England and Wales Court of Appeal (Civil Division) Decisions

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE UPPER TRIBUNAL (LANDS CHAMBER)
[2016] UKUT 223 (LC)

		Royal Courts of Justice Strand, London, WC2A 2LL
		24/01/2018

Lord Justice Lewison:

1. Part I, Chapter II of the Leasehold Reform, Housing and Urban Development Act 1993 ("the Act") gives lessees of flats holding under long leases the right to acquire a new lease of the flat in question at a peppercorn rent on payment of a premium. This appeal from the Upper Tribunal (Morgan J and Mr Andrew Trott FRICS) ("the UT") concerns how that premium is to be calculated. The decision of the UT is at [2016] UKUT 223 (LC); [2016] L & TR 32.
2. The particular flat with which we are concerned is Flat 3, 36 Elm Park Road London SW3, which was the subject of one of three applications before the UT. One of the other flats in issue was Flat 5, 17 Cranley Gardens which to some extent was treated as the lead case when the UT came to consider the evidence.
3. The right to appeal to this court from the UT is confined to a point of law arising out of their decision: Tribunal Courts and Enforcement Act 2007 s 13 (1). At a renewed hearing before Kitchin LJ, Mr Mundy was given permission to appeal on one ground; and his application for permission to appeal on two other grounds was adjourned for consideration by the full court. Permission to appeal on a fourth ground was refused.
4. I should mention at this point that the test for the grant of permission to appeal from the UT when exercising an original jurisdiction is the ordinary first appeals test. Although the Appeals from the Upper Tribunal to the Court of Appeal Order 2008, prescribing the second appeals test, is apparently unqualified in its application to all appeals from the UT to the Court of Appeal, the statutory power under which the Order was made (section 13 (6) of the Tribunal Courts and Enforcement Act 2007) is limited to appeals from decisions of the UT which were themselves decisions on appeal from the First Tier Tribunal: see Nwankwo v Secretary of State for the Home Department [2018] EWCA Civ 5. There has been some uncertainty in the past about whether the 2008 Order applies to all appeals from the UT, so it is as well to set the record straight.
5. The statutory provisions providing for the calculation of the premium are set out in Part II of Schedule 13 to the Act. The premium consists of three components:

i) The diminution in value of the landlord's interest in the flat consequent on the grant of the new lease;

ii) The landlord's share of marriage value; and

iii) Compensation for any other loss that the landlord will suffer as the result of the grant of the new lease.

6. The third component of the premium played no part in the UT's decision.
7. Under paragraph 3 of Schedule 13 the diminution in value of the landlord's interest is the difference between:

i) The value of the landlord's interest in the flat prior to the grant of the new lease and

ii) The value of his interest in the flat once the new lease is granted.

8. In essence the value of the landlord's interest before the new lease is granted consists of (a) the right to receive the rent and other sums payable under the lease and (b) his right to possession of the flat at the end of the term. Both these rights must be given a capital value. Because the right to possession will not materialise until the end of the term the landlord is not entitled to exploit the vacant possession value of the flat immediately. Thus the present value of that right is determined on the basis that the vacant possession value is deferred. In the case of the valuation of the landlord's interest before the grant of the new lease it is deferred to the expiry date of the existing lease; and in the case of the valuation of the landlord's interest after the grant of the new lease, it is in theory deferred to the expiry date of the new lease (although in practice once the new lease has been granted at a peppercorn rent the landlord's interest will be worthless). The value of the landlord's interest is its open market value, but Schedule 13 paragraph 3 of the Act requires a number of assumptions to be made. Among these is the assumption that:

"Chapter I and this Chapter confer no right to acquire any interest in any premises containing the tenant's flat or to acquire any new lease."

9. There is no dispute about this element of the premium.
10. The dispute concerns marriage value. Marriage value is defined by Schedule 13 which (so far as relevant) provides as follows:

"Landlord's share of marriage value

4 (1) The marriage value is the amount referred to in sub-paragraph (2), and the landlord's share of the marriage value is 50 per cent of that amount.

(2) … the marriage value is the difference between the following amounts, namely—

(a) the aggregate of—

(i) the value of the interest of the tenant under his existing lease,

(ii) the value of the landlord's interest in the tenant's flat prior to the grant of the new lease, and

(iii) …; and

(b) the aggregate of—

(i) the value of the interest to be held by the tenant under the new lease,

(ii) the value of the landlord's interest in the tenant's flat once the new lease is granted, and

(iii) ….

(3) For the purposes of sub-paragraph (2)—

(a) the value of the interest of the tenant under his existing lease shall be determined in accordance with paragraph 4A;

(aa) the value of the interest to be held by the tenant under the new lease shall be determined in accordance with paragraph 4B;

(b) the value of any such interest of the landlord as is mentioned in paragraph (a) or paragraph (b) of sub-paragraph (2) is the amount determined for the purposes of paragraph 3(1)(a) or paragraph 3(1)(b) (as the case may be); and

(c) ….

4A (1) Subject to the provisions of this paragraph, the value of the interest of the tenant under the existing lease is the amount which at the relevant date that interest might be expected to realise if sold on the open market by a willing seller (with neither the landlord nor any owner of an intermediate leasehold interest buying or seeking to buy) on the following assumptions—

(a) on the assumption that the vendor is selling such interest as is held by the tenant subject to any interest inferior to the interest of the tenant;

(b) on the assumption that Chapter I and this Chapter confer no right to acquire any interest in any premises containing the tenant's flat or to acquire any new lease;

(c) on the assumption that any increase in the value of the flat which is attributable to an improvement carried out at his own expense by the tenant or by any predecessor in title is to be disregarded; and

(d) on the assumption that (subject to paragraph (b)) the vendor is selling with and subject to the rights and burdens with and subject to which any interest inferior to the existing lease of the tenant has effect.

(2) It is hereby declared that the fact that sub-paragraph (1) requires assumptions to be made in relation to particular matters does not preclude the making of assumptions as to other matters where those assumptions are appropriate for determining the amount which at the relevant date the interest of the tenant under his existing lease might be expected to realise if sold as mentioned in that sub-paragraph."

11. Paragraph 4B contains similar provisions relating to the valuation of the tenant's interest after the new lease has been granted.
12. The point at issue in this appeal is what is known in the jargon as "relativity". There are two aspects to this concept. First, it is used to determine the value of the freehold with vacant possession. Because sales of flats with vacant possession are almost invariably sales of leasehold interests, a valuer needs to convert values derived from sales of leases into a value to be ascribed to the freehold. The relationship, expressed as a percentage, between the value of a lease of a given term and the freehold is one form of relativity, known in the jargon as "real world relativity". Second, the concept is used to describe the relationship between the value of a leasehold interest in the real world and the value of the same interest making the assumption required by paragraph 4A (1) (b). The problem arises because in the real world most sales of leasehold flats are sales of leases to which rights under the Act attach, whereas the Act requires those rights to be disregarded.
13. The problem is not a new one; and it has been recognised in many decisions of the UT (and before it the Lands Tribunal). It has been exacerbated in recent years by the abolition of the previous residence requirements under the Act and other changes which have extended the benefit of the Act to all (or almost all) long lessees. A number of methods of determining relativities have been considered by tribunals, none of which has proved to be without flaws. The most common method has been by way of relativity graphs, produced by a number of firms of chartered surveyors active in the market; but they differ in their results. These graphs plot the relationship in terms of value between a lease of a given length and the freehold. The holy grail would be a method of determining relativity which is both reliable and simple to apply. The tribunal has encouraged the Royal Institution of Chartered Surveyors to produce a standardised graph. In response the RICS established a working group, chaired by Jonathan Gaunt QC, to attempt to produce definitive graphs that could be used to establish relativities. Unfortunately, the attempt was unsuccessful, and the best that could be done was to draw together the various relativity graphs with details of the underlying data. More details about this can be found in the decision of the UT in Kosta v Trustees of the Phillimore Estate [2014] UKUT 319 (LC), [2014] L & TR 25 at [133] to [135].
14. In the present case the UT considered a number of relativity graphs, and commented on them in detail in Appendix C to their decision. It was unable to give unqualified endorsement to any of them. The most influential of these graphs was the Gerald Eve graph. In addition to the relativity graphs produced by chartered surveyors the UT also considered a different model (which I understand to have been a piece of computer software) known as the Parthenia model. It is necessary to say something about both the Gerald Eve graph (which the UT described as the "industry standard") and the Parthenia model.
15. The UT described the origins of the Gerald Eve graph as follows. Both Gerald Eve and John D Wood & Co were instructed by the Grosvenor Estate during the mid 1970s to advise on leasehold reform valuations arising from the Leasehold Reform Act 1967. Gerald Eve began keeping a schedule by the end of 1975 of the settlements they had negotiated, including details of relativity. Grosvenor relied upon John D Wood & Co's advice about the acceptability of vacant possession values and upon Gerald Eve's advice about the other component parts of the statutory valuation. Sometimes the make-up of the final valuation was agreed explicitly with the leaseholder but in most cases the component parts of the valuation were apparent from explicit agreements reached earlier in the valuation. In some cases, nothing was agreed other than the settlement price. The same procedure also applied to instructions from the Cadogan Estate.
16. The schedule comprised settlements reached by both the Grosvenor and Cadogan Estates. It was shared with John D Wood & Co and also the valuers acting on behalf of leaseholders. Following an adverse decision in the LVT Grosvenor decided to make the schedule more widely available; and Mr Pope of Gerald Eve eventually produced a table of relativities at 5 year intervals which was converted into a graph by the Gerald Eve drawing office. The Grosvenor Estate made the graph available to anyone asking for it from October 1996 onwards. The settlement schedule has been maintained and enlarged over subsequent years. The more recent settlements reached by Gerald Eve generally supported the pattern shown by the original 1996 graph and although the graph has been constantly reviewed it has not been changed.
17. One of the concerns that the UT had about the reliability of the Gerald Eve graph was that because the relativities shown by that graph had not changed since the graph was first compiled, it no longer reflected relativities as they were in 2014. Structural changes in interest rates and rates of investment returns, changes in the nature of the market such as an influx of foreign buyers, and changes in the institutional and legal structure of the residential market all suggested that the Gerald Eve graph overstated the relative value of a lease by comparison with the value of a freehold. These concerns were summarised at [28] to [32] of Appendix B to the UT's decision and reiterated in Appendix C at [64].
18. The Parthenia model is based on a statistical technique called hedonic regression analysis. This technique starts from the proposition that a good (in this case the lease of a flat) has a value which is made up of many component parts (size, location, physical condition etc). One of those components is the length of the lease. The hedonic regression analysis seeks to isolate that particular characteristic and attribute value to it. The UT accepted that a well-specified hedonic model is one approach to the determination of relativities in a particular time period. The Parthenia model was based on transactions which took place between 1987 and 1991, before the coming into force of the Act, with the objective of eliminating the effect of the Act on relativity.
19. The Parthenia model was subjected to sustained criticism by the expert witnesses called by the landlords. The UT recorded those criticisms in Appendix B to its decision. It rejected the Parthenia model for reasons that I will describe.
20. The UT arrived at its valuation by the following steps.
21. First it determined the freehold vacant possession value ("FHVP") of the flat in question. In the case of Flat 5 this was agreed between the valuers. In the other two cases there was a dispute. The UT considered a number of comparable transactions, adjusting them to reflect differences in physical characteristics, transaction dates and so on. In fact the comparable transactions were all sales of leases, so the UT had to make an adjustment to arrive at the FHVP. In the case of one of the flats both valuers had made that adjustment by using another relativity graph called the Savills 2002 enfranchiseable relativity graph. This was a graph produced by Savills, based on valuations of 240 leasehold properties of varying lease length. The valuations valued (a) the open market value of the stated lease length assuming an enfranchiseable lease (b) the open market value of the stated lease length assuming a non-enfranchiseable lease and (c) the freehold value. The UT found that this graph, although based solely on the opinions of valuers, was in common use by valuers at the relevant time. In the case of Flat 3 the UT went through a similar exercise. Again the underlying comparables were leasehold transactions in the real world. The landlord's valuer adjusted them by using the Savills 2002 enfranchiseable graph. The tenant's valuer used the Parthenia model, with a further adjustment for "Act rights". The UT rejected the Parthenia model and therefore did not find the tenant's valuer's adjustments helpful. They preferred the Savills 2002 enfranchiseable graph.
22. Having determined the FHVP, the UT went on to determine the value of the existing leases with rights under the Act. Again, in the case of Flat 5 that was agreed between the valuers. In the case of the other two flats, the UT used the Savills 2002 enfranchiseable graph in reverse, by applying it to the FHVP as found in order to arrive at the proper value for the lease in question having regard to the length of the unexpired term.
23. The next stage in the UT's decision was to determine the value of the leases on the assumption that they did not attract rights under the Act. This stage in the valuation also took as its starting point the FHVP as determined. The question for the UT was how to adjust that determined value in order to arrive at the value of the leases on the assumption that they did not attract rights under the Act. In the case of Flat 5 the tenant's valuer again used the Parthenia model applied to the FHVP as determined to arrive at the adjusted value which, again, the UT rejected. The landlord's valuer considered a number of different kinds of evidence. None of these consisted of any graph. Based on these various sources of evidence the UT decided that in the case of the lease of Flat 5, which had an unexpired term of 41.32 years, the appropriate deduction from the value of the lease in the real world, in order to adjust for the assumed lack of rights under the Act, was 10 per cent. As a cross-check the UT considered the Gerald Eve graph. That produced a slightly higher figure than the figure that the UT had reached based on the evidence. It adopted the lower of the two figures as being one in which it had confidence. That conclusion on the evidence was consistent with the UT's perception that the Gerald Eve graph overstated the relative value of a leasehold interest on the assumption that it had no rights under the Act attached to it. In the case of Flat 3 it considered two different methods. One was to apply the Gerald Eve relativity graph. This produced a relativity of 47.2 per cent. That gave a figure for the value of the lease without rights under the Act. The other was to apply the Savills 2002 enfranchiseable graph for real world relativity. That gave a real world relativity figure of 61.5 per cent which the UT had to adjust to account for the assumed lack of rights under the Act. Having regard to the valuation evidence about deductions to reflect the lack of rights under the Act that it had heard in relation to Flat 5 (where they had made a deduction of 10 per cent), it made a larger deduction of 20 per cent in the case of Flat 3 to reflect a shorter unexpired lease term. Having considered these two approaches it took the one that produced the lower figure. That was the figure produced by the Gerald Eve graph which, as mentioned, the UT considered overstated relativities. Finally, the UT made a further adjustment to reflect the fact that the lease of Flat 3 had an onerous ground rent.
24. It will be noted that the second stage in the UT's determination, namely the value of the leases with rights under the Act, did not feed directly into the valuation of the leases without rights under the Act, because the latter was derived from the FHVP, adjusted for relativity, to which the UT then applied a further deduction. Rather, the UT used the former valuation in order to test the relativity models placed before it in evidence. In the course of its consideration of the evidence the UT considered and rejected the Parthenia model, which I have mentioned. This was the method of assessing relativity on which the tenants relied.
25. The main problem that the UT found with the Parthenia model can be simply stated. In the case of Flat 5 the value of the lease in the real world (with rights under the Act) was agreed. The agreement was based on an actual sale of the very same lease in the market within a week of the valuation date. It was also agreed that a lease with rights under the Act was more valuable than a lease without rights under the Act. However, when the Parthenia model was applied to that agreed value, it produced a figure for the same lease without rights under the Act which was higher than the value of the same lease with rights under the Act. Given the two points that had been agreed, that was an impossible result. As the UT put it, the Parthenia model was "a clock which strikes 13". The Lands Tribunal reached a similar conclusion in Nailrile Ltd v Earl Cadogan [2009] 2 EGLR 151 at [217]. This was not, however, the only criticism of the Parthenia model that the UT had. The UT set these out in Appendix B at [61]. Two of them were singled out for special mention. One was highly technical and need not be discussed. The other was that as in the case of the Gerald Eve graph structural changes in interest rates and rates of investment returns, changes in the nature of the market such as an influx of foreign buyers, and changes in the institutional and legal structure of the residential market all suggested that relativities had changed between the dates of the transactions on which the Parthenia model was based and the valuation dates.
26. Mr Johnson QC, on behalf of the tenant, argued that the comparison between the value of the lease in the real world (with rights under the Act) and the value of the lease without such rights as shown by the Parthenia model was an illegitimate comparison. The two interests are fundamentally different. The legal effect of the assumptions that the Act requires precludes the valuer (and the UT) from having regard to any leasehold transaction in the real world where the lease attracts rights under the Act. This is the ground upon which he has been given permission to appeal.
27. I do not accept that submission. In the first place I cannot see that it raises a point of law at all. Whether to accept or reject the Parthenia model (unless perverse) was a question of fact for the UT. In view of the sustained criticism of the Parthenia model by the experts called on behalf of the landlord, there was ample evidence upon which the UT could rely.
28. Second, property valuation usually proceeds by way of comparison with appropriate adjustments. If I may repeat something I have said before (Marklands Ltd v Virgin Retail Ltd [2003] EWHC 3428 (Ch), [2004] 2 EGLR 43 at [9]):

"Valuation essentially proceeds by analogy. The valuer looks for an analogue that is as close as possible to that which he has to value, and which has been the subject matter of a real transaction. He then works on the premise that if the subject matter of his valuation were to be the subject of a similar transaction, it would command the same value as the analogue. Since the analogue will never be identical to the subject matter of the valuation, the valuer will have to make adjustments to the value revealed by the analogue in order to reflect the differences between the analogue and the subject matter of his own valuation. In the case of a property valuation, the analogues are usually called "comparables". In a property valuation, typical adjustments will reflect differences between the comparables in location, terms of letting and so on."

29. These adjustments are essentially a matter of valuation judgment. The fewer the differences there are between the comparable and the subject of the valuation, the greater the weight that can be given to the comparable. As Carnwath LJ put it in Earl Cadogan v Sportelli [2007] EWCA Civ 1042, [2008] 1 WLR 2142 at [87]:

"Mr Jourdan did not, I think, challenge the tribunal's conclusion that the assumed market was substantially different from the real market mainly because of the longer term nature of the landlord's security. Once that is accepted, the degree of difference and its relevance to the valuation must be a matter of judgment for the tribunal. The apparent paradox implied by Mr Jourdan's second question proves nothing. The comparison is only surprising if one assumes a direct relationship between the no-Act and subject-to-Act valuations. Unless the comparison is of like with like, the degree of difference cannot of itself show irrationality." (Emphasis added)

30. Sometimes it will be difficult to adjust for differences in location, difference in time, differences in physical condition and so on. But in the case of Flat 5 there was no need to make any of those adjustments because the analogue or comparable was the very same lease of the very same flat sold almost exactly at the valuation date. The only adjustment that needed to be made was the adjustment to reflect the difference between a lease with rights under the Act and a lease without such rights. There is nothing legally impermissible in making an adjustment to reflect the terms of a statutory assumption. This has certainly been the practice of the Upper Tribunal (and before it the Lands Tribunal) for many years: see for example Arrowdell Ltd v Coniston Court (North House) Ltd [2007] RVR 39 at [39]; Nailrile Ltd v Earl Cadogan at [228]. The practice was also approved by this court, in relation to fair rents under the Rent Act 1977, in Curtis v Chairman of the London Rent Assessment Committee [1999] QB 92. Moreover, in the field of commercial leases it is commonplace for valuers to make adjustments to reflect the fact that some tenants have security of tenure under Part II of the Landlord and Tenant Act 1954 while others do not.
31. Mr Johnson seized on the last two sentences in the passage from Carnwath LJ's judgment that I have quoted. He submitted that Carnwath LJ was saying that there was no direct relationship between the no-Act and subject-to-Act valuations, and that therefore a comparison of the two was illegitimate as a matter of law. I do not consider that he was. What he was saying is that whether there is such a relationship is a matter of valuation judgment, and that it was not irrational for the Lands Tribunal to conclude in that case that there was no sufficient relationship to make market evidence useful. I do not consider that Carnwath LJ was expressing a proposition of law at all.
32. Third, the first requirement of paragraph 4A is the assumption of a sale in the open market. This is a familiar concept in valuation. The starting point in such a valuation is that things are to be taken as they are in reality on the valuation date, except to the extent that the instrument postulating the hypothetical transaction requires a departure from reality. In the old cases this is summarised in the Latin phrase rebus sic stantibus. In the more modern cases it has been described as the principle of reality. Mr Johnson accepted that this principle was applicable.
33. The following points amplify the reality principle. All are well supported by authority (which I described more fully in Harbinger Capital Partners v Caldwell [2013] EWCA Civ 492, although that case was not cited to us):

i) The hypothesis of a sale is only a mechanism for enabling one to arrive at a value of a particular property for a particular purpose. It does not entitle the valuer to depart from the real world further than the hypothesis compels. The various hypotheses must be taken no further than their terms make strictly necessary. It is necessary to adhere to reality subject only to giving full effect to the hypothesis.

ii) Giving effect to the hypothesis may require a legal impediment to the implementation of the hypothesis to be ignored or treated as overridden; but only to the extent necessary to enable the hypothesis to be effective.

iii) The world of make-believe should be kept as near as possible to reality: reality must be adhered to so far as possible. The valuer should depart from reality only when the hypothesis so requires.

iv) Although the sale is hypothetical there is nothing hypothetical about the market in which it takes place.

34. The principle of reality is also relevant when interpreting a hypothesis required to be made.
35. In my judgment there is nothing in the positive requirement in paragraph 4A (1) (b) to assume that no rights under the Act attach to the lease or the premises in which the tenant's flat is situated that forbids a valuer from looking at transactions in the real world in order to assist in determining value on the required assumptions. Moreover, the logic of Mr Johnson's argument is that real world transactions cannot be lawfully used in determining any value on the equivalent assumption to that in paragraph 4A (1) (b). But as we have seen, (a) the valuation of the freehold must be carried out on the same assumption (b) in arriving at its valuation of the freehold the UT used transactions from the real world and (c) there is no complaint about either the process or the result. In addition that value is one of the components from which marriage value is derived. The tenant's valuer himself applied the Parthenia model to that value. Thus Mr Johnson's argument is internally inconsistent. He attempted to escape from this by submitting that in practice it would make very little difference to FHPV and was pragmatically justified. However, if the point is a good one it would not be right to condone legally impermissible valuations on purely pragmatic grounds. A further objection to the argument is that if leasehold transactions in the real world are eliminated from consideration, it is difficult to see any evidence that would be admissible before a tribunal in order to demonstrate the value of any interest. Even the Parthenia model needs to be applied to a base value; and if real world transactions cannot be used to determine that base value, there is nothing to which the model can be applied.
36. Mr Johnson goes on to argue that the reason why the value of the lease in the real world is an impermissible comparator is that the market itself had been "corrupted". It had been corrupted because, as the UT held, it had been influenced by the "industry standard" Gerald Eve relativity graph. The current method of valuing leases without rights under the Act was thus circular, self-reinforcing, self-perpetuating and unreliable. The UT itself made its valuations on the basis of what it described as the "least unreliable" evidence; and that is not a position that can be sustained. It is important to repeat at this point that an appeal to this court lies only on a question of law.
37. The UT made findings about how the Gerald Eve graph had impacted on the market. At [105] it found:

"… there was agreement between the valuers that short or medium term leases which have the benefit of the 1993 Act are sold at a price which reflects the value of an extended lease less the estimated cost of the extension. …It was also agreed that, for this purpose, in the market the estimated price for the extension is based upon a graph of relativities for leases without rights under the Act and, in particular, a graph … which is usually known as the Gerald Eve graph." (Emphasis added)

38. At [145] it found:

"We also record that over the years, the Gerald Eve graph has been extensively used both for the purpose of settling and arguing claims to enfranchisement but it is also used in the market place. Of course, existing leases without rights under the 1993 Act are not typically for sale in the market place but when parties are negotiating for the sale and purchase of an existing lease with rights under the 1993 Act, one or both parties to the negotiation will have regard to the relativity shown by the Gerald Eve graph when calculating the premium which would be expected to be payable for an extended lease and the amount of that premium will then influence the price paid for the existing lease. Accordingly, the Gerald Eve graph has not been simply a valuation tool used in relation to disputed claims to enfranchise, but it has also influenced the performance of the market in relation to the sales of existing leases with rights under the 1993 Act." (Emphasis added)

39. However, it qualified that finding to some extent in Appendix C at [58]:

"If market transactions involving leases with rights under the 1993 Act relied exclusively on the relativities shown by the GE graph, then there would be no reason for relativities for leases with rights under the 1993 Act to have changed between 2002 and 2015 (the dates of the two Savills enfranchisable graphs) since the GE graph has not changed over that period. Nonetheless, a comparison of the Savills 2002 graph and the Savills 2015 graphs does suggest that there has been a change in relativities for leases with rights under the 1993 Act. This may be due to either the difference in the methodology of construction of the Savills 2002 and 2015 graphs, or a real change in relativity in the market. Any such market change is not reflected in the GE graph; this would suggest that the GE graph is not relied upon by the market to quite the extent which has been suggested (although we do accept that the market has been influenced by the GE graph)."

40. It is clear from these findings that to describe the market as "corrupted" is an overstatement.
41. If the Gerald Eve graph is used in real market transactions, and has influenced the performance of the market, on what legal justification can it be said that transactions in the real market (albeit influenced by the Gerald Eve graph) can be ignored? The market may not be perfect but it is still the market. As Hoffmann J explained in Electricity Supply Nominees Ltd v London Clubs Ltd [1988] 2 EGLR 152:

"The open market may be a false market in the sense that it is based upon false assumptions, but it is still the open market."

42. Sometimes markets behave irrationally. The Tulip mania of the mid-seventeenth century, the South Sea Bubble of the early eighteenth century, the railway mania of the mid-nineteenth century and the dot-com bubble of the late twentieth century are well-known examples. Even in the absence of these extreme examples, markets are often influenced by what John Maynard Keynes called "animal spirits". In my judgment there is no legal justification in a case like this for ignoring real market transactions. Accordingly, I consider that the UT was correct to say at [142] (substantially repeated at [166]):

"The Gerald Eve graph was a real market circumstance which influenced market behaviour. The Upper Tribunal is required by statute to determine the market value of an asset in a real market which behaved in that way. It is not the function of the Upper Tribunal to tell the market how it ought to behave in the future and we have certainly no power to replace real market forces at past valuation dates by some other forces which we might consider ought instead to have operated."

43. Mr Johnson next argued that the Act required it to be assumed that the market was one in which no one had rights under the Act. This is a ground on which he needs permission to appeal. The relevant assumption is:

"that Chapter I and this Chapter confer no right to acquire any interest in any premises containing the tenant's flat or to acquire any new lease."

44. This is a limited assumption confined to premises containing the tenant's flat. Although in some cases judges have spoken of a "no-Act world," that is inaccurate. The phrase originates in the jargon of compulsory acquisition: see Daejan Properties Ltd v Weeks [1998] 3 EGLR 125 at 128F. But it cannot displace the express words of the Act. That has been recognised by many decisions of the UT (and before it the Lands Tribunal) (see, for example Arbib v Earl Cadogan [2005] 3 EGLR 139 at [82]; 82 Portland Place (Freehold) Ltd v Howard de Walden Estates Ltd [2014] UKUT 133 (LC) at [13]; Kosta v Trustees of the Phillimore Estate [2014] UKUT 319 (LC), [2014] L & TR 25 at [98]); and by this court in Carey-Morgan v Trustees of the Sloane Stanley Estate [2012] EWCA Civ 1181, [2012] HLR 47 at [8]. In some cases the tribunal has adopted the phrase "no-Act building" which is a better shorthand for the statutory hypothesis.
45. Mr Johnson argued that the assumption could be read as extending to the whole world, and that such a reading would better give effect to the intention of Parliament. He placed heavy reliance on a passage from Lord Hoffmann's speech in Earl Cadogan v Sportelli [2008] UKHL 71, [2010] 1 AC 226 at [17] in which Lord Hoffmann said that the hypothetical sale takes place in a market undisturbed by the existence of compulsory power.
46. I do not accept this argument. First it is not the natural meaning of the words, and we should adopt the natural meaning of the words unless that produces a result which is nonsensical or inconsistent with the intention of the legislature. Second, it overlooks the fact that in paragraph 4A (1) (b) Parliament has specifically delineated the geographical extent to which the assumption applies. Third, it would conflict with the primary instruction to assume a sale in the open market. Fourth, it would produce an unworkable result by eliminating all (or almost all) available evidence from consideration. Nor do I consider that the passage from Lord Hoffmann's speech (which was a dissenting speech) assists the argument. Lord Hoffmann was speaking in general terms, and the point now under discussion was not in issue.
47. The final point that Mr Johnson takes is that an assumption of a "no- Act world" should be made under paragraph 4A (2) (described by Lord Walker in Sportelli at [35] as "puzzling"):

"It is hereby declared that the fact that sub-paragraph (1) requires assumptions to be made in relation to particular matters does not preclude the making of assumptions as to other matters where those assumptions are appropriate for determining the amount which at the relevant date the interest of the tenant under his existing lease might be expected to realise if sold as mentioned in that sub-paragraph."

48. This is also a ground on which he needs permission to appeal. In my judgment there are a number of insuperable hurdles facing this argument. First the point was not advanced before the UT, and consequently there was no evidence about what additional assumptions were appropriate and what their effect on value might have been. Second, what is an "appropriate" assumption must be a question of valuation judgment. Since the UT was never asked to consider this question, it has not ruled one way or the other. It is not simply a question of law on which this court can rule. Third, the fundamental principle underlying paragraph 4A (1), and incorporated by reference into paragraph 4A (2), is that the sale takes place in the open market. The open market means the real market (subject to the assumptions required to be made). As I have said, although the transaction and the parties to it are hypothetical, there is nothing hypothetical about the market in which it takes place. To assume a different market in which no one has rights under the Act would contradict that fundamental principle.
49. There is a further consideration. In front of the UT it was expressly conceded that the "no-Act" assumption was confined to the building containing the tenant's flat and did not extend to the whole world. The evidence proceeded on that basis. If either way of putting Mr Johnson's point were a good one the whole course of evidence would have been different. As Mr Johnson accepted, the parties would have to start all over again. That is, in itself, a powerful reason for refusing permission to appeal.
50. There are two concluding points that I should make. First, the UT held at [265] that the Parthenia model should not be put forward in future cases. It is common ground that the reason why the UT (rather than the FTT) decided the dispute was so that there would be an authoritative test case of the validity of the Parthenia model. One of the functions of the UT is to provide definitive guidance to tribunals on valuation matters: Earl Cadogan v Sportelli [2007] EWCA Civ 1042, [2008] 1 WLR 2142 at [99]. In my judgment the UT was well within the scope of its functions in ruling out future use of the Parthenia model in its current form. Second, we were told by Mr Jourdan QC that at the invitation of the Government the Law Commission is to consider the simplification of valuations under the Act. It may be, therefore, that the holy grail will one day be found.
51. I would dismiss the appeal on the single ground for which permission has been given and refuse permission to appeal on the other two grounds. In respect of the latter two grounds, on which I would refuse permission to appeal, I would nevertheless give permission for this judgment to be cited under para. 6.1 of the Practice Direction (Citation of Authorities) [2001] 1 WLR 1001.

Lord Justice Peter Jackson:

52. I agree.

Lady Justice Arden:

53. I also agree.