On Tower Ltd v AP Wireless (II) (New Zealand Farm) [2022] EW Misc 6 (CC) (26 August 2022)

“I agree with Mr Thornton-Kemsley that where a rent is to be determined under section 34 of the 1954 Act the adoption of the structured approach resorted to in Hanover Capital is only necessary where reliable transactional evidence is missing. This case is being decided sixteen months after Hanover Capital and on different evidence. If the evidence of new lettings of bare sites is of sufficient quality and quantity to enable clear conclusions to be drawn, as I believe it is in this case, it is unnecessary to adopt the structured approach previously identified.”

Both experts relied on a comparative valuation based on market evidence of comparable transactions and on a Hanover Capital structured valuation.

Mr Stott’s analysis

On behalf of On Tower, Mr Stott provided a useful review of the different approaches to new lettings taken by the major operators and infrastructure providers in the telecoms sector. I summarise his evidence in relation to each of the four largest participants.

EE typically agrees total rents of between £750 and £1,000 for new sites. These totals are usually the aggregate of a base rent and a “commercial top-up”. In addition it offers premiums which it refers to as early completion incentive payments or “ECIPs”. Mr Stott said that in the last year these payments had become a standard figure of £15,000 offered in almost all cases subject to completion of the transaction by a fixed date (although he was aware of cases where the payments were made despite the transaction being delayed for reasons beyond the control of the site provider). Mr Stott explained that EE’s willingness to make these ECIP payments was related to commercial pressure it was under to deliver new sites for the emergency services network (ESN) programme. This, and the fact that New Zealand Farm was not an ESN site, led Mr Stott to disregard ESN deals as not being representative of a market in which he assumed prospective tenants would not be under similar pressures.

Mr Stott said that MBNL tends to renew agreements at rents of £1,000 to £2,000 which again are expressed as a base rent plus a “commercial top-up”. It too offers premiums which vary from site to site and range from £1,750 to £25,000.

Cornerstone seeks to limit the rents it is prepared to agree on renewals or for new sites to the level it considers properly reflects a valuation undertaken in accordance with the Code. It then offers premiums ranging between £1,450 and £20,000 to secure lettings at those rents.

On Tower was said not to be in the market for new sites. When renewing leases of existing sites it also seeks to base the rents it agrees on its own assessment of consideration under paragraph 24 of the Code, but it authorises its agents to increase this figure to a higher level in order to reach agreement. For renewals at greenfield sites with no alternative use value the average annual rent agreed by On Tower in 2021 was said by Mr Stott to be around £1,750. It seeks to complete lease renewals without making capital payments, although it has been prepared to do so on some occasions. As an alternative to capital payments it has been prepared to agree a reducing level of rent over a short period of years to achieve a less painful transition from old Code to new Code levels.

Mr Stott emphasised that capital payments were not a feature of every transaction and that they varied in size. The range of premiums paid is apparent from all of the evidence; as for their prevalence, Mr Stott suggested in his written evidence that “capital payments are not readily available” notwithstanding that the evidence heard by the Court in EE v Morriss was that they were “almost invariably paid”.

Despite Mr Stott’s disagreement with the evidence and the Court’s conclusions in EE v Morriss his own evidence demonstrated that premiums are almost invariably paid by EE for its ESN sites. They were also paid in eight of the nine Cornerstone transactions he identified (in two of these cases the capital payment was the only payment made) and in five out of nine MBNL transactions. His view was that such payments were necessary to incentivise otherwise unwilling site providers and that they should be disregarded when assuming a new letting between a willing landlord and a willing tenant as section 34 required. In his experience almost all site providers were unwilling to transact at a rent determined in accordance with paragraph 24 and needed to be induced to do so by an additional capital sum. He applied this analysis across the board and did not distinguish between private individuals on the one hand, and local authorities or commercial organisations for whom the provision of improved electronic communications might have been thought to be attractive. One example was provided in the evidence of an agreement for a new letting of a greenfield site in West Yorkshire by a local authority which was intended to replace a site on a redundant building which was being redeveloped to provide a new supermarket. Mr Clark suggested that there was no reason for the operator to incentivise the local authority in that case and that the agreed premium was simply part of the financial package which the parties negotiated. Mr Stott did not agree. He considered there was no reason why operators in the open market would agree to pay a premium or a rent enhanced because no premium was paid, because he assumed that operators would not be under the sort of commercial pressures that were experienced in reality and would not be in any particular hurry to complete the deal.

Mr Stott next identified nine new lettings of bare, greenfield sites completed by Cornerstone since the introduction of the Code. He focused on five of these which had been completed between August 2020 and October 2021. He considered the most recent of these transactions to be the best comparable, the letting of a rural site at Speycoe Lane in Longford, Derbyshire at an annual rent of £425. That letting happened to be at the lowest rent of any of Mr Stott’s comparables, but the tenant was also paid a premium of £6,000 on completion and a disturbance payment of £2,000 while the site was being constructed.

Mr Stott calculated that the average annual rent of his five comparables was £766, and that where capital payments were agreed these reflected an average annual equivalent of £541, providing a combined rent and annualised premium equivalent to £1,307. For the reasons previously mentioned Mr Stott disregarded the capital sums paid at Speycoe Lane and the capital payments which featured in the other transactions and had regard only to the agreed annual rent.

In order to compare these transactions to New Zealand Farm Mr Stott applied a discount of 10% to the rent in each case except Speycoe Lane to reflect what he considered to be the inferior access to the New Zealand Farm site. He applied a further discount of 20% to each rent to reflect what he took to be the lack of a set down area at New Zealand Farm. In the case of the site at Speycoe Lane this 20% adjustment produced a comparable annual rent of £350, which Mr Stott adopted as his base rent for the site. He made no allowance, either positive or negative, for the fact that the tenant was obliged by the agreement at Speycoe Lane to install a new gate and a new road to provide access to the site.

Mr Stott then added an annual sum of £332, in lieu of a one-off payment for professional fees which he would expect to be paid by the tenant on any Code transaction (calculated on the basis of a payment of £3,500 with a yield of 5% over the 15-year term). This produced an annual rent of £682, which Mr Stott rounded up to £700.

Despite his conviction that capital payments were not a proper reflection of market value and ought in principle to be ignored when undertaking a comparative valuation, Mr Stott acknowledged that the evidence suggested that landlords were generally prepared to accept lower annual rents where a premium was paid but required a higher rent in cases where no capital payment or only a modest payment was offered. In fact the evidence he provided about Speycoe Lane gave a slightly different picture. An email from one of the agents who had negotiated the letting explained that in that case the site provider had asked for a higher annual rent but Cornerstone had insisted instead that the premium should be increased. I take that to be an example of a case where the operator was prepared to pay more for a site than its own assessment of a paragraph 24 valuation but wished to preserve a low headline rent figure by increasing the capital payment, even where the site provider would have preferred a lower premium and a higher annual rent. That is typical of what appears, from the evidence in this and in other cases, to be a common approach amongst operators, who seek to restrict the headline rent figures which they are prepared to agree, while nevertheless being prepared to pay significantly more than the headline figure by adding “commercial top-ups”, “early completion incentives”. No explanation for that approach has been provided, but one obvious inference is that it is done so that those negotiating new rents, or giving evidence to courts or tribunals, can maintain that anything above their assessment of a paragraph 24 rent is a concession or an incentive and does not truly reflect the appropriate rent which should be agreed or determined.

Mr Stott suggested that the evidence of comparable transactions involving new lettings was not sufficiently plentiful or detailed to be relied on in isolation. He therefore referred additionally to evidence of Code lease renewals. He acknowledged the limitations of this evidence identified in previous judgments of the Court, and in the end the material he supplied did not feed directly into his valuation. Nevertheless, his experience of lease renewals of rural greenfield sites by MBNL and Cornerstone was that rents were generally in the range £1,000 to £1,400 with capital payments averaging £6,300 for Cornerstone and £8,500 for MBNL. On Tower more often agreed rents at around £1,750 and was less likely to offer an additional capital sum.

Mr Stott also undertook a valuation following the Hanover staged approach. He originally assumed that the site had no alternative use value and arrived at a figure of £632 (including a contribution to professional fees), which he rounded to £650. In his oral evidence he increased this total figure to £1,150 (including contribution to fees) because of the evidence of an email from Mr Ellis confirming his interest in taking a lease of the site at a rent of £1,200 a year in order to get rid of the mast in the grounds of his estate and within view of his house. Mr Stott did not think that £1,200 was a realistic alternative use value and he assumed that, on reflection, Mr Ellis would be prepared to pay only 25% of the figure he had originally suggested.

Averaging his comparable and staged valuations Mr Stott came to the conclusion that the annual rent which would be agreed in the open market between a willing landlord and tenant on the assumptions required by section 34 was £675. He acknowledged that for most sites a rent at that level would be “improbably low”, but he considered that the nature of the landlord’s interest in this case, which was restricted to the demised site itself, justified an assessment which was “significantly lower than the general tone of the market”. That was a surprising conclusion since he had made no allowance for that factor in his comparative valuation, which was based on lettings by landlords who, as at Speycoe Lane, owned other land surrounding the site. The only allowance in his Hanover staged valuation was to omit what would otherwise have been a sum of £100 to account for “burdens” placed on the landlord by the transaction. Had that allowance not been made Mr Stott’s Hanover valuation would have been at £732 and had he applied the same rounding and averaging the final result would have been a figure of £700 or £725, which would still have been significantly lower than the general tone of his own evidence.

When he came to give his oral evidence Mr Stott abandoned his reliance on a comparative valuation. He preferred instead to base his view exclusively on his revised Hanover valuation of £900. That was because, as he explained it, a comparative valuation was only relevant where a site had no alternative use value. Where, as here, there appeared to be a special purchaser, a “site specific analysis” was required.

Cross examined by Mr Clark, Mr Stott did not accept that the New Zealand Farm site would attract competitive bidding if it was offered to the open market. That was because operators do not bid against each other for sites, and an operator would not pay more for a site than its own assessment of the paragraph 24 consideration. In his view the rent determined under section 34 of the 1954 Act should be the same as a rent determined under paragraph 24 of the Code, except that an additional sum would be added to reflect professional fees which would be payable as compensation under the Code but not under section 34.

Finally, Mr Stott considered that since the date for assessing the interim rent post-dated the coming into force of the Code, it should be set at the same level as the rent for the new lease.

Mr Peat’s analysis

Mr Peat relied only on a comparative valuation and although he did not consider that the structured approach was helpful now that more transactional evidence was available he nevertheless carried out a structured valuation of the sort discussed in Hanover.

Mr Peat provided evidence of rents agreed for a number of 1954 Act renewals since the coming into force of the Code. Some of these transactions included premiums and when these were rentalised the average annual rent at four sites renewed in 2022 and 2021 was £4,000. Mr Peat considered that these transactions were an unreliable guide, because the parties were in an existing relationship and the context was therefore unlike the hypothetical circumstances of a new letting which section 34 required to be assumed. He also considered that the levels of rent were “artificially low” and reflected the relative strength of the tenants with huge portfolios of sites and the weakness of small landlords. For that reason he eventually placed little or no weight on them in his final valuation.

Mr Peat did take account of historic transactions pre-dating the Code, which he uprated for inflation using the retail prices index; he considered that these transactions supported an annual rent of £7,000. He acknowledged the comments in EE v Morriss and in decisions of the Upper Tribunal about the utility of historic comparables but took the view that where a site was being offered in the open market (rather than within the constraints of paragraph 24 of the Code) and where it was likely to attract interest from infrastructure providers who were not Code operators (such as Cellnex or AP Wireless), the market would be competitive and historic levels of rent would be a valid point of reference. Site providers would be aware that the value of their site to an operator was reflected in old Code rental levels and would bear those levels in mind when considering what they were prepared to accept. Mr Peat nevertheless considered that his valuation of £7,000 was supported by other evidence even if old Code transactions were ignored.

Mr Peat placed greatest weight on what he referred to as “general market data” (although the data on which this part of his evidence was based was specific to one site and one infrastructure provider and comprised an estimate of the cost of constructing a new mast and the return which could be expected from it by an infrastructure provider who attracted two operators to the site). In what Mr Peat called a properly functioning market he considered that there would be competition for this site between operators and infrastructure providers (including those already present at the site): it was an important location with planning permission in the green belt which was already shared and hence known to be suitable for a number of operators. He considered that the income that an infrastructure provider would expect to make from the mast, or alternatively the saving that an operator would achieve by providing their own mast, would support a rent of £9,900 a year.

Taking account of each of these different approaches, and after undertaking a reality check by reference to the current income from the site (including site share payments), Mr Peat arrived at an annual rent on the comparative basis of £7,000.

Although he did not regard it as a satisfactory approach, Mr Peat also prepared a valuation on the Hanover model. He identified two alternative uses: for storage, and acquisition by the owner of New Zealand Farm to regain control of the whole Estate, thereby enhancing their own security and amenity. He originally incorporated a figure for the loss to the site provider of the opportunity of putting up their own mast. Mr Clark did not rely on that aspect of Mr Peat’s valuation and, accepting that its admissibility turned on a point of law, Mr Peat did not press the point. Omitting the loss of opportunity element from the calculation, the resulting annual figure was £1,400. This figure was underpinned by an indication from Mr Ellis, the owner of New Zealand Farm, that he would be prepared to pay a rent of £1,200 for a lease of the site which would ensure that there would be no mast. Mr McGhee QC questioned why someone would pay such a rent when the site might be the subject of a notice from an operator claiming Code rights but, in agreement with Mr Peat, I do not think that is a legitimate objection where the purpose of the first stage of a Hanover valuation is to identify the value of the site for a use other than for telecommunications.

Mr Peat also thought that an interim rent of £5,750 would be appropriate, but that was on the assumption that the existing “pay-away” (currently £3,700 a year) would continue to apply while the tenancy was continued under the 1954 Act.

Mr Peat was cross examined by Mr McGhee QC about his experience of premiums in the current letting market. He agreed that they were not offered in every case and that they were usually offered to get a landlord interested in doing a deal and to ensure an early conclusion to a transaction. In his experience they were still paid if the site provider refused to agree to an early completion condition and on behalf of his own clients he had never agreed to a conditional premium. Nor could he see why a site provider who was willingly offering his site to the market would be prepared to accept less for it than a site provider who was reluctant to let.

Discussion and conclusion

I begin with Mr Stott’s evidence that the site would be let by a willing landlord in the open market at a rent of £375 plus a contribution towards costs. He acknowledged that a rent at that level would be “improbably low” yet he persisted in it until he finally abandoned his comparative valuation in favour of an enhanced Hanover valuation.

I did not find any of the changing stages of Mr Stott’s analysis persuasive. His original figure of £375 plus costs was so far from being a credible valuation that it called his objectivity into question. The base figure almost represented a return to the derisory rents offered by operators to unrepresented site providers in the earliest days of the Code.

Mr Stott’s original figure (net of professional fees) was significantly lower than any rent agreed in any of the transactions in evidence. It was only half of the bottom end of the range of rents (£750 to £2,500) which he said was typically paid for new sites or renewals. It was not supported by his own comparables and depended on the lowest rent agreed (Speycoe Lane) which was massaged down by 20% because of the allegedly more favourable arrangements for a set down area than at New Zealand Farm. There was no substance in that discount, and no evidence at all that the erection of a mast would be more or less difficult on any of the sites, yet Mr Stott applied the same discount to all of his comparables no matter how much space was available. He made no attempt to reflect the benefit to the site provider of the capital sums and compensation, totalling £8,000, paid in addition to the rent, nor the benefit of the operator’s agreement to resurface the road and install a new gate.

Mr Stott’s insistence that capital payments are irrelevant and must be ignored was dogmatic and insupportable. The evidence establishes beyond doubt that capital payments are very widely available and that there is a relationship between rents and premiums such that if no premium were to be offered the transaction would be likely to be completed at a higher rent. Sometimes, as at Speycoe Lane, it is only the operator’s insistence on a low rent and a higher premium which causes transactions to be structured in the way they are. There is no reason why a site provider who was willing to let a site without a premium would be prepared to do so at a figure which ignored the payments so generally available to others letting more reluctantly. Nor is there any reason why operators, prepared to pay significant capital sums to ensure that transactions proceed smoothly, would refuse to pay comparable amounts for a site offered on the open market which they wished to secure.

Mr Stott initially justified his improbably low valuation by reference to the assumption that the site provider owned no other land and so would not be inconvenienced by the presence of a mast. Had he considered that was a point of valuation significance one would have expected an allowance to have been made to reflect it in his comparative valuation. No such allowance was made, and the suggested justification for Mr Stott’s unrealistic figure was an unconvincing afterthought. The willing lessor identified in section 34 is an abstraction and is not APW. I am far from convinced that it is permissible to rely on a particular characteristic of the actual site provider, rather than of the site, in order to drive down the price which would otherwise be agreed in the market, and no legal submissions were made in support of Mr Stott’s approach. Even if such a discount is permitted by section 34, Mr Stott provided no evidence that it is encountered in reality.

Mr Stott eventually recognized that his comparative valuation and his revised Hanover assessment were irreconcilable. Rather than question his analysis, he then abandoned his comparative valuation altogether, despite it being an evidence-based approach which ought in principle to be preferred, and despite his own initial preference for it.

Mr Stott said that he was aware of 250 new lettings of greenfield sites with no alternative use value since the commencement of the Code. In EE v Morriss I explained why, with the availability of real market evidence, it ought no longer to be necessary when valuing under section 34 to resort to the artificial structured approach resorted to out of necessity in Hanover. A section 34 valuation is an open market valuation and there is no need to adopt the speculative Hanover approach. Mr Stott’s abandonment of his own comparative valuation and his exclusive reliance on an approach which has nothing to do with the open market hypothesis meant that none of his opinion evidence was relevant to the exercise required by section 34.

Mr Peat’s opinion that a rent of £7,000 was the appropriate rent was undermined by his misunderstanding of the statutory valuation hypothesis in two important respects.

First, he wrongly assumed that the notional transaction would take place in a “functional market” and not in the real market for telecommunications sites. Section 34 requires that the subject of the valuation be taken to be offered by the site provider to the open market, but it does not require that other sites should be assumed to be offered in the same way rather than, as actually happens, taken by operators identifying the sites they want and negotiating off market with the site provider.

Secondly, Mr Peat assumed that On Tower and the two operators who are currently on the site would be very keen to return because they would be suffering from the creation of a “not-spot” in their network coverage. That is not a permissible assumption. Section 34(1)(a) requires that the valuation should disregard any effect on rent of the fact that the tenant or its predecessors in title has been in occupation of the holding. Section 34(1)(b) requires that there be disregarded any goodwill attached to the holding by reason of the carrying on of the tenant’s business. These statutory disregards prohibit any allowance in the valuation for the fact that On Tower has previously occupied the site or has conducted a business which has enhanced its rental value.

Both of those misconceptions were important to Mr Peat’s assessment of rent because they contributed to his expectation that there would be significant competition between those who already have an interest in the site if it was offered to the market on a new lease. As Mr McGhee QC conceded in closing argument, there is no reason to assume that the site would be of interest to only one operator, but that is different from the expectation which underpinned Mr Peat’s evidence that there would be particular competition between operators with an established need for a presence in this location.

Despite these difficulties, Mr Peat’s evidence was thoughtful and conscientious although his final figure was significantly too high. His assumption that an informed site provider would be aware of pre-Code rents and would take them to be a measure of the true value of sites to operators was coherent and his treatment of such rents as a point of reference rather than a direct comparable was moderate. His focus on “economic drivers” was less persuasive and a more straightforward analysis of the evidence of comparable transactions would have been more helpful.

The most useful part of Mr Stott’s evidence was his review of how different operators behave in the current market and the rents and other sums they are prepared to pay for new sites. He provided details of deals done by Cornerstone for nine new greenfield sites and by MBNL for nine replacement sites where notices to quit had been given on existing sites in the vicinity. All of these leases were for terms of 10 years.

Six of the Cornerstone transactions post-dated August 2020; four included a capital payment of between £6,000 and £7,500 in addition to the rent. In the other two cases the consideration was paid as a one-off capital sum with only a peppercorn rent being reserved. Converting all sums other than site-specific compensation to an annual equivalent over the 10-year term (at 5%), produces a range from £637 (Mullany Business Park) to £1,971 (Prime Four Ltd - also a business park).

Mr Stott’s evidence of MBNL transactions was all from 2021 and included five greenfield sites, of which three were let on terms which included a capital payment in addition to an annual rent. Converting all sums other than compensation to an annual equivalent in the same way as the Cornerstone transactions produces a range from £750 pa (Focal Community Centre and Slacky Lane) to £2,295 (land at Garnet Lane next to the A64).

Mr Peat provided evidence of three lease renewals agreed for greenfield sites in 2020 and 2021 by Airwave, EE and Arqiva. The equivalent annual payments ranged from £1,300 to £3,750. His evidence of three renewals under the 1954 Act by EE, EE/H3G and Arqiva in 2020 and 2021 included capital payments in each case and ranged on an annual equivalent basis from £2,324 to £5,090. None of these transactions was at arm’s length because they were between parties who were already in a landlord and tenant relationship, but they provide evidence of the levels at which settlements are reached between such parties which would be available to a reasonably well informed and advised site provider.

A properly advised landlord offering a site to the market at large and looking to receive an annual rent which was not supplemented by a premium would expect to better a rent restricted by the artificial assumptions dictated by paragraph 24 of the Code. They would look around at the transactions taking place and would see that Code consideration is routinely enhanced by annual “commercial payments” and capital sums. They would recognise the substantial saving in time and money which an operator would achieve by doing a consensual deal rather than by using the Code. So too would the operators to whom the site would be likely to be attractive.

Based on the range of evidence presented in this case I am satisfied that a negotiation between willing parties in a competitive market would be in the space between £2,000 and £3,000. Because of the site’s proximity to an important arterial road and its relative ease of access it would be towards the top of that range, settling at £2,750. To that would be added the annual equivalent of professional fees. Mr Stott assumed that these would be recouped by the site provider over the full 15-year term at an annual equivalent rate of £332. I do not think a site provider would be willing to agree to that; it is not what happens in practice (fees are invariably reimbursed up front). The inclusion in the new lease of a tenant’s right of termination would increase the risk that later payments would not be received. Annualising a typical contribution to fees of £3,000 over a 10-year period (at 5%) would add £450 to the annual rent, and that is what I would expect willing parties to agree. In aggregate they would be likely to settle on an annual rent of £3,200.

Neither party made submissions on the appropriate level of interim rent but after a draft of this judgment was released to them they agreed that, in this case, the interim rent should simply follow the new rent.

	[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	English and Welsh Courts - Miscellaneous
You are here: BAILII >> Databases >> English and Welsh Courts - Miscellaneous >> On Tower Ltd v AP Wireless (II) (New Zealand Farm) [2022] EW Misc 6 (CC) (26 August 2022) URL: http://www.bailii.org/ew/cases/Misc/2022/6.html Cite as: [2022] EW Misc 6 (CC)

English and Welsh Courts - Miscellaneous