Cobalt Data Centre 2 LLP Cobalt Data Centre 3 LLP v The Commissioners for HM Revenue and Customs: [2020] UKUT 0356 (TCC) [2020] UKUT 0356 (TCC)
[3]INCOME TAX– Expenditure on relevant interest in a building – apportionment – amount qualifying for enterprise zone allowances[5]COBALT DATA CENTRE 2 LLP COBALT DATA CENTRE 3 LLP Appellants -and- THE COMMISSIONERS FOR HER MAJESTY’S REVENUE AND CUSTOMS Respondents[6]TRIBUNAL MR JUSTICE ZACAROLI JUDGE JONATHAN RICHARDS[7]Sitting in public by way of remote video hearing treated as taking place at The Rolls Building, Fetter Lane, London from 17 to 19 November 2020[8]Nicola Shaw QC and Michael Jones, instructed by Macfarlanes LLP for the Appellants[9]David Ewart QC, Edward Waldegrave, Emma Pearce and Laura Ruxandu, instructed by the General Counsel and Solicitor to HM Revenue and Customs, for the Respondents © CROWN COPYRIGHT 2020 DECISION Introduction This decision is an addendum to our Decision Notice dated 15 November 2019 (the “Decision”) in which we determined a number of issues arising in the taxpayers’ appeals against HMRC’s closure notices and their claim for judicial review. We refer to the Decision for the relevant background. We adopt the same definitions and abbreviations used in the Decision. Recap of the Decision In the Decision we concluded that: the expenditure incurred by the Developer in relation to the construction of the Data Centres was incurred under a contract entered into prior to 19 February 2006; the LLPs were carrying on business with a view to profit for the purpose of s863 of ITTOIA (albeit that the principal purpose of the LLPs was to obtain the benefit of the EZAs for their members); and the Price paid by the LLPs was not wholly paid for the relevant interest for the purposes of s296 of CAA 2001 with the result that the LLPs were not entitled to EZAs on the whole of the Price. For the reasons set out at [286] to [291] of the Decision, however, we deferred consideration of how much of the Price was paid for the relevant interest. The relevant interest in this case included the rights acquired by the LLPs in April 2011 under the Golden Contract (being the right to have the Data Centres built and fitted out). In this decision, we will refer to these as the “GC Rights”. We characterised the component elements of the consideration for which the Price was paid – in addition to the GC Rights – as follows (see [240] of the Decision): Expenses support arrangements (“ESAs”); Rental support arrangements (“RSAs”); Capital repayment support arrangements (“CRSAs”); and The discharge of the Arranger’s Fee. Of these, we accepted that the RSAs were to be regarded as part and parcel of the relevant interest and so ancillary to it (see [250] of the Decision). It therefore followed that any part of the Price paid for the RSAs was paid for the relevant interest under s296 and so attracted EZAs. We concluded (see [253] of the Decision) that the ESAs, by contrast, were an asset separate from the relevant interest. Accordingly, amounts paid for ESAs did not attract EZAs. However, in our determination of the judicial review claim, we concluded that HMRC had to apply their published practice and grant the LLPs EZAs on the part of the Price that was paid for ESAs. The Yearly Sums payable by the Developer to the LLPs (described at [230], [240] and [241] of the Decision) comprised the ESAs and RSAs. The LLPs adduced evidence as to the value of the Yearly Sums (by way of Mr Watson’s expert report). That evidence, which was unchallenged and which we accepted in the Decision, was that the present value, as at April 2011, of the Yearly Sum was £10.89 million in respect of DC2 and £7.86 million in respect of DC3. In the Decision, we also accepted Mr Watson’s evidence as to the value of the GC Rights. He valued the Data Centres on three different bases: On the assumption that the Data Centres were completed, fully fitted and let to grade ‘A’ tenants on a 15-year lease at rent of £170 psf uplifted in accordance with RPI, then DC2 was worth £76,360,000 and DC3 was worth £54,520,000; On the assumption that the Data Centres were completed, fully fitted and let to grade ‘A’ tenants on a 15-year lease at rent of £170 psf, but on the basis of a 30-month rent-free period, then DC2 was worth £63,730,000 and DC3 was worth £45,500,000; and On the assumption that the Data Centres were completed, fully fitted and available with vacant possession, then DC2 was worth £51,275,000 and DC3 was worth £36,835,000. Mr Watson concluded, however, that the reasonable value of the interest acquired by the LLPs (which effectively involved a commitment from the Developer to deliver, fit out and market the Data Centres with associated rental support arrangements) was equal to the net present value (as at April 2011) of the Data Centres, completed and fully fitted but with vacant possession, plus the net present value (as at the same date) of the Yearly Sum. For DC2, that was £48.335 million + £10.89 million = £59.22 million. For DC3 it was £34.725 million + £7.86 million = £42.58 million. Mr Watson’s valuations ignored any impact which there may have been from the fact that the Data Centres, being located in an EZ, would entitle a purchaser to EZAs. The Arranger’s Fee had a fixed value of £9,555,716 (exclusive of VAT) for DC2 and £6,587,670 (exclusive of VAT) for DC3. The issues now to be determined We received further witness statements from each of Mr Fielding and Mr Pulford and both of them attended for cross-examination. We also heard expert evidence as follows. Mr Douglas Smith of CBRE (who gave evidence on behalf of the LLPs at the first hearing) provided a further report addressing the question of the extent, if any, to which the value of the GC Rights as at April 2011 was increased by the fact that the Data Centres were located in an EZ. Mr Adrian Williams of EY LLP provided an expert report (on behalf of the LLPs) addressing the question of the value of the CRSAs. Mr Ian Mackie of Berkeley Research Group provided an expert report (on behalf of HMRC) addressing both of those questions. Each of the experts attended for cross-examination. The task now before us is to determine how much of the Price that the LLPs paid to the Developer qualifies for EZAs. Both parties are content to proceed on the basis that the attribution is to be made under s296 of CAA 2001 by asking how much of the Price was paid “for” the relevant interest. They were agreed that s562 of CAA 2001 is not relevant to the determination of that issue. They also agree that the provisions of s356 of CAA 2001 are not applicable as s356 deals with a different situation, where it is clear that an amount has been paid for a single relevant interest but some of that expenditure qualifies for allowances and some does not: for example where a purchaser pays a price entirely for a “relevant interest” in a building but that building consists partly of residential units, which do not qualify for EZAs, and partly of offices, which do. That said, both parties agreed with the Upper Tribunal’s observation at [286] of the Decision that it was unlikely to make any practical difference whether the attribution was performed under s296 or s356. In essence, the task involves determining how much of the Price was paid for “qualifying” assets consisting of the GC Rights, the RSAs and, because of our conclusion in the judicial review, the ESAs. As to the other elements acquired by the LLPs:(1) no valuation exercise is necessary in relation to the Arranger’s Fee, as it was in a fixed amount; and(2) as we explain below, the parties disagreed as to whether it was necessary to reach a standalone value for the CRSAs, in order to determine how much of the Price was paid for the qualifying assets. Both parties were agreed however that, while an analysis of the market value of any of the elements that the LLPs acquired could be relevant and instructive, that market value would not in itself determine how much was paid “for” the relevant asset since, conceptually, it is possible for someone to pay more, or less, for a particular asset than its market value. Both parties were also agreed that, in practice, the market value of the GC Rights would be enhanced by the fact that the Data Centres were located in EZs although they differed as to (i) the “starting value” to which that enhancement should be applied and (ii) the amount of the appropriate enhancement. Both parties also emphasised that the submissions made at the present hearing were without prejudice to submissions they may wish to make in any appeal against the Decision. With that background, the LLPs argued that it is necessary to adopt the following approach: As Step 1, identify an appropriate market value for each element of the Price. As Step 2, identify the respective contribution that each element made to the total Price paid. As Step 3, stand back and consider whether the answer produced is just and reasonable in the particular circumstances of the case in order to conclude what was paid “for” each element. The LLPs applied that approach as follows in relation to DC2 (adopting a similar approach in relation to DC3 which we will not set out) in reliance on the expert evidence adduced by them: At Step 1: The market value of the relevant interest should be determined by reference to how the value of that asset would be assessed by an investor. An investor would (and the LLPs did) assess that value on the basis that the Data Centres were fully fitted and let to a Grade A tenant on a 15-year lease at a rent of £170 psf uplifted in accordance with RPI. Ignoring the effect of EZAs, the market value of DC2, had it been let on such terms as at April 2011, was £76,360,000, as set out in Mr Watson’s expert report. It was appropriate to uplift that value by a factor of two-thirds to reflect the benefit of EZAs so that the market value of the relevant interest was £127,260,000 in relation to DC2. The valuation of DC2 on the assumption that it was fully let and income-producing already took into account the effect of RSAs with the result that no separate value needed to be allocated to RSAs. The market value of the ESAs relating to DC2 was £2,001,638 (a figure that includes no uplift to reflect the Data Centres’ location in an EZ). The market value of the CRSAs was £27,565,363.50 (no uplift for location in an EZ being necessary). The market value of the benefit of the Developer’s agreement to discharge the Arranger’s Fee was £9,555,716 using the VAT-exclusive figure, and with no uplift in respect of location in an EZ. At Step 2, since the aggregate market value of the individual components set out at Step 1 (£166,382,447.50) exceeds the total Price that CDC2 paid to the Developer, it is necessary to “scale back” each figure determined at Step 1 by an appropriate factor reflecting the “relative share” of that factor. This is a purely mathematical exercise and results in the following values for each element: £117,567,106 for the relevant interest (including the right to RSAs); £1,848,932 for ESAs; £25,465,818 for CRSAs; and £8,827,895 for the discharge of the Arranger’s Fee. At their Step 3, the LLPs conclude that the answer produced at Step 2 represents a just and reasonable apportionment of the Price in all the circumstances. Accordingly, they reason that CDC2 is entitled to EZAs on £119,616,038 (being the aggregate of the £117,567,106 that they attribute to the combination of the relevant interest and the RSAs and the £1,848,932 they apportion to the ESAs). HMRC’s preferred approach, which we also illustrate by reference to the figures for DC2, proceeds as follows: HMRC take as their starting point the market value of the GC Rights, being the right to receive a fully fitted, but unlet, DC2 as at April 2011 of £48.335m. Relying on Mr Mackie’s expert evidence, the premium attributable to the fact that DC2 was located in an EZ was £7.8m, an increase of 16.1%. Accordingly, the market value of the GC Rights as at April 2011, including the effect of EZAs, was £56.1m. The Upper Tribunal has already determined that the market value of the right to receive Yearly Sums is £10.89m. HMRC apply a similar 16.1% uplift to that figure so as to result in a combined value of the RSAs and the ESAs comprised in the Yearly Sum of £12.7m. Accordingly, the aggregate market value of the GC Rights, the RSAs and the ESAs including the effect of EZAs was £68.8m. Conceptually, it is possible for someone to pay more than market value for an asset. However, there was no basis for concluding that the LLPs had done so in this case as there was no evidence that the parties had allocated any particular part of the Price to particular elements of the “package” of rights that CDC2 obtained. Moreover, the Decision set out a rejection of HMRC’s arguments, made by way of analogy to the decision of the Supreme Court in Tower MCashback LLP1 v HMRC [2011] STC 1144, to the effect that the way in which certain elements of the Price were used provided an indication as to what those elements were paid “for”. Therefore, in HMRC’s submission, there was little alternative to concluding that the price attributable to the relevant interest was the market value of that interest. CDC2 was thus entitled to claim EZAs by reference to aggregate expenditure of £68.8m (£56.1m representing the market value of the GC Rights plus £12.7m in respect of the Yearly Sum). On HMRC’s primary case, there was no need to calculate individual values for the CRSAs or the obligation to discharge the Arranger’s Fee. However, HMRC set out an alternative case if, contrary to their submissions, it was necessary to apportion the Price among all of the rights that the LLPs acquired. That case can be summarised as follows: The aggregate market value of the GC Rights, the RSAs and the ESAs was £68.8m for the reasons noted at [16] above. The VAT-exclusive amount of the Arranger’s Fee was £9.6m. That amount should be uplifted by 16.1% (consistent with the approach proposed in relation to the Yearly Sums set out at [16(2)] above) so as to produce a figure of £11.15m. Mr Mackie’s preferred method of valuing the CRSAs involves a “residual value approach” and produces a figure of £72m. The aggregate of the market values set out at (1) to(3) above was £151.95m, somewhat less than the purchase price of £153.7m. HMRC adopt a similar scaling approach to that adopted by the LLPs and summarised at [15(2)] above, thus attributing £69.6m as paid “for” the GC Rights and the Yearly Sum. Alternatively, if the CRSAs are valued on an “income approach”, they would have a value of £59m. Applying the same scaling approach, £76.08m is attributed as paid for the GC Rights and Yearly Sums. The parties’ respective approaches therefore give rise to the following areas of disagreement that we will use to structure the discussion that follows: The parties do not agree on the appropriate starting point that should be used when deciding on the “uplift” to be applied when ascertaining the market value of the relevant interest taking into account the Data Centres’ location in an EZ. The LLPs start with Mr Watson’s “fully let and income-producing value” (£76,360,000 in the case of DC2). HMRC start with his “vacant possession” value (£48,335,000 in the case of DC2). Whatever the correct starting point, the parties do not agree on the appropriate uplift to apply to reflect the value of EZAs. In reliance on Mr Smith’s expert evidence, the LLPs argue for an uplift of 66.67%. HMRC, in reliance on Mr Mackie’s expert evidence, argue for an uplift of 16.1%. The LLPs consider that a full apportionment is required involving both the determination of the market value of all elements of the package of rights that they acquired and an allocation of the Price to all constituent elements of that package. HMRC argue that no such full apportionment is required since, in the circumstances of this case, the amount of the Price that was paid “for” the GC Rights, the RSAs and the ESAs can only be the market value of those rights taking into account the Data Centres’ situation in an EZ. If a full apportionment is needed, the parties do not agree on how the CRSAs should be valued for the purposes of that apportionment. The LLPs, in reliance on Mr Williams’s evidence, argue for an “income approach” that treats the CRSAs as a complicated financial instrument and seeks to value the cash flows associated with that financial instrument. HMRC, in reliance on Mr Mackie’s evidence, argue primarily for a “residual value” approach to valuing the CRSAs or, in the alternative, an income-based approach to valuation that is very different from that applied by Mr Williams. If a full apportionment is needed the parties do not agree on the elements of the Price to which an uplift must be applied reflecting the Data Centres’ location in an EZ. HMRC contend that all elements are uplifted, whereas the LLPs apply the uplift only to some. The appropriate starting point for valuing the GC Rights The LLPs contend that the correct starting point is the first, and highest, of Mr Watson’s valuations (£76,360,000 for DC2, which assumed the Data Centres were completed, fully fitted out and let on long term leases at a rent of £170 psf uplifted in accordance with RPI). Mr Smith contended that this is the correct starting point because the LLPs were acquiring the Data Centres to be constructed with the benefit of rental support and the expectation that the Developer would secure occupational tenants either during or after the construction period. He said that investors would have considered they were purchasing an investment property, and would have taken the same approach as if they were purchasing an investment property in a non-EZ location, that is by looking at value based on the income to be derived from an occupational letting of the property as if completed. This approach was “better aligned with what would likely have been anticipated to be the future outcome that sometime after completion, the Data Centres would be let and fully occupied.” In contrast, the other bases on which Mr Watson valued the Data Centres ignored the wider arrangements and represented the development at an interim stage, not representative of the investment the investors would have considered they were acquiring. HMRC contend that the correct starting point is the third, and lowest, of Mr Watson’s valuations (£48.335 million for DC2) which assumed that the Data Centres were completed, fully fitted out and available with vacant possession. As a preliminary point, we note Mr Watson’s opinion that neither his first nor second basis of valuation reflected what the LLPs were acquiring. They were acquiring the right to have Data Centres built and fully fitted out, together with associated rental support arrangements. It was his opinion, therefore, that his assessment on a vacant possession basis plus the value of the rental support provided “a reasonable opinion of the value of the Subject Properties, ignoring EZAs at the date of investment.” (It is important to note that Mr Watson’s valuation of what he called “rental support” was in fact the present value as at April 2011 of the right to receive the Yearly Sums, which encompassed what we characterised in the Decision as RSAs and ESAs). We accepted Mr Watson’s conclusions as to valuation in the Decision, and neither party is now permitted to go behind that conclusion. This does not, however, preclude the LLPs from contending that the “starting point” for considering the value of the GC Rights taking into account the benefit of EZAs is Mr Watson’s highest valuation. We consider, in agreement with HMRC, that the appropriate starting point is Mr Watson’s third basis of valuation and not (as the LLPs and Mr Smith contend) the first basis of valuation because Mr Watson’s third basis is what most accurately reflects the value of the GC Rights as they existed at April 2011. In the first place, Mr Smith’s contention that investors would look to the value of the Data Centres once fully fitted out necessarily involves taking into account the benefit of the RSAs and the ESAs. When the LLPs acquired the GC Rights, the Data Centres were not even built; still less let at a rent of £170 psf. Therefore, it is only with the benefit of the ESAs and RSAs that the assumption that the Data Centres would generate income based on a rent of £170 psf could be made. In fact, Mr Watson’s first basis of valuation (as at April 2011) assumes much more than this, because the Yearly Sum would not result in an income equivalent to fully let Data Centres at a rent of £170 psf for a number of years . The ESAs and RSAs are, however, separate from the GC Rights. They are rights that were bargained for in addition to the GC Rights. The Yearly Sum, for example, was expressly stated in the Services Agreement to be in consideration for, among other things, the Price under the SDA (see the Decision at [230]). While we have concluded in the Decision that for the purposes of s296 payment for the Yearly Sum is to be regarded as payment “for” the relevant interest (either as a matter of statutory construction or as the result of our conclusion on the judicial review claim), the Yearly Sum (comprising the RSAs and the ESAs) is nevertheless separate from the GC Rights. We do not think, in assessing the market value of the GC Rights, it is relevant to have regard to the outcome the investors either hoped for or anticipated with the benefit of the additional financial support arrangements for which they were also paying the Developer within the Price. Second, Mr Smith’s evidence that an investor would value the GC Rights by reference to the income to be derived from an occupational letting of the property does not of itself explain why Mr Watson’s highest basis of valuation should be used. All of Mr Watson’s valuations, including the lowest “vacant possession” valuation, were arrived at by reference to the income that could be derived from occupational lettings. The “vacant possession” basis simply discounted the assumed rental income to take into account the fact that it would only be received in the future. Third, it would be wrong to adopt as the starting point a valuation of the GC Rights that includes the benefit of the Yearly Sum, because that would already take into account at least part of the benefit of the EZAs (the very effect of which Mr Smith was tasked with valuing). In his first report, Mr Smith noted that the additional benefits available through EZ status would typically be shared by the various parties involved, including the developer (who would have an expectation of higher profit), the investor (who would seek a greater return due to the higher risk of the location) and the occupier (who would seek a larger incentive). In an internal HMRC memo prepared by Mr David Cooper in July 1998 (on which the LLPs placed reliance in the first hearing) it was noted that HMRC accepted that devices (which inflated the acquisition price) such as developer’s rent were simply part of the package of incentives needed to get an investor to acquire property in EZs. They are possible in an EZ because of the availability of EZAs. Partly for this reason, we reject the submission made by Ms Shaw QC in oral argument that, while accepting that Mr Watson’s “fully let and income producing” valuation was possible only because of the presence of RSAs and ESAs, the LLPs’ approach took full account of that factor since it did not, in its determination of the market value of the relevant interest, ascribe any additional value to RSAs or ESAs. The LLPs’ approach involves double counting as it applies an uplift (to take account of EZAs) to something (the RSAs and ESAs) which themselves reflect part of the benefit of the availability of EZAs. Mr Smith considers that investors would not look to apportion the calculation of the price between differing elements or components, but would consider the investment as a whole. That may well be correct as a statement of how investors would appraise the whole investment opportunity being offered. However, we have already determined in the Decision that as a matter of objective fact the investment opportunity involved the LLPs acquiring a package of distinct elements such that the whole of the Price was not paid for the relevant interest so that it is necessary to decide how much of the Price was paid for different elements or components. The fact that investors would not have regarded the financial support arrangements as a separate asset to the GC Rights does not, therefore, assist in the determination of the objective market value of those GC Rights. Accordingly, we consider - in agreement with Mr Mackie – that while investors may well have considered what the value of the Data Centres would be once they were fully let, that additional value was not part of the GC Rights as at the date they were acquired in April 2011. The adjustment necessary to take into account the effect of EZAs (which we will address in the next section) therefore needs to be applied by reference to Mr Watson’s “vacant possession” basis of valuation which produces a value of £48,335,000 in the case of DC2 and £36,835,000 in the case of DC3. Adjustment to take account of EZAs It is common ground that some adjustment is necessary to take account of the fact that the Data Centres were located in an EZ so that a purchaser could potentially acquire the benefit of EZAs by reason of s296. Mr Smith and Mr Mackie agree that this is because the seller of a building within an EZ will be aware of the value of the EZAs generated by the sale and would be in a position to bargain for a share of that value. In their joint statement, the two experts agree that the value of the tax benefit should be shared between the investors and the developer, as well as other parties involved in the wider development including the landowner and occupiers. Mr Smith’s evidence is that if, in 2011, he had been asked by a prospective purchaser what price he would have expected as a result of the developments being located in an EZ, he would have considered an increase of two-thirds above the non-EZ value of the asset to be appropriate. Mr Smith considers that a two-thirds uplift represents a fair split between the investor and developer taking into account the various factors affecting, respectively, the developer and the investor: On the investor’s side, the demand, risk and tax rate; and On the developer’s side, limited supply, costs, expertise, developer guarantees of the sort provided in this case and time and effort expended in respect of the development. His view is that, taking into account those factors, an investor would be looking for a “value advantage” of 20%. This is explained by the following simple example: The property has a value, ignoring EZ benefits of £120 (the “non-tax value”); A two-thirds uplift (£80) would result in a price of £200 (the “pre-tax price”); Assuming that the investor has a marginal tax rate of 50%, and assuming that the whole of the pre-tax price qualifies for EZAs, then the investor will have incurred a net cash cost of £100 (the “post-tax price”); Thus for a net cash cost of £100 the investor has acquired an asset with a non-tax value of £120, being a “value advantage” of 20%. On Mr Smith’s approach, the value of the EZAs is split 80% for the developer and 20% for the investor. He supported this conclusion by reference to his experience in advising on transactions in EZs. Significantly for what follows, he accepted that in most cases the transactions on which he had advised involved the purchase of the building being funded, in part, by way of debt and the provision of financing arrangements involving security (e.g. cash collateralisation) and rental support arrangements similar to those present in this case. Moreover, Mr Smith’s calculations proceeded on the basis that the EZAs should be valued by assuming a 50% marginal tax rate since that was the tax rate applicable to most individual investors who chose to invest in EZ opportunities. Mr Mackie favours a 50:50 split of the benefit of the EZAs. This assumes equality of negotiating strength, which according to Mr Mackie is an assumption commonly used when assessing a price between hypothetical market participants. He applies this split, however, to the value of the EZAs after discounting for various factors. These include: The litigation risk that the EZAs may not be allowed, which he identifies as 20%; Liquidity risk, because of the requirement to hold the asset for a period of at least seven years, which he estimates at 5%; An adjustment for the discounted value of lost plant and machinery allowances, which he values at £4.4m. Mr Mackie considers that, if the effect of EZAs is valued using a “market value” approach, a hypothetical acquirer of the GC Rights should be assumed to be a company that was subject to the 26% rate of corporation tax that prevailed at the time. He justifies that approach by his observation that the overwhelming majority of commercial real estate in the UK is held by companies. On this basis Mr Mackie, applying a “market value” approach to EZAs, arrives at a value for DC2 taking into account the availability of EZAs of £50.1 million. As we have noted above, however, HMRC’s case is that the value of DC2 taking into account the availability of EZAs was £56.1 million. This is based on Mr Mackie’s alternative calculation, based on “equitable value”, which assumes an investor with a marginal tax rate of 50%. While, in light of this, it is strictly unnecessary to reach a conclusion on the point, we agree with Mr Smith that the valuation should assume that investors had a tax rate of 50%. It may be true that most commercial real estate in the UK is held by companies who are subject to rates of corporation tax that are lower than rates of income tax, but Mr Mackie accepted in cross examination that the typical market participant in an EZ investment was an LLP or syndicate of higher-rate taxpayers. In considering how any adjustment is to be applied, there is conceptually little to choose between making an adjustment to the proportion of the EZA benefits which the investor would be prepared to share with the seller, or making a deduction from the value of those benefits before applying a 50:50 split. For the sake of simplicity we adopt the second approach. It follows that the ascertainment of the market value of the GC Rights with the benefit of the EZAs involves the following three steps: First, it is necessary to determine the correct approach to valuing the benefit that is available to be split. That will involve determining whether the value of the EZAs should be discounted to reflect matters such as uncertainty as to their availability or countervailing matters, such as plant and machinery allowances. Second, it is necessary to determine, in percentage terms, the appropriate split of those benefits as between buyer and seller. Third, it is necessary to determine an increase to Mr Watson’s “vacant possession” basis of valuation that results in the available benefit set out at (1) being shared in the percentages set out at (2). The total value of EZAs available to be shared As regards factors which Mr Mackie considered to reduce the value of EZAs set out at [37] above: It is common sense that a hypothetical purchaser of an asset in an EZ would recognise that, if the chosen means of sharing EZAs involved adding an amount to the up-front purchase price, the purchaser alone (to the exclusion of the seller) would take all of the risk that the EZAs would not be allowed. We accept that a hypothetical purchaser of the GC Rights in 2011 would have been aware that there was at least some risk that EZAs would not be allowed. The investors in the LLPs were undoubtedly taking the risk that, because the LLPs were acquiring a package of assets in addition to the relevant interest, the whole of the Price would not be characterised as qualifying expenditure. That, however, is irrelevant to the present question, because a purchase of the GC Rights at market value would not have presented that risk. Nevertheless, a hypothetical purchaser of the GC Rights at market value in 2011 would still be taking some risk that EZAs would not be available given the timing of these transactions. The EZ at the Site came to an end in 2006. A purchaser of the GC Rights could only successfully claim EZAs if the Developer had incurred expenditure on construction “under” a Golden Contract. There was room for doubt as to whether this requirement was met. That said, we accept the LLPs’ arguments that Mr Mackie did not have the expertise to suggest that a purchaser would discount the value of the EZAs by a factor of 20% to reflect this risk. Mr Mackie also acknowledged that investors would not typically discount EZAs to reflect perceived risk about their availability, but noted that this particular investment was different from the norm because the EZ had expired by the time the investment was made. On balance we conclude that a market investor would make some adjustment to factor in the risk of EZAs not being available. There is no scientific basis for calculating the amount of such deduction. It is not a question upon which the evidence of the experts can assist. It requires us to reach an evaluative judgment as to the view investors were likely to take at the time as to the quantum of the legal risk arising from the timing of the transaction. Litigation risk is notoriously difficult to quantify but we consider in all the circumstances that investors would have regarded the risk as relatively low such that an appropriate discount is 10%. As to the suggestion that a hypothetical investor could have claimed plant and machinery allowances on acquisition of a building outside an EZ, so that the Data Centres’ location in an EZ conferred only an incremental benefit, the LLPs contended that it was impermissible to compare the value of the EZAs with whatever benefit the investors might have achieved under some other transaction. We see the force of HMRC’s point that if a purchaser of GC Rights would in principle have been entitled to plant and machinery allowances (had there been no EZAs available), the incremental benefit that such a purchaser would obtain from the Data Centres being located in an EZ consists not of 100% of the EZAs, but the difference between the value of the EZAs and the value of plant and machinery allowances that could have been obtained even if the Data Centres were located outside a EZ. Mr Smith’s opinion was that an investor in an EZ would not typically make any adjustment in relation to capital allowances forgone. His experience was limited however, as we have already noted, to transactions involving additional financial arrangements similar to those in this case, where investors were likely to have focused on the overall return (in particular, as in this case, the size of the tax benefit received upon implementation of the transaction) rather than on arriving at a price for the relevant building. Mr Mackie’s experience, on the other hand, was limited to advising corporate buyers who were considering the option of buying an asset within or outside an EZ. For such buyers, the differential in the available allowances would be a factor in their decision. On balance, we conclude that since it was not Mr Mackie’s evidence that the differential between capital allowances and EZAs was ever taken into account by investors seeking to acquire assets in an EZ (who would not be considering an alternative acquisition outside an EZ) the adjustment he suggests is not one we should make in this case. As to the third of Mr Mackie’s adjustments, we do not accept that the hypothetical purchaser would have discounted the value of the GC Rights by reference to the fact that they were required to hold the asset for at least seven years. Mr Mackie accepted that a period of seven years was not untypical for holding a property intended as an investment. We conclude, therefore, that a hypothetical investor, appreciating that the EZAs were uncertain given the expiry of the EZ and that a developer receiving an enhanced up-front purchase price for GC Rights would not be sharing in that risk, would have been prepared to treat only 90% of the value of EZAs as available to be shared. The proportions in which EZAs would be shared We do not accept Mr Smith’s conclusion that a hypothetical purchaser of the GC Rights would be prepared to pay 66.67% more for those rights so as to retain the benefit of just 20% of those EZAs. First, as we have already noted, we consider that Mr Smith’s conclusion suffered from the limitation that it was based on his experience of other transactions that, almost without exception, involved a package of bank finance, rental support and rental guarantees. Therefore, while Mr Smith has experience of purchasers paying a total price that was 66.67% above what would have been the value of a building had it been located outside an EZ, that provides relatively weak evidence as to the market value of the building since, given the presence of rental support, bank finance and rental guarantees, it is quite possible that those other purchasers were, like the LLPs, acquiring other assets, in addition to the building, that had a significant market value. Second, we see no reason why a purchaser would, if it acquired the GC Rights separate from the other package of assets, be prepared to agree to pricing that resulted in it retaining only 20% of the EZAs. Such pricing would make no sense since, if an investor purchased only the GC Rights from the Developer, the Developer would be doing nothing (apart from selling the GC Rights) to secure EZAs for the purchaser and would be taking no risk in the availability of those allowances. A benefit split of 80:20 in favour of the Developer would, in our judgment only make sense if the Developer was providing something other than the GC Rights. As discussed in the next section, the extra benefit that the Developer provided in the context of this transaction included the RSAs, ESAs and CRSAs which, between them, meant that the LLPs were insulated from much of the economic risk associated with their acquisition of the Data Centres and their borrowings under the Bank Winter Loans which meant that they could pay a price for their relevant interests that was significantly in excess of market value. We prefer, therefore, Mr Mackie’s assessment, that the benefit would be split 50:50, recognising that this is based on an assumption (commonly used in assessing market value on the basis of what would be agreed between hypothetical market participants) of equal negotiating strength. The LLPs objected that the Developer would have the greater bargaining strength, given the timing of the transaction, as it was investors’ last chance to take advantage of EZAs. We think that this point cuts both ways, however, because it was also the last chance for the Developer to enter into any sale of the rights it already held in a way which enabled it to share in the benefit of EZAs. We also do not accept the LLPs’ objection that Mr Mackie was wrong to rely on the fact that investors would, acting rationally, be disincentivised from paying any more than the market value (without the benefit of EZAs) because for every additional £1 they spent, they would only receive a tax benefit of 50p. Mr Mackie was not suggesting that an investor would, on this basis, be irrational to pay any more than the market value without the benefit of the EZAs. Given it was common ground that the benefit of the EZAs would be shared to some extent between a hypothetical buyer and seller, it would be necessary to pay more than the market value without the benefit of EZAs in order to secure the acquisition. He was merely pointing out that (in the absence of financing arrangements of the kind that exist in this case) there is no net benefit to an investor of paying more than that amount, so that it would be irrational for an investor to do so save to the extent that it was necessary in order to secure the acquisition . The LLPs also objected that it was wrong, in identifying the extent to which a purchaser would be prepared to share the EZAs with the seller, to calculate the benefit to the seller by reference to the difference between the market value (with the benefit of EZAs) and the market value (without that benefit). That was because the Developer in this case was required to pay more to the Contractor, to have the Data Centres built and fitted out, than their market value once completed. We reject this objection on the basis that the relevant task is to identify market value by reference to the amount which a hypothetical buyer would pay over and above the value of the asset in its hands once purchased by reason of the availability of EZAs. On the basis of our findings in the Decision, it so happens that the value of (for example) DC2 in the hands of CDC2 (£48,335,000) was less than the amount the Developer was obliged to pay the Contractor to build it and fit it out (£54,845,150 – see [38] of the Decision). If the transaction the Developer intended to enter into was a plain sale of the GC Rights then that would simply represent a piece of poor business by the Developer, which would be irrelevant to the market value, assessed objectively, of the asset being sold. In any event, in this case the Developer’s intention was to enter into a much more complex transaction, at a much higher price than one based on a sale of the GC Rights at market value. For that reason, too, the fact that the Developer was obliged to pay more to the Contractor than the market value of the Data Centres without the benefit of the EZAs is irrelevant. Finally, the LLPs also objected that Mr Mackie’s approach was to enhance the value of the relevant interest by the smallest amount necessary to equalise the parties’ position. While Mr Mackie did appear to accept that characterisation of his use of an algorithm to arrive at the uplift in value, it did not make any sense given that the approach he adopted was one which was intended to identify a valuation which encompassed a 50:50 split of the value of the EZAs. In any event, we have not adopted an approach which is intended to arrive at the “smallest amount”, but have sought to calculate (in the next section) the single price which would result in a 50:50 split of the relevant benefits. We therefore consider that, for the purposes of determining the market value of the GC Rights, viewed separately from other constituents of the package that the LLPs purchased, a 50:50 split of the benefit of the EZAs is appropriate. The calculation of the value of the GC Rights The third step of those identified in [40] above is purely arithmetic (having established that Mr Watson’s third valuation is the correct starting point for the value of the GC Rights without the benefit of EZAs). It is complicated because the value of the EZAs is a function of the price paid, but it is possible to do so by means of a formula. What is required is to identify that increase, x, to be paid by way of addition to the market value of the GC Rights of £48.335m which, in the case of DC2, satisfies the following equation : x=0.5 x 0.9 x 0.5 (£48.335m + x) The solution of that equation is that x = £14,032,742 yielding a total purchase price of £62,367,742 for DC2. Solving a similar equation for DC3 yields a total purchase price of £47,529,032 for DC3. The accuracy of the solution for DC2 can be demonstrated as follows: £62,367,742 (our calculation of the purchase price that would achieve a 50:50 benefit split) reflects a premium of £14,032,742 over the vacant possession value of £48,335,000; The EZAs available to a purchaser at this price (using an assumed tax rate of 50%) would be £31,183,871; 90% (being the amount of the EZAs which is to be shared equally) is therefore £28,065,484; 50% that sum is £14,032,742. That is equal to the uplift in the price paid over the vacant possession value and thus results in 50% of 90% of the EZAs being shared with the seller. Another way of looking at this calculation is that, using the figures for DC2, it results in the Developer obtaining a premium for sale of the GC Rights of £14,032,742 which represents 45% of the total EZAs that would be available of £31,183,871 – in other words a benefit split of 45:55 in favour of the LLPs. We regard that as an appropriate sharing of the benefit given that the LLPs were taking all of the risk as to the availability of EZAs. Approach to identifying how much of the Price was paid for the relevant interest In light of the conclusions reached above, the value of the elements comprising the package of assets and rights for which the Price was paid in relation to DC2 can be updated as follows: The GC Rights: £62,367,742 The combined value of RSAs and ESAs: £10.89 million. The logic of HMRC’s submissions would appear to be that only the aggregate of these two figures would attract EZAs. However, in their skeleton argument, HMRC adopted an approach that was more beneficial to the LLPs. They accepted that the combined value of the RSAs and the ESAs should themselves be increased to reflect the availability of EZAs by a factor equal to that used to uplift the value of the GC Rights. While this involves the same double-counting we have identified at [27], the case the LLPs had to meet was that advanced by HMRC and it would not be fair at this stage to take a different approach. Accordingly, applying the same uplift to that which we have concluded applies to the market value of the GC Rights, the amount paid for the relevant interest consists, in the case of DC2, of £76,415,842 being the aggregate of: £62,367,742 (the market value of the GC Rights reflecting the value of EZAs, representing a 29% increase to the market value of those rights ignoring EZAs); plus £14,048,100 (Mr Watson’s valuation of the Yearly Sums also increased by factor of 29%). HMRC contend that it is this amount that was paid "for the relevant interest” on the basis that the rest of the Price was referrable to the Arranger’s Fee and the CRSAs. The LLPs contend, on the other hand, that it is necessary to identify a market value for the CRSAs because it is only then that a true apportionment of the Price (insofar as it might differ from the aggregate amount of each of the separate elements) can be undertaken, by rateably adjusting the amount of each element. Ms Shaw QC submitted that this is consistent with the approach taken by Vinelott J in Bostock v Totham [1997] STC 764. That case concerned the apportionment of the price paid as between (1) expenditure on the construction of a building within an EZ (which attracted capital allowances) and (2) the value of the land itself (which did not attract allowances). Section 21(3) of the Capital Allowances Act 1990 provided that the sum paid on the sale of the relevant interest was “…deemed to be reduced by an amount equal to so much thereof as, on a just apportionment, is attributable to assets representing expenditure other than expenditure in respect of which an allowance can be made under this Part.” The Special Commissioners found that the land value was £200,000 and that the construction costs of the building were £333,174. The aggregate of those figures was £533,174. The consideration paid on the sale was, however, £1,131,735 (a difference of £598,561). The taxpayer contended that a “just apportionment” required the land value to be deducted from the sale price, leaving the surplus value, or developer’s profit, to be attributed to the cost of construction of the building. Vinelott J rejected that approach, on the basis that it seemed to involve “no apportionment at all”. Instead, he ordered that the surplus be attributed rateably between the land value and the cost of construction. In our judgment, for the following reasons, we consider that HMRC’s approach outlined at [58] above is to be preferred. As a preliminary point, we left open in the Decision the nature of the apportionment exercise to be carried out. At [290] of the Decision, we concluded that (1) the whole of the Price was not paid for the relevant interests and (2) “the Price is to be apportioned between” the relevant interest, the ESAs, the CRSAs and the Arranger’s Fee. We had not, at that stage, received any submissions as to how the portion of the Price that was paid for the relevant interest was to be ascertained. Bostock had not been cited to us. At [291], we specifically invited submissions on, among other things, how, as a matter of law, the Price should be apportioned between the relevant interest and the other rights and benefits. The exercise in this case is different from that required to be undertaken in Bostock. The relevant statutory provision in that case expressly required a “just apportionment” between the separate elements of the asset (land together with the building on it) that was acquired. Here, in contrast, the task is to identify the amount that was paid for the relevant interest. While an apportionment in the Bostock sense might be instructive in some cases in determining the amount that was paid “for” the relevant interest, whether it does so in a particular case depends upon the nature of the transaction and the relationship between the component elements provided in consideration for the purchase price. The nature of the transaction and the relationship between the component elements in this case make a Bostock apportionment inappropriate. First, while we concluded in the Decision that for the purposes of s.296 CRSAs constituted a separate part of the consideration for which the Price was paid, they are not a separate and distinct asset in the same sense that the land and buildings in Bostock were separate assets. That is because the fundamental benefit the CRSAs provided to the LLPs was the ability to pay a higher purchase price for the GC Rights (at least three times higher than their market value in the hands of the LLPs) and thus claim a substantially higher amount of EZAs. Of the Price, CDC2 raised £46.1 million from investors and the remainder (£107.6 million) was financed by the Bank Winter Loan. The only “commercially practical way” of securing the financing from Bank Winter was by cash-collateralising it. Integral to the cash-collateralisation of the Bank Winter Loan were the terms relating to the Developer Loan and its subsequent conversion into equity, i.e. the substantial elements of the CRSAs. In simple terms (as explained to investors in the IMs) for every £30 contributed by the investors they would receive initial tax relief of £50. In relation to DC2 the additional EZAs potentially amounted to £53.5 million (being one-half of the additional £107 million that was included in the Price as a result of the Bank Winter Loan). Put another way, the value of the EZAs to be obtained was a function of the Price: the greater the Price, the greater the amount of the EZAs. In those circumstances we consider that it is simply inapposite to regard the additional rights and benefits provided by the Developer which enabled the Price to be increased as a distinct asset which must be valued as part of an apportionment in the Bostock sense. In our judgment, it is appropriate to view the amount of the Price paid (in relation to DC2) that exceeded £76,415,842 as being paid “for” those additional rights and benefits. This reasoning applies equally to the CRSAs and the amount paid to discharge the Arranger’s Fee, although in the latter case there is no valuation exercise necessary because it was in a fixed sum. Second, as we explained in the Decision ([266] to [267]), while in many cases there is no necessary correlation between the market value of an asset and the amount paid for it (because the price is the product of negotiation between the two particular parties), in this case there was no negotiation in respect of the Price. The LLPs made assumptions as to the market value of the Data Centres and relied on the other benefits negotiated under the contract (being the ESAs, the RSAs and the CRSAs) to ensure that value flowed back to them from the Developer to the extent that the Price turned out to be an overpayment. It is more consistent with this approach to value the relevant interest by reference to its market value at the time of the transaction, with the remainder of the Price (save for that relating to the Arranger’s Fee) being allocated to the various aspects of the transaction that insulated the LLPs from loss. Third, the calculation of the market value of the GC Rights already factors in the profit element the Developer would expect to receive from a sale of an asset located in an EZ. A simple sale of GC Rights would involve no particular complexity, and thus no particularly onerous work for the Developer, nor any significant risk for it. In the actual transaction, the Developer received a Price that was considerably in excess of the market value of GC Rights even taking into account the effect of EZAs. The clear inference is that this “super-premium” is referable to elements other than the sale of the GC Rights, including in particular the CRSAs. To the extent, therefore, that the residue of the Price (after deducting the GC Rights, the Yearly Sum and the Arranger’s Fee) contains an element of developer profit, there is a rational basis for apportioning that element of profit to the CRSAs. We therefore reject the LLPs’ case that an apportionment of the kind summarised at [59] is necessary. Rather, in our judgment, HMRC are obliged to treat the following amounts as paid “for” the relevant interest (taking into account our conclusions in relation to the judicial review) and allow the LLPs’ claims for EZAs accordingly: £76,415,842 in the case of CDC2. £57,668,432 in the case of CDC3. We have considered carefully whether, in the light of this conclusion, we should nevertheless go on to make findings as to the market value of the CRSAs on the basis that they are a separate asset. We have concluded, however, that seeking to arrive at a value for the CRSAs viewed in isolation is ultimately of little assistance in determining how much the LLPs paid for the GC Rights, the RSAs and the ESAs. For the reasons we have already given in concluding that an apportionment in the Bostock sense is inappropriate, the enhanced EZAs acquired as a result of paying a substantially increased price themselves produced a significant benefit which the parties sought to share between them. An aspect of that benefit-sharing was the Developer’s provision of CRSAs. The CRSAs cannot, therefore, sensibly be viewed in isolation from the entire package of rights that the LLPs acquired with the result that any attempt to value them as a standalone asset provides little insight as to the extent of the Price that was paid for the relevant interest. We have, therefore, concluded that it is unnecessary for us to consider the expert evidence of Mr Williams and Mr Mackie as to what value might be ascribed to the CRSAs on the basis that they were a separate asset so that an apportionment in the Bostock sense could be carried out. Nor do we consider, given the points that we have made above, that a consideration of the market value of the CRSAs viewed in isolation would provide any form of valuable “sense check” of the conclusions we have reached. Disposition The Decision already contains our final conclusions relating to the LLPs’ claims for judicial review and no further amplification of those conclusions is needed. The LLPs’ statutory appeals against HMRC’s closure notices are allowed in part with the LLPs each being entitled to EZAs on the amounts set out at [70] above.JUDGE JONATHAN RICHARDS RELEASE DATE: 14 December 2020
Cited in 1 later judgment