Day and another v Hosebay Limited

before

Lord Phillips
Lord Walker
Lord Mance
Lord Clarke
Lord Wilson
Lord Sumption
Lord Carnwath

JUDGMENT GIVEN ON

10 October 2012

Heard on 16, 17 and 18 July 2012

Appellant Edwin Johnson QC Oliver Phillips (Instructed by Pemberton Greenish LLP)	Respondent Stephen Jourdan QC Anthony Radevsky (Instructed by Bircham Dyson Bell)
Appellant Jonathan Gaunt QC Katharine Holland QC (Instructed by Speechly Bircham LLP)	Respondent Anthony Radevsky Mark Sefton (Instructed by Wallace LLP)

LORD CARNWATH (with whom Lord Phillips, Lord Walker, Lord Mance, Lord Clarke, Lord Wilson and Lord Sumption agree)

Introduction

1. The Leasehold Reform Act 1967 is on its face a statute about houses, not commercial buildings. The buildings with which we are concerned were originally designed and used as houses, but at the relevant date were used entirely for commercial purposes, one for offices, the other (in the judge’s words) as a “self- catering hotel”. In both cases the courts below felt constrained to hold that they were “houses” within the meaning of the 1967 Act, with the consequence that the lessees were entitled to “enfranchise”, that is, to acquire the freeholds compulsorily from their lessors on the terms fixed by the Act.

2. In the Court of Appeal [2010] EWCA Civ 748; [2010] 1 WLR 2317 Lord Neuberger of Abbotsbury MR regretted this result. He saw it as the probably unintended consequence of amendments made by the Commonhold and Leasehold Reform Act 2002, removing the previous residence requirements. However, he felt bound to apply his view of the relevant provisions as they stood after those amendments, rather than to decide what “the legislature would have said if it had fully appreciated the consequences…” (para 57).

3. From the material we have been shown, he was clearly right to think that his interpretation did not reflect Parliament’s intentions. The thinking behind the 2002 legislation is apparent from the preceding Draft Bill and Consultation Paper “Commonhold and Leasehold Reform” (Cm 4843), published by the Lord Chancellor in 2000. It included proposals for the introduction of an entirely new form of tenure, known as “Commonhold”, and for amendment of the existing provisions relating to leases of flats (under the Leasehold Reform, Housing and Urban Development Act 1993) and of houses (under the 1967 Act). The first paragraph of the Introduction leaves no doubt that its purpose was to address perceived flaws in the “residential leasehold system” (p 107), not in the leasehold system more generally.

4. In relation to flats, the government’s view was that the residence tests under the 1993 Act were too restrictive, for example, in excluding someone subletting a flat, or occupying a flat as a second home. The residence requirement would therefore be abolished; but, to “restrict the scope for short-term speculative gains”, it would be replaced by a rule requiring the qualifying tenant to have held the lease for at least two years (pp 155-6).
5. A similar approach was proposed for leases of houses under the 1967 Act:

“This would bring the residence test for houses in line with the proposals for flats. It would allow long leaseholders of second homes to benefit and would also enable leaseholders who lease houses through a company to enfranchise. Furthermore, as in the case of flats, it would restrict the scope for short-term speculative gains (p 189).”

There is no evidence then or thereafter of any ministerial or parliamentary intention to extend the scope of the Act more generally, or in particular to confer statutory rights on lessees of buildings used for purely non-residential purposes.

6. Although the 1967 Act like the 1993 Act is in a sense expropriatory, in that it confers rights on lessees to acquire rights compulsorily from their lessors, this has been held not to give rise to any interpretative presumption in favour of the latter. As Millett LJ said of the 1993 Act:

“It would, in my opinion, be wrong to disregard the fact that, while the Act may to some extent be regarded as expropriatory of the landlord’s interest, nevertheless it was passed for the benefit of tenants. It is the duty of the court to construe the 1993 Act fairly and with a view, if possible, to making it effective to confer on tenants those advantages which Parliament must have intended them to enjoy.” (Cadogan v McGirk [1996] 4 All ER 643, 648)

By the same token, the court should avoid as far as possible an interpretation which has the effect of conferring rights going beyond those which Parliament intended.

Statutory definition

7. Section 2(1) defines “house” in the following terms:

“… ‘house’ includes any building designed or adapted for living in and reasonably so called, notwithstanding that the building is not structurally detached, or was or is not solely designed or adapted for living in, or is divided horizontally into flats or maisonettes; and-

1. where a building is divided horizontally, the flats or other units into which it is so divided are not separate ‘houses’, though the building as a whole may be; and

2. where a building is divided vertically the building as a whole is not a ‘house’ though any of the units into which it is divided may be.”

8. In the present cases, nothing turns directly on the qualifications introduced by the word “notwithstanding” (which I shall refer to as “the proviso”). We are concerned with the main part of the definition, which raises two separate but overlapping questions: (i) is the building one “designed or adapted for living in”?

(ii) is it a “house … reasonably so called”? Both questions remain live in Hosebay; in Lexgorge the first has been conceded in favour of the lessees.

9. The two parts of the definition are in a sense “belt and braces”: complementary and overlapping, but both needing to be satisfied. The first looks to the identity or function of the building based on its physical characteristics. The second ties the definition to the primary meaning of “house” as a single residence, as opposed to say a hostel or a block of flats; but that in turn is qualified by the specific provision relating to houses divided horizontally. Both parts need to be read in the context of a statute which is about houses as places to live in, not about houses as pieces of architecture, or features in a street scene, or names in an address book.

The facts

10. The first case (Hosebay) concerns three properties, 29, 31, and 39 Rosary Gardens, South Kensington, London SW7. They were originally built as separate houses as part of a late Victorian terrace forming the west side of Rosary Gardens. The current leases of Nos 29 and 39 were granted in 1966 for terms expiring in December 2020, subject to covenants for their use as “16 high class self-contained private residential flatlets”. The current lease for No 31 was granted in 1971 for a term expiring in December 2030, subject to a covenant restricting its use to that of “a single family residence” or “a high class furnished property for accommodating not more than 20 persons”. It was common ground that the current use, which had begun some time before 1981, was not in accordance with the covenants.

11. It was unclear from the evidence when the premises had been converted to their present layout. The judge (para 83) proceeded on the basis that the conversions “may well” have been carried out substantially before the current

leases were granted in 1966 and 1971. Although there was no evidence as to the actual purpose of the conversions, the Master of the Rolls “on the balance of probabilities” inferred (principally from the lack of documentation in the hands of the landlords to indicate otherwise) that they had been for the uses described in the leases (para 37).

12. Hosebay Ltd acquired all three leases in 1996. On 23 April 2007 it served notices on its landlords under section 8 of the 1967 Act to acquire the freeholds of the three properties.

13. Judge Marshall QC found that the three properties were at the relevant date being used together to provide “short term accommodation for tourists and other visitors to London”, or what she described as a “self-catering hotel” (paras 8 and 19). Each of the three properties had been “fully adapted to provide individual rooms for letting out” (para 9), with the exception of two rooms in No 31, one of which was used for office and reception purposes, and the other for storage. The great majority of the rooms could be described as “rooms with self-catering facilities”. Each room had between one and four beds, furniture, and limited storage space, cooking facilities, and small “wet rooms” with shower, basin and WC. Fresh bed linen and room cleaning, but no other services, were provided to those staying in the rooms.

14. On these facts, the judge concluded that each of the three properties was physically “adapted for living in” even though the current use was itself too transient to qualify as such. The Court of Appeal agreed. I quote the Master of the Rolls:

“33… My primary reason for that conclusion is that, in order to determine whether premises are adapted for living in, one looks at the most recent works of adaptation, and assesses objectively, whether they resulted in the property being adapted for living in…

36.  In this case, I consider that the effect of the most recent works of conversion to the three properties, if they were works of adaptation, adapted those properties for living in. Ignoring one or two rooms, each room in the three properties is a self-contained unit of accommodation, with its own basic small shower room/WC, and its own even smaller and more basic cooking facilities. As Moore- Bick LJ pointed out in argument, the rooms are entirely appropriate for letting to students on three year degree courses, and, as Mr Johnson rightly accepted, if they had been, all the rooms, and therefore the three buildings, would have been used for living in.

Even if, as Mr Johnson argued and I am prepared to assume without deciding, the current use of the three properties is not for living in, that certainly does not mean that, viewed objectively, the three properties were not adapted for living in.”

15. The judge and the Court of Appeal held also that the properties were “houses reasonably so called”, as the Master of the Rolls explained:

“externally, each of the three properties has the appearance of being a relatively large town house; internally, each of the three properties has been converted so that almost every room can be used as a self- contained unit for one or more individuals, with cooking and toilet facilities. … I find it hard to see how the judge could be faulted for concluding that, even if each of the three properties might be called something else as well, they could each reasonably be called a house.” (para 38)

16. The other case (Lexgorge) relates to 48 Queen Anne Street, in Marylebone, London W1. It was built in the early 18th century as a house comprising five floors including basement, in a terrace of substantial houses. It was occupied for that purpose for many years until 1888, when it began to be used for commercial purposes.

17. Coming to more recent times, planning permission was granted in December 1949 for conversion of the second and third floors into a self-contained maisonette, and there is some evidence that it was implemented. However, from about 1961, all four upper floors were used as offices, and they were so used when the notice was served under the Act on 4 March 2005. The whole building was still in office use in June 2005. However, by the time of the trial in October 2009, when the judge inspected the property, the upper two floors were in use for residential purposes. The office use of the lower floors continued.

18. The current lease was granted in 1951 for a term of 110 years. The lease described the property as a “messuage or residential and professional premises”, and restricted its use (subject to landlords’ consent) to “self-contained flats or maisonettes” on the upper two floors, professional offices on the first and ground floors, and in the basement “storage … and lavatory … in connection with other parts of the demised premises”. In 1978, the lease was acquired by Lexgorge Ltd. At the time of the notice the office use of all floors had become “established”, and therefore lawful for planning purposes, although in breach of the lease as respects the upper floors. The building is listed as a building of special architectural or

historic interest (grade 2); English Heritage’s records describe it as a “Terraced House”.

19. In this case, as already noted, it is conceded by the lessors that at the material date the premises, although used for offices, were still at least in part “designed or adapted for living in”. It was held by the judge (Judge Dight) and by the Court of Appeal that it was a “house reasonably so called”, and therefore within the definition. The Master of the Rolls said:

“53. If the upper two floors of the property had been empty, I have little doubt but that the property could reasonably have been called a house, bearing in mind its external character and appearance (a classic town house in London’s West End), its internal character and appearance at least on the upper two floors (which were, as I understand it, substantially as constructed), the description of the property in the lease as ‘messuage or residential or professional premises’, and, to the extent that it is relevant, the terms of the lease (restricting the use of the upper two floors to residential). I find it hard to see why the fact that the upper two floors had been used (even for many years) as offices (in contravention of the terms of the lease) should wreak such a change that the property could no longer reasonably be called a house.”

The authorities

20. The first relevant case under the Act was Lake v Bennett [1970] 1 QB 663. However, I find it helpful to start from an authority in a different statutory context, Lord Denning MR’s judgment in Ashbridge Investments Ltd v Minister of Housing and Local Government [1965] 1 WLR 1320. The case related to compulsory acquisition of two properties for the purpose of slum-clearance under the Housing Acts. The level of compensation would vary significantly depending on whether the property was or was not a “house”. In the absence of a statutory definition of “house”, Lord Denning adopted the following formula: “a building which is constructed or adapted for use as, or for the purposes of, a dwelling” (p 1324). In Lake v Bennett¸ he suggested that the draftsman of the 1967 Act definition had “adopted these words, but added the limitation ‘reasonably so called’” (p 670).

21. Ashbridge itself concerned two adjoining buildings in the same terrace, which had been designated for compulsory purchase, the first (No 17) as an unfit “house”, the second (No 19) as a building other than a house. The buildings were very similar in appearance; both had been designed as shops with rear living rooms and living quarters above, but neither was in current use for living purposes. No

17, which had undergone no structural alterations, was held by the Minister to have “retained its identity as a dwelling”. No 19, by contrast, was held to have “lost its identity as a dwelling”, following structural alterations involving the extension of the shop into the rear living area (p 1325). The latter decision was described in the Court of Appeal as “extraordinary” (p 1327, per Harman LJ), but that did not undermine the validity of the decision in relation to No 17. Lord Denning’s formula can be seen as his way of expressing the present “identity” (in the inspector’s words), or perhaps function, of a building not currently in use, defined by reference to the purpose of its construction or subsequent adaptation.

22. Lake v Bennett itself concerned a three storey house, the ground floor of which had been converted into a shop. There was no issue as to the first part of the definition, as it was clear that the building was at least in part adapted for use for living in. The Court of Appeal held that notwithstanding the commercial element, the building as a whole was a house “reasonably so called” and was therefore within the scope of the 1967 Act.

23. The reasoning of Lake v Bennett was adopted and extended by the House of Lords in Tandon v Trustees of Spurgeons Homes [1982] AC 755, which remains the leading House of Lords authority on this part of the definition. Unfortunately the reasoning of the single majority speech of Lord Roskill, although carrying the unqualified support of Lord Scarman and Lord Bridge, is not without difficulty. Further, the case needs to be read in its factual context. As in Lake v Bennett, the main problem was to reconcile the statutory recognition (under the “proviso”) that the building need not be “solely” designed or adapted for living in, with the need for the building as a whole to be a house “reasonably so called”. This is not a problem in the present cases.

24. At the end of his judgment Lord Roskill referred with approval to Lake v Bennett, which he welcomed as “stating a principle and [confining] the question of fact to a narrow area” , and from which he “deduced” the following three “propositions of law” :

“… (1) as long as a building of mixed use can reasonably be called a house, it is within the statutory meaning of ‘house’, even though it may also reasonably be called something else; (2) it is a question of law whether it is reasonable to call a building a ‘house’; (3) if the building is designed or adapted for living in, by which, as is plain from section 1(1) of the Act of 1967, is meant designed or adapted for occupation as a residence, only exceptional circumstances, which I find it hard to envisage, would justify a judge in holding that it could not reasonably be called a house. They would have to be such that nobody could reasonably call the building a house.” (p 767)

25. Although expressed as “propositions of law”, they do not in my view offer much assistance as such, at least beyond the facts of the case. The first proposition was in terms directed to a building in mixed residential and commercial use. Such a building could plausibly be described either as a house with a shop below, or as a shop with a dwelling above. That was enough to show that it could “reasonably” be called a house. That proposition cannot in my view be applied more generally. The mere fact that a building may be described as a “house” for other purposes (for example, in the English Heritage list) is not enough to bring it within this part of the definition.

26. The second proposition, that what is a house “reasonably so called” is a question of law, is not easy to extract from the judgments in Lake v Bennett. Lord Denning described the judge’s negative answer to that question as “an inference from primary facts” depending “in part at least on the true interpretation of the words ‘reasonably so called’”, and one with which the court could interfere if it was a conclusion to which the judge “could not reasonably come” ([1970] 1 QB 663, 671). Salmon LJ described it as “partly a question of fact but also a question of law as to the true construction of the meaning of the word ‘house’ in this Act.” (p 672). Elsewhere Lord Roskill himself had accepted counsel’s submission that the definition of “house” was “a mixed question of fact and law” ([1982] AC 755, 765), but he saw it as one in which, in the interests of consistency, the question of fact should be confined “within narrow limits”: p 767. More modern authorities have leant against such conceptual debates (see, for example, Lord Hoffmann, in Moyna v Secretary of State for Work and Pensions [2003] UKHL 44, [2003] 1 WLR 1929, paras 25-27). In any event, none of these formulations throws much light on how the question should be answered in any particular case.

27. The third proposition is again in terms hard to extract from Lake v Bennett. Lord Denning described the case before them as a “typical case”, but thought that difficult issues might arise in other cases: [1970] 1 QB 663, 671. He did not suggest that, in such cases, an affirmative answer to the first question would lead to any presumption in respect of the second. The examples given in the judgments (pp 671, 672) of cases that would not satisfy the second test – a block of flats, the Ritz Hotel or Rowton House (a working men’s hostel) – can hardly be described as “exceptional”. Rather than a free-standing proposition of law, deduced from Lake v Bennett, this proposition seems more an expression of Lord Roskill’s own view as to the correct policy approach to a building of the kind before him, which was adapted at least in part for occupation as a residence.

28. It may be that the real difference between the majority and the minority in Tandon came down to one of policy. Lord Wilberforce (in the minority) thought it clear that the building could not reasonably be called a “house”; it was rather a “mixed unit consisting in part of a shop and in part of a dwelling”, and as such was not within the policy of the Act: [1982] AC 755, 760. For Lord Roskill (in the

majority) Parliament had made clear that such mixed units were not in principle to be excluded. He noted that such small shops combined with living accommodation were a familiar feature of towns and villages across the country (p 766). In this he echoed the view of Salmon LJ (Lake v Bennett [1970] 1 QB 663, 672), who thought that a tenant living above a shop in the circumstances of that case was “obviously the sort of person to whom the legislature intended to give security of tenure”. Such policy considerations do not assist the lessees in this case. For the reasons already given, policy if anything points the other way.

29. Of more significance for present purposes is the relative lack of weight given by the majority to the appearance of the buildings as a factor in answering the second question. Lord Fraser of Tullybelton (in the minority) had regarded appearance as “the main element in the character of a building”: [1982] AC 755, 762. He attached particular weight to the photograph which showed “a shop in a row of shops”, in contrast with the “converted house” in Lake v Bennett; to him it was “obvious from the photograph” that the building could not reasonably be called a house (p 763). That, however, was not the approach of the majority. Lord Roskill had apparently accepted that in determining the “character” of the building for these purposes, physical appearance could be relevant, as also its history and the terms of the lease (p 766). However, those factors played no detectable part in the final decision. The determinative points were that the proportion of residential use, even if only 25%, was “substantial” (p 766), and that a tenant occupying such a building as his residence was within what was perceived to be the scope of the protection intended by Parliament (p 766). Those factors were enough to bring the case within the “principle” established by Lake v Bennett notwithstanding the differences from that case in relation to the original design and physical appearance of the respective buildings.

30. The only other relevant authority at the highest level is the much more recent decision in Boss Holdings Ltd v Grosvenor West End Properties Ltd [2008] UKHL 5; [2008] 1 WLR 289. The House of Lords held that a building previously designed or adapted for living in remained a house, even though at the material time it was not only disused but in parts “stripped out to the basic structural shell” (para 24). In contrast to Tandon this case was concerned solely with the first question. It was not in dispute that if that question was answered in the affirmative the building qualified as a house “reasonably so called”.

31. As will be seen I do not regard the case as determinative in either of the present appeals. However, some comment is desirable, in view of the change of view of Lord Neuberger on one aspect of his leading speech. He had proposed the following grammatical analysis of the relevant words of the statutory definition:

“18. In my judgment, the words ‘designed or adapted for living in’, as a matter of ordinary English, require one first to consider the property as it was initially built: for what purpose was it originally designed? That is the natural meaning of the word ‘designed’, which is a past participle. One then goes on to consider whether work has subsequently been done to the property so that the original ‘design’ has been changed: has it been adapted for another purpose, and if so what purpose? When asking either question, one is ultimately concerned to decide whether the purpose for which the property has been designed or adapted, was ‘for living in’.

19. The notion that the word ‘designed’ in section 2(1) is concerned with the past is reinforced by the later words in the same section ‘was or is [not] solely designed or adapted’. The use of the past tense is striking in a section which contains a number of verbs only in the present tense. In my judgment, the expression is to be construed distributively: thus, the word ‘was’ governs ‘designed’, and the word ‘is’ governs ‘adapted’. The present tense is appropriate for ‘adapted’ because, as Lord Scott of Foscote pointed out in argument, there could have been several successive adaptations, and it is only the most recent which is relevant. The word ‘was’ is in any event difficult to reconcile with Grosvenor’s case (as accepted by the judge and the Court of Appeal), as it would be irrelevant whether the property could have been fit for residential occupation at any time in the past.”

32. Later in his speech, he considered the implications of this analysis for other cases, including how the definition should apply to a property which had been designed for living in, but had subsequently been adapted to another use. As a matter of “literal language”, he thought such a property would be within the definition. If, as appeared, “designed” and “adapted” were alternative qualifying requirements, a building which had been designed as a house would remain within the definition in spite of its adaptation to other uses. Such a conclusion, he accepted, might seem surprising, but it could have been more readily understandable when taken with the residence requirement in the original Act (para 26).

33. It was on this latter point that, as Master of the Rolls in the present case, he has had second thoughts. It had been put directly in issue by the tenants in Hosebay, who argued (as they have in this court) that because the buildings were originally designed for “living in”, that was sufficient to bring them within the definition, regardless of any subsequent adaptation to other uses. On reconsideration, Lord Neuberger felt bound to reject the argument. Although the “literalist meaning” of “designed or adapted” was that either alternative would do,

that was “not by any means what the words naturally convey”. His earlier thoughts had been based on “an over-literalist approach to the language used by the legislature”: [2010] 1 WLR 2317, para 31. In his revised view, a building originally designed for living in, but adapted for some other purpose, was not “designed or adapted for living in”, unless subsequently re-adapted for that purpose (para 40).

34. I have no doubt, with respect, that Lord Neuberger’s second thoughts on this point were correct. Context and common sense argue strongly against a definition turning principally on historic design, if that has long since been superseded by adaptation to some other use. However, that approach may also have implications for the earlier part of his grammatical analysis in Boss Holdings (see para 31 above). The expression “was or is designed or adapted” is, as he says, to be read “distributively”: that is, as equivalent to “was designed or is adapted”. While that may support the view that the word “designed” is directed to the past, the same cannot be said of the expression “is adapted”. Nor (pace Lord Scott) is that grammatically the same as “was most recently adapted”. Logically that expression can only be taken as directed to the present state of the building.

35. Once it is accepted that a “literalist” approach to the definition is inappropriate, I find myself drawn back to a reading which accords more closely to what I have suggested was in Lord Denning’s mind in Ashbridge [1965] 1 WLR 1320, that is a simple way of defining the present identity or function of a building as a house, by reference to its current physical character, whether derived from its original design or from subsequent adaptation. Furthermore, I would not give any special weight in that context to the word “adapted”. In ordinary language it means no more than “made suitable”. It is true that the word is applied to the building, rather than its contents, so that a mere change of furniture is not enough. However, the word does not imply any particular degree of structural change. Where a building is in active and settled use for a particular purpose, the likelihood is that it has undergone at least some physical adaptation to make it suitable for that purpose. That in most cases can be taken as the use for which it is currently “adapted”, and in most cases it will be unnecessary to look further.

36. That interpretation does not of course call into question the actual decision in Boss Holdings. The basis of the decision, as I understand it, was that the upper floors, which had been designed or last adapted for residential purposes, and had not been put to any other use, had not lost their identity as such, merely because at the material time they were disused and dilapidated. It was enough that the building was partially “adapted for living in”, and it was unnecessary to look beyond that: see [2008] 1 WLR 289, para 25. That reasoning cannot be extended to a building in which the residential use has not merely ceased, but has been wholly replaced by a new, non-residential use.
37. Finally I must refer to Prospect Estates Ltd v Grosvenor Estate Belgravia [2008] EWCA Civ 1281; [2009] 1 WLR 1313. The Court of Appeal held that a building which had been designed and built as a house, but which for many years had been used almost wholly as offices, was not a house within the definition. As in Tandon the case turned ultimately only on the second question, whether the building was a “house reasonably so called”. The facts were much closer to those of the present cases. The leading judgment was given by Mummery LJ.

38. The building had been built in the 1850s as a house for residential occupation, but since 1958 it had been used substantially (88.5% of the floorspace) for office purposes. Under the most recent lease granted in 1972 the use was restricted to offices on all floors, except the top floor which was limited to use as a flat for a director or senior employee of a business occupying the offices below. It was accepted by the lessors that there had been insufficient works of adaptation to conclude that it had ceased to be designed for living in (Mummery LJ, para 9), but they challenged the judge’s conclusion that it was “a house reasonably so called”. That had been based, as the “overwhelmingly significant factor”, on the fact that the building was “designed for living in” and that “its structure and appearance have (largely) remained unchanged” (para 8). Mummery LJ held that the judge had given too much weight to those factors, and insufficient weight to “the prescriptive terms of the lease, the actual uses of the building and the relative proportions of the mixed use at the relevant date” (para 20). Goldring LJ, agreeing, found it impossible to accept that a “building can reasonably be called a house although no one can lawfully live in virtually 90% of it” (para 23).

39. In the present case, the Master of the Rolls ([2010] 1 WLR 2317, para 43) questioned the weight placed on that case by counsel for the present appellants in Hosebay:

“There can be no doubt that the external and internal appearance of the properties are highly relevant factors on this issue, and it is clear from the Prospect Estates case [2009] 1 WLR 1313 that, in so far as user is significant, the permitted use under the lease is a relevant factor. In those circumstances, even assuming that actual use is also relevant, I find it hard to see how it can be sensibly said that each of the three properties cannot ‘reasonably [be] called’ a ‘house’. To hold otherwise would involve concluding that the actual user, even where it involved people occupying virtually all the rooms in the building for relaxing, sleeping, cooking and washing, albeit on a short term basis, trumped all the other factors to the extent of disabling the building from being able to be a ‘house … reasonably so called’.”

40. He also doubted the decisive weight placed by Goldring LJ on the terms of the lease. He thought the thrust of the judgments in Lake v Bennett [1970] 1 QB 663 and the opinion of Lord Roskill in Tandon [1982] AC 755 was that the question was to be determined “essentially by reference to [the building’s] external and internal physical character and appearance” (para 46). He was not convinced that it would occur to most people, asked whether a building could reasonably be called a house, “to ask about the permitted use under any lease, or that they would be influenced if told what the permitted use was” (para 47). He suggested that the ratio of Prospect Estates should be treated as being “limited to a case where residential use is either prohibited entirely, or restricted to a very small part of the building, and the actual use accords with that” (para 49).

41. As will be apparent from my earlier analysis of Tandon, I cannot agree that Lord Roskill regarded “external and internal physical character and appearance” as the determining factors. I agree with the Master of the Rolls that the terms of the lease as such should not have been treated as the major factor. However, in so far as Mummery LJ treated the use of the building, rather than its physical appearance, as determinative, his approach was in my view entirely consistent with the reasoning of the majority in Tandon as I have explained it. I consider that Prospect Estates [2009] 1 WLR 1313 was rightly decided, and that the ratio need not be limited in the way the Master of the Rolls proposed.

The present cases

42. I turn to consider the application of these principles to the present appeals, which I can deal with briefly.

43. I would allow the appeal in Hosebay on the grounds that a building which is wholly used as a “self-catering hotel” is not “a house reasonably so called” within the meaning of this statute. As appears from para 38 of their judgment (quoted above), the contrary view of the Court of Appeal turned on two main points: (i) the external appearance of each property as a town house; (ii) the internal conversion to self-contained units, with cooking and toilet facilities. I find it difficult with respect to see the relevance of the second point to this part of the definition, which only arises in relation to a building which is in some sense adapted for living in under the first part. It is not suggested that the building is divided in a way which comes within the proviso. The first point, for the reasons given in my analysis of Tandon, should not have been given determinative weight. The fact that the buildings might look like houses, and might be referred to as houses for some purposes, is not in my view sufficient to displace the fact that their use was entirely commercial.
44. In these circumstances I find it unnecessary to reach a concluded view on the application of the first part of the definition in this appeal. I agree with the appellants (and the judge) that “living in” means something more settled than “staying in”; and that the present use does not qualify as such. There is more room for debate, however, whether the premises are to be taken as “adapted” solely for such use, to the exclusion of longer term occupation. The Court of Appeal, as I understand it, were influenced not only by the consideration that the rooms might be used (for example) for longer term student occupation, but also that their current layout probably dates from earlier adaptation to the uses described in the leases, which could well be regarded as sufficiently settled to qualify as “living in”. One of the values of the two-part definition is that it becomes unnecessary to resolve such narrow factual issues.

45. In Lexgorge I would also allow the appeal on similar grounds. A building wholly used for offices, whatever its original design or current appearance, is not a house reasonably so called. The fact that it was designed as a house, and is still described as a house for many purposes, including in architectural histories, is beside the point. In this case no issue arises under the first part of the definition. It is unnecessary to consider whether the concession in that respect was rightly made, although it is possible that it was based on a wider interpretation of Boss Holdings [2008] 1 WLR 289 than my own analysis would have supported.

46. In summary, I would allow both appeals, and hold that neither building was on the relevant date a “house” within the meaning of section 2 of the 1967 Act.