Leasehold Reform, Housing and Urban Development Act 1993 – Lease extension – Statutory valuation assumption – Valuation of existing leases of flats without benefit of rights under 1993 Act – Use of graphs of relativity – Whether “Parthenia model” producing accurate results – Alternative approaches to valuation – Issues determined accordingly
The three joined cases concerned applications, under section 48 of the Leasehold Reform, Housing and Urban Development Act 1993, for a determination of disputed matters in relation to claims by the long leaseholders of three flats (flat 3, flat 11 and flat 5) to acquire an extended lease pursuant to Chapter II of Part I of the Act. The application was made by the landlord in the first two cases and by the leaseholder in the third. The matter in dispute was the premium payable by each leaseholder for the extended lease, pursuant to section 56 of, and Schedule 13 to, the 1993 Act. Each of the flats was located in south-west London and the existing leases had unexpired terms of 23, 37.71 and 41.32 years respectively.
The first-tier tribunal (FTT) referred the applications to the Upper Tribunal to determine an issue as to the appropriateness of using a particular model (the Parthenia model) for the purpose of determining the value of the existing lease of each flat, on the statutory assumption that Chapter I and Chapter II of Part I of the 1993 Act conferred no right to acquire any interest in any premises containing the lessee’s flat or to acquire any new lease.
Such a valuation was acknowledged to be problematic because leases currently sold in the open market in the real world did have the benefit of rights under the 1993 Act, so that there was little or no available market evidence, for comparison purposes, involving the sale of existing leases without such rights. The promoters of the Parthenia model had collected market data for leases in an earlier period, from 1987 to 1991, and, applying a statistical method called “hedonic regression” to that data, claimed to have isolated the effect on value of the single variable of lease length. They had sought in that way to calculate the value of each individual property comprised in their data relative to the freehold vacant possession (FHVP) value, by representing it as a percentage (relativity) of the FHVP value. Their findings were set out in a graph of relativity where the X axis was the length of the lease and the Y axis was the relativity percentage. The Parthenia model tended to produce a higher relativity for a lease of a particular length when compared with the result of using other possible methods of determining relativity.
The parties were agreed that it was appropriate to use another graph, the Savills 2002 enfranchisable graph, when considering the value, relative to FHVP, of real world comparables consisting of leases that did have rights under the 1993 Act (real world relativity).
Held: The issues were determined accordingly.
(1) For the purposes of valuing the existing leases, the 1993 Act required the assumption that a hypothetical asset, namely a lease without rights under the Act, was available for sale in the real market. It was clear that the existing lease without rights under the Act was a less valuable interest and would not command a value higher than the value of the existing lease with rights under the Act. In that way, the value of the existing lease with rights under the Act provided a ceiling for the value of the existing lease without rights under the Act.
That consideration ruled out the use of the Parthenia model, since, in the case of flat 5, the value of £2,232,450 indicated by that graph for the existing lease without rights under the 1993 Act was higher than the agreed value of £2m with those rights. The result that it produced was impossible. It followed that the Parthenia model could not be regarded as a reliable tool and should be rejected. If the Parthenia model did not work for flat 5, then it could not be used in any other case to determine the value of an existing lease without rights under the Act. Even if the application of the model produced an answer in another case that was not impossible, that did not mean that the model had been shown to be reliable.
Hedonic modelling was characterised by the need for the researcher to make choices, such as the data set to use, the functional form of the model and the choice of hedonic attributes, which might be constrained by the limitations of the data set, to explain the variable of interest. While the use of a substantial number of actual transactions that were unaffected by the 1993 Act was a potentially attractive feature of the Parthenia model, that model could not fairly be described as an objective analysis of those transactions given the significant exercise of subjective judgment that the model required. The Parthenia model produced results that were incompatible with the market evidence and therefore did not pass the test of producing results that were observed in practice. That might be due to technically deficiencies in the model and also to economic and market changes since 1987-1991, which had seen a general lowering of relativities over time.
(2) In the case of flat 5, the value without rights under the 1993 Act was best arrived at by considering comparable properties that did have such rights and then making an appropriate deduction, taking into account previous Upper Tribunal decisions on the same issue. It was common for a valuer to be asked to determine the amount of an uplift or a deduction to reflect a particular feature of a property. In some cases, a valuer would have a great deal to guide him as to the appropriate adjustment to make. In other cases, he might have less to go on, but it was not impermissible in such cases for an experienced valuer to make a valuation judgment. For a flat with an unexpired term of 41.32 years, the appropriate deduction was 10%; that produced a value of £1.8m on the assumption that the lease did not have rights under the Act: Trustees of the Eyre Estate v Saphir [1999] 2 EGLR 123, Chelsea Properties Ltd v Earl Cadogan [2007] PLSCS 197, Nailrile Ltd v Earl Cadogan [2009] RVR 95; [2009] 2 EGLR 151, Earl Cadogan v Cadogan Square Ltd [2011] UKUT 154 (LC); [2011] 3 EGLR 127 and 82 Portland Place (Freehold) Ltd v Howard de Walden Estates Ltd [2014] UKUT 133 (LC); [2014] PLSCS 275 considered.
In the real world at the valuation date, the value of an existing lease without rights under the Act was often estimated by using the Gerald Eve graph. There was an element of circularity in that approach since it involved making a deduction from a value, for an existing lease with rights under the Act, that itself made an assumption about the matter that was to be determined. Nonetheless, the Gerald Eve graph was a real market circumstance that influenced market behaviour. The Upper Tribunal was required to determine the market value of an asset in a real market that behaved in that way. It was not its function to tell the market how it ought to behave in the future, and it had no power replace real market forces at past valuation dates by some other forces that it thought should have operated instead. The figures produced by the Gerald Eve graph broadly supported the valuation of £1.8m already reached and that valuation should be adopted accordingly.
(3) In the case of flat 11, using the Gerald Eve graph produced a value of just over £370,000 for the existing lease without rights under the 1993 Act. Alternatively, using the Savills 2002 graph for real world relativity, and making the same 10% deduction as for flat 5 to reflect the value without rights under the 1993 Act, produced a value of £385,534.
However, it was not possible to give an unqualified endorsement to the use of either the Gerald Eve graph or the Savills 2002 graph. In particular, there was reason to think the relevant market forces at the valuation dates in 2014 would have been different from the relevant market forces at the times at which the two graphs were prepared. The relativities shown by those two graphs might well be higher than was appropriate for the market that existed at the valuation dates. That suggested that the figures produced using those graphs were too high rather than too low. The tribunal would adopt the figure produced by the Gerald Eve graph for flat 11 because it was the lower of the two; that did not involve an unqualified endorsement of that method but simply represented the least unreliable figure of those available on the evidence in the case.
Finally, a further deduction of £10,424 should be made to reflect the fact that flat 11 had an onerous ground rent; that produced a valuation of £359,731, which was then rounded up to £360,000.
(4) The valuation of flat 3 should be carried out in a similar way to that already discussed in relation to Flat 11. The Gerald Eve graph indicated a valuation of £448,400. Using the Savills 2002 graph, and making a 20% deduction to reflect the shorter unexpired term of 23 years, produced a valuation of £467,400. The tribunal again adopted the figure produced by the Gerald Eve graph as the lower, and least unreliable, of the two figures. After making a deduction of £13,302 to reflect the fact that flat 3 also had an onerous ground rent, the resulting value of the existing lease without rights under the 1993 Act was £435,098, rounded off to £435,000.
(5) Although the Upper Tribunal had not been able to arrive at a method of valuation that would be clear, simple and predictable as to its future application, the following matters night be of use in future cases.
Although it was necessary to focus on how the market actually performed in the past at the relevant valuation date, it was possible that it might perform differently in the future. In future, less weight might be given in the market to a particular graph or a new graph might emerge. If those new developments affected market behaviour, then they would have to be taken into account when assessing market forces. It was also conceivable that decisions of the tribunals might similarly influence valuers and, in turn, influence parties in the market.
In some, or perhaps many, cases in the future, it was likely that there would have been a market transaction around the valuation date in respect of the existing lease with rights under the 1993 Act. If the price paid for that market transaction was a true reflection of market value for that interest, then that market value would be a very useful starting point for determining the value of the existing lease without rights under the 1993 Act. It would normally be possible for an experienced valuer to express an independent opinion as to the amount of the deduction that would be appropriate to reflect the statutory hypothesis that the existing lease did not have rights under the 1993 Act.
The more difficult cases in the future were likely to be those where there was no reliable market transaction concerning the existing lease with rights under the 1993 Act at or near the valuation date. In such cases, valuers would need to consider adopting more than one approach. One possible method was to use the most reliable graph for determining the relative value of an existing lease without rights under the 1993 Act. Another method was to use a graph to determine the relative value of an existing lease with rights under the 1993 Act, and then to make a deduction from that value to reflect the absence of those rights on the statutory hypothesis. If those methods threw up different figures, it would then be for the good sense of the experienced valuer to determine what figure best reflected the strengths and weaknesses of the two methods that had been used.
(6) The authority of the Savills 2002 enfranchisable graph had to some extent been eroded by the emerging Savills 2015 enfranchisable graph. While the 2015 graph was still subject to some possible technical criticisms, that graph was likely to be beneficial if those criticisms could be addressed and removed If a version of that graph emerged, not subject to those technical criticisms and based on transactions rather than opinions, it might be that valuers would adopt that revised graph in place of the Savills 2002 graph. If that were to happen, valuers and tribunals might have more confidence in a method of valuation for an existing lease without rights under the 1993 Act which proceeded by two stages: first, using the new graph to adjust the FHVP value for the property to the value of the existing lease with rights under the 1993 Act; and, second, making a deduction from that value to reflect the absence of rights under the 1993 Act on the statutory hypothesis.
Anthony Radevsky (instructed by Pemberton Greenish LLP) appeared for the applicant landlord in the first and second cases; Stephen Jourdan QC and Julia Petrenko (instructed by CMS Cameron McKenna LLP) appeared for the respondent landlord in the third case; Philip Rainey QC and Cecily Crampin (instructed by Collins Benson Goldhill LLP appeared for the leaseholders in all three cases.
Sally Dobson, barrister
