before the UT. It was not a ground of appeal from the FTT and did not form part of his skeleton argument before the UT.

The UT rejected that submission (in para. [67]) because, among other reasons, the whole case had been argued before the FTT on the basis that the meaning of “level” included the composition and concentration of gases in a particular environment. The point is linked to HMRC’s failed applications before the UT and this court to amend the grounds of appeal because it is the reason why HMRC contends that the applicable heading is Heading 9027 rather than Heading 9026.

The UT acted well within their power in refusing to entertain that new and very late argument. It is not, as I said, a ground of appeal to this Court and I would refuse to entertain it both for that reason and because, not having been raised before the FTT, there has been no opportunity for Honeywell to adduce relevant evidence on the point, whether by way of expansion of Mr. Townsend’s evidence or by adducing expert evidence.

I turn to the only outstanding issue which is whether the UT was entitled to conclude, as a matter of fact, that the Device satisfied the description in Heading 9026. Having regard to the principles I have extracted from the cases set out above, the UT would only have been wrong to have done so if it was bound by findings of fact of the FTT to conclude that the Device’s measuring function was “theoretically conceivable but highly improbable”. The FTT never addressed that issue directly because they did not have a correct analysis of the law in mind. It is obvious, however, that the Device’s measuring function was neither simply theoretically conceivable nor highly improbable: as the FTT themselves stated, the Device not only had a measuring function but the alarm function was necessarily dependent on that measuring function, and the objective characteristics of the Device included a removable memory card on which measurements would be recorded and could be analysed. The fact that, as the FTT concluded, the Device would not have been acquired but for its alarm function is neither here nor there on this point.

As I have said earlier, the FTT concluded (in para. 38) that the intended use, and actual use, of the Device was the alerting of its wearer to the presence of noxious levels of gas in a confined space by different types of alarm signal and that “alerting [was] the only thing the device [was] intended to be used for”.

It is clear that the FTT did not intend to convey, by those quoted words, that the only function performed by the Device was alerting. In the very next paragraph (para. [39]) the FTT said that they also found that one of the things the Device did in order to be able to give its alerts was measuring, and it noted that HMRC did not dispute that measuring was one of the Device’s “functions”. What was important, from the FTT’s perspective, was their finding (in the last sentence of para. [39]) that “the measurement is a means to an end, not an end in itself; measurement is not its intended use”. That was echoed later in the FTT’s decision when (in para. [91]) they said that “the measuring is subordinate to the alarm function”.

It is also clear that the FTT did not intend to say, and did not find, that the measuring function had no value or use in addition to its alerting function. In describing the Device (in para. [35]) the FTT stated, as findings of fact, that: “Optionally the user may log the readings and the Device stores several months of continuous data on a removable memory card”. Furthermore, the FTT also concluded (in para. [92]):

“Any measurements the device makes that may be of interest to the operator’s employer will we presume be scrutinised when the operator has come out of the confined space or will be examined and analysed on a computer into which the device’s memory card has been transferred”.

Weaving together these different findings, it seems clear that what the FTT were saying was that, although the Device had a measuring function, which was necessary for the operation of the alerting function, and although it had a capability for recording measurements for future examination and analysis, a person would only acquire the Device because it had an alarm function, and in that sense the measuring was subordinate to the alarm function. That was sufficient to knock out the Device from Heading 9026 on the FTT’s view of the law (in para. [80]) that the items listed in Heading 9026 are ones whose only function and use are to measure the level of gas etc.

On any footing, it is plainly impossible to say, from the objective characteristics of the Device, that the FTT made a finding of fact that the Device’s measuring function was “theoretically conceivable but highly improbable”. To contend otherwise would be to enter into a world of complete unreality. The Device could not work without the measuring function. Indeed, HMRC’s current stance, together with that of other EU Member States, that the proper categorisation is under Heading 9027 only makes sense if measuring is one of the functions of the Device.

HMRC challenges on this appeal the statement of the UT (in para. [24]) that, for the purposes of classification, marketing materials and a product’s targeted use are not to be taken into account. HMRC also challenges the statement of the UT (in para. [53]) that the finding of the FTT that the measurement display was a means to an end, not an end in itself and the only use was for alerting, was nothing more than an “assumption”.

I agree with HMRC that the UT were wrong to say that, for the purposes of classification, marketing materials and a product’s targeted use are always irrelevant. That is apparent from what I have said earlier. I also agree with HMRC that the UT were not entitled to interfere with the FTT’s decision insofar as the FTT found as a fact that a person would only acquire the Device because it had an alarm function, and in that sense the measuring was subordinate to the alarm function.

For the reasons I have given, however, those errors on the part of the UT are irrelevant to the analysis since the only question of fact that was material to classification under Heading 9026 was whether the Device had a measuring function which was no more than “theoretically conceivable but highly improbable”. The objective characteristics of the Device and the findings of the FTT plainly show that the Device’s measuring function was more than that. Any other conclusion would not merely be inconsistent with HMRC’s current stance that Heading 9027 is the applicable heading. It would be perverse.

On the above analysis, no issue arises on the GIRs under the second ground of appeal.

For those reasons, I would dismiss this appeal.

It is unfortunate that his matter comes before this court in circumstances where the issue raised will prospectively become outmoded by the Device subsequently being allocated to Heading 9027. But that particular allocation was not raised before, or debated at all in, the First-tier Tribunal; and it was not permitted (for entirely understandable reasons) to be debated, on amendment, in the Upper Tribunal. I entirely agree with the Master of the Rolls that the Upper Tribunal gave cogent reasons for refusing to allow the point to be raised by amendment at that stage as a new ground of appeal; and that there is also no proper basis for permitting such a point to be pursued in this court.

However, as between the parties the matter still may have some potential relevance: for example, it may bear on the issue of reclaims, if not also, possibly, on other matters.

In my view, the First-tier Tribunal correctly directed itself as to the relevant principles at paragraph 70 of its decision. It said, in summary:

The First-tier Tribunal thus fully understood that the essential exercise was an objective one; subjective considerations had no part to play.

That being so, the task for the First-tier Tribunal was to find and evaluate the relevant facts and to reach its conclusion in the light of the applicable principles accordingly. It is, indeed, a general principle in this context that tariff classification is a matter of factual evaluation in the light of the features and properties of the item to be classified.

In my opinion, the Upper Tribunal made a clear error of law in holding that marketing materials and a product’s targeted use were not to be taken into account. There are numerous authorities from the European jurisprudence to contrary effect. That erroneous approach then infected the Upper Tribunal’s approach to appraising the First-tier Tribunal decision: including its saying, somewhat dismissively, that the First-tier Tribunal’s finding that measurement was not an intended use was “made on nothing more than an assumption as to the use to which the Device was to be put” and saying that that finding “went beyond an assessment of the objective characteristics.” However, that had not been an unwarranted assumption on the part of the First-tier Tribunal. On the contrary, it had been an evidence - based conclusion, in significant part drawn (properly) from the evidence of marketing materials and targeted use.

I need not say more on this aspect, however, as I understand that the Master of the Rolls and Sales LJ are agreed that the Upper Tribunal wrongly directed itself in this respect. Where they differ is as to what the consequence is. Putting it another way, was the Upper Tribunal’s error of law itself material? The Master of the Rolls is of the view that that error on the part of the Upper Tribunal ultimately is “irrelevant”.

Once, however, it is accepted that the First-tier Tribunal had not misdirected itself as to the applicable legal principles, I rather struggle to see how it is to be said that the First-tier Tribunal’s factual evaluation nevertheless was not one properly open to the First-tier Tribunal or that the Upper Tribunal’s decision (substituting, albeit from an erroneous basis, its own conclusion) was one then open to the Upper Tribunal.

The First-tier Tribunal had concluded that the only intended use of the Device was as an alarm: and that “alerting is the only [its emphasis] thing the Device is intended for”: paragraph 38. It reiterated, at paragraph 77, that: “the only intended use (and in fact the only conceivable actual use) of the Device is as an instrument for alerting its operator….” In this regard, Ms Sloane submitted, in line with the Respondent’s Notice, that such findings were perverse. I understand the Master of the Rolls to accept that.

Of course it is to be accepted – and the appellants have never disputed – that one of the objective properties or functions of the Device was that of measuring the concentration of gases. That, indeed, was an inherent and essential aspect of the Device: there could be no alarm function without a prior measurement of the concentration of gases to initiate the alarm. But that does not of itself determine what the Device was for. The question ultimately, as the First-tier Tribunal rightly appreciated, for the purposes of Heading 9026 as compared with Heading 8531, was just that: was the Device for measuring or checking the flow etc of liquids or gases? The First-tier Tribunal found that the measuring function was entirely subordinate to the alarm function. It in terms found, rejecting the respondent’s case on this, that the measuring function was simply a means to an end: the end being the alerting function. Overall, therefore, the specific factual finding was that the measurement or checking of gases was not what the Device was for. On that basis, it did not fall within Heading 9026.

It is maintained on behalf of the respondent, however, that such a conclusion downplays the fact that, as a matter of its objective characteristics, the Device also was able to memorise, record and display measuring information and that it moreover could be disabled from its alerting function. That indeed was precisely one of the points emphasised before the First-tier Tribunal. Its finding, however, was that that particular capacity or characteristic was of no real materiality and that such a capacity or characteristic was not part of the intended use of the Device: the First-tier Tribunal explaining why it so found. Ms Sloane’s argument in effect came down to an assertion that, overall, the First-tier Tribunal had no option other than to accept the respondent’s case on this. I cannot agree with that. It was precisely what was in issue and what, on the evidence, the First-tier Tribunal had to evaluate and decide.

Moreover, the ultimate conclusion by the First-tier Tribunal amply accords, as I see it, with the approach indicated in various of the European authorities. Uses of a particular device which are found to be but a theoretical possibility are on any view to be disregarded: see, for example, among other cases, the Sysmex case (cited above). And that is what the First-tier Tribunal had in terms found to be the case here, on the evidence.

In my opinion, therefore, the fact that the Device here also had the capacity to store and record measuring information and the capacity to have the alerting function disabled did not necessarily mean that measurement was one of its intended uses, let alone its principal purpose or intended use. I cannot accept that the First-tier Tribunal’s conclusion on this as a matter of fact was perverse or was one that no reasonable tribunal could properly reach. It is a very strong thing to set aside a fully reasoned decision of an expert tribunal on such a basis; and I can see no sufficient justification for doing so here. As stated by Longmore LJ in Bowater v North West London Hospitals NHS Trust [2011] EWCA Civ 63 at paragraph 19 (in the context of an employment case but in words of general application) an appellate tribunal “must not, under the guise of a charge of perversity, substitute its own judgment for that of the [fact-finding tribunal].”

Accordingly I would for my part allow the appeal on this basis. I would set aside the Order of the Upper Tribunal and restore the Order of the First-tier Tribunal. In reaching this conclusion I have of course anxiously considered the judgments, in draft, of the Master of the Rolls and Sales LJ. As will be gathered, I agree, with respect, with the reasoning and the conclusion of Sales LJ. I also agree with him that this is a very fact-specific exercise on a narrow point, by reference to the case as presented before the First-tier Tribunal.

That makes it strictly unnecessary, so far as I am concerned, to rule on the further point then arising and which the Master of the Rolls considers decisive. That is whether (contrary to the First-tier Tribunal’s actual findings) the asserted object and potential use for recording and displaying measuring information was at any rate separate from the main object or intended use as an alerting device in a way which (contrary to the First-tier Tribunal’s actual findings) could not be described as theoretical or conceptual: and thus ultimately caused the Device to be within Heading 9026: which then, by operation of Note 1(m), would prevail. This was the approach of the Upper Tribunal, which I understand the Master of the Rolls favours.

I will, however, albeit relatively briefly, say why I have reservations as to the correctness of this.

On any view, the respondent’s argument that checking or measuring the level of gases was the principal function of the Device had failed conclusively on the facts before the First-tier Tribunal. Nor (even on its mistaken legal approach) did the Upper Tribunal make such a finding. It simply, as I read its decision, proceeded on the footing that measuring was at least one purpose and one of the intended uses. To the extent that Ms Sloane, as I understood her, sought to maintain before us that measuring the level of gases was indeed the principal purpose and principal function of the Device that is now untenable: although it is perhaps revealing that that seems to have been the respondent’s primary case advanced before the First-tier Tribunal.

That being so, I think that, on this scenario, there still are real doubts as to whether the Device could fall within Heading 9026. I entirely understand the point that, where there are competing functions or uses, the Notes and GIRs are potentially available and the need to resort to a “principal function” approach becomes less obvious; and I agree also that a number of the authorities in the European jurisprudence, as cited to us, are decisions where the Explanatory Notes feature.

However, as I read them, a number of the authorities do also invoke a principal function or main intended use principle in order to help assess what particular classification heading is appropriate to the particular product in question: such principle operating independently of, albeit concurrently with, Explanatory Notes or GIRs.

Thus in the Hauptzollamt Hannover v Amazon case (cited above) the Court of Justice accepted that, while Amazon’s Kindle device had a dictionary function, that did not make the device an electronic dictionary for classification purposes: since that was not the “principal function” of a Kindle. It is true that Note 3 to Section XVI featured in that case; but the statements of principle there set out seem to go wider than that. The cases of Neckerman, Sysmex and Sony Corporation (cited above), among others, also all contain statements of principle to like effect: again, statements of principle not based on Notes or GIRs. For example, in Sysmex (at paragraph 32) this was said, in unqualified terms:

“… according to the case law, in order to be classified under the tariff-heading relating to a use, the product to be classified need not be solely or exclusively intended for that use. It suffices that that use is the main use for which the product is intended…”

Accordingly, even on this alternative basis I would query if the respondent could have succeeded here, unless it could have established that it was the (or, possibly, a) main intended use of the Device to measure or check the level of gases. And that it had not established.

That said, I would, in agreement with Sales LJ, allow the appeal on the ground that the First-tier Tribunal made findings and drew conclusions properly open to it on the evidence; and that there was no sufficient basis in law for the Upper Tribunal substituting its own view of the matter.

I am grateful to the Master of the Rolls for setting out the legal framework for the BTI decision in this case so clearly. As he explains, HMRC were refused permission by the Upper Tribunal to introduce for the first time on appeal an argument in favour of classification of Honeywell’s Gas Alert Micro 5 under heading 9027 in the CN and there is no good ground to impugn that decision in this court. Thus, for procedural reasons, the appeal in this court falls to be determined on what is an artificial and limited basis, where the choice of classification headings in relation to the Device is to be taken as limited to the binary choice between heading 8531 and heading 9026, even though it can now be seen that the correct applicable heading is quite possibly heading 9027. Nonetheless, it is appropriate for this court to decide the case, framed as it is before us and as it was before the FTT and the Upper Tribunal as involving an artificial choice of allocation between heading 8531 and heading 9026, rather than treating it as a purely academic exercise. This is because Honeywell seeks to maintain that it has the benefit of the Upper Tribunal decision in its favour and therefore may rely on the principle of res judicata to insist on a tariff classification under heading 9026 for its goods pursuant to the relevant BTI. To explain the significance of this, the tariff rate under heading 9026 is zero, the rate under heading 8531 is 2.2% and the rate under heading 9027 is 2.5%.

The Master of the Rolls would dismiss the appeal on the basis that he regards the decision of the Upper Tribunal as one which it was entitled to make for the reasons it gave. In respectful disagreement with him, I consider that the Upper Tribunal erred in its approach and that its reasoning was flawed. The questions then arise whether its error was material and whether the decision of the FTT should be upheld or the case remitted for further consideration. In my view, the Upper Tribunal’s error was material. I also consider that the FTT’s decision to classify the Device under heading 8531 was lawful, in the context of the artificially limited choice of classification either under heading 8531 or heading 9026 with which it was confronted. Therefore, there is no good reason to remit the case for further consideration. I would allow the appeal, with the consequence that the FTT’s decision stands.

I should explain that I agree with the observations of the Master of the Rolls at paras. [67] to [69] above regarding the principles which emerge from the case-law of the CJEU in relation to the interpretation and application of the Tariff Regulation and the CN. However, in my opinion, in this case the FTT was lawfully entitled to find on the evidence before it and in the context of the artificial binary choice presented to it that the Device fell under heading 8531 and not under heading 9026. There was therefore no scope for the application of Note 1(m) in Chapter XVI of the Regulation, as relied upon by Honeywell and the Upper Tribunal.

Some aspects of the factual and procedural background are of particular significance. As the Master of the Rolls has explained, Honeywell asked HMRC for a BTI for its Gas Alert Micro 5. In its application Honeywell described the item as “a gas monitoring device which is carried on the person (portable) and used by people who work in confined spaces and may have reason to come into contact with potentially dangerous toxic gases … The units contain both an audible, visible and a vibration alert mechanism” and asked for it to be classified under heading 9026 in the CN, rather than heading 8531. Honeywell stated that it believed that the principal function was that of gas detection/measurement; that the sound or visual signalling aspect was subordinate to and just a by-product of this; and that the product was frequently used for its display capability rather than for test purposes, and the user might never view or hear any of the alarm features.

However, HMRC issued a BTI dated 19 December 2013 classifying the goods under heading 8531 and this BTI was affirmed by HMRC by a decision dated 16 October 2013 after a review requested by Honeywell (“the BTI”). By a letter dated 18 January 2017, Honeywell was informed that the BTI ceased to have effect from 1 January 2017. What is in issue, therefore, is the classification of the goods for the period until 31 December 2016.

Honeywell exercised its right of appeal to the FTT against the BTI, contending that the goods should be classified under heading 9026 rather than heading 8531. As explained above, the hearing before the FTT proceeded on the basis that there was a binary choice to be made, between classification of the goods under heading 9026 or under heading 8531. The possibility of classification under heading 9027 was not raised.

HMRC’s case was that, contrary to Honeywell’s claim that the principal function of the Gas Alert Micro 5 was detection/measurement, its function or principal function was as an alarm. The name of the product indicated this. The relevant Harmonised System Explanatory Note for heading 8531, at para. (G), indicated that items within it included “Electric vapour or gas alarms, consisting of a detector and a sound or visual alarm, to warn of the presence of hazardous gaseous mixtures (e.g. natural gas, methane)”. HMRC argued that although the device detected the amount of gases, this was a function in service to its ultimate purpose, namely the indication by the alarm when a dangerous level of gases had been reached. The ultimate, main or overriding purpose of the device was to provide a warning to persons working in confined spaces, having first detected the gases, rather than solely to measure the pressure or volume of a gas.

In further support of their case, HMRC also relied on the published manuals for use of the Device and Honeywell’s product description, available on its website. The Quick Reference Guide stated that the device was “designed to warn of hazardous gas levels above user-defined alarm setpoints”. The Manual included a Safety Information notice, which stated “Use the detector only as specified in this user manual, otherwise the protection provided by the detector may be impaired.” It also included a recommendation that the sensors should be bump tested before each day’s use, “to confirm their ability to respond to gas by exposing the detector to a gas concentration that exceeds the alarm setpoints. Manually verify that the audible and visual alarms are activated.” As emphasised by the FTT at [28], there was a prominent warning that extreme caution should be used when disabling a sensor, as then it “cannot detect and alarm against the applicable gas.” On a fair reading of the Manual as a whole, it is directed to ensuring that the device is maintained so that the alarm function is effective, as the witness for Honeywell, Mr Townsend, agreed and the FTT found: [25]. Although it appears the alarm function can be switched off and the device also has a visual display, there is no section of the Manual addressed to use of the device other than as an alarm. The product description appeared under the heading, “Protect yourself” and was set out at paras. [17] and [18] of the FTT’s decision.

The FTT referred in detail to the evidence of Mr Townsend at [21]-[32]. It found that the device in issue required the operator to enter a confined space with the Device on their person: [23]. Mr Townsend agreed under cross-examination that although the alarm could be disabled at the discretion of the user, it was not best practice to disable the alarm detection sensors in a gaseous environment: [27]-[28]. Although he was concerned to promote Honeywell’s case on the appeal, he also agreed that the alerting function was “one of the main applications of the device” and that the emphasis in all the literature issued by Honeywell was on protection and safety: [25]. As noted above, the FTT found (as it plainly was entitled to) that this was in any event clear from the literature itself.

Having considered all the evidence before it in some detail, the FTT found that the Device has the physical characteristics and properties of an alerting device: [37]. It also found that the intended and actual use of the Device is the alerting of the wearer to the presence of noxious gases in a confined space by at least one and usually two or three different types of alarm signal, visual, audible and vibrating; in answer to the question, “what is the device for?”, the FTT found that “it is to do that alerting”; and it also found “that alerting is the only thing the device is intended to be used for”: [38]. It also found that although the Device measured the quantity of gas in the environment by reference to time, that measurement “is a means to an end, not an end in itself; measurement is not its intended use”: [39]. At para. [40] the FTT referred to Mr Townsend’s evidence that the alarms could be disabled, but set it against the information about safety checks in the Manual requiring a check that the alarms were working before entering a confined space, and said “we find it inconceivable, and undoubtedly more likely than not, that this particular device … would ever be used by an operator in a confined space without any alarms enabled, even if it was possible to disable them all”: [40].

Honeywell, pursuant to its respondent’s notice, contends that these findings were perverse. In my view, that is an unsustainable contention. In my opinion, there was ample evidence, as referred to above, which entitled the FTT to make the findings of primary fact which it did, in particular regarding the intended use of the Device, and to arrive at the overall evaluative conclusion in light of those findings that, as between heading 8531 and heading 9026, the device should be classified under heading 8531 and could not properly be classified under heading 9026.

I note, as did the FTT, that if the FTT had concluded that the Device was capable of being assigned to both of these headings, then Note 1(m) in Section XVI of the Regulation would have required that priority should be given to classification under Chapter 90, so that it was allocated to heading 9026 in preference to heading 8531. Note 1(m) was set out by the FTT at [43] and the argument for Honeywell based upon it was referred to at [59], [95] and [98]-[100]. If the Device falls within the scope of both heading 8531 and heading 9026, Note 1(m) prescribes how it should be classified. There would be no need or justification for recourse to the tie-break provisions in Rule 3 of the General Rules.

However, in my judgment, the findings made by the FTT regarding the intended use of the Device, as appeared objectively from the literature issued by the respondent in relation to the device (and as emphasised by the FTT at [77], [78] and [80] in its analysis in the section of its decision headed “Discussion”), were such that it was entitled to find that the Device did not fall within heading 9026. It explained that this was its view, e.g. at [99]. In applying the text of heading 9026 in the context of its factual findings, the FTT was entitled to make the evaluative judgment that the Device was not an instrument “for measuring or checking the … level … of gases”: see [40] and [80]. This is on the basis, in particular, of the FTT’s finding that “no employer would send an operator into a confined space with this device to measure the level of dangerous gases without the alerting functions being operative”: [80]. That was a finding which the FTT was entitled to make on the evidence before it. In the light of that and its other findings, the FTT was entitled to make the evaluative judgment which it did in deciding that the Device did not fall under heading 9026.

The Upper Tribunal held that the FTT erred by taking into account Honeywell’s marketing materials and its targeted use and therefore found that in making this evaluation it went beyond an assessment of the objective characteristics and properties of the Device itself, which was all that was permissible under EU law: see [9], [24] and [53]-[55]. Unfortunately, however, I consider that the Upper Tribunal has itself erred in its analysis. I think it is clear from relevant EU case-law that the marketing literature and manuals issued by a producer of an item are themselves part of the objective materials to which it is legitimate and appropriate to have regard when considering the application of the tariff headings.

This point emerges clearly in the judgment of the CFI in Case T-243/01 Sony Computer Entertainment Europe Ltd v Commission of the European Communities [2003] ECR II-4189. That case concerned the allocation of Sony’s PlayStation®2 device for playing video games, which was also capable of being used as a basic computer, as between heading 8471 (“Automatic data-processing machines [etc]”) and heading 9504 (which included “Video games of a kind used with a television receiver”). The relevant Section notes prescribed that headings in Chapter 84 did not apply to articles within Chapter 95 - i.e. the equivalent of Note 1(m) in the present case. The CFI accepted that heading 8471 did apply. The question for it, therefore, was whether heading 9504 applied as well, since if it did it would be the applicable heading by virtue of the relevant note.

The CFI found that heading 9504 did apply. There was no definition of “video games”, so the CFI held, following Case C-395/93 Neckermann Verstand [1994] ECR I-4027, at [6]ff, and Case C-338/95 Wiener SI GmbH [1997] ECR I-6495, [13]-[14], that “it was appropriate to look for the objective characteristic of those goods which tended to distinguish them from others in the use for which those goods were intended” and to assess in that regard whether the goods were used mainly for the purpose given by a tariff heading, even though it might also be possible to envisage another use for them: see [110]. It was appropriate to consider as video games “any products which are intended to be used, exclusively or mainly, for playing video games, even though they might be used for other purposes”: [111]. In making that assessment, the CFI had regard to “the manner in which the PlayStation®2 is imported, sold and presented to the public” and to promotional material which indicated how it was marketed and sold to consumers, namely as a video game console: [112]-[113]. These materials and the way in which the device was configured showed that it was intended for use mainly for playing video games, even though it might be used for other purposes: [112]-[113].

Therefore it is in my view clear that the FTT in our case was fully entitled to take into account the manuals and other information about the product presented by Honeywell to consumers in the way that the FTT did. Conversely, the Upper Tribunal was wrong to leave these materials out of account when undertaking its own assessment of classification as between heading 8531 and heading 9026. Such material forms part of the objective characteristics and properties of the goods in question for the purposes of applying the classification headings in the tariff Regulation. The relevance to tariff clarification of the objective manner in which an item is presented to consumers or users is also confirmed by the judgment of the CJEU in Joined Cases C-288/09 and C-289/09 British Sky Broadcasting Group [2011] STC 1519, at [77]-[79]. Indeed, given the importance for tariff classification under various headings of the use to which an item is intended to be put, it seems to me that it would be most odd and contrary to principle to leave out of account the way in which consumers are encouraged to use the item in question by materials placed into the public domain and objectively verifiable for the purposes of tariff classification.

Although the basis on which the Upper Tribunal allowed the appeal against the decision of the FTT was wrong, as I would hold, Honeywell says that the decision of the FTT can nonetheless be seen also to have been wrong. I disagree.

First, as the Master of the Rolls has explained at para. [68] above, there is a principle of EU law in assessing the application of tariff headings that one should ignore uses of an item which are conceivable but are in reality “no more than a theoretical possibility”: see Case C-480/13 Sysmex Europe GmbH ECLI:EU:C:2014:1097, [32] and [42]. To use a domestic law analogy, this is a de minimis type principle. In the Sysmex Europe case, a substance for testing for the presence of white blood cells was found to be classifiable under heading 3822 (laboratory reagents) rather than heading 3212 (dyes and other colouring matters), because its use as a colouring matter was no more than a theoretical possibility. This meant that its use as a laboratory reagent constituted its exclusive use for the purposes of classification, so that it had to be assigned to heading 3822: [42]-[43].

On the finding properly made by the FTT in our case, that the Gas Alert Micro 5 was intended for use only as an alarm – and in the context of the binary choice between heading 8531 and heading 9026 which the FTT had to make - it is my view that the FTT was entitled to find that it should be classified under heading 8531 and not under heading 9026.

On the findings made by the FTT, there was no real possibility that the Device would be used separately as a device for measuring the level of gas. At most, that could only be regarded as a theoretical possibility.

That left the distinct argument by Honeywell that in order to operate as an alarm, it was necessary for the Device to go through a process of measuring the gases in its environment and that it should therefore be regarded as an item falling under heading 9026. However, so far as that is concerned, as I have noted, the FTT was entitled to find that the sole intended use of the Device was as an alarm. That being so, I consider that in making its evaluative assessment the FTT was not required to go further and to disaggregate that objective in a wholly artificial manner for the purposes of tariff classification and say that it included within it as a distinct objective the measurement of the level of gases.

Further, in the context of the artificial binary choice which the FTT had to make as between heading 8531 and heading 9026, the FTT was entitled to rule that on proper interpretation of heading 9026, in order to fall within it the item in question had to be intended for use as a measuring instrument, as the main or principal purpose for using it. On a comparison of the language of heading 9026 and that of heading 8531, it is my view that this is the force of the word “for” in the phrase, “Instruments and apparatus for measuring or checking the … level … of … gases”, particularly in light of the examples then given in heading 9026 itself. This interpretation is also supported by the general approach to application of the tariff headings set out in the Sony case referred to above. By combination of the particular language used in heading 9026 and that approach, I consider that for Honeywell’s case to succeed before the FTT it was incumbent on it to show that the main or principal use of the Device was to measure the level of gas, which it failed to do.

This analysis leaves out of consideration the impact of heading 9027 upon proper interpretation and application of the tariff Regulation, as both the FTT and the Upper Tribunal did because of the way the debate before them was framed. It is also worth mentioning that the precise language of heading 9026 and its context, reading that heading as a whole together with relevant HSENs, is different from the language and context of heading 9027. So it is in any event doubtful that arguments based on heading 9027 could provide a helpful guide as to the choice between application of heading 8531 and heading 9026, in the context of the artificial exercise of interpretation and application of the tariff Regulation which arose in making the binary choice which fell to the FTT in this case.

It is fair to say that the FTT went too far in interpreting heading 9026 to mean that the items covered by it were ones “whose only function and use is to measure the level etc of gas etc”: [80]. In my view, for the purposes of application of heading 9026 on the interpretive exercise undertaken by the FTT, it would have been sufficient if that was the main function of the Gas Alert Micro 5. Just as in the Neckermann case items of clothing were properly to be classified as pyjamas because that was their main intended use, even though they could also be used in other ways, I consider that a device could be said to be an instrument for measuring the level of gases if that was its main use, even if it might be used for other things. But plainly, on the findings made by the FTT, that was not the main use or function of the Gas Alert Micro 5. Faced with the arguments before it about how to characterise the use or function of the Device, the FTT was entitled to have regard to the main or principal object for which it was to be used, namely as an alarm rather than for measuring the level of gas.

For the reasons I have given, my view is that the decision of the Upper Tribunal cannot stand and that the appeal should be allowed. In light of the findings and evaluation made by the FTT, which were properly open to it on the evidence before it, and in view of the artificial binary choice it had to make, its decision was a proper and lawful one. There is no reason to remit the case to the FTT or the Upper Tribunal for further consideration.

I would, however, wish to emphasise in closing the narrow ambit of what we have had to decide, which is itself a reflection of the narrow ambit of the decisions of both the FTT and the Upper Tribunal. It now appears that the proper classification for the Gas Alert Micro 5 may be under heading 9027. No such argument was presented to the FTT and its introduction was ruled out by the Upper Tribunal. Therefore, none of the FTT, the Upper Tribunal and this Court have been in a position to consider arguments directed to such a classification. I would allow the appeal on the very narrow basis that, given the binary choice it had to make between classification under heading 8531 and under heading 9026, the decision of the FTT was a lawful one. I express no view about what position might now obtain, e.g. if HMRC seek to withdraw the BTI or to support the outcome of the case and the amount of duty payable in respect of the goods in issue in any further dispute between them and Honeywell on some other basis of determination regarding the correct duty payable in respect of the goods.