Urenco Chemplants Limited & Anor v The Commissioners for His Majesty's Revenue and Customs

This case concerns the availability of capital allowances for expenditure on the construction of a specialised facility for the treatment and management of highly toxic and radioactive waste in the civil nuclear industry, by a process known as “deconversion”. The disputed expenditure was incurred by a company in the Urenco group called Urenco Chemplants Limited, in its yearly accounting periods ended 31 December 2011 to 31 December 2015 inclusive. The facility in question is at Capenhurst in Cheshire, and is known as a “Tails Management Facility” or “TMF”.

Construction of the TMF at Capenhurst was substantially completed in 2018, at a total cost of approximately £1 billion. The treatment for capital allowance purposes of most of this expenditure was agreed between Urenco and the Revenue, but disputes remained in relation to claims for allowances of approximately £192 million. These disputes led to appeals to the First-tier Tribunal (“FTT”) brought by Urenco Chemplants Limited against two discovery determinations and four closure notices made and notified by the Commissioners for HM Revenue and Customs (“HMRC”) in December 2017, covering the five years in issue. There was also a further appeal by another company in the group, Urenco UK Limited, in respect of consequential losses surrendered to it by Urenco Chemplants Limited.

Nothing turns on the separate identity of companies in the Urenco group, and I shall therefore follow the FTT and the Upper Tribunal (“UT”) in generally referring to the two appellant companies, as well as to the corporate group, as “Urenco”.

The FTT (Judge Jonathan Cannan) released its decision (“the FTT Decision”) on 7 August 2019, after a four-day hearing in December 2018 and an earlier one-day site visit on 14 September 2018, when Judge Cannan said that he was “shown all aspects of the TMF” before it became operational: see the FTT Decision, [2019] UKFTT 522 (TC), at [9]. The parties were represented before the FTT, as they have been subsequently, by Mr Jonathan Peacock KC leading Mr Michael Ripley for Urenco, and Mr Jonathan Bremner KC leading Mr Edward Waldegrave for HMRC. I express the gratitude of the court to them all.

The FTT held that Urenco was not entitled to capital allowances for any of the disputed expenditure, because (a) most of it was not incurred “on the provision of plant or machinery” within the meaning of section 11 of the Capital Allowances Act 2001 (“CAA 2001”), and (b) the whole of the expenditure was, in any event, excluded from allowances by section 21 of CAA 2001, on the grounds that it was expenditure “on the provision of a building”. The FTT also considered whether any of the expenditure was saved from the effect of section 21 by certain exceptions, or “carve-outs”, contained in List C set out in section 23 of CAA 2001 (“List C”). The FTT concluded that none of the exceptions upon which Urenco sought to rely was applicable. Accordingly, Urenco’s appeals were all dismissed.

Urenco then appealed to the UT (Mellor J and Judge Thomas Scott), which released its decision (“the UT Decision”) on 28 January 2022: see [2022] UK UT 00022 (TCC). It is notable that Urenco did not challenge any of the FTT’s findings of fact, nor did either side contend that the FTT had misunderstood the legal principles which govern the question of what constitutes “plant” for the purposes of CAA 2001. Nevertheless, the UT detected what it considered to be material errors of law in the FTT’s treatment of the issues relating to the “provision of plant” and the “provision of a building” under sections 11 and 21 of CAA 2001. The UT therefore set aside the FTT Decision and remitted the case to the FTT to remake the relevant decisions on those two issues. The UT also dismissed Urenco’s appeal on the potential applicability of Items contained in List C.

Both sides now appeal to this court against aspects of the UT Decision. HMRC pursue three grounds of appeal, for which permission was granted by the UT. In short, the grounds are that the UT was wrong in law:

Urenco appeals on two grounds, with permission granted for the first by the UT, and for the second by this court (Lewison LJ) on 16 May 2022. The first ground relates to Items 1 and 4 of List C, and contends that, properly construed, they apply to expenditure “on the provision of” those items, and not merely to expenditure “on” them as the UT held. The second ground relates to Item 22 in List C, which provides a carve-out for expenditure on “[t]he alteration of land for the purpose only of installing plant or machinery”; it contends that the UT wrongly held Item 22 did not apply to the disputed assets on the facts found by the FTT.

At this stage it is unnecessary to give more than a short summary of the basic facts, mainly taken from the “background” section of the UT Decision at [3] to [10].

The Urenco group provides about 30% of the global enriched uranium supply for the civil nuclear industry. The group has uranium enrichment plants in the UK, Germany, the Netherlands and the USA. The UK facility is at Capenhurst.

Depleted uranium hexafluoride, or “Tails”, is a radioactive and highly toxic by-product of uranium enrichment. The TMF at Capenhurst was built to process Tails safely by way of “deconversion”. Tails are not only radioactive and toxic, but also unstable and highly corrosive. The deconversion process carried out in the TMF involves removing the fluorine content of the Tails in the form of hydrofluoric acid, leaving a stable but still radioactive uranium oxide compound which can be stored more easily. The hydrofluoric acid which is produced can be sold for industrial use unless it has any radioactive contamination, in which case it is safely destroyed. The uranium oxide is stored at the TMF.

The TMF at Capenhurst processes Tails from Urenco’s facilities in the UK, Germany and the Netherlands. It has eight to ten operators on shift at any one time. Certain areas of the site are designed to be unoccupied save for necessary inspection and maintenance purposes.

There are five key facilities at the TMF, which operate as follows:

As one would expect, Urenco was obliged to comply with stringent regulatory and Health and Safety requirements in respect of the design, construction and operation of the TMF. The FTT made detailed findings of fact about these matters in the FTT Decision, at [24] to [39]. Thus, for example, the Capenhurst site was required to have a nuclear site licence issued by the Office for Nuclear Regulation, and Urenco had to meet safety objectives framed in terms of outcomes which had to be achieved, such as maximum permitted levels of radiation dosages. The FTT also explained, at [26]:

“In order to satisfy the safety objectives, it has been necessary to construct certain “safety significant structures”. The purposes of safety significant structures are:

i)

To provide radiation shielding, blocking the path of radiation, and/or

ii)

To provide containment, preventing the release of radioactive particles, and/or

iii)

To support machinery, equipment and various structures to ensure that they will continue to perform their safety functions in the event of a 1-in-10,000 year earthquake, known as “seismic qualification”.”

It is also important to note what the FTT said about the external appearance of the TMF, at [39]:

“From the outside the various facilities comprising the TMF give the appearance of a single modern industrial type building albeit with different roof heights. The [cylinder handling facility] and the [uranium oxide store] give the appearance of being large warehouse type structures. This is the effect in part of the external cladding described below. It was a condition of the planning permission that the facilities have the appearance of a modern business park. The site is very close to local community facilities. Despite appearances, each of the facilities under consideration is structurally independent. The cladding can be removed and replaced.”

The FTT helpfully appended to the FTT Decision a series of simplified block diagrams, showing the various facilities and components. The diagrams identify the main elements of each structure, and the main equipment and machinery in that structure. Colour-coding distinguishes between assets and safety- significant structures which HMRC had accepted as qualifying for capital allowances, and those which were in dispute. As the colour-coding reveals, most of the items in dispute are “safety significant structures” shown shaded in grey. Broadly speaking, these are made up of the shielding structures within each of the five main facilities at the TMF, such as reinforced concrete walls, roofs and floor slabs. By contrast, there is no dispute in relation to the major items of machinery and equipment which perform the key functions at each stage of the deconversion process, such as the cranes, autoclaves, kilns, condensers and storage vessels, all of which are agreed to be plant.

The UT also appended the same diagrams to the UT Decision. Since they may readily be consulted by reference to either Tribunal decision, we have not thought it necessary to append them to our judgments as well.

The legislation relating to capital allowances in force during the years under appeal was contained, as it still is today, in CAA 2001. This was principally a consolidation statute, the purpose of which, according to its long title, was “to restate, with minor changes, certain enactments relating to capital allowances”. CAA 2001 also formed part of the Tax Law Rewrite project, and as such falls to be interpreted in accordance with the principles reviewed by Lord Carnwath JSC in R (Derry) v Revenue and Customs Commissioners [2019] UKSC 19, [2019] 1 WLR 2754, (“Derry”) at [7] to [10]. The previous consolidation statute was the Capital Allowances Act 1990 (“CAA 1990”), to which important amendments, excluding certain categories of expenditure from the expression “expenditure on the provision of machinery or plant”, were made by section 83(7) of, and Schedule AA1 to, the Finance Act 1994 (“FA 1994”). This statutory history is relevant to the issues raised on Urenco’s cross-appeal, including the proper construction of what are now Items 1 and 4 in List C (contained in section 23 of CAA 2001).

The general rules governing the availability of capital allowances for plant and machinery are contained in section 11 of CAA 2001:

“11 General conditions as to availability of plant and machinery allowances

(1)

Allowances are available under this Part if a person carries on a qualifying activity and incurs qualifying expenditure.

(2)

“Qualifying activity” has the meaning given by Chapter 2.

(3)

Allowances under this Part must be calculated separately for each qualifying activity which a person carries on.

(4)

The general rule is that expenditure is qualifying expenditure if –

(a)

it is capital expenditure on the provision of plant or machinery wholly or partly for the purposes of the qualifying activity carried on by the person incurring the expenditure, and

(b)

the person incurring the expenditure owns the plant or machinery as a result of incurring it.

(5)

But the general rule is affected by other provisions of this Act, and in particular by Chapter 3.”

The main relevant provisions of Chapter 3 (of Part 2 of CAA 2001) are as follows:

“21 Buildings

(1)

For the purposes of this Act, expenditure on the provision of plant or machinery does not include expenditure on the provision of a building.

(2)

The provision of a building includes its construction or acquisition.

(3)

In this section, “building” includes an asset which –

(a)

is incorporated in the building,

(b)

although not incorporated in the building (whether because the asset is moveable, or for any other reason), is in the building and is of a kind normally incorporated in a building, or

(c)

is in, or connected with, the building and is in list A.

List A

Assets treated as buildings

1.

Walls, floors, ceilings, doors, gates, shutters, windows and stairs.

2.

Mains services, and systems, for water, electricity and gas.

3.

Waste disposal systems.

4.

Sewerage and drainage systems.

5.

Shafts or other structures in which lifts, hoists, escalators and moving walkways are installed.

6.

Fire safety systems.

(4)

This section is subject to section 23.

22 Structures, assets and works

(1)

For the purposes of this Act, expenditure on the provision of plant or machinery does not include expenditure on –

(a)

the provision of a structure or other asset in list B, or

(b)

any works involving the alteration of land.

List B

Excluded structures and other assets

…

7.

Any structure not within items 1 to 6 other than –

(a)

a structure (but not a building) within Chapter 2 of Part 3 (meaning of “industrial building”),

…

(2)

The provision of a structure or other asset includes its construction or acquisition.

(3)

In this section –

(a)

“structure” means a fixed structure of any kind, other than a building (as defined by section 21(3)), and

(b)

“land” does not include buildings or other structures, but otherwise has the meaning given in Schedule 1 to the Interpretation Act 1978.

(4)

This section is subject to section 23.

23 Expenditure unaffected by sections 21 and 22

(1)

Sections 21 and 22 do not apply to any expenditure to which any of the provisions listed in subsection (2) applies.

(2)

…

(3)

Sections 21 and 22 also do not affect the question whether expenditure on any item described in list C is, for the purposes of this Act, expenditure on the provision of plant or machinery.

List C

Expenditure unaffected by sections 21 and 22

1.

Machinery… not within any other item in this list.

…

4.

Manufacturing or processing equipment;…

…

22.

The alteration of land for the purpose only of installing plant or machinery.

…”

In the light of the legislative structure set out above, it seems to me that the logical starting point in a dispute about the availability of capital allowances for expenditure on plant or machinery is to consider whether the general conditions in section 11 of CAA 2001 are satisfied. In particular, it is necessary to ask whether the expenditure is “on the provision of plant or machinery” within the meaning of section 11(4)(a). If it is not, the claim for capital allowances falls at the first hurdle. If, however, the expenditure is on the provision of plant or machinery, and if the other conditions in section 11 are all satisfied, it is then necessary to go on to consider whether the prima facie entitlement to capital allowances under section 11 is removed or modified by any other provisions of the Act, including sections 21 to 23 in Chapter 3 of Part 2.

In the present context, on the assumption that the section 11 hurdle has been overcome, the next question is whether the disputed expenditure is “on the provision of a building” within the meaning of section 21(1). What constitutes a “building” for these purposes is partly elucidated by the remainder of section 21, including the partial definition of “building” in subsection (3) read (where appropriate) with List A. If the answer is that the expenditure falls within section 21(1), as being expenditure on the provision of a building, then the claim will fall at this second hurdle, unless it is saved by section 23, to which section 21 is expressly made subject: see section 21(4).

Section 23 begins, in subsection (1), by disapplying sections 21 and 22 to any expenditure which falls within section 23(2), which is not material in the present case. However, section 23(3) then further provides that sections 21 and 22 “also do not affect the question whether expenditure on any item described in List C is… expenditure on the provision of plant or machinery”. Accordingly, expenditure on any of the 33 items described in List C is saved from the exclusionary effect of section 21, even if it constitutes expenditure on the provision of a building, although the basic qualifying conditions in section 11 must still be satisfied: it is only the exclusions in sections 21 and 22 which are removed where section 23(3) applies.

Before turning to the grounds of appeal, it is convenient to record some significant matters which have never been, or are no longer, in dispute.

First, it was common ground before both Tribunals, as it was in this court, that all of the conditions of availability of plant and machinery allowances in section 11 of CAA 2001 are satisfied in relation to the disputed expenditure, apart from the question whether the expenditure was “on the provision of plant or machinery”. Thus, it is agreed that Urenco (or, more precisely, Urenco Chemplants Limited) at all material times carried on a qualifying activity, in the form of a trade: see section 15(1)(a) of CAA 2001. It is also agreed that the disputed expenditure is all of a capital nature, that it was incurred wholly or partly for the purposes of Urenco’s trade, and that Urenco owned the plant or machinery as a result of incurring the expenditure.

Secondly, it is now common ground that the exclusion by section 22(1)(a) of expenditure on the provision of a structure or other asset in list B does not apply, because any items which would otherwise constitute a structure within Item 7 of List B would fall within the saving in sub-paragraph (a) of Item 7 for structures which are “industrial buildings” within Chapter 2 of Part 3. Nor could the exclusion in section 22(1)(b) for “any works involving the alteration of land” apply, because it is now established that there is no overlap between section 22(1)(a) and (b), and if expenditure is not excluded by section 22(1)(a) it is unnecessary to consider whether it is excluded by section 22(1)(b): see the decision of this court in Revenue and Customs Commissioners v SSE Generation Ltd [2021] EWCA Civ 105, [2021] STC 369. Accordingly, neither side has founded any arguments before us on the provisions of section 22.

Thirdly, there was no appeal by either side against the conclusions of the FTT on the important question of how the assets comprising the five key facilities should be identified. This issue was discussed by Judge Cannan in the FTT Decision, at [65] to [72]. As he was fully entitled to do, he reached conclusions which did not coincide with either side’s submissions and were in various respects intermediate between the “piecemeal” approach advocated by Urenco and the “single entity” approach preferred by HMRC. The Judge’s conclusions were set out at [72], where he said (omitting some of the detail):

“72.

I accept that the disputed components in each facility are closely physically connected. To a large extent they support each other and work together in providing radiation shielding, containment and seismic qualification, or a combination of those three functions. These structures all have a separate visual identity, especially when they are considered without the cladding. In my view to a large extent each structure comprises a whole. Each structure can be described as a “safety significant structure” in its own right. Looking at each structure in turn:

(1)

The items which comprise the [cylinder handling facility] can readily be seen to form a separate structure and to function as such. Mr Nicholson [one of Urenco’s witnesses] described it as a monolith, by which he meant that the walls and roof act together to provide support and seismic qualification. The structures provide radiation shielding for the environment, and for operators within, and containment. It is seismically qualified as a whole. In my judgment it would be artificial to consider the raft slab, the walls, the roof or other components as having separate identities… The only exceptions to this are the internal radiation shield walls where seismic qualification is merely incidental to the purpose of shielding operators in the CHF; the raised platforms or plinths which perform a specific function of supporting rails at the correct height for transportation purposes; and the stairs and access platforms which are intended only to service the crane…

(2)

The items which comprise the vaporisation facility can readily be seen to form a separate structure and to function as such. That structure includes the first-storey concrete box which Mr Nicholson also described as a monolith. The concrete box has a containment function, a shielding function and is seismically qualified as a whole. It would be artificial to consider the slabs, external walls and internal walls as having separate identities. The only exceptions to this are the stairs and access platform which are for the maintenance and inspection of equipment…”

Broadly similar conclusions were then reached in relation to the items comprising the kiln facility, the condenser facility and the uranium oxide store respectively.

The meaning of “plant” in section 11 of CAA 2001, and predecessor fiscal legislation, has never been the subject of statutory definition. Parliament has been content to leave its meaning to be elucidated and developed by judges in a long series of cases, beginning with the decision of the Court of Appeal in Yarmouth v France (1887) 19 QBD 647, where the issue was whether a vicious horse was a “defect in the condition of…plant” used in the business of a wharfinger for the purposes of the Employers Liability Act 1880. Holding that it was, Lindley LJ said at 658:

“There is no definition of plant in the Act; but, in its ordinary sense, it includes whatever apparatus is used by a business man for carrying on his business, - not his stock-in-trade which he buys or makes for sale; but all goods and chattels, fixed or moveable, live or dead, which he keeps for permanent employment in his business.”

A century later, in Wimpy International Ltd v Warland (Inspector of Taxes) [1988] STC 149 (“Wimpy”), where the issue was whether expenditure on various improvements to premises for the purposes of their use as fast food restaurants qualified for capital allowances under section 41(1)(a) of the Finance Act 1971, Hoffmann J described Lindley LJ as the “celebrated pioneer cartographer” of this territory; and after quoting the above passage from Yarmouth v France, he continued, in an illuminating passage which has frequently been followed in later cases, at 170-171:

“It is important to notice the various discriminations which are stated or implied in this description. First, it excludes anything which is not used for carrying on the business. Secondly, it excludes stock-in-trade both expressly and because, although used for the purposes of the business, its use lacks permanence. Thirdly, it excludes things which are not “apparatus…goods and chattels, fixed or moveable, live or dead” or not employed in the business. This excludes the premises or place in or upon which the business is conducted.

Before going any further I must say something about the third distinction and the way in which the courts in subsequent cases have refined the boundary between plant and premises. The words “apparatus…goods and chattels, fixed or moveable, live or dead” might suggest that the distinction turns upon whether the item is a chattel or fixture on the one hand or a building or structure on the other. This was the view of the minority in the House of Lords in IRC v Barclay, Curle & Co Limited [1969] 1 WLR 675, 45 TC 221. But the majority held that even a building or a structure (in that case a dry dock) could be plant if it was more appropriate to describe it as apparatus for carrying on the business or employed in the business than as the premises or place in or upon which the business was conducted. By this test a swimming pool used in connection with the operation of a caravan park has been held to be plant, in Cooke (Inspector of Taxes) v Beach Station Caravans Ltd [1974] 1 WLR 1398, while conversely in Benson (Inspector of Taxes) v Yard Arm Club Ltd [1979] 1 WLR 347, a ship used as a floating restaurant, although a chattel, was held not to be plant because it was the place in which the business was conducted: see Lord Lowry in IRC v Scottish & Newcastle Breweries Ltd [1982] 1 WLR 322 at 333.

It will be seen, therefore, that although the three distinctions in Yarmouth v France… each involves a test which can be called functional, they are subtly different from each other. If the item is neither stock-in-trade nor the premises upon which the business is conducted, the only question is whether it is used for carrying on the business. I shall call this the “business use” test. However, under the second distinction, an article which passes the “business use” test is excluded if such use is as stock-in-trade. And under the third distinction, an item used in carrying on the business is excluded if such use is as the premises or place upon which the business is conducted. The fact that an item may pass the “business use” test but fail what I may call the “premises” test is central to this case.”

Wimpy went to the Court of Appeal, reported at [1989] STC 273, where this court agreed with Hoffmann J’s conclusion that the disputed items (with one minor exception) qualified as plant. After referring to a number of cases, Fox LJ said at 279:

“In the light of the authorities, the position appears to me to be this. There is a well established distinction, in general terms, between the premises in which the business is carried on and the plant with which the business is carried on. The premises are not plant. In its simplest form that is illustrated by Lord Lowry’s example of the creation of atmosphere in a hotel by beautiful buildings and gardens on the one hand and fine china, glass and other tableware on the other. The latter are plant; the former are not. The former are simply the premises in which the business is conducted.

The distinction, however, needs to be elaborated, for present purposes, by references to Lord Lowry’s further formulation, namely that the fact that different things may perform the same function of creating atmosphere is not relevant: one thing may function as part of the premises and the other as part of the plant. Thus, “something which becomes part of the premises instead of merely embellishing them is not plant except in the rare case where the premises are themselves plant.” The latter part of those observations is a reminder that it is not sufficient to say that something is part of the real property. It can still be plant as the Barclay Curle and Beach Station Caravans cases show. Moreover, the test is not whether the item is a fixture. Central heating apparatus must, I think, be a plant. But there may be cases in which the degree of affixation is a matter to be taken into consideration.”

Fox LJ added, at 280:

“It is proper to consider the function of the item in dispute. But the question is what does it function as? If it functions as part of the premises, it is not plant. The fact that the building in which a business is carried on is, by its construction particularly well-suited to the business, or indeed was specially built for that business, does not make it plant. Its suitability is simply the reason why the business is carried on there. But it remains the place in which the business is carried on and is not something with which the business is carried on.

…

I would agree with Hoffmann J that the question is whether it would be more appropriate to describe the item as part of the premises rather than as having retained a separate identity.”

Another influential case exploring the distinction between plant and premises was Carr (Inspector of Taxes) v Sayer [1992] STC 396, (1992) 65 TC 15, which concerned expenditure incurred on the construction of specialised quarantine kennels for cats and dogs brought into the UK. In his judgment, Sir Donald Nicholls V-C identified the relevant law in a number of propositions, at 402:

“First,… plant carries with it a connotation of equipment or apparatus, either fixed or unfixed. It does not convey a meaning wide enough to include buildings in general. The premises whether an office or a factory or a warehouse or whatever, at which or in which a business is carried on would not normally be understood as intended to be embraced by the expression “machinery or plant”.

…

Second, the expression “machinery or plant” is apt to include equipment of any size. If fixed, a large piece of equipment may readily be described as a structure, but that by itself does not take the equipment outside the range of what would normally be regarded as plant. The equipment does not cease to be plant because it is so substantial that, when fixed, it attracts the label of a structure or, even, a building.

…

Third, and this follows from the above, equipment does not cease to be plant merely because it also discharges an additional function, such as providing the place in which the business is carried out. For example, when a ship is repaired in a dry dock, the dock also provides the place where the repair work is carried out. That is no more than the consequence of the extensive size of a piece of fixed plant.

Fourth, and conversely, buildings, which I have already noted would not normally be regarded as plant, do not cease to be buildings and become plant simply because they are purpose-built for a particular trading activity. Such a distinction would make no sense. Thus, the stables of a racehorse trainer are properly to be regarded as buildings and not plant. A hotel building remains a building even when constructed to a luxury specification… Similarly with a hospital for infectious diseases. This might require special lay-out and other features but this does not convert the buildings into plant. A purpose-built building, as much as one which is not purpose-built, prima facie is no more than the premises on which the business is conducted.

Fifth, one of the functions of a building is to provide shelter and security for people using it and for goods inside it. That is a normal function of a building. A building used for those purposes is being used as a building. Thus a building does not partake of the character of plant simply, for example, because it is used for storage by a trader carrying on a storage business. This remains so even if the building has been built as a specially secure building for use in a safe-deposit business. Or, one might add, as a prison. ”

Applying those principles, Sir Donald Nicholls V-C held that the permanent quarantine kennels which the taxpayers had constructed were not plant, but were “the premises at which and in which the business is conducted”: ibid.

Further instructive examples of cases where purpose-built structures of a specialised nature have been held to fall on the premises side of the line are:

On the other side of the line, there is a relatively small number of cases where a complex structure, viewed as a whole, has been held to function as plant in the taxpayer’s business, and thus to pass the premises test. The classic example of this in the English authorities is the dry dock case, IRC v Barclay, Curle & Co Ltd [1969] 1 WLR 675 (HL), where the House of Lords decided by a bare majority (Lords Hodson and Upjohn dissenting) that the cost of excavating and constructing a dry dock with direct access to the river Clyde was expenditure on plant for the purposes of the predecessor provision in section 279(1) of the Income Tax Act 1952. For the majority, Lord Reid said at 679:

“It seems to me that every part of this dry dock plays an essential part in getting large vessels into a position where work on the outside of the hull can begin, and that it is wrong to regard either the concrete or any other part of the dock as a mere setting or part of the premises in which this operation takes place. The whole dock is, I think, the means by which, or plant with which, the operation is performed.”

Other examples on the plant side of the line include:

In view of the significance attached by Mr Peacock KC to the Wangaratta case, I will set out the core passage in the reasoning of McTiernan J. After giving a detailed description of the dyehouse, and the functions of its component parts, and referring to various Australian and English authorities, he referred at page 10 to the then very recent decision of the House of Lords in the Barclay, Curle case, and to Lord Donovan’s view that the dry dock constituted plant because it was “in the nature of a tool of the taxpayer company’s trade”.

McTiernan J continued:

“I am of opinion that the appellant’s dyehouse is “in the nature of a tool” in the trade and does “play a part” itself in the manufacturing process. It is much more than a convenient setting for the appellant’s operations. It is an essential part in the efficient and economic operation of the appellant’s business. The complex ventilation system including the cavity wall does more than merely clear the atmosphere. Its structure is an active tool in preventing spoiling of material, and in enabling the operatives to carry out their tasks. It would be completely unnecessary in almost every other industry and quite useless to any buyer except a dyer. The protective coatings and tiling are essential in preserving the whole “tool”. It is as unreal to dissect the paint or tile from its foundation as it is to separate the paint from a workman’s tool of trade. The drains do not just remove waste liquids, they remove volatile liquids which would disrupt the process as much as vapours escaping from the vats. If boiling liquids were left uncovered in the building, in vats or drains the whole process would quickly become unworkable. I think therefore that the dyehouse should be regarded as a single unit of plant and not a collection of bricks, mortar, paint, timber etc, each of which is to be separately examined. It is not merely a special factory; it is a complex whole in which every piece is essential for the efficient operation of the whole. I would however except from the description of “plant” what might be referred to as the external “cladding” of the dyehouse, that is the external walls including the single walls at the east and west ends and the roof as distinct from the ceiling, but not the controlled louvres or the cowlings in the roof. The cladding really does nothing more than exclude the elements; and, whilst I am not convinced of the validity of this distinction, nevertheless it is clearly supported by prior decisions on this sort of question. ”

It can be seen from this passage that McTiernan J so concluded because he considered it appropriate to regard the dyehouse as a whole, constituting a single unit of plant in which “every piece is essential for the efficient operation of the whole”. As Peter Gibson LJ later observed, in the Attwood case at 1175:

“McTiernan J… held that a dye-house was a complex whole of which every piece was essential for the efficient operation of the whole and was plant, save for the external walls and roof which served only to provide protection against the elements. Again, as it seems to me, that structure with that significant exception passed the premises test.”.

To bring matters up to date, I must also refer to the valuable recent decision of this court in Cheshire Cavity Storage 1 Ltd v Revenue and Customs Commissioners [2022] EWCA Civ 305, [2022] STC 622 (“Cheshire Cavity”). The leading judgment was delivered by Lewison LJ, with whom Baker and Whipple LJJ agreed. Lewison LJ identified the issue on the appeal, at [1], as:

“whether the taxpayers are entitled to capital allowances in respect of the expenditure incurred on the introduction of water into salt bearing rock so as to dissolve the rock and create an impervious cavity, typically in the shape of a teardrop (“leaching”), and the displacement of the resulting brine by the introduction of gas (“de-brining”) so as to permit the storage of gas in the cavity.”

Lewison LJ went on, at [3], to give important guidance, derived from the authorities, about the role of the fact-finding tribunal (here the FTT) where there is a dispute about the availability of capital allowances:

“In order to qualify for capital allowances, the expenditure must be expenditure on the provision of plant. It is clear from the authorities that a decision whether something is or is not plant is a question of fact or a question of fact and degree. In some cases, it is possible to take either view, and in such a case the decision of the fact-finding tribunal cannot be impugned. Since the question whether something is or is not plant is a question of fact, or a question of fact and degree, it is necessary to pay close attention to the facts of previous cases. In some cases the court has upheld the decision of the fact-finding tribunal on the basis that it was entitled to find as it did. In such a case, it does not follow that the fact-finding tribunal would have made an error of law if it had decided the question differently.”

It follows from this passage, and from the well-established proposition that cases exist where “it is possible to take either view”, that the question of what constitutes plant is not a hard-edged question of law to which there is always only one right answer. Accordingly, it is always necessary to have close regard to the facts found by the FTT, which is a specialist tribunal, and where the question is one of fact and degree, to the FTT’s evaluation of those facts.

Returning to this theme at the end of his judgment, Lewison LJ gave further guidance, under the heading “The test in this court”:

“84.

Whether it is “more appropriate” to describe the item as apparatus or premises is clearly a value judgment. As Jacob LJ said in Procter & Gamble UK v Revenue and Customs Commissioners [2009] EWCA Civ 407, [2009] STC 1990, 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.”

85.

Similarly, in Re Sprintroom Ltd, Prescott v Potamianos, Potamianos v Prescott [2019] EWCA Civ 932, [2019] 2 BCLC 617, [2019] BCC 1031 this court said at [76]:

“So, on a challenge to an evaluative decision of a first instance judge, the appeal court does not carry out a balancing task afresh but must ask whether the decision of the judge was wrong by reason of some identifiable flaw in the judge’s treatment of the question to be decided, “such as a gap in logic, a lack of consistency, or a failure to take account of some material factor, which undermines the cogency of the conclusion”.”

Cheshire Cavity also decided an important question of law, which has some bearing on the present case. Appearing for the taxpayers in that case, Mr Peacock advanced a key submission which Lewison LJ recorded at [24] in these terms:

“When an asset has both plant and premises/setting functions, the asset should be classified as plant unless it is merely premises/setting. It is not a question whether the asset predominantly performs a premises function rather than a plant-like function. Nor it is a question whether an asset is more appropriately described as premises or setting on the one hand, or plant on the other. As soon as it is found that an asset has any plant-like function, it qualifies as plant without more, unless it is specifically excluded by legislation.”

In order to examine this submission, Lewison LJ then embarked upon a full and illuminating review of the main authorities, before concluding at [81] that, even if the cavities could be said to have a “plant-like function”:

“I do not find in the authorities solid support for Mr Peacock’s submission that if something performs any “plant-like function” it is necessarily plant. On the contrary it is a question of fact and degree; that is to say an evaluative exercise. One applicable test in performing that exercise is “the premises test” as formulated by Hoffmann J. ”

It will be recalled that, in Wimpy, Hoffmann J had formulated the premises test, by reference to the dry dock in Barclay, Curle, in terms of whether “it was more appropriate to describe it as apparatus for carrying on the business or employed in the business than as the premises or place in or upon which the business was conducted” (emphasis supplied).

To conclude, in the absence of any material misdirection of law, a decision by the fact-finding tribunal on the appropriate way to describe a disputed structure is an evaluative conclusion which must normally be respected by an appellate court or tribunal. That is particularly so because an appeal lies from the FTT to the UT, and from the UT to this court, only on points of law: see sections 11 and 13 of the Tribunals, Courts and Enforcement Act 2007.

Having identified the relevant assets, the FTT considered the question “whether the assets function as plant” in the FTT Decision at [73] to [99]. In what follows, paragraph numbers refer (unless otherwise stated) to the FTT Decision. The conclusion, summarised in [99], was:

“… that the structures of the kiln facility and the condenser facility operate as plant, together with the various plinths identified. Otherwise the structures and disputed assets are merely part of the setting in which the plant and machinery functions and are not plant.”

The FTT therefore considered that two of the five main structures in dispute operated as plant, together with certain “plinths”. The relevant plinths were (a) the raised platforms in the cylinder handling facility which “perform the specific function of enabling rails to be at the correct height to enable transportation of cylinders to and from the vaporisation facility”, and (b) the concrete plinth supporting the hopper in the kiln facility: see [98](3) and (11) respectively. In addition, although omitted from the summary in [99], the FTT held that the air sealed crane access hatch in the roof of the kiln facility, for the purpose of installing the hydrolysis chamber, was plant: see [98](9).

The FTT reached these conclusions by the following stages.

First, the FTT reviewed the authorities on the meaning of plant, and derived from them the following propositions set out in [89]:

“(1)

Plant can comprise large structural items.

(2)

There is a distinction to be made between a structure which is merely the setting in which a trade is carried on and a structure which constitutes the apparatus with which the trade is carried on.

(3)

The function of plant in a trade can be active or passive. For example, moveable partitioning might be said to perform its function passively but it may still be plant.

(4)

Premises do not fall to be regarded as functioning as plant simply because they have been designed to satisfy the particular requirements of the business in question.

(5)

A structure which is merely the setting in which a business is carried on is not plant.

(6)

If a structure is both the setting and the means by which the business is carried on then it will be plant.

(7)

An item that might otherwise be described as a building is likely to be a place in which the business is carried on and not plant, but not necessarily so.

(8)

It is important to be careful and precise in analysing the function of the item for the purpose of distinguishing premises from plant.”

As I have already said, no issue is taken by either side with this summary, apart from proposition (6) which it is now clear is too rigidly stated in the light of the Cheshire Cavity case. At the date of the FTT Decision, the Cheshire Cavity litigation had still to reach the FTT, and the prevalent view may well have been that a dual-purpose structure which is both the setting and the means by which the business is carried on must be characterised as plant, rather than the correct proposition that it remains a question of fact and degree, and of which is the more appropriate description.

Secondly, the FTT summarised the rival submissions at [92] to [93]. On behalf of Urenco, Mr Peacock submitted that each of the disputed structures and other assets performed a function in the trade and was properly to be treated as plant, even if in the case of the structures they might also be described as the premises in which the trade is carried on. He relied, in particular, on the decision in Wangaratta. For HMRC, Mr Bremner submitted that the structures were merely the setting in which the deconversion process takes place, and they do not perform any function in that process. The FTT then rejected a specific submission by Mr Bremner that the radiation hazards which needed to be controlled at the TMF were generated by the presence of Tails and the uranium oxide, rather than by any processes carried on at the TMF. The FTT considered that submission to be “too simplistic” and accepted the evidence of Mr Nicholson that “processing the Tails gives rise to new and additional risks.”

Thirdly, the FTT identified the trade carried on by Urenco at the TMF, at [95]:

“It may be described as the deconversion of Tails so as to produce and store uranium oxide and to produce hexafluoric acid for sale as an industrial material. All the processes carried out at the TMF are directed towards those ends.”

There is no challenge by either side to this description of Urenco’s trade.

Judge Cannan continued (ibid):

“I consider that the safety functions of shielding, containment and seismic qualification are properly viewed as part of the setting in which the trade is carried out. Shielding and containment are akin to preventing noxious fumes or odours escaping from a processing plant. In Wangaratta the dyehouse and the apparatus within it were treated as a complex whole in which every element including the structure was essential for the efficient operation of the whole. The structure did not just provide the setting but was part of the dyeing process, removing volatile gases and liquids which would otherwise adversely affect the dyeing process. That is not the case here. The safety significant structures provide a safe setting for the processes to be carried out. Without the structures the actual processes could be carried on efficiently, although I accept that is entirely theoretical because the regulatory environment would not permit it to happen. But the regulatory environment is not in my view relevant to whether an asset performs a function in the trade. It cannot be said that in providing shielding and/or containment the structures have any function in the actual processing of Tails which is carried out by the plant and machinery in the TMF.

96.

As far as seismic qualification is concerned, in a sense it is incidental to ensuring the integrity of the shielding and containment functions of each facility. It represents the standard and method of construction required to maintain shielding and containment in an extreme seismic event.”

Fourthly, the FTT rejected a submission by Mr Peacock that all the features of the structures were “an essential and necessary part of the trade processes”, performing a “trade function and not simply a premises function”. The Judge said, at [97]:

“I do not accept that argument. In my view the expenditure cannot be regarded as part of the cost of installation of the plant and machinery within the structures merely because that plant and machinery could not safely be used without it.”

Judge Cannan then stated his view, at the beginning of [98], that “there must be some other function performed by the structures in the trade if they are to be treated as plant”. He considered each disputed structure and asset in turn, in the 15 numbered sub-paragraphs of [98], giving his reasons for coming to the conclusions which I have already summarised. For example, in relation to the cylinder handling facility structure, he said:

“(1)

The functions of the CHF structure are shielding and containment of radioactivity. It has the appearance of a building, with four walls and a roof enclosing a substantial volume of space. The walls and roof are not required or intended to provide shelter for material, equipment, machinery or operators. The roof cladding is intended to provide shelter, not for the items inside the CHF but for the concrete roof that would otherwise be damaged by the effects of standing rainwater. In my view the shielding and containment functions are functions of premises and not functions in processing the Tails. The structure is purpose built to house the plant and machinery required to carry out the processing of the Tails. Such a structure would not sensibly be used in any other context but fundamentally it simply provides a safe and secure setting in which the Tails are processed. It is not part of a complex whole in the same way that the dyehouse was in Wangaratta.”

By contrast, he came to the opposite conclusion in relation to the structure of the kiln facility:

“(7)

The functions of the kiln facility structure are containment, support for the kiln, the hopper and associated equipment, and enabling the use of gravity to receive uranium oxide in the basement and thereafter to use a hopper for packing the uranium oxide. It provides shelter to the equipment and operators within, which is a function of premises, although this is incidental to the functions described above. It seems to me that this structure is not just specifically designed as the setting for plant and equipment within it, but also to hold items of equipment at specific levels to take advantage of gravity in the process of deconversion and the packing of uranium oxide. As such, I consider that it does fulfil a function in a similar way to the grain silos in Schofield and the dyehouse in Wangaratta. I am satisfied that it falls within the common law meaning of plant.”

Having dealt with the structure of the kiln facility, the Judge then considered the separate assets within it. He concluded that the stairs and access platforms, together with certain access hatches in the mezzanine and first floor to enable kiln filters to be changed, did not qualify as plant, but that the crane access hatch in the roof for the purpose of installing the hydrolysis chamber, together with a concrete plinth supporting the hopper, did: see sub-paragraphs (8) to (11).

By way of a final example, the FTT concluded that the uranium oxide store functioned as premises, commenting in sub-paragraph (14):

“The UOS is used for storage of uranium oxide and by analogy with what was said in Carr v Sayer, a structure does not take on the character of plant simply because it is used for storage by a trader carrying on a storage business, even where that storage business is highly specialised.”

The alleged errors of law in the decision of the FTT

Urenco appealed to the UT against the FTT Decision on six grounds, of which the first relates to this part of the case. It was formulated in these terms:

“(1)

The [FTT] Decision misclassifies the safety functions of assets used in shielding, containment and seismic qualification as being merely “part of the setting”, as opposed to being assets used in the business and having plant-like functions.”

See the UT Decision at [29].

The UT considered this ground of appeal at [45] to [87] of the UT Decision. It began with an extensive review of the principles to be derived from the case law, with the benefit of the then recent decision of the UT in the Cheshire Cavity case. The UT then (a) set out the FTT’s main findings of fact on the regulatory regime governing the construction and operation of the TMF, (b) explained the FTT’s classification of the disputed items, (c) provided a helpful summary of the FTT’s conclusions on the plant issue at [70], and (d) summarised the rival submissions on the appeal, before coming to a section headed “Discussion” at [75].

The UT considered that the “essential thought processes and reasoning” of the FTT in relation to the plant issue were to be found within [95] to [97] of the FTT Decision, most of which I have quoted at [53] to [55] above. The UT then noted that the FTT had analysed the case law correctly, subject to the qualification required by the Cheshire Cavity case. The question for the UT was therefore “whether the FTT misapplied or misunderstood those principles in a relevant respect in its reasoning and conclusions at [95] – [97].” Reference was made to the speech of Lord Lowry in Inland Revenue Commissioners v Scottish & Newcastle Breweries Ltd [1982] 1 WLR 322 (HL), where at 327 he recognised that a decision of the tribunal of fact in a case of the present type “is a decision on a question of fact and degree and cannot be upset as being erroneous in point of law unless [they] show by some reason they give for a statement they make … that they have misunderstood or misapplied the law in some relevant particular.”

Having thus directed themselves, the UT then identified what they considered to be three errors of law in the FTT’s treatment of the question.

The first alleged error was that the FTT had failed to apply the relevant test to Urenco’s business “as it is actually carried on”. The UT stated, at [80]:

“we consider that the FTT misunderstood or misapplied the functionality test at [95] when it concluded in setting out its reasoning that (1) without the safety significant structures the actual processes carried on at the TMF could still be carried on efficiently, although that was entirely theoretical because the regulatory environment would not permit it to happen, and (2) the regulatory environment was not relevant to whether an asset performed a function in the trade.”

In response to Mr Bremner’s argument that it would be wrong to look at any part of [95] of the FTT Decision in isolation, the UT continued:

“82.

We agree that the FTT recognised the applicable regulatory regime. We accept that the weight to be given to that factor in making the functionality assessment was a matter for the FTT. We also think that the FTT were right to observe that performance of a safety function cannot convert an item which is merely premises into plant. However, we think that the error into which the FTT fell was essentially to confuse the relevance of the regulatory/safety aspects in assessing the functionality of the Disputed Items with their relevance to the nature of the business in which functionality was to be assessed. That was what led it to state that without the safety structures the relevant processes “could still be carried out efficiently”, when in fact the effect of the regulatory constraints was that those processes – and therefore the relevant business which Urenco in fact carried on – could not permissibly be carried out at all.”

…

83.

The same confusion lies behind the FTT’s assertion that “the regulatory environment is not… relevant to whether an asset performs a function in the trade”. We observe that here the FTT is going much further than Mr Bremner’s assertion (with which we agree) that a safety function cannot convert premises into plant; it is saying that the safety function is simply not relevant, and so presumably should be given no weight. In any event, the critical point is that if the trade is itself shaped and determined by the regulatory environment, then that environment must necessarily be relevant to assessing functionality “in the trade”.”

The second error identified by the UT was, in summary, that the FTT had focused unduly on the “actual processing” carried out at the TMF. The following extracts from [84] and [85] of the UT Decision should be sufficient to show the thrust of the UT’s reasoning:

“84.

We also accept the force of Mr Peacock’s argument that in its essential reasoning the FTT appears to concentrate unduly in determining functionality on the “actual processing” carried out at the TMF.

…

85.

It is not clear whether the FTT’s focus on processing and “actual processing” amounts to, or stems from, a failure to apply the principle established by case law… that the function of plant may be active or passive. If so, that would be an error of law. We think it is more likely that it results from the drawing of a false dichotomy between items used in activities which might be said to be “actively” involved in processing Tails at the TMF and those which enable the carrying on of the entire range of activities in compliance with regulatory requirements.

…

Reading the FTT’s critical reasoning in the context of the decision as a whole, we consider that the FTT did adopt an approach to functionality which assumed that an item falling within the latter category was indeed as a result less plant-like and more like premises. Particularly given the nature of Urenco’s business activities at the TMF, we consider that this was an error, as it misunderstood or misapplied the relevant principles as applicable to those activities as actually carried on.”

The third alleged error identified by the UT is set out in [86]:

“86.

The approach by the FTT which we have described also informed its conclusion that, contrary to Mr Peacock’s submissions, Wangaratta actually supported its decision that most of the items were not plant. If the FTT had adopted the correct approach, a much closer comparison to the “complex whole” in that case would in our view have been called for. Even if Mr Bremner is correct that in principle containment and safety are typical functions of premises, these were not typical premises and this was not a typical trade.”

The UT considered that the errors which it had detected were material, in the sense that the FTT’s decision on this issue, in relation to those items which it had found not to constitute expenditure on the provision of plant, might have been different but for the errors: see the UT Decision at [171-2]. Accordingly, the UT decided to set aside the FTT’s decision on those items and to remit the case to the FTT to remake the relevant decisions: see [175] to [177].

The issue is now the subject of HMRC’s second ground of appeal to this court.

HMRC’s basic submission in support of this ground of appeal is that the FTT was fully entitled to reach the conclusions it did on which of the disputed assets constituted “plant” for the purposes of section 11(4)(a) of CAA 2001, and that in doing so the FTT committed no error of law. Accordingly, it was not open to the UT to intervene. HMRC rely on the statements of principle in Cheshire Cavity, which I have already reviewed. Since the FTT had in substance identified the correct legal principles, and in the absence of any challenge to the FTT’s specific findings of fact on Edwards v Bairstow [1957] AC 14 principles, the UT should have respected the FTT’s conclusions.

HMRC go on to submit that none of the alleged errors of law identified by the UT is made out. With regard to the FTT Decision at [80], HMRC say the UT has overlooked the express finding by the FTT that “the safety functions of shielding, containment and seismic qualification are properly viewed as part of the setting” in which Urenco’s trade is carried out: see the FTT Decision at [95]. This was a conclusion the FTT was fully entitled to reach, leading to its conclusion that three of the facilities were merely functioning as purpose-built buildings. Furthermore, the FTT had the regulatory environment well in mind, as shown by its detailed findings of fact, and the recognition in [95] that the safety significant structures “provide a safe setting for the processes to be carried out.” Read in context, the FTT was clearly seeking to test whether the relevant structures were appropriately to be regarded as forming part of the relevant process, as opposed to forming part of the setting in which that process was carried out.

In performing this exercise, say HMRC, it is unrealistic to suppose that the FTT lost sight of the nature of the actual business that Urenco was carrying on, or of the impact of the regulatory requirements on the design and structure of the facilities. The FTT was also well aware that the function of plant in a trade can be passive as well as active: see, for example, the third of the propositions derived by the FTT from the authorities at [89](3). There is no rule of law that an item which has a passive function in a trade must be characterised as plant. The relevant question is always whether the item is functioning as plant or as premises. Furthermore, the FTT was entitled to distinguish Wangaratta, on the basis that “the dyehouse and the apparatus within it were treated as a complex whole in which every element including the structure was essential for the efficient operation of the whole”. Contrary to the criticisms of the UT, it was not incumbent on the FTT to perform “a much closer comparison” to the complex whole in Wangaratta.

In his oral submissions to us, Mr Bremner argued that Wangaratta should in any event be treated with some caution. Apart from the fact that it was a first-instance decision in Australia, it was decided over half a century ago (in 1969), before the more sophisticated development of the premises test in the English authorities beginning with Wimpy. Mr Bremner therefore suggested that it should be treated as a decision on its own facts, and that if anything in it conflicted with the later English authorities, we should not follow it.

On behalf of Urenco, Mr Peacock submits that the reasoning of the FTT in [95] of the FTT Decision is seriously flawed. The argument is neatly encapsulated in Urenco’s skeleton argument, with reference to the final sentences of [95]:

“In this way, the FTT disregarded the regulatory framework of the nuclear industry and confined “Plant” to those assets which would be involved, in an entirely theoretical world, in the actual processing of Tails to the exclusion of assets that provided for the only safe way of processing Tails in the real world. This is wholly unrealistic; the determination of whether an item has a “plant-like” function must be determined in the real world, reflecting actual use in what was the only permitted operating environment.”

Urenco further submits that, by confining relevant functions to those involved in the “actual processing” of Tails, the FTT erred in law in two respects. First, there is no requirement in the notion of “plant” that the asset be directly used in the trade process concerned. It is enough that the asset has an apparatus function because, in one way or another, it constitutes the means by which the trade is carried on. For example, an office desk and chair are plant even if only used in the back office of a manufacturing business. This is recognised in the legal principle that plant need not have an active function, but may be passive. Secondly, Urenco’s trading activities at the TMF are not limited to the “actual processing” of Tails. The whole purpose of the TMF is to meet Urenco’s safety obligations by de-converting Tails and processing and handling the relevant by-products.

As to Wangaratta, Mr Peacock submits that support for Urenco’s case can be found in McTiernan J’s identification of the role of the ventilation system in removing poisonous gasses and thus providing a safe working environment. This was one type of plant-like function which the judgment in that case rightly relied upon. More generally, Mr Peacock reminded us that Wangaratta has been cited with approval in a number of leading English authorities, including the speech of Lord Lowry in Schofield and the judgment of Peter Gibson LJ in Anduff.

My starting point in considering this ground of appeal is to emphasise the reluctance with which an appellate court or tribunal should interfere with an evaluative conclusion reached by the tribunal of fact on the question whether a disputed asset is plant. Although the meaning of “plant” in what is now section 11(4)(a) of CAA 2001 is ultimately a question of law, the application of the legal test to the facts of a particular case is a question of fact, or sometimes a question of fact and degree, and there are cases in which it is possible to take either view without erring in law: see Cheshire Cavity at [3]. In particular, evaluation of the “premises test” formulated by Hoffmann J in Wimpy is a question of fact and degree, as illustrated by cases such as Gray, Anduff and Bradley: ibid at [81] to [83].

In a case such as the present where the tribunal of fact has in substance correctly identified the governing principles of law, the reluctance of an appellate court to interfere should, if anything, be even stronger. Nevertheless, to characterise a question as one of fact and degree does not automatically confer immunity from legal challenge. As this court rightly recognised in Cheshire Cavity at [85], referring to the Sprintroom case [2019] EWCA Civ 932, at [76], the decision of the first instance judge may be “wrong by reason of some identifiable flaw in the judge’s treatment of the question to be decided”. The examples there given are “a gap in logic, a lack of consistency, or a failure to take account of some material factor, which undermines the cogency of the conclusion”.

With this guidance in mind, I turn at once to what seems to me to be Urenco’s strongest ground of challenge to the FTT’s treatment of this question, namely the passages in the FTT Decision at [95] which are said to show that the FTT fell into the twin errors of

I would accept that these passages could have been better phrased, and that the point which the FTT was seeking to make in them is not immediately apparent. Read in context, however, I do not accept that they evince any error of law, let alone a material one. The FTT was plainly well aware of the actual trade carried on by Urenco at the TMF, because it accurately identified that trade at the beginning of [95] itself. The FTT was also fully aware of the imperative need for the design and structure of the five facilities at the TMF to satisfy stringent environmental and safety criteria. The means by which these criteria were satisfied are reflected in the FTT’s functional analysis of the component parts of each of the five facilities in [98]. However, I agree with HMRC that the crucial finding of fact in [95] is “that the safety functions of shielding, containment and seismic qualification are properly viewed as part of the setting in which that trade is carried out”.

That was in my judgment an evaluative conclusion which the FTT was entitled to reach. In its view, as elaborated in [98], the shielding and containment functions of the three facilities which the FTT found not to constitute plant were functions performed as part of the setting for the trade, albeit a setting of a highly specialised nature. Since each of those three structures shared the premises-like characteristics of having foundations, four walls and a roof, and since the FTT had already decided that the structure of each facility should in principle be regarded as a separate asset, the FTT was entitled to conclude that each of the three facilities constituted premises within which the trade was carried on. As the FTT said, in [98](1), of the structure of the cylinder handing facility:

“Such a structure would not sensibly be used in any other context but fundamentally it simply provides a safe and secure setting in which the Tails are processed. It is not part of a complex whole in the same way that the dyehouse was in Wangaratta”.

Similarly, the vaporisation facility structure was “essentially a concrete box which provides a setting for the autoclaves and associated equipment”, while the shielding and containment functions of the uranium oxide store, together with its steel roof cladding to enable humidity control, were “functions of premises rather than functions in the processing of Tails”: see [98](5) and (14).

That is the broader context within which the potentially problematic sentences in [95] must be placed. On a fair reading, I consider that the FTT was doing no more in those sentences than:

Having disposed of the alleged errors of law in [95] of the FTT Decision, I consider, with respect to the UT, that there is no real substance in the other two main errors which they diagnosed. The criticism that the FTT concentrated “unduly” on the “actual processing” carried out at the TMF seems to me no more than a disagreement with how the FTT chose to approach and evaluate the totality of the evidence. Similarly, I am unable to detect any error of law in what the FTT said about Wangaratta in [95] and [98](1) and (7). Moreover, it seems clear to me that the FTT was seeking to distinguish Wangaratta in [95], and not (as the UT say in [86]) to find support in it for the conclusion that most of the disputed items were not plant. In the light of the FTT’s full description of Wangaratta at [77], I have no doubt that the FTT understood the essential reasoning in that case, and also appreciated that the result turned on the application of what was later to be called the premises test to unusual facts. Contrary to the view of the UT, I think that the FTT’s treatment of Wangaratta was more than adequate, and that it would have been a fruitless exercise to undertake a detailed comparison between the unusual facts of that case and the even more unusual facts of the present case. After all, the purpose of reviewing the authorities is to extract the principles of law which explain how they were decided, not to perform detailed factual comparisons.

I would, however, reject Mr. Bremner’s submission that Wangaratta should be marginalised as a decision confined to its own facts, or even not followed. It seems to me to deserve its place in the standard litany of the case law as an example of how a complex structure, viewed as a whole, may sometimes be appropriately regarded as having the function of plant in the taxpayer’s trade, in much the same way as the dry dock in Barclay, Curle. Nor should we forget that such a conclusion was reached by the FTT in our case in relation to two of the disputed structures: the kiln facility and the condenser facility. HMRC do not challenge either of those conclusions.

As a cross-check on this part of the case, I find it helpful to try to imagine how the FTT, on a remitter, might reasonably be expected to approach and perform its task in a significantly different way, with the benefit of the guidance and criticisms contained in the UT Decision at [78] to [87]. I confess that I find it hard to imagine the FTT coming to a different conclusion, which is perhaps another way of saying that, to my mind, the FTT’s conclusions are essentially of an evaluative nature, where an appellate court or tribunal should generally resist the temptation to interfere.

Accordingly, I would allow HMRC’s appeal on Ground 2.

It is convenient at this point to consider HMRC’s third ground of appeal, which is confined to expenditure on one part of the vaporisation facility, namely the walls and first-floor raft slab. The UT detected an error of law in the treatment of these items by the FTT, which had concluded that they had a “premises type function” and formed part of the setting in which the Tails were processed. The UT therefore set aside the FTT’s conclusion and included these items in the remitter to the FTT for reconsideration. HMRC now appeal to this court, arguing that the FTT was entitled to conclude that the items constituted part of the premises. The focus of this ground is on the question whether the FTT was obliged to conclude that the expenditure on the walls and floor slab was expenditure on the “provision” of plant, because that part of the structure provided the necessary support for certain items of pipework which are admittedly plant.

It is important to appreciate that the walls and slab are integral components of the reinforced “concrete box” which encloses the ground floor of the vaporisation facility. The relevant pipework is attached to steel supports which are fastened to the upper part of the side walls and the underside of the first-floor slab, described in the relevant diagram as “steel supports for plant and pipework”. The floor slab also provides support for items of equipment at first-floor level which are agreed to be items of plant.

The FTT’s findings about the vaporisation facility are contained in [98](5):

“The function[s] of the vaporisation facility structure, excluding the upper steel storey and the ground floor lean-to, are shielding and containment. It is essentially a concrete box which provides a setting for the autoclaves and associated equipment. It also provides shelter for the autoclaves which were not designed for outside use, although they could have been. Looking at the structure as a whole, its functions are essentially the functions of premises rather than functions in the processing of Tails. The walls and first-floor raft slab also provide support for pipework necessary for the processing of Tails. In my view that is also essentially a premises type function. As with the [cylinder handling facility], I regard the vaporisation facility as part of the setting in which the Tails are processed. I do not accept that it can be regarded as expenditure to make the plant in the facility usable”.

The UT considered this issue as part of a ground of appeal which contended that the FTT’s decision unduly restricted expenditure which qualified as being “on the provision of” plant and machinery. At [89], the UT recorded Mr. Peacock’s argument that “expenditure on the delivery and installation of plant or which is necessary to make plant usable qualifies for allowances as being expenditure “on the provision of” plant and machinery.” The UT then pointed out that there is no statutory definition of the phrase “on the provision of” in the context of what is now section 11(4)(a) of CAA 2001, and referred to a “helpful review of the leading authorities” provided by the Upper Tribunal (Adam Johnson J and Judge Richards) in Inmarsat Global Ltd v Revenue and Customs Commissioners [2021] UKUT 59 (TCC), [2021] STC 713, at [66] to [74]. As that review makes clear, the leading authorities on this question are two decisions of the House of Lords, Barclay, Curle and Ben-Odeco Ltd v Powlson (Inspector of Taxes) [1978] 1 WLR 1093 (“Ben-Odeco”).

The relevant issue in Barclay, Curle was whether the costs of excavation of the land where the dry dock was constructed was expenditure on the “provision” of plant. This question was addressed by each of the majority of three who considered that the dry dock itself constituted plant. Lord Reid said, at 680:

“So, the question is whether, if the dock is plant, the cost of making room for it is expenditure on the provision of the plant for the purposes of the trade of the dock owner. In my view, this can include more than the cost of the plant itself because plant cannot be said to have been provided for the purposes of the trade until it is installed: until then it is of no use for the purposes of the trade. This plant, the dock, could not even be made until the necessary excavating had been done. All the Commissioners say in refusing this part of the claim is that this expenditure was too remote from the provision of the dry dock. There, I think, they misdirected themselves. If the cost of the provision of plant can include more than the cost of the plant itself, I do not see how expenditure, which must be incurred before the plant can be provided, can be too remote.”

To similar effect, Lord Guest said, at 686:

“The excavation was a necessary preliminary to the construction of the dry dock and, in my view, was covered by the provision of plant under Section 279. “Provision” must cover something more than the actual supply. In this case it includes the excavation of the hole in which the concrete is laid”.

See too the speech of Lord Donovan at 691.

The issue in Ben-Odeco was whether financing costs undertaken by the taxpayer to fund the acquisition of an oil rig was expenditure on the “provision” of the oil rig, which was admittedly an item of plant. The House of Lords answered this question in the negative. The leading speech was delivered by Lord Wilberforce, who pointed out at 1115 that if the argument for the taxpayer were accepted, a trader who borrowed money would obtain a larger allowance then one who paid for the provision of plant out of its own resources, which “may amount to treating an investor worse than a speculator”. Lord Wilberforce then said:

“The words “expenditure on the provision of” do not appear to me to be designed for this purpose. They focus attention on the plant and the expenditure on the plant, not limiting it necessarily to the bare purchase price, but including such items as transport and installation, and in any event not extending to expenditure more remote in purpose. In the end the issue remains whether it is correct to say that the interest and commitment fees were expenditure on the provision of money to be used on the provision of plant, but not expenditure on the provision of plant and so not within the subsection”.

I note, for completeness, that a further review of these authorities has recently been carried out by this court on an appeal from the UT in the Inmarsat case: see [2022] EWCA Civ 1076, STC 1246, at [61] to [63] (Newey LJ).

In the light of these principles, the UT considered that the FTT must have misdirected itself in law in its treatment of the walls and slab. The UT reasoned as follows, at [103]:

“The FTT held that the Vaporisation Facility itself was not plant, on the basis that it was part of the setting. It is not made explicit whether the FTT’s decision in relation to the supporting walls and slab was reached on the basis that they were not plant, or on the basis that the expenditure was not on the provision of plant. However, the rejection of Mr. Peacock’s argument that the expenditure qualified as it was to make plant usable strongly indicates that it was reached on the latter basis. On that basis, we consider that the FTT’s reasoning indicates that it misdirected itself as to the law, because the fact that the walls and slab themselves performed a premises type function was not material to whether expenditure on those items was on the provision of plant. If the expenditure fell within the principles we describe above, then the fact that the expenditure happened to result in physical items which performed a premisses function would not render it ineligible. Of course, because the FTT did not consider that the Vaporisation Facility was itself plant, it would follow that no expenditure on its provision would itself qualify (unless it could be shown to be on the provision of some other item of plant). Nevertheless, we consider that the FTT made an error of law…. in reaching its decision in relation to the walls and slab”.

In response to this argument, HMRC submit that, ultimately, there is only one statutory question for the purposes of section 11(4)(a): was the expenditure “on the provision of” plant or machinery? That was the question addressed by the FTT, leading to Judge Cannan’s clear conclusion in [98](5) that:

“I regard the vaporisation facility as part of the setting in which the Tails are processed. I do not accept that it can be regarded as expenditure to make the plant in the facility usable”.

The final sentence indicates that the FTT was well aware of the line of authority upon which Urenco relies. Indeed, at [90] the FTT had said it was common ground that expenditure “on the provision of” plant includes installation costs, referring to Ben-Odeco and quoting the passage in Lord Wilberforce’s speech which says that items such as transport and installation are included, but not “expenditure more remote in purpose”. In short, HMRC submit that the FTT’s conclusion was one of fact, or alternatively a value judgment, which (in either case) it was fully entitled, and indeed correct, to reach.

I agree with HMRC. The point is ultimately a short one, but I would find it paradoxical if the walls and floor slab, which so clearly form part of the setting within which the items of equipment in the vaporisation facility operate, were to qualify as plant merely because steel supports for some of those items are fastened to them. To my mind, the natural conclusion to draw is that this feature of the vaporisation facility reflects its role as a specialised setting for the operations carried out within it, in much the same way as the specialised structure of the electrical substation in Bradley. As Blackburne J said in that case, at 1084:

“The fact that features of the structure were carefully designed to accommodate the equipment within does not convert what is otherwise plainly the premises in which the activity is conducted into the plant or apparatus with which that activity is conducted”.

The facts of the present case are in my opinion very far removed from the examples given in the authorities of parasitical expenditure which qualifies for allowances because it is incurred in installing an item of plant or otherwise enabling it to function.

For these short reasons, I would also allow HMRC’s third ground of appeal.

As I have already explained, the question whether an item of disputed expenditure is disqualified by section 21 as being “on the provision of a building” arises only if the item has satisfied the test of being expenditure on the provision of plant or machinery within section 11(4)(a). The section 21 “building” test is therefore likely to be of particular importance as a further hurdle which must be overcome in cases where the disputed item has at least some premises-like characteristics, but nevertheless functions as plant in the taxpayer’s trade, not having been disqualified by application of the “premises test”.

It is common ground that the legislative purpose of the amendments to the capital allowances legislation enacted by FA 1994, and restated in Chapter 3 of Part 2 of CAA 2001, was to promote certainty for both taxpayers and the Revenue, and to prevent further erosion of the boundary between buildings and structures on the one hand, and plant on the other hand, which had been brought about by a series of court decisions, while at the same time entrenching the effect of those decisions and of established Revenue practice: see the comments of Rose LJ (as she then was) in the SSE case [2021] EWCA Civ 105, [2021] STC 369, at [19], and the decision of the UT (Judges Herrington and Brannan) in that case, [2019] UKUT 332 (TCC), [2020] STC 107, at [66].

The UT in SSE referred to Ministerial statements made by Mr Stephen Dorrell MP, as Financial Secretary to the Treasury, when moving Government amendments to clause 110 of the Finance Bill 1994 in Standing Committee A of the House of Commons on 10 March 1994 (Hansard, 10 March 1994, column 602):

“Clause 110 introduces a schedule containing new rules which provide that buildings, structures or land, with certain exceptions, cannot qualify for capital allowances as plant or machinery. These new rules are not intended to change the treatment of assets that qualify as plant at present, as a result of court rulings. The intention behind the legislation is to clarify and strengthen the boundary between buildings and structures on the one hand, and plant on the other. The boundary has been eroded over the years by a number of court cases which have reclassified certain expenditure on buildings and structures as being expenditure on plant. That has affected Exchequer receipts and has created uncertainty about where the boundary lies”.

It is a striking feature of section 21 that, although there is now a blanket exclusion of buildings from the scope of “plant or machinery”, Parliament has chosen not to provide a full definition of “building” in this context. Some assistance, however, may be gained from the partial definition in section 21(3), which tells us that certain assets are included in the term, or (if contained in List A) are to be “treated as buildings”. Thus, the term includes assets which are fixtures (“incorporated in the building”, subsection (3)(a)), or which are of a similar nature but not fixtures (subsection (3)(b)), or which are “in, or connected with, the building” and in List A (subsection (3)(b)). The List A assets comprise standard physical features of many ordinary buildings (“walls, floors, doors, gates, shutters, windows and stairs”); utility services for water, electricity and gas; systems for waste disposal, sewerage, drainage, and fire safety; and “shafts or other structures in which lifts, hoists, escalators and moving walkways are installed”.

It is therefore tolerably clear from the wording of section 21 itself that the meaning of “building” in this context requires a focus on the physical features of the relevant structure or premises, as well as the services and systems which enable it to function in the taxpayer’s trade or business. This in turn suggests that a consideration of both structure and function may be required in deciding what kind of structure or premises answer to the description of “building”. Furthermore, the fact that Parliament has chosen to use an ordinary word in everyday use, which is not a legal term or art, indicates that the test was intended to be simple to operate and relatively unsophisticated.

Although “building” is not a term of art, I consider that Parliament must be taken to have been aware of, and endorsed, the guidance on the meaning of the word which can be derived from the judgment of Sir Donald Nicholls V-C in Carr v Sayer in 1992, not long before the enactment of the 1994 amendments. It will be recalled that one of the principles identified by the Vice-Chancellor in the predecessor legislation then in force was that plant “does not convey a meaning wide enough to include buildings in general”. He also observed (65 TC 15, at 23) that:

“one of the functions of a building is to provide shelter and security for people using it and for goods inside it. That is a normal function of a building. A building used for those purposes is being used as a building. Thus, a building does not partake of the character of plant simply, for example, because it is used for storage by a trader carrying on a storage business. This remains so even if the building has been built as a specially secure building for use in a safe-deposit business. Or, one might add, as a prison”.

So, the provision of shelter and security are typical features of a building, and as Sir Donald Nicholls also observed (ibid):

“A purpose-built building, as much as one which is not purpose-built, prima facie is no more than the premises on which the business is conducted”.

It is true that in the cases where the “premises test” has been applied, a contrast is not normally drawn between “buildings” and “structures”. Rather, the contrast is between “plant” on the one hand, and buildings or structures on the other. It is, therefore, a new and relevant feature of the 1994 amendments that there is a separate blanket exclusion for expenditure on the provision of a structure in what is now section 22 of CAA 2001, subject to various exceptions. For the purposes of section 22, “structure” is defined as meaning “a fixed structure of any kind, other than a building (as defined by section 21(3))”: see section 22(3)(a). However, this point is of little assistance in determining what constitutes a “building” for the purposes of Section 21. The wording of section 22 simply makes it clear that there are fixed structures which are not buildings for the purposes of the Chapter, and that while every building is likely to be a fixed structure, the converse is not always true. Examples of fixed structures which are not buildings might include, for example, pylons, wind turbines, mobile telephone masts, or bus shelters.

It is common ground that there is no previous authority on the meaning of “building” in section 21. That is why I have concentrated on the statutory language of the section and the remainder of the Chapter of which it forms part, together with such guidance as may be gleaned from the case law on the “premises test” in relation to plant. We were also referred to a few cases which have discussed the meaning of “building” in other statutory contexts. For the most part, they seem to me to provide little useful guidance, because the question is always so context specific. I would, however, make an exception for the judgment of Lord Neuberger of Abbotsbury MR (with whom Moore-Bick and Etherton LJJ agreed) in R (Ghai) v Newcastle City Council [2010] EWCA Civ 59, [2011] QB 591. The claimant in that case was an orthodox Hindu, who asked the local authority to dedicate land for traditional open air funeral pyres. The local authority refused, on the basis of legislation relating to cremation contained in the Cremation Act 1902 and associated regulations. The relevant issue for present purposes was whether open air funeral pyres fell within the definition of a crematorium in section 2 of the 1902 Act, as a “building fitted with appliances for the purpose of burning human remains”.

In dealing with this issue, Lord Neuberger MR reasoned as follows:

“21.

On behalf of the Secretary of State, Mr Swift contended that a structure could only be a “building” within the Act if it was “an inclosure of brick or stonework, covered in by a roof”. This contention was supported by three arguments, namely (i) the view of Lord Esher MR in Moir v Williams [1892] 1 QB 264, 270 that this was “what is ordinarily called a building”, (ii) the desirability of having a clear and simple meaning for the word, as breach of the Act would be a criminal offence, and (iii) the need to ensure that cremations could not be seen by the general public. I turn to consider those three arguments in turn.

22.

The first argument is based on the normal meaning of the word “building”. The meaning of the word “building”, or, to put the point another way, determining whether a particular structure is a “building”, must depend on the context in which the word is used. Interpreting a word in a statute or a contract, or indeed in any other document, can, of course, only be sensibly done by considering the context in which it is being used. However, where, as is the case here, the word is one which is used in ordinary language and has no established special legal or technical meaning and is not defined in the document in question (in this case, the Act), one can usefully take as a starting point the word’s ordinary meaning.

…

24.

Particularly as it appears that Lord Esher’s statement as to the “ordinary” meaning of the word “building” may be treated as some sort of authoritative guidance as to the normal meaning of the word, I take this opportunity to say that it would be wrong to see it as having any such effect. In my opinion, the word “building” in normal parlance is naturally used to describe a significantly wider range of structures than would be included within Lord Esher’s “inclosure of brick or stonework, covered in by a roof”.

25.

There are many wooden or other structures not made of “brick or stonework”, such as chalets, stables, or industrial sheds, and there are many structures which are not “inclosures”, such as wood-drying stores, bandstands, or Dutch barns, all of which, on the basis of the normal use of the word, are “buildings”. Other structures come easily to mind, such as the pyramids or the colosseum, which are buildings in normal parlance, but do not fall within Lord Esher MR’s “ordinary” meaning. So, too, at least some prefabricated structures, particularly if attached to a concrete, or similar base, are naturally described as buildings.

26.

Deciding what a word means in a particular context can often be an iterative process, and the ultimate decision should not be affected by whether one starts with a prima facie assumption as to the meaning of the word and then looks at the context, or one starts by looking at the context and then turns to the word. However, if one approaches the issue by making a preliminary assumption as to the meaning of a word such as “building”, then, in agreement with what Etherton LJ said in argument, I do not think that it would be right to take a somewhat artificially narrow meaning of the word, and then see whether the context justifies a more expansive meaning. It is more appropriate to take its more natural, wider, meaning and then consider whether, and if so to what extent, that meaning is cut down by the context in which the word is used”.

It is also relevant to note what Lord Neuberger MR went on to say, at [33]:

“At least in general, it appears to me that, both in principle and in practice, it is inappropriate for the court to seek to define a word or expression used in a statute, where the legislature has not done so. It would virtually be a judicial encroachment onto the legislative function. Judicial guidance on such an issue, through the court’s reasoning in a case where the meaning of a word is in issue, is inevitable, and, it is to be hoped, helpful. But a conscious and unnecessary definition of the word by the court is another matter”.

The FTT considered the meaning of “building” in section 21 at [103] to [110]. After pointing out, correctly, that there is no statutory definition of the term, and no binding authority on its meaning in this context, Judge Cannan then accepted Mr Peacock’s submissions for Urenco, as recorded in [104], commenting at [105] that they “do not appear to be controversial”:

“104.

Mr Peacock submitted that the term “building” is an ordinary English word which must be defined within its particular context and against the relevant statutory background. In that sense he submitted that it was a “flexible term”. He noted that it was implicit in s22 that whilst a building is a fixed structure, not all fixed structures are buildings. He submitted that the typical functions of a building are to provide shelter and security to person or things contained within. However, he submitted that not all structures which provide shelter are buildings, giving the example of the grain silos in Schofield where shelter and security were not the sole or main functions of the structure. A further example in a different statutory context related to stone walls known as beals built in the middle of fields to shelter sheep (Morrison v IRC [1915] 1 KB 716).”

I pause to note that Morrison concerned the valuation for the purposes of certain duties of an agricultural hereditament in the northern Pennines, and the issue was whether the drystone beals were “buildings” of which the farm ought to be deemed to be divested within the meaning of section 25(2) of the Finance (1909-10) Act 1910. In that context, Rowlatt J said, at 722:

“It is quite clear that the expression “buildings” does not mean everything that can by any means be described as built; it means buildings in a more narrow sense than structures, because there are other structures of a limited class which under the terms of the sub-section may also be taken into consideration. In my judgment, as a more or less limited sense has to be placed upon the word “buildings”, - a sense limited in the direction which I have indicated - the only way to construe it is by looking at the nature of the property which is being dealt with. It is impossible to hold that the question whether a thing is or is not a building depends solely on the character of the workmanship that is put into it…. The character of the erection and the nature of the property on which it is and its function on that property must all be looked at.”

Rowlatt J concluded that the walls in question were not “buildings”, in a statutory context which, like ours, drew a distinction between buildings and structures.

The FTT then recorded the submission of Mr Bremner, for HMRC, that “building” must be “given a meaning consistent with everyday usage”, and that the question “whether or not something constitutes a building depends not on its function… but on its inherent characteristics”. Otherwise, said Mr Bremner, section 21 would simply be repeating the common law test for plant.

The FTT agreed that “building” should be given a meaning consistent with its ordinary usage, but disagreed that function is irrelevant. Judge Cannan said, at [107]:

“I agree with Mr Peacock that the function of a structure will be a factor, but is not determinative. The inherent characteristics of a structure must be seen in the context of the function of the structure. Those functions might include providing shelter and security. The common law test for plant considers the function of the asset, in particular its function in the trade. Section 21 in my view requires consideration of the nature and characteristics of a structure including whether or not the functions it is intended to perform are typical functions of a building”.

The FTT then referred to the guidance given by Sir Donald Nicholls V-C in Carr v Sayer, observing at [109]:

“The Vice Chancellor gave the example of a prison, but depending on the trade one might also wish to keep noise, dangerous fumes or other material or unpleasant odours contained. In one sense a building might be designed and function to keep people, animals or things in, as much as to keep the elements or other things out.

110.

On Mr Peacock’s case, the question of whether a structure is a building only arises once it has been determined that it satisfies the common law definition of plant. As such it must be treated as having a plant-like function in the trade and one then looks to see whether it also functions as a building. However, in my view it is not solely a question of how the structure functions. It is also a question of the characteristics of the structure. For example, does it have the form of a building? A structure which has four walls and a roof might naturally be described as a building, whatever specialist function it might have in any trade”.

Having thus directed himself, Judge Cannan considered the disputed assets in turn at [112] to [126]. His conclusion was that each of the five main structures is a building within section 21, and that other separately identified assets are incorporated in or connected with those buildings, either by virtue of section 21(3)(a) or by virtue of section 21(3)(c) read with List A. With regard to the cylinder handling facility and the uranium oxide store, he accepted at [112] that their “predominant purpose” was to provide radiation shielding and containment, and to reflect specific nuclear safety requirements, but “those characteristics do not mean that the structures are not buildings”. Similarly, Judge Cannan accepted that “the predominant function of the vaporisation facility, the kiln facility and the condenser facility is either to support machinery and other equipment or to provide radiation and/or containment rather than shelter”, but again “that does not mean they are not buildings”: see [113].

The Judge also had well in mind that most of the external cladding was “for planning or aesthetic reasons to disguise the structures and to make them look like modern business units”, but the same could not be said “for the roofs of the CHF, the kiln facility and the UOS or the cladding of the condenser facility”: see [114].

By way of example, the FTT then reasoned as follows in relation to the cylinder handling facility:

“115.

The CHF has four walls and a roof and encloses a substantial volume of space. The roof cladding protects the shield roof from the elements, namely standing rainwater. Otherwise, the roof and walls are not intended to provide shelter to material, equipment, machinery or operators inside the CHF. The walls and roof contain radiation inside the structure protecting the environment and the people outside. In my view the CHF does have the inherent characteristics of a building, namely it has walls and a roof. It also functions as a building in containing things. When one looks at the CHF, with or without the cladding, it looks like a building. Overall, I consider that in everyday terminology it is naturally described as a building.

116.

The internal radiation shield walls and the stairs and access platforms are connected with the building and are properly viewed as walls, floors and stairs within List A. As such, expenditure on those items is treated as expenditure on a building. The raised platforms or plinths are in my view incorporated in the building and by virtue of s21(3)(a) expenditure on those items cannot [be] treated as expenditure on plant”.

The FTT then reached similar conclusions in relation to the other four facilities and their contents. In relation to the condenser facility, which was one of the two the FTT had found constituted “plant”, the FTT regarded it as “very much at the margin … whether it would naturally be described as a building”, but since one of its functions was to contain hazardous fumes, it fulfilled one of the functions of a building, “and with four walls, a roof and internal floors it gives the appearance of a building”. On balance, therefore, the FTT was satisfied that it is properly described as a building: see [122].

I will say at once that, in my judgment, the FTT directed itself almost impeccably on this question, and came to conclusions which it was fully entitled to reach. As to the law, the FTT appreciated that the meaning of an ordinary English word like “building” may vary depending on its context, and here both structural characteristics and function must be considered. Although the FTT did not refer to the Ghai case, it wisely resisted the temptation to seek to define an everyday word which Parliament had deliberately left undefined, or at least only partially defined. It also seems clear to me that, in its discussion at [104] to [110], the FTT was in substance adopting an iterative process of the kind recommended by Lord Neuberger MR in Ghai at [26].

If I have one slight criticism of the FTT’s approach, it is that Judge Cannan might have derived more support than he appears to have done, at any rate explicitly, from the actual wording and immediate statutory context of section 21 itself; but that does not begin to amount to a material error of law, because such consideration could only have reinforced his correct rejection of HMRC’s then submission that function was irrelevant. Furthermore, once the FTT had rightly appreciated that considerations of both structural form and function were relevant to the decision whether each of the disputed structures was a building, the evaluation of those matters is a matter deliberately entrusted by Parliament to the common sense of the tribunal of fact. In my judgment, this is an area where undue sophistication should be avoided, and the unvarnished simplicity of the statutory language must be respected.

If I may say so, I consider that the UT fell into the trap of over-complicating this part of the case, and of purporting to detect errors of law in what were, in truth, no more than the FTT’s evaluative conclusions of fact and degree. So, for example, the UT considered that a purposive construction of “building” in section 21 requires considerations of function to be “particularly important”, and “appearance or physical characteristics should not be the primary determinant of the line in the sand which sections 21 to 23 were intended to draw”: see the UT Decision at [130]. In my view, this formulation impermissibly seeks to elevate matters of weight and evaluation into a principle of law which cannot fairly be derived from the statutory context.

I am similarly unable to accept the UT’s criticism, at [131], that the FTT erred in law “in placing emphasis in its reasoning for each of the facilities on their physical appearance or characteristics and how they would be naturally described”. In my view, this was a legitimate approach for the FTT to adopt, once it had correctly directed itself on the law. Nor, in my opinion, was it helpful for the UT to canvass, at [129], a spectrum of possible everyday meanings of “building”, when none of those meanings has any firm basis in the statutory language, and they illustrate no more than the obvious point that, depending on the factual context, the weight to be given to physical and functional considerations will vary.

To conclude, I am unpersuaded that this part of the FTT Decision is vitiated by any material error of law, and in my judgment the UT was wrong to hold otherwise. In reaching that conclusion, the UT itself erred in law.

Accordingly, I would also allow HMCRC’s first ground of appeal.

This ground of Urenco’s cross-appeal raises a short, but important, question about the interpretation of section 23(3) and List C in CAA 2001. The issue is whether Items 1 and 4 in List C, construed in their context, save from the effect of sections 21 and 22 expenditure “on the provision” of those items, or merely expenditure “on” them. If the latter is the correct interpretation, the result, according to HMRC, would be that while capital allowances can be claimed for the actual cost of machinery or processing equipment within the scope of Items 1 or 4, no allowances can be claimed for the ancillary costs of transport, installation, etc which must be incurred before the machinery or processing equipment can be safely used on site. Such expenditure, it is said, would be on the provision of the machinery or equipment, not on the machinery or equipment itself.

This distinction is, of course, one which is well recognised in the case law on what constitutes expenditure “on the provision of plant or machinery” for the purposes of the basic test in section 11(4) of CAA 2001. I have already referred to some of the leading authorities on what is meant by the “provision” of plant or machinery in that context, including the decisions of the House of Lords in Barclay, Curle and Ben-Odeco. It is important to note, however, that for the purposes of section 11 the statutory language which has to be construed is “on the provision of plant or machinery”. The test is, therefore, a composite one, and the question which has to be answered in respect of any disputed items is whether money has been spent “on [its] provision”. It is the unpacking of that language which has led to the distinction in the authorities between expenditure “on” the item, and expenditure on its “provision”.

The critical part of section 23 is subsection (3), which for convenience I will repeat:

“(3)

Sections 21 and 22 also do not affect the question whether expenditure on any item described in List C is, for the purposes of this Act, expenditure on the provision of plant or machinery”.

Accordingly, the expenditure which is to be unaffected by sections 21 and 22 is described as expenditure “on any item” contained in List C, and the saving effect is achieved by saying that sections 21 and 22 “do not affect the question” whether such expenditure is, for the purposes of CAA 2001, “expenditure on the provision of plant or machinery”. The final words take the reader back to the basic test in section 11(4), and repeat its composite wording, thus making clear that inclusion in List C does not alone guarantee the availability of capital allowances. The basic test in section 11(4) must still be satisfied.

List C itself is headed “Expenditure unaffected by sections 21 and 22”. There is then a list of 33 numbered items. Items 1 to 21 inclusive all describe particular types of asset, without any prefatory words (although it is common ground that the preposition “on” must be supplied, by reference back to subsection (3)). Items 1 and 4, with which we are directly concerned, are “Machinery… not within any other item in this list” and “Manufacturing or processing equipment…” respectively. Other examples are “Refrigeration or cooling equipment”, “Strong rooms in bank or building society premises”, “Decorative assets provided for the enjoyment of the public in hotel, restaurant or similar trades”, “Swimming pools (including diving boards, slides and structures on which such boards or slides are mounted” and “Movable buildings intended to be moved in the course of the qualifying activity”.

Item 22, which is the subject of Urenco’s second ground of appeal, is “The alteration of land for the purpose only of installing plant or machinery”.

By contrast, the remaining 11 items in List C are all introduced by the words “The provision of”. They run from Item 23, “The provision of dry docks”, to Item 33, “The provision of fixed zoo cages”. In relation to these items, the word “on” must again be supplied from section 23(3), so the overall effect is to replicate the composite “on the provision of” test which applies for the purposes of section 11. The question which divides the parties is whether the omission of any express reference to “the provision of” Items 1 to 21 means that expenditure on the “provision” of those items cannot qualify, although expenditure “on” them clearly does. The issue does not arise in relation to Item 22, because it would not make sense to talk of expenditure on the provision of the alteration of land. In relation to Items 1 to 21, however, the question is of real importance, because the costs of “providing” the relevant asset in its desired location may be very substantial, for example the cost of excavating the ground to install a swimming pool, or the cost of placing a telecommunications satellite in orbit.

On the face of it, it seems implausible that Parliament would have intended to draw a distinction of this nature in relation to Items 1 to 21 of List C. What rational legislative purpose could be served, for example, by permitting capital allowances to be claimed for expenditure on the provision of dry docks (such as the excavation costs in Barclay, Curle), but not expenditure on the excavation of a site for a swimming pool, or expenditure on the transport and installation costs of machinery, processing equipment, refrigeration equipment or fire alarm systems? It is notable that counsel for HMRC have not attempted to propound any rationale for the distinction, or to point to any admissible aids to construction which might support it. In particular, there is nothing in the detailed explanatory notes to CAA 2001 which provides any support for HMRC’s case. (For the admissibility of such explanatory notes, see R (Project for the Registration of Children as British Citizens v Secretary of State for the Home Department [2022] UKSC 3, [2022] 2 WLR 343, at para [30] per Lord Hodge DPSC).

The summary at the beginning of the explanatory notes to CAA 2001 states that:

“3.

The main purpose of the Capital Allowances Act is to rewrite tax legislation relating to capital allowances so as to make it clearer and easier to use.

4.

The Act also makes some minor changes to the legislation. These are within the remit given to the Tax Law Rewrite Project and the Parliamentary process for the Bills it produces”.

The explanatory note to section 23, headed “Expenditure unaffected by sections 21 and 22”, then says:

“169.

This section is based on column 2 of Table 1 in paragraph 1, column 2 of Table 2 in paragraph 2 and paragraph 1(3) of Schedule AA1 .... It also makes two minor changes.”

With regard to List C, paragraphs 174 to 176 of the explanatory notes state that:

“174.

Subsection (4) includes List C. This is made up from column 2 from both Table 1 and Table 2 of Schedule AA1 to CAA 1990. It also includes the items in paragraph 1(3).

175.

The merger of the columns involves a minor change. In Schedule AA1 the columns apply differently:

•

whether a building is plant is unaffected by the Schedule for assets in column 2 of Table 1;

•

whether a structure is plant is unaffected by the Schedule for assets which are within either column 2 of Table 1 or column 2 of Table 2 (paragraph 2(3) of Schedule AA1 provides this rule).

176.

Merging the Tables in this Act in principle increases the range of expenditure on buildings which is unaffected by the exclusion of buildings from the definition of plant. See Change 2 in Annex 1.”

In the relevant part of Annex 1, the explanatory note on “Change 2” states that:

“Sections 21 to 23 are based on Schedule AA1 to CAA 1990. But they differ from the Schedule in that they simplify and extend the provisions about what items are treated as unaffected by the express exclusions from what can be plant or machinery.

Schedule AA1 was inserted in CAA 1990 by section 117 of FA 1994 against the background of burgeoning case law extending the meaning of “plant”. The Schedule was intended, so far as practicable, to call a halt to this process. In general terms, it contains provisions excluding things from being machinery or plant and other provisions about things that are unaffected by the exclusions.

…

An extension of the provisions treating items as “unaffected” by the express exclusions therefore means that more items can be treated as plant or machinery. This in turn favours the taxpayer, by potentially extending the range of circumstances in which the taxpayer may obtain plant and machinery allowances.

In more detail, list C in section 23 merges column 2 of Table 1, column 2 of Table 2 and paragraphs 1(3)(b) to (e) and 5(2) of Schedule AA1 into a single list of items that are unaffected by the express exclusions in sections 21 and 22.

This changes the literal effect of paragraphs 1(3) and 2(3) of Schedule AA1. Paragraph 1(3) treats the question whether expenditure on the items in paragraphs (b) to (e) is expenditure on plant or machinery as unaffected by paragraph 1(1) only. Paragraph 2(3) treats the items in column 2 of Table 2 as unaffected by paragraph 2(1) only”.

It is apparent from these explanatory notes that the intended overall purpose of section 23 and List C was to merge various provisions previously contained in Schedule AA1 to CAA 1990, as inserted by the 1994 amendments, and at the same time to introduce a minor change which was deliberately intended to favour the taxpayer by extending the disregard for items contained in the new merged list so that, for each item, the disregard embraces the whole of sections 21 and 22. It would indeed be surprising if, in the same context, Parliament had deliberately legislated in such a way as to preclude the giving of any allowances at all for expenditure on the provision of more than half the items contained in List C. Not only would such a change have gone well beyond the minor changes envisaged as acceptable under the Parliamentary process for Bills introduced pursuant to the Tax Law Rewrite Project, but, even more surprisingly, not a word of explanation or warning would have been given in the extensive explanatory notes for CAA 2001 which (as paragraph 1 of the notes records) had “been prepared by the Tax Law Rewrite Project at the Inland Revenue in order to assist readers of this Act and to help inform debate on it”.

Against this background, there must in my view be a strong suspicion, even without recourse to the predecessor legislation, that something has gone wrong with the drafting of section 23(3) and List C, or at least that those provisions should, if at all possible, be construed in a way that does not draw a distinction between items contained in List C depending on whether the expenditure is “on” them or on their “provision”. In those circumstances, common sense also suggests that it must be permissible to look back to the predecessor provisions in Schedule AA1 to CAA 1990 to see if they throw any light on the problem.

Schedule AA1 to FA 1994 is headed “Exclusions from expenditure on machinery or plant”. Paragraph 1 then deals with “Buildings”, and paragraph 2 with “Structures, assets and works”. So far as relevant, paragraph 1 provides:

“1(1). For the purposes of this Act expenditure on the provision of machinery or plant does not include any expenditure on the provision of a building.

(2)

For the purposes of this Schedule “building” includes any asset in the building –

(a)

which is incorporated into the building, or

(b)

which, by reason of being movable or otherwise, is not so incorporated, but is of a kind normally incorporated into buildings;

and in particular includes any asset in or in connection with the building included in any of the items in column 1 or column 2 of the following Table (“Table 1”).

(3)

Sub-paragraph (1) above does not affect the question whether expenditure on the provision of –

(a)

any asset falling within column 2 of Table 1,

… is for the purposes of this Act expenditure on the provision of machinery or plant”.

Table 1 has two columns, the first of which is headed “Assets included in the expression “building””, and the second of which is headed “Assets so included, but expenditure on which is unaffected by the Schedule”. Column 2 contains 16 numbered items, which in substance correspond with Items 1 to 16 in List C (although not always in the same order). The important point is that expenditure on the provision of those items was clearly included, by virtue of the express wording of sub-paragraph 1(3) (“expenditure on the provision of… any asset falling within column 2 of Table 1”). It is also relevant to note that the heading to column 2 uses the phrase “expenditure on which”, which must in this context be equivalent to, or shorthand for, “expenditure on the provision of which”.

Paragraph 2 of Schedule 1AA is organised in a generally similar way, but with some significant differences:

“2.

(1) For the purposes of this Act expenditure on the provision of machinery or plant does not include any expenditure on –

(a)

the provision of structures or other assets to which this paragraph applies, or

(b)

any works involving the alteration of land.

(2)

This paragraph applies to any structure or other asset which falls within column 1 of the following Table (“Table 2”).

(3)

Sub-paragraph (1) above does not affect the question whether –

(a)

any expenditure falling within column 2 of Table 2, or

(b)

any expenditure on the provision of any asset of a description within any of the items in column 2 of Table 1,

is for the purposes of this Act expenditure on the provision of machinery or plant.”

Like Table 1, Table 2 has two columns, the first headed “Structures and assets”, and the second “Expenditure which is unaffected by the Schedule”. Column 2 contains 12 numbered items, beginning with “Expenditure on the alteration of land for the purpose only of installing machinery or plant”. This corresponds with the future Item 22 in List C. Items 2 to 12 in column 2 then correspond with Items 23 to 33 in List C, save that each begins with the words “Expenditure on the provision of”. Again, therefore, it is clear beyond argument that expenditure on the provision of the listed items is included; and, for good measure, the draughtsman has also expressly provided, in paragraph 2(3)(b), that paragraph 2(1) does not affect the question whether expenditure on the provision of any asset in column 2 of Table 1 is expenditure on the provision of machinery or plant. This therefore reinforces the point that expenditure on the provision of the Table 1 items, which include the future Items 1 and 4 in List C, was expressly included in the statutory carve out from disqualified expenditure on the provision of a building.

With the benefit of hindsight, it is not difficult to see what went wrong when the draughtsman of section 23 and List C in CAA 2001 attempted to merge column 2 of Table 1 and column 2 of Table 2 (together with certain other provisions) “into a single list of items that are unaffected by the express exclusions in Sections 21 and 22” (to quote from the explanatory note on Change 2, set out at [134] above). Inadvertently, the draughtsman reproduced the substance of the items listed in column 2 of Table 1, but not the wording of paragraphs 1(3) and 2(3)(b) which made it clear that the saving extended to expenditure on the provision of those items, as well to expenditure on them.

The differences in the structure of Tables 1 and 2, and their governing paragraphs, also explain the otherwise strange contrast between the wording of Items 1 to 16 in List C, which make no express reference to “provision”, and the wording of Items 23 to 33, which do.

If my analysis is correct thus far, the next question is whether it is possible to construe section 23 and List C in a way which gives effect to Parliament’s evident intention, as elucidated by recourse to the 1994 precursor legislation. I consider that there are at least two ways in which the language of section 23 and List C may legitimately be so construed.

The first, and in my view preferable, solution is to construe the words “on any item” in section 23(3) as equivalent in meaning, in this particular context, to “on the provision of any item”. In a suitable context, I see no reason why expenditure “on” an item should not include expenditure on its provision, or (alternatively) why it should not be regarded as shorthand for the composite phrase “on the provision of” which lies at the heart of the judge-made law on capital allowances reflected in section 11(4) of CAA 2001. That this is a perfectly possible, and to my mind natural, use of language is in my view borne out by the heading to column 2 of Table 1 in Schedule 1AA to CAA 1990: see [138] above. I would add that these column headings appear to me to be an integral part of the Tables as drafted, and not comparable with side notes which are merely added for ease of reference and receive no Parliamentary scrutiny; but even if that is wrong, the column headings still form part of FA 1994 as enacted.

The second, and more radical, solution is to exercise the power, which the court undoubtedly has, to correct clear cases of drafting mistakes as a matter of statutory construction. As Lord Nicholls explained in Inco Europe Ltd v First Choice Distribution [2000] 1 WLR 586, at 592:

“This power is confined to plain cases of drafting mistakes. The courts are ever mindful that their constitutional role in this field is interpretative. They must abstain from any course which might have the appearance of judicial legislation. A statute is expressed in language approved and enacted by the legislature. So, the courts exercise considerable caution before adding or omitting or substituting words. Before interpreting a statute in this way, the court must be abundantly sure of three matters:

(1)

the intended purposes of the statute or provision in question;

(2)

that by inadvertence the draughtsman and Parliament failed to give effect to that purpose in the provision in question; and

(3)

the substance of the provision Parliament would have made, although not necessarily the precise words Parliament would have used, had the error in the Bill been noticed.

The third of these conditions is of crucial importance. Otherwise, any attempt to determine the meaning of the enactment would cross the boundary between construction and legislation: see per Lord Diplock in G Jones v Wrotham Park Settled Estates [1980] AC 74,105-106.”

For an example of the application of this principle to a tax statute, see the decision of this court in Pollen Estate Trustee Company Ltd v Revenue and Customs Commissioners [2013] EWCA Civ 753, [2013] 3 All E R 742. The leading judgment in that case was delivered by Lewison LJ, with whom McFarlane and Laws LJ agreed. At [26], Lewison LJ also quoted the well-known words of Lord Reid in Luke v IRC [1963] AC 557 at 577:

“To apply the words literally is to defeat the obvious intention of the legislation and to produce a wholly unreasonable result. To achieve the obvious intention and produce a reasonable result we must do some violence to the words. This is not a new problem, though our standard of drafting is such that it rarely emerges. The general principle is well settled. It is only where the words are absolutely incapable of a construction which will accord with the apparent intention of the provision and will avoid a wholly unreasonable result, that the words of the enactment must prevail”.

In the present case, if one needs to go that far, I consider that the three conditions stated by Lord Nicholls in Inco Europe are clearly satisfied. First, the evident purpose of List C was to preserve allowances on expenditure which would have qualified for relief both before and after the enactment of FA 1994. Secondly, a drafting error has obviously been made in transposing the items from the second columns of Tables 1 and 2 in Schedule AA1 to CAA 1990. Thirdly, it is clear what Parliament was seeking to achieve in substance, and it is easy to supply wording which would have corrected the error. This could be done either by inserting the words “the provision of” between “on” and “any item” in section 23(3) of CAA 2001, or, more laboriously, by inserting the words “The provision of” before each of Items 1 to 21 in List C.

It remains to consider whether it is legitimate, in the present case, to have recourse to the statutory antecedents of section 23 and List C as an aid to their construction. Given the clear and cogent assistance which such recourse provides, and the detailed references to the previous legislation contained in the explanatory notes, I would be dismayed if this were forbidden by binding authority. Fortunately, however, that is not the case. In Derry, loc. cit., the Supreme Court gave valuable guidance on the correct approach to interpretation of a Tax Law Rewrite statute: see the judgment of Lord Carnwath JSC at [7] to [10], and the further observations of Lady Arden JSC at [84] to [90]. The guidance emphasises that, in construing either a consolidating statute (such as CAA 1990) or a Rewrite statute (such as CAA 2001), it would in general be wrong for the court to refer back to antecedent legislation: see [9] and [87]. However, the fact that this is the general rule clearly allows for possible exceptions in an appropriate case. Moreover, in Farrell v Alexander [1977] AC 59, which is the leading case on the construction of consolidating statutes, Lord Wilberforce, while also affirming the general rule, recognised at [73] that recourse to antecedents “should only be had when there is a real and substantial difficult or ambiguity which classical methods of construction cannot resolve.”

In my judgment, the construction of section 23 and List C as they stand in CAA 2001 does give rise to a real and substantial difficulty, because of the striking differences in the drafting of the Items in List C, and the sheer implausibility of Parliament having wished to draw a distinction between expenditure “on” the items in the first part of the List and expenditure on the “provision” of those items. The absence of any explanation in the explanatory notes, which themselves refer back to the predecessor legislation and appear to indicate an intention to replicate it, subject only to specified minor changes, can only reinforce the nagging sense that something must have gone wrong in the drafting. Accordingly, this appears to me to be a classic instance of a case where recourse to the antecedent legislation is not only permitted, but is essential if justice is to be done.

I should add that, in reaching this conclusion, I have had in mind the cautionary note recently sounded by the Supreme Court in NCL Investments Ltd v HMRC [2022] UKSC 9, [2022] 1 WLR 1829, at [44] to [47]. As I read those paragraphs, however, they cast no doubt on the guidance given by Lord Wilberforce in Farrell v Alexander, which must in my view apply with at least equal force to a Tax Law Rewrite statute such as CAA 2001.

The two Tribunals below accepted HMRC’s argument that the language of section 23 and List C is clear, and that Parliament must have intended to draw a distinction between expenditure on certain assets, and expenditure on the provision of those assets: see the FTT Decision at [141] to [144] and the UT Decision at [150] to [167]. Apart from considering the wording of section 23 to be “plain”, the UT at [166] thought that “a literal construction of Items 1 and 4 … produces a result which is not illogical or absurd”. The UT was “not persuaded that Parliament made a clear and obvious error in relation to the first sixteen Items in List C” (ibid). For the reasons which I have given, I respectfully disagree, and I would therefore allow Urenco’s appeal on this ground. If the other members of the Court agree, I would accordingly remit the question to the FTT for reconsideration.

At the remitted hearing, the FTT should in my view consider the application of Items 1 and 4 of List C, correctly construed, and should also be free to determine which assets potentially fall within the ambit of those Items. The FTT should not necessarily be confined to consideration of those items which were the subject of argument before the UT, and it should rule on any objections which HMRC may raise to the consideration of particular assets. The procedural history of this question is a little complex, as Urenco did not include reliance on Items 1 and 4 in its original grounds of appeal to the FTT, and only raised the issue shortly before the hearing. The FTT granted permission at the hearing for Urenco to amend its grounds of appeal, taking the view that HMRC were not materially prejudiced by Urenco’s late reliance on Items 1 and 4: see the FTT Decision at [129] to [140]. On the appeal to the UT, there was further argument on the question of which of the disputed assets could fairly be considered for the purposes of this ground, and Mr Peacock, in response to a request by the UT, produced a list of potentially relevant assets which was longer than those by reference to which permission to appeal had been granted by the FTT. In the event, the UT found it unnecessary to determine any of these procedural issues, because it rejected Urenco’s case on the construction of Items 1 and 4.

Neither side has suggested that this court should attempt to resolve the procedural issues, and I therefore consider that the appropriate course is to include them in the remitter to the FTT. I would hope that Urenco and HMRC will be able to reach agreement on the precise scope of the remitter, and the assets in respect of which the application of Items 1 and 4 is to be considered by the FTT. If agreement cannot be reached, the FTT should give appropriate directions for resolution of such matters as remain in dispute.

Item 22 of List C saves from the disqualifications in sections 21 and 22 of CAA 2001 expenditure on “The alteration of land for the purpose only of installing plant or machinery”. As with the other items in List C, the expenditure in question must still satisfy the basic conditions in section 11 if it is to qualify for capital allowances. No issue arises, however, about expenditure on the “provision” of this item, because (as I have already said) the alteration of land is not something that can meaningfully be provided. Expenditure is either on the alteration, or it is not.

It is common ground that “land”, in Item 22, has the meaning provided by Schedule 1 to the Interpretation Act 1978 (“the 1978 Act”), “unless the contrary intention appears”: see section 5 of the 1978 Act. As defined in Schedule 1, the word “includes buildings and other structures, land covered with water, and any estate, interest, easement, servitude or right in or over land.” The important point is the inclusion of “buildings and other structures” in the meaning of “land”. As Mr Peacock pointed out to us, “land” has a different meaning in section 22, where it is defined for the purposes of that section as not including buildings or other structures, although it otherwise has the meaning given by the 1978 Act: see section 22(3)(b).

It is also worth noting that, when Item 22 was first enacted, as item 1 in column 2 of Table 2 in Schedule AA1 to CAA 1990, “land” was defined for the purposes of that Schedule by reference to the 1978 Act, but again with the omission of the words “buildings and other structures”: see paragraph 5(3) of Schedule 1AA.

In the light of that fact, and the absence of any explanatory note or other admissible material to explain why the change was made, there may perhaps be scope for an argument that Parliament intended the modified 1978 Act definition (excluding buildings and other structures) to apply for the purposes of Item 22 in List C, as well as for the purposes of section 22, and that this was another drafting error made in transposing the provisions of Schedule 1AA into Chapter 3 of Part 2 of CAA 2001. Support for such an argument might also be found in the fact that the word “land” appears nowhere else in section 23 and List C, apart from its solitary appearance in Item 22. However, no such argument was advanced by either side, although counsel for HMRC do contend in their written submissions on this part of the case that the same result is achieved on the basis that, read in context, a contrary intention appears within section 5 of the 1978 Act, so that “land” in Item 22 must be interpreted as excluding buildings and other structures.

I have considerable sympathy with that argument, but it was not developed orally before us, and I would not wish to reach a firm conclusion on it without also hearing argument on the question of a possible drafting mistake. Since, as I shall explain, resolution of the point is unnecessary to our decision on this ground of appeal, I am content to proceed on the assumption that Urenco is right to submit that “land” in Item 22 has its unmodified 1978 Act meaning, and thus includes buildings and other structures.

In their written submissions, counsel for Urenco provide a helpful example of how Item 22 would operate in a simple case:

“An uncontentious example of the operation of Item 22 is the installation of a glasshouse within List C, Item 17. If a taxpayer lays foundations in order to receive the glasshouse, the taxpayer will have altered the land. In the absence Item 22, those groundworks could be caught by s.22(1)(b) as ‘works involving the alteration of land’. The effect of Item 22 is to ensure that the expenditure on those groundworks nevertheless qualifies for allowances (so long as receiving the glasshouse is the only purpose of the groundworks).”

I have no difficulty with that example, but problems arise as soon as one considers expenditure on the alteration of buildings and structures. Urenco does not shrink from advancing the bold argument that the construction of all the constituent parts of the TMF involved alteration to the “land”, in its extended sense, and that if (as the FTT found in relation to all of the main facilities) the disputed assets are “buildings”, and thus prima face disqualified by section 21, the whole of the relevant expenditure on those assets is then saved by section 23 and Item 22. Urenco submits that the sole purpose of the alterations was the installation of machinery and processing equipment, and that altering land to create a location for the plant and machinery to be used safely is an ordinary installation purpose which does not prevent Item 22 from applying.

The FTT did not accept this argument, finding that the relevant buildings were not constructed “for the purpose only of installing plant or machinery”. Having decided that it was unnecessary to determine the “pure question of law” about the meaning of “land” in Item 22, the FTT said, at [151]:

“Even if Urenco are right, they must still satisfy me that the structures were constructed solely for the purpose of installing plant or machinery. That is clearly not the case here. Mr Peacock submitted that each of the disputed assets was designed and constructed solely with a view to enabling the installation and safe operation of the TMF. Expenditure incurred because it is necessary to create a location for the plant and machinery to be used safely is part of the installation purpose. I do not accept that submission. They were constructed in part at least to protect operatives, the public and the environment and to provide premises which house the plant and machinery. Not for the purposes of installation.”

The decision of the FTT on the purpose or purposes for which the relevant alterations of land were made was in principle a decision on a question of fact which would normally be immune from challenge on an appeal confined to questions of law. Undeterred, however, Urenco argued in the UT that the FTT’s decision involved an error of law, in that the separate safety purposes identified by the FTT were all installation purposes, and they could not logically be dissociated from the actual process of installation. The UT was unimpressed by this argument, saying at [145]:

“The FTT regarded the purposes described in the final two sentences of [151] [of the FTT Decision] as not being “for the purpose only of installing” plant and machinery. We consider that they were justified in doing so. Mr Peacock’s proposed construction of “installing” would give an extremely wide meaning to the term, and afford little weight to the word “only”. No case was put forward why a purposive construction would or should lead to such a result. We agree with the Upper Tribunal in SSE that “installation” is apt to describe a process of integrating one thing into another, and not the construction or manufacture of an asset before it is installed. We also agree that the use of the word “only” makes clear that the saving in Item 22 was intended to be a limited one.”

I respectfully agree with the UT that the saving in Item 22 must have been intended by Parliament to be a limited one. That is clear not only from the use of the word “only”, but also from the focus on the installation of plant or machinery, and the requirement that the relevant alteration of land must have no other purpose. The problem with Urenco’s much broader conception of what “installation” means in this context is that, in a case of the present type, the saving in Item 22 would in practice swallow up the basic disqualification of the provision of buildings in section 21. Where buildings are excluded from the general prohibition, specific language is used in List C, which narrows down the scope of the exception. Thus, the immediately preceding Items in the List cover expenditure on:

“20.

Buildings provided for testing aircraft engines run within the buildings.

21.

Moveable buildings intending to be moved in the course of the qualifying activity.”

In order to contain Item 22 within reasonable bounds, therefore, the alteration of land must in my view be for the sole purpose of installing items of plant or machinery which have a separate existence, and which need to be fitted in place within the building. The wording of Item 22 is not apt to include cases where the purposes to be served go beyond the installation of extraneous pieces of equipment, and still less to include cases where the alterations consist in works of construction of a structure or building which, when completed, is intended to function as a single item of plant in its own right.

In common with the UT, I find assistance on this point from the reasoning of the UT in the SSE case. The relevant passages are set out in the UT Decision at [142], from which I quote the following extracts:

“127.

The OED defines “install” as “place (an apparatus, system, etc.) in position for service or use”. We accept that the case law does not limit the term to simply taking a prefabricated asset and placing it in position…However, in the case law which we have reviewed, the common theme is the process which involves the integration, often with a degree of complexity, of an article or articles which have already been made into another article, structure, building or even the land itself. In none of the cases that we have been referred to has the term been held to include the creation of an item of plant in situ.

128.

…It seems to us that Item 22 in List C is confined to items which need to be installed separately from the process of manufacture or construction.”

As the UT in the present case went on to observe at [143], when SSE reached this court, Rose LJ said at [69] that she saw the force of the UT’s reasoning on the question, and did not dissent from it, although she declined to come to a concluded view in the abstract “given the elastic nature of the words used in these provisions”.

Given the high level of generality at which this part of the case was argued before us, I too would be reluctant to express a concluded view on the precise scope of the concept of “installation” in Item 22. Nevertheless, I feel little doubt that a restricted meaning, along the lines indicated by the UT in SSE, is appropriate, and that it cannot have the wide meaning for which Urenco contends. On that basis, I consider that the FTT was fully justified in identifying non-installation purposes which meant that the sole purpose test in Item 22 was not satisfied, and I would also respectfully endorse paragraph [145] of the UT Decision (quoted above).

For these reasons, I would dismiss the second ground of Urenco’s cross-appeal.

Accordingly, if the other members of the court agree, I would allow HMRC’s appeal on all three grounds, and I would also allow Urenco’s cross-appeal on the first ground, but not on the second ground which I would dismiss.

I agree.

I also agree.