COURT OF FINAL APPEAL, HKSAR
CHIEF JUSTICE GEOFFREY MA
JUSTICE PATRICK CHAN PJ
JUSTICE R.A.V. RIBEIRO PJ
JUSTICE KEMAL BOKHARY NPJ
LORD WALKER OF GESTINGTHORPE NPJ
21 DECEMBER 2012
Chief Justice Ma
This appeal concerns the question of the Government rent that is payable on construction sites used for development or redevelopment. That such sites are subject to the payment of Government rent can be seen from Regulation 2 of the Government Rent (Assessment and Collection) Regulation, Cap 515A.
For the reasons contained in the judgment of Lord Walker of Gestingthorpe NPJ, I would dismiss the appeal. The decisions of this Court in Commissioner of Rating and Valuation v Agrila Ltd (2001) 4 HKCFAR 83 and in the present appeal are to be regarded as providing the necessary guidance as to the assessability and assessment of Government rent in relation to construction sites used for development or redevelopment
Justice Chan PJ
I agree with the judgment of Lord Walker of Gestingthorpe NPJ.
Justice Ribeiro PJ
I agree with the judgment of Lord Walker of Gestingthorpe NPJ.
Justice Bokhary NPJ
I am in entire agreement with the judgment of Lord Walker of Gestingthorpe NPJ, and add only a word of my own. The nil or nominal rateable value for which Mr John Howell QC contends on the appellant’s behalf is obviously unrelated to reality. Among the ways in which Mr Howell skillfully endeavoured to cope with that difficulty in his path was to place reliance on a passage in Lord Asquith of Bishopstone’s speech in East End Dwellings Ltd v Finsbury Borough Council  AC 109. That passage, which appears at pp 132- 133, reads:
If you are bidden to treat an imaginary state of affairs as real, you must surely, unless prohibited from doing so, also imagine as real the consequences and incidents which, if the putative state of affairs had in fact existed, must inevitably have flowed from or accompanied it. One of these in this case is emancipation from the 1939 level of rents. The statute says that you must imagine a certain state of affairs; it does not say that having done so, you must cause or permit your imagination to boggle when it comes to the inevitable corollaries of that state of affairs.
Such a statement must nowadays be read subject to the imperative of purposiveness in statutory interpretation. Turning from the general to the particular, I begin by noting that in the present case, the “imaginary state of affairs”, to use Lord Asquith’s expression, is the one that arises whenever the nature of the tenement concerned is such that it would never in real life be taken on a year to year tenancy. It arises because of the statutory requirement that the rateable value of all tenements, even those of that nature, be ascertained in terms of the hypothetical year to year tenancy laid down by s 7(2) of the Rating Ordinance, Cap 116.
Needless to say, that rating hypothesis was not enacted for the purpose of creating unreality. Obviously, its objective is uniformity. And that means uniformity in the starting point from which to proceed when ascertaining rateable values. No such value is ultimately ascertained, as opposed to distorted, if purposiveness is laid aside and artificiality is carried to the point of a substantial value being artificially reduced to nothing or next to nothing. In my view, Lord Asquith’s statement quoted above does not compel anything of that kind.
Lord Walker of Gestingthorpe NPJ
This appeal is the latest stage in protracted litigation concerned with the rent payable under Government leases of land which is in course of development or redevelopment. The problem came before this Court in 2001 in two consolidated appeals, one brought by the Commissioner of Rating and Valuation (“the Commissioner”) and the other (by way of cross-appeal) by Agrila Ltd and 58 other companies having the same, or a similar interest to that of Agrila Ltd: Commissioner of Rating and Valuation v Agrila Ltd (2001) 4 HKCFAR 83 (“Agrila”). This Court allowed the Commissioner’s appeal and dismissed the cross-appeal from the decision of the Court of Appeal  1 HKC 175, which was on an appeal from a decision of the Lands Tribunal issued on 20 February 1998. The Tribunal had, with the consent of the parties, restricted its decision to answering nine preliminary questions of law (a tenth question had been formulated but was disposed of by agreement). Best Origin Ltd (“Best Origin”), the appellant in the present appeal, is one of the other 58 companies which were parties to Agrila.
The questions answered by the Tribunal, and subsequently considered by the Court of Appeal and this Court, related to the validity, interpretation and application of statutory provisions under which rents payable under Government leases of land in course of development or redevelopment are to be calculated by reference to rateable value. There is no longer any issue as to the validity of these provisions, but their correct interpretation and application is still contested. Agrila raised issues of great public importance and considerable difficulty. Several of those issues have been clearly and authoritatively answered, leaving no room for further dispute. But the preliminary nature of some of the questions raised in Agrila inevitably led to a degree of generality in the answers given by this Court. In the present appeal, by contrast, there is a detailed and comprehensive decision of the Tribunal given after it had considered a large volume of oral and documentary evidence.
It is best to start with a very general explanation of the problem. In Hong Kong the law of rating is based on the English law of rating, which has a long history, as Tang Ag CJHC remarked at the beginning of his judgment in the Court of Appeal. It has (in striking contrast to other forms of taxation, which are subject to frequent legislative intervention) shown remarkable stability over the years. Instead the courts have undertaken the task of interpreting and adapting the traditional statutory formula so as to apply it to modern conditions. All this was explained by Viscount Hailsham LC in Railway Assessment Authority v Southern Railway Company  AC 266, 273, cited by Tang Ag CJHC in para 2 of his judgment.
On this occasion the legislature has intervened, and the courts have to work out the full implications of the intervention. It has for a long time been a fundamental proposition of rating law (as Sir Anthony Mason NPJ put it in Agrila, para 48 (99C), citing Lord Radcliffe in London County Council v Wilkins (Valuation Officer)  AC 362, 380) that:
Building sites themselves are not treated as hereditaments [the English rating equivalent for tenements] while the work of building is in progress.
That is because a building site is not in rateable occupation, a term of art indicating occupation which is
beneficial in the sense of being of value to the occupier, and
not too transient.
These requirements apply under the Rating Ordinance: Yiu Lian Machinery Repairing Works v Commissioner for Rating and Valuation  HKDCLR 32, 39, cited by Sir Anthony Mason NPJ in Agrila at para 47 (98I-99B).
It is now clear and no longer in dispute, as a result of this Court’s decision in Agrila, that s 2 of the Government Rent (Assessment and Collection) Regulation, Cap 515A (“the Rent Regulation”), made under s 34 of the Government Rent (Assessment and Collection) Ordinance, Cap 515 (“the Rent Ordinance”), is valid and changes (though only for the purposes of Government rent) the “fundamental proposition” that a building site is not a rateable tenement. Section 2 (set out in para 24 below) is a “deeming” provision under which a building site is to be treated, for the purposes of the Rent Ordinance, as if it were a rateable tenement. But, as Sir Anthony Mason NPJ observed in Agrila, para 49 (99F-G), s 2 says nothing about how the rateable value is to be ascertained. That question was raised as point 4 before the Tribunal in Agrila, and it received an answer, in general terms, in the judgment of Sir Anthony Mason NPJ, with which the other members of this Court agreed. But it is argued that the Tribunal made an error of law in applying that guidance in the present case.
As is apparent from the differences of opinion between the Court of Appeal and this Court in Agrila, it is not easy to see how the deeming provision in s 2 of the Rent Regulation is intended to fit in with the scheme of the rules as to the ascertainment of rateable value set out in ss 7 and 7A of the Rating Ordinance, Cap 116. Those rules themselves contain some provisions of a hypothetical character, in particular the requirement to ascertain “the rent at which the tenement might reasonably be expected to let, from year to year”. Lord Wilberforce commented on the corresponding words in English rating statutes, in Dawkins (Valuation Officer) v Ash Brothers and Heaton Ltd  2 AC 366, 385:
The interpretation of these words has come over the years to be invested with a good deal of learning, and even of mystery, but up to a point there is nothing very difficult about them. It is upon the delimination of the ‘actual,’ on the one hand, and the ‘hypothetical,’ on the other, that the argument in the present case takes its shape.
The same general comment might be made about the argument in this appeal.
The Rating Ordinance
The Rent Ordinance and the Rent Regulation operate, to a large extent, by reference to the Rating Ordinance, and with a similar scheme for assessment. It is therefore best to start with the Rating Ordinance. The scheme of assessment for rating purposes is that under ss 11 to 14 of the Rating Ordinance the Commissioner must, at the direction of the Chief Executive, prepare a valuation list every year. The Chief Executive has power under s 11(1)(b) to designate a date (“the relevant date”) by reference to which the values of tenements are to be ascertained. The present practice is for the valuation list to come into force on 1st April based on valuations as at 1st July (the relevant date) in the preceding year. The list must set out the rateable value of each tenement (which is under s 2 the basic unit of property on which rates are to be levied). By s 18 rates are to be levied on each tenement as a prescribed percentage of its rateable value. Rating uses the concept of rateable value in order to set a standard as between one rateable tenement and another. In the Court of Appeal Tang Ag CJHC cited the observations on this point made by Lord Pearce in Dawkins  2 AC 366, 381:
Rating seeks a standard by which every hereditament in this country can be measured in relation to every other hereditament. It is not seeking to establish the true value of any particular hereditament, but rather its value in comparison with the respective values of the rest.
Sections 7 and 7A are of central importance to this appeal. They direct how rateable value is to be assessed. Section 7 is as follows:
Section 8, 8A and 9 (relating to machinery, plant and advertising stations) are not relevant to this appeal, except for one point on s 8 taken by Mr Holgate QC (for the Commissioner) as part of a fallback submission.
Subsections (2) and (5) of s 7A are as follows:
Subsection (1) was repealed in 1987 and subsections (3) and (4) are not relevant to this appeal.
Section 36 contains exemptions from liability to rates for various categories of land and buildings, including agricultural land. The effect of the exemption is that tenements in these categories are not rateable tenements. Sections 37 to 40 lay down the procedure for challenging decisions of the Commissioner. There is a right of appeal to the Lands Tribunal, with a possible further appeal on the ground of error of law.
The Rent Ordinance and the Rent Regulation
The complex background to the enactment of the Rent Ordinance is described in the judgment of Sir Anthony Mason NPJ in Agrila, paras 52 to 64 (100A-102D). He referred to the Crown Leases Ordinance, Cap 40 (enacted in 1973), the Joint Declaration of 19 December 1984, the Rent Conditions uniformly applied between 1985 and 1997, the New Territories Leases (Extension) Ordinance, Cap 150 (enacted in 1988) and the Basic Law, promulgated on 4 April 1990 and coming into force on 1 July 1997. He also referred (para 74, 104A-C) to matters which are a legitimate guide to statutory interpretation under Pepper v Hart  AC 593. It is not necessary to repeat all these background matters, but some reference will be made to them below.
Section 3 of the Rent Ordinance identifies the leases to which it applies. They are referred to as “applicable leases”. It is common ground that it applies to the lease of Inland Lot 8874, Electric Road (“the site”) which is the subject-matter of this appeal.
By s 6(1) the lessee under an applicable lease is liable to pay, as Government rent, “an annual rent of an amount equal to 3 per cent of the rateable value of the land leased”. By s 6(2)(a) the lessee is treated as having covenanted to pay the Government rent as assessed “under and in accordance with this Ordinance” – that is the Rent Ordinance, not the Rating Ordinance. As Sir Anthony Mason NPJ put it in Agrila, para 19 (94A-D), the two statutory regimes are distinct, but a concept originating in the Rating Ordinance is applied by the Rent Ordinance for a different purpose.
Sections 10 and 11 of the Rent Ordinance contain provisions similar to those of ss 11 to 14 of the Rating Ordinance. The Commissioner prepares the Government Rent Roll (“the Roll”) each year at the direction of the Chief Executive, who has power to designate the date (“the valuation date”) by reference to which rateable values are to be assessed for rent purposes. The present practice, matching that under the Rating Ordinance, is for the Roll to come into force on 1st April based on valuations as at 1st July (the valuation date) in the preceding year. This appeal relates to the first Roll, for which the interval between the dates was rather longer. The first Roll came into force on 28 June 1997, based on valuations as at 1 July 1996.
By s 7(1) the rateable value of land which is let under an applicable lease is the aggregate of the rateable values of the tenements comprised in that land. By s 7(2) rateable values are to be those set out in the Roll (or in an interim valuation, which is not relevant to this appeal). Section 8(1) gives the Commissioner power to value land held under an applicable lease in order to ascertain its rateable value for the purposes of s 7, and by s 8(2) the Rating Ordinance “applies to the ascertainment of rateable values under this Ordinance subject to any specific provisions of this Ordinance”.
Sections 16 to 27 contain a procedural code for challenges to decisions of the Commissioner. Proposals and appeals made by the lessee in respect of “identical tenements” (as defined in s 18 of the Rent Ordinance) may be made only under the Rating Ordinance. Proposals and appeals in respect of non-identical tenements are to be made under the Rent Ordinance. Building sites are non-identical tenements, because they are not included in valuation lists for the purposes of rating under the Rating Ordinance. Under s 17 of the Rent Ordinance a lessee may make a proposal for reduction of a rateable value shown on the Roll. The Commissioner then serves a decision notice under s 25. By s 26 a dissatisfied lessee may appeal to the Lands Tribunal, whose powers are laid down in s 27. There is a further appeal to the Court of Appeal, but only on a point of law, under s 11 of the Lands Tribunal Ordinance, Cap 17.
Section 34 of the Rent Ordinance gives the Chief Executive power to make regulations enabling rateable values to be ascertained for “land, including interests held under applicable leases”. The Rent Regulation was made under this power. Section 2 of the Rent Regulation provides as follows:
Section 4 applies to land before redevelopment (defined as development after a previous development of land subject to a lease, so that it applies where a building has been constructed and then demolished during the term of the same lease). The rateable value is to be taken as the rateable value of the site as if the demolished building were still standing. This is called the last available rateable value (“LARV”). Sections 3 and 5 contain corresponding provisions for land that is partially developed or partially redeveloped. In Agrila this Court upheld the validity not only of s 2 but also that of ss 4 and 5 and (by implication) s 3.
This issues on this appeal
In Agrila this Court was concerned with all nine of the questions which the Tribunal had raised for preliminary rulings on points of law. On this appeal the Court is concerned with only one of the questions (the fourth), and with the full implications of the answer to that question given in the judgment of Sir Anthony Mason NPJ in Agrila. But it may be helpful to set out the first question and answer also; both these questions and answers are set out in para 117 (113A-C, 113I-114D) of his judgment:
Whether the Commissioner is required or empowered by s.8 of the Rent Ordinance or regulation 2 of the Rent regulations to ascertain the rateable value of leased land before or during development otherwise than in accordance with sections 7 and 7A of the Rating Ordinance (and the rebus sic stantibus rule) whether or not it is liable for assessment to rates under the Rating Ordinance.
When making a valuation under s.8 of the Rent Ordinance of leased land before or during development what assumptions, whether using the contractors or another basis of valuation, the Commissioner is required or empowered by law to make as to:-
The general issue in this appeal is whether the Tribunal made an error of law in its understanding and application of the principles embodied in, or underlying, the answer that this Court gave to point 4 in Agrila. That general issue has been refined into five more particular issues set out in para 1 of the court’s order dated 1 December 2011 granting leave to appeal. That part of the order is set out (with some small clerical amendments) in the next paragraph. But this detailed formulation of particular issues, some of them (such as the rebus sic stantibus principle) of a technical character, must be read subject to an important caveat. The court’s task is essentially an exercise in statutory construction. The task is to determine the legislative intention as expressed in the text of the Rent Ordinance, the Rent Regulation and the parts of the Rating Ordinance which they incorporate by reference. Rating law has a long history, as already noted. But respect for intellectual freight from the past must not obstruct a purposive approach which looks for the aims and scheme of the legislation as a whole. The five particular issues identified in the order granting leave to appeal are interwoven strands of the one essential issue of statutory construction.
The order states the issues as follows:
The site was bought by Best Origin, a subsidiary of Sino Land Company Ltd, at a Government land auction on 11 December 1996. Best Origin paid a premium of $760 m. The term of the Government lease expires on 30 June 2047. At the time of the sale there were some unoccupied buildings on the site, that is the old Causeway Bay police station and its accommodation block. The new owner was required to demolish these buildings within nine months and to erect a new building to be completed by the end of 2000. The new building was to have a total floor area of not less than 11,000 square metres and the maximum floor area (having regard to the Building (Planning) Regulations) was 18,339 square metres.
The completed development was described by the Tribunal in para 10 of its decision:
Demolition of the buildings on the site began before 28 June 1997 and was completed very soon after that date. The agreed statement of facts before the Tribunal included (in paras 38 to 40) the following matters:
The agreed statement of facts (dated 18 October 2006) also identified a number of factual issues, and matters of professional judgment as to valuation assumptions, on which the parties had not been able to agree as at that date. But since then agreement has been reached on further issues and matters of judgment. It is therefore unnecessary to summarize more of the statement. The parties are to be congratulated that they have been able to reduce the area of dispute in this way.
In the Roll for 1997-98, coming into force on 28 June 1997, the rateable value originally entered was $37,050,000. On the same day Best Origin made a proposal under ss 16 and 17 of the Rent Ordinance to alter the entry. By a notice of decision issued on 20 February 1998 under s 21 of the Rent Ordinance the Commissioner reduced the rateable value to $29,640,000. On 18 March 1998 Best Origin gave notice of appeal under s 26.
On 18 June 1998 the Civil Litigation Unit of the Department of Justice furnished the solicitors acting for Best Origin with a copy of a short valuation report on the site prepared by the Commissioner. It valued the site, as at the valuation date, at $741,000,000 and arrived at the rateable value of $29,640,000 by decapitalization at 4 per cent. A note to the valuation stated:
The decapitalization rate adopted is dependent on the type of development being undertaken. Generally the following rates have been adopted:-
office/commercial 4%; and
So the Commissioner was relying “from day 1” (as it was put in the Tribunal’s decision, para 136(3)) on the contractor’s basis (“CB”) method of valuation, as explained and considered further below.
The proceedings before the Tribunal
The proceedings before the Tribunal (Lam J, President, and Mr W K Lo, Member) occupied a considerable time. There were six weeks of oral hearings during October and November 2006; four further days of oral hearings in December 2006 and March 2007; and then final written submissions, completed by August 2007. Each side put forward reports by several different experts, but not all of them were cross-examined at length. The principal valuation experts relied on by Best Origin were Mr George Doran and (after Mr Doran’s retirement) Mr Lynch. The Tribunal did not accept their conclusions but its rejection of their evidence was because of the faulty legal basis of their instructions. Mr John Charman, Mr David Lee and Professor Gordon Hughes also gave evidence on behalf of Best Origin; the latter’s evidence was quite severely criticized by the Tribunal. The principal expert for the Commissioner was Mr Tang Ping-kwong; Professor Andrew Baum, Mr Nicholas Brooke and Mr Laurence Hatchwell also gave evidence on her behalf.
The Tribunal’s decision was published on 25 February 2008. It runs to 324 paragraphs. Paragraphs 1 to 126 are the work of the President, with the first 86 paragraphs discussing and reaching conclusions about the legal principles to be applied, paras 87 to 121 commenting on some of the expert evidence, and paras 122 to 126 commenting on a strike-out summons issued by the Commissioner (on which no order has been made, either by the Tribunal or by the Court of Appeal). Paragraphs 127 to 316 are the work of Mr Lo, and are almost exclusively concerned with the technicalities of the valuation exercise. The final eight paragraphs are written by the President, and deal with the disposal of the appeal. The rateable value of the site for 1997-98 was to be reduced from $29,640,000 to $26,880,000, and the Commissioner was directed to amend the Roll accordingly.
There is not in the Tribunal’s decision any discernible difference of opinion between the President and Mr Lo. Their contributions to the decision are clearly intended to be read as a whole. But it is convenient, in considering such a lengthy written decision, to have a general outline of its arrangement.
In paras 7 and 8 of its decision the Tribunal cited the answer given by this Court to point 4 in Agrila (para 25 above) and observed that this was the first case in which the Tribunal had to apply the guidance in Agrila. The Tribunal went on to observe (para 18) that the first issue was to determine the implications of the agreed fact that the mode and character of the occupation of the site was that of a development site.
That was indeed the first and essential issue of law to be resolved, and the Tribunal’s discussion and resolution of it occupy paras 19 to 86 of the written decision. It is essentially the only issue on the present appeal, although the terms of the order granting leave to appeal to this Court (para 27 above) show that it can be separated out into a number of sub-issues. It would not be helpful to attempt any brief summary of the Tribunal’s reasoning in this part of its decision. It will be better to consider, point by point, the sustained attacks on its reasoning put forward in the written and oral submissions made on behalf of Best Origin.
There is no longer any dispute as to the more technical questions of valuation methods and judgment that were addressed mainly by Mr Lo. It is sufficient to record that the Tribunal approved the use of the CB method (described in professional guidance notes set out in para 137 of the decision). It rejected various arguments put forward by Best Origin in contending that the $760 m premium paid for the site, even after adjustment, was not the correct starting-point for the CB calculation. It took the correct starting-point to be an adjusted amount of $611 m (para 206). It then gave detailed consideration to the appropriate decapitalization rate and decided that the rate of 4.4 per cent put forward by Mr Tang was reasonable and appropriate (para 235). 4.4 per cent of $611 m is $26,884,000, but in the final disposal (para 322) it was rounded to four significant figures, that is $26,880,000.
The proceedings in the Court of Appeal
Best Origin appealed to the Court of Appeal (Tang Ag CJHC, Kwan JA and A Cheung J) which on 19 November 2010 dismissed the appeal after a five-day hearing ( 1 HKC 488). The main judgment was given by Tang Ag CJHC, with whom the other members of the court agreed.
As with the decision of the Lands Tribunal, it would not be helpful to attempt a brief summary of the judgment of Tang Ag CJHC. It will be better to consider it, so far as necessary, by reference to the written and oral submissions made on behalf of Best Origin. But before coming to those submissions point by point, it will be useful to take a broader look at two topics of central importance to the argument, that is the rebus sic stantibus principle and the rating hypothesis of a yearly tenancy.
The rebus principle
“Rebus sic stantibus” is a Latin phrase with the literal meaning “things standing thus” or, in rather more colloquial English, “as things stand”. It has been familiar to lawyers for centuries, but originally in the context of the law of treaties, a branch of public international law. In that context it refers to the doctrine (well-recognized in principle, but controversial in its practical application) that any treaty must be taken to contain an implied term that it may cease to bind the parties after a fundamental change of circumstances. In that context it goes back to the founding fathers of international law, including Grotius in the 17th century and Vattel in the 18th century. It is still relevant today, and is reflected in Article 62 of the 1969 Vienna Convention on the Law of Treaties (see for instance Huang, The Doctrine of Rebus Sic Stantibus in International Law – a Dissertion (Shanghai 1935); McNair, Law of Treaties (1961) pp 681-691; Oppenheim, International Law, 9th ed (1996) pp 1304-1309).
The first known use of the expression in the context of rating is in the judgment of Cockburn CJ in R v Fletton Overseers (1861) 3 E&E 450, 465. In this context it is a convenient shorthand expression by which to identify the general effect of statutory provisions such as those found in the three paragraphs of s 7A(2) of the Rating Ordinance: that is assumptions as to
the state of the tenement being the same,
relevant factors affecting the mode or character of occupation of the tenement being the same, and
the state of the locality being the same.
In Hong Kong the provisions have a further secondary function, that is to avoid the need for projection of values by bridging the gap (typically nine months) between the relevant date and the date when the valuation list comes into force. As was explained by the Lands Tribunal (President: HH Judge Cruden) in China Light & Power Co Ltd v Commissioner of Rating and Valuation  RA 475, 482-483, this was the purpose of the introduction of s 7A in 1981.
In the context of rating the expression is therefore no more than a convenient shorthand for referring to statutory assumptions which are (as appears from many of the authorities cited in argument) applied in a fairly flexible way in order not to depart too far from the realities of the situation. (These assumptions, as well as the assumption of a yearly tenancy, are part of the rating hypothesis, so that there is some overlap with the next section of this judgment.) The fact that the shorthand is expressed in an ancient classical language should not create the impression that it embodies some fundamental principle of justice. Its function is more mundane, that is to provide some reasonably straightforward statutory assumptions in order to simplify the task of official valuers (and the task of tribunals reviewing their decisions) in administering a tax charged on an annual basis. The provisions set limits on the need for valuers and tribunals to inquire into the likely economic effect of future alterations in the structure of a tenement, or in the use to which it is put, or of possible future changes in the neighbourhood where the tenement is situated.
That is the purpose of the provisions. They do not require the tribunal to disregard what Lord Pearson, in Dawkins  2 AC 366, 393 referred to as “a present probability of a future happening”, because “the present probability affects the present value of the hereditament” (in that case part of the hereditament, a factory, was likely to be demolished for road-widening purposes within about a year).
The provisions must also be applied so as to take account of the intrinsic character of the tenement in question. Mr Howell QC (for Best Origin) mentioned as examples a tenement consisting of a gravel pit or a landfill site. If a gravel pit or a landfill site is in active use it is inevitable, and intrinsic to its character as a tenement, that it will change appreciably in the course of a year, as gravel is extracted or waste is put in. The same is true of an active development site, as the site is agreed to have been in 1997-98. Within the period of four years after the Government land auction (with the 1997-98 rental year starting shortly after the beginning of that four-year period) the site went from one with derelict police buildings awaiting demolition to a new 35-storey building for retail and office use. Throughout the period it must have been in a state of almost ceaseless change.
The rating hypothesis: the problem of the yearly tenancy
“The rating hypothesis” is not a term of art, but it is often used to identify the hypothetical test laid down by statute as the basis for the ascertainment of rateable values. That test is laid down in Hong Kong by the joint effect of ss 7(2) and 7A(2) of the Rating Ordinance (and in England, by similar statutory provisions). The general legislative purpose of the test is to achieve uniformity in valuation for rating purposes, and fairness as between one ratepayer and another. As Hodson LJ said in Humber Ltd v Jones (Valuation Officer) (1960) 6 RRC 161, 166:
It is clear that it is necessary to set up some standard. The actual rent paid would not be an adequate standard for rating purposes, because it might be complicated by there being a small rent because of the payment of a premium, or perhaps because it was an old rent fixed a long time ago when the economic value of the hereditament was low.
The part of the test laid down in s 7A(2), that is the rebus principle, has already been considered. It is now necessary to consider the part of the test laid down in s 7(2), and in particular the yearly nature of the tenancy which it postulates: “the rent at which the tenement might reasonably be expected to let, from year to year, ....” on stated assumptions as to the obligations of the landlord and the tenant respectively under the tenancy. It is the statutory requirement of a hypothetical tenancy from year to year that has caused great difficulty in this case, as it has in numerous other reported and unreported decisions for more than a century. In the paragraphs that follow the expression “the rating hypothesis” is directed primarily to the yearly character of the tenancy postulated by s 7(2).
At common law a yearly tenancy can be terminated by either party on six months’ notice expiring at the end of a year of the tenancy. The rating hypothesis disregards statutory interventions in the common law of landlord and tenant, whether they take the form of restrictions on the level of rent or protection of security of tenure (Poplar Metropolitan Borough Assessment Committee v Roberts  2 AC 93). So in relation to premises used for heavy industry or for the purposes of a public utility such as a power station, involving large-scale investment in expensive infrastructure and plant, there is often great difficulty in the application of the hypothesis. For instance Willmer LJ said in Humber Ltd v Jones (Valuation Officer) (1960) 6 RRC 161, 171:
Bearing in mind that the statutory hypothesis is a tenancy from year to year, everybody agrees that, in the world as it is, no sane manufacturer would take a tenancy from year to year of a factory in which he is going to install valuable machinery and for which he is going to assemble a skilled labour force to work for him.
Lord Millett NPJ applied these observations to the Lamma Island power station in Hong Kong Electric Co Ltd v Commissioner of Rating and Valuation  4 HKC 509, para 146.
The same considerations apply to a development site on which major construction work is proceeding. In this case the Tribunal, echoing the observations just cited, stated (para 29):
Obviously, there are tenements that no tenant would take up a tenancy from year to year in the real world. Lord Pearce [in Dawkins  2 AC 366, 381-382] gave the examples of sewage works, portion of railway lines. Development sites fall within the same category. In the real world, no developer with his right mind will take up a site for development based on a yearly tenancy.
Since the 19th century courts have struggled with this problem. One theme that emerges strongly is that although a yearly tenancy can be terminated at relatively short notice, it will unless terminated continue from year to year, indefinitely. One of the earliest observations on this point, repeatedly cited in later authorities, is that of Cockburn CJ in Great Eastern Railway Co v Haughley (1866) LR 1 QB 666, 679:
But I think it is one thing to start with the assumption that you are dealing with a tenancy from year to year, and another thing to say that the hypothetical tenant, in calculating what he can reasonably pay as rent for the premises, is necessarily to assume that his tenancy would not last beyond a year. I think the possibility of its longer duration is one of the surrounding circumstances which the tenant from year to year would take into account.
It is unnecessary to multiply citations on this point. But there are two decisions of the House of Lords that are of particular interest. One is the case of Railway Assessment Authority v Southern Railway Company  AC 266. Mention has already been made of the speech of Viscount Hailsham LC, part of which was quoted by Tang Ag CJHC in para 2 of his judgment in the Court of Appeal. But it bears repetition, not least because it goes on to refer to the CB method of valuation, adopted by the Tribunal in this case, as an appropriate solution to the problem. After describing the early history of rating in England Viscount Hailsham continued:
My Lords, this method of assessing liability for rates was no doubt very suitable for the simple conditions which prevailed in the reign of Queen Elizabeth; but when it was sought to apply it to modern conditions, and especially to the assessment of great public utility undertakings, such as railways, waterworks, gasworks and the like, whose operations might easily extend over great areas, stretching far beyond the limits of any individual parish, the greatest difficulty was found in adapting the statutory provisions to the facts of such a case. The Courts, more than once, appealed to the Legislature to relieve them from the burden thus cast upon them; see, for example, the judgment of Lord Campbell in Reg. v. Great Western Ry. Co. [(1851) 15 QB 379, 396, 397] and the judgment of Wightman J. in Reg. v. West Middlesex Waterworks Co. [(1859) 1 E&E 716, 728]. Unfortunately, the Legislature did not see its way to intervene and the Courts were compelled to evolve a system of calculation, with the assistance of the expert advisers to the rating authorities and to the undertakings, which involved a number of very difficult assumptions, but which received the approval of your Lordships’ House on a number of occasions, and which became recognized as the standard method of assessing the rateable value of hereditaments in these cases. So far as the railways were concerned, the method adopted was to divide the undertaking into those portions which were regarded as directly productive of profit, such as the permanent way, and those portions which were regarded as only indirectly productive, such as stations and other hereditaments of that kind. The latter were usually valued by taking the cost of erection and allowing a percentage upon that cost, on the theory that the companies would not have erected them unless they had regarded them as worth at least a fair average rate of interest upon the money invested in their construction. This method was commonly described as the ‘contractor’s basis.’
The other case that calls for further mention is Dawkins  2 AC 366. The ratepayers owned a factory, part of which was compulsorily acquired for road-widening. It was then let to the ratepayers on a yearly tenancy, which was to be terminated when the road works started. So there was an actual yearly tenancy which really was likely to be terminated in the near future, as well as the hypothetical yearly tenancy postulated by the rating hypothesis. The House of Lords was split, but the majority decided that the prospect of early termination, in the real world, should be recognized in applying the hypothesis.
Lord Guest, who was one of the dissenting minority, said at p 380:
Some difficulty may have been caused by undue emphasis on the expression ‘year to year’ as if this was limited to a yearly tenancy. The expression ‘taking one year with another’ which appears in the Valuation (Metropolis) Act, 1869, quoted in Poplar Metropolitan Borough Assessment Committee v Roberts  2 AC 93 and treated as having the same effect as a tenancy ‘from year to year’, makes it clear that, although as a matter of valuation a yearly tenancy is to be assumed, the hypothesis is that it will be of indefinite duration.
Lord Wilberforce (in the majority) made a thorough review of the old cases and referred to “the practical approach” of the judges. He also referred to “taking one year with another” as having always been regarded as an identical test. He concluded at p 387:
So we should regard the words ‘from year to year’ as meaning no more than that the tenancy is not a fixed or definite one; it is one of indefinite duration, determinable by notice, but not, I would think, according to the technicalities governing the giving of notice in tenancies of this kind.
Lord Pearson (also in the majority), after citing the observations of Cockburn CJ in Great Eastern, saw no inconsistency with the rebus principle (at p 393):
It was also said on behalf of the appellant that equality of rating requires that each hereditament should be valued as it now is – rebus sic stantibus – and the prospect of a future partial destruction of it must be disregarded. But it seems to me that this point can be turned against the appellant. In the expression rebus sic stantibus which are the res? In other words, which are the factors to be taken into account in order to produce equality of rating? There is, in this case, a present probability of a future happening, and the present probability affects the present value of the hereditament.
These and other authorities cited to the Court show that over the years the courts and tribunals have understood and applied the statutory hypothesis in a practical and flexible way so as to assess rateable values as fairly and uniformly as possible. There have been a few cases in which it has been held that a rateable tenement (or hereditament) has a nil rateable value. The court was shown two: Black v Oliver  1 QB 870 and Hoare (Valuation Officer) v National Trust  RA 391. These were both cases on very unusual facts: at one extreme a house that was unfit for human habitation, but occupied by an unfortunate tenant who might lose her claim to be rehoused if she moved out; and at the other extreme historic mansions owned by the National Trust, which on the rating hypothesis was the only possible tenant for the buildings, which were very expensive to maintain. The extreme nature of these examples serves to emphasize that despite the difficulty of applying the hypothesis to some types of tenement, tribunals and courts do their best to apply it in a practical way, and it is very unusual for a rateable tenement or hereditament to be found to have no rateable value. The well-known case of Consett Iron Co v Assessment Committee for North Western Area of Durham  AC 396 is a striking example of this. A loss-making coalmine, with no prospect of early economic recovery, was held to have a significant rateable value, despite the yearly tenancy postulated by the rating hypothesis.
Both leading counsel agreed that there is some difference between the Hong Kong case law and the English case law as to how far the rebus principle permits a valuer to contemplate physical changes in a tenement (on the one hand) and changes in its use (on the other hand), and the interaction between these matters. On the authorities as they stand that appears to be so, although the point has not yet been definitively considered by this Court. It is not necessary to go into it on this occasion. The point is of some general importance, but in the context of this appeal it is a matter of fine tuning. It cannot be determinative of the more fundamental issue now before the Court. So far as it has anything to contribute to the resolution of this appeal, I would proceed on the basis that the Hong Kong approach is rather more liberal.
The rating hypothesis: how far should it be taken?
In Hoare (Valuation Officer) v National Trust  RA 391, 408 Schiemann LJ stated,
The statutory hypothesis is only a mechanism for enabling one to arrive at a value for a particular hereditament for rating purposes. It does not entitle the valuer to depart from the real world further than the hypothesis compels.
Similarly in the same case Peter Gibson LJ drew attention to other areas of law, such as compulsory purchase and taxation, in which statutory hypotheses are regularly encountered. He cited what Hoffmann LJ said about valuation in a capital transfer tax case, Inland Revenue Commissioners v Gray  RVR 129, 136:
It cannot be too strongly emphasized that although the sale is hypothetical, there is nothing hypothetical about the open market in which it is supposed to have taken place.
The same applies, equally emphatically, to the hypothetical yearly tenancy postulated by the Rating Ordinance. It must take account of real market conditions: Hong Kong Electric Co Ltd v Commissioner of Rating and Valuation  4 HKC 509, 549, para 147(d), where Lord Millett NPJ set out with approval the Tribunal’s summary.
So Hoare is authority for the proposition that the rating hypothesis should not be taken further than is necessary, and the decision of the House of Lords in Dawkins is to the same general effect. But how, in this context, is necessity to be judged? Mr Howell cited a passage from the speech of Lord Asquith in East End Dwellings Co Ltd v Finsbury Borough Council  AC 109, 132-133:
If you are bidden to treat an imaginary state of affairs as real, you must surely, unless prohibited from doing so, also imagine as real the consequences and incidents which, if the putative state of affairs had in fact existed, must inevitably have flowed from or accompanied it .... The statute says that you must imagine a certain state of affairs; it does not say that having done so, you must cause or permit your imagination to boggle when it comes to the inevitable corollaries of that state of affairs.
Lord Asquith’s epigrammatic saying has often been cited, one such citation being by the Lands Tribunal in China Light & Power Co Ltd v Commissioner of Rating and Valuation  RA 475, 485. But I respectfully doubt whether it is entirely consistent with the purposive approach that courts now take on issues of statutory construction. Any statutory hypothesis has an element of “makebelieve” (the expression used by Willmer LJ in Humber (1960) 6 RRC 161,171), since it is treating some state of affairs as real regardless of whether it exists, or even if it is known not to exist, in the real world. Identifying the “inevitable corollaries” of a hypothetical state of affairs may be problematical.
The trend of more recent authorities can be seen in cases such as Inland Revenue Commissioners v Metrolands (Property Finance) Ltd  1 WLR 637, in which Nourse J said:
When considering the extent to which a deeming provision should be applied, the court is entitled and bound to ascertain for what purposes and between what persons the statutory fiction is to be resorted to.
This emphasis on statutory policy and purposes was approved by the Court of Appeal and the House of Lords in Marshall v Kerr (1993) 67 TC 56,  1 AC 148, although reference was also made to inevitable consequences and incidents, language clearly derived from East End Dwellings.
This reasoning was developed by Neuberger J in Jenks v Dickinson  STC 853, 878:
It appears to me that the observations of Peter Gibson J, approved by Lord Browne-Wilkinson, in Marshall indicate that, when considering the extent to which one can ‘do some violence to the words’ and whether one can ‘discard the ordinary meaning’, one can, indeed one should, take into account the fact that one is construing a deeming provision. This is not to say that normal principles of construction somehow cease to apply when one is concerned with interpreting a deeming provision; there is no basis in principle or authority for such a proposition. It is more that, by its very nature, a deeming provision involves artificial assumptions. It will frequently be difficult or unrealistic to expect the legislature to be able satisfactorily to [prescribe] the precise limit to the circumstances in which, or the extent to which, the artificial assumptions are to be made.
This passage was approved by the Supreme Court of the United Kingdom in Commissioners for HM Revenue and Customs v DCC Holdings (UK) Ltd  1 WLR 44, para 39 (and para 40 gives an example of apparently inevitable consequences which could not sensibly have been intended).
The statutory purpose of the rating hypothesis is, as already noted, to provide a pattern or template for ascertaining rateable values, in a uniform and reasonably simple way, and on a yearly basis. It is important to note that it is not only the hypothetical tenancy, but also the actual process of assessment, that is on a yearly basis. Mr Holgate pointed out that even in a situation in which the hypothetical tenant could expect to remain as tenant for a long time (such as Humber (1960) 6 RRC 161, 169, where there was evidence that “if the landlord gave notice to the tenant in a case of this kind, he would know that he would never find another tenant to go into that factory”), the Rating Ordinance requires a rateable value to be fixed for every rateable tenement on a yearly basis. The prospective length of the hypothetical tenancy, and events that may or may not happen in the course of its duration, are relevant only if and so far as they affect the current rating year; similarly in the next rating year, and so on. The more remote such future happenings are, in time or in probability (or both), the less plausible is the argument for treating them as within the scope and legislative purpose of the rating hypothesis. It will be necessary to come back to this in considering various disputed points on the Tribunal’s decision.
The Rent Regulation revisited
In Agrila the Court of Appeal  1 HKC 175 held that s 2 of the Rent Regulation was nugatory in its effect. This Court reversed that conclusion. The legislative purpose of s 2 is, as Sir Anthony Mason NPJ put it (para 49, 99E-F) “to overcome the problem that building sites are not rateable tenements for the purposes of the Rating Ordinance”.
Mr Howell, without seeking to challenge that statement of the section’s legislative purpose, pointed out that the section’s scope is not limited to development sites. In the section (the text of which is set out at para 24 above) the word “development” (defined in s 1) is not actually used at all (although the word does appear in the heading of that section), but the word “developed” (used twice) must be intended to have a meaning corresponding to the definition, and it is reasonably clear from the context that it refers to the completion of a development. The words of the deeming provision, “as if the leased land were a tenement liable for assessment to rates,” make sense only in relation to land which would, in the absence of the deeming provision, not be liable for assessment to rates. It could apply, at least in theory, to agricultural land and certain other categories of land or buildings which under s 36 of the Rating Ordinance are exempt from liability for assessment, and so are not rateable tenements. But it must be borne in mind that these provisions are relevant only in relation to land comprised in applicable leases (as defined in s 3 of the Rent Ordinance). In practice, the importance of s 2 is in relation to active development sites.
That is confirmed by the legislative history of the Rent Regulation, to which Sir Anthony Mason NPJ referred in paras 67 to 74 (102J-104C) of his judgment under the principle in Pepper v Hart  AC 593. Each side sought to derive some advantage from this material. Mr Howell pointed to the closing remarks of the Secretary for Planning, Environment and Lands in opposing Mr Arculli’s amendment (Hansard for 25 June 1997, pp 812-813):
There is no question of the Administration redefining the rateable value. The purpose of section 2 is simply to enable the Administration to ascertain the rateable value for the purpose of determining the amount of Government rent payable as required by the Joint Declaration and the Basic Law, yet leaves room for the situation where a minimal rateable value or [nil-agreed correction] rateable value is ascertained in which case it is accepted that no Government rent would be payable.
Mr Holgate pointed out that the original wording of s 2 had provided for the rateable value of newly granted sites to be 5 per cent of the market value of the land (p 811) and that the current wording (which Mr Arculli’s amendment proposed to change) followed closely the general lease conditions (p 812 - these conditions are referred to in para 58 (101B-E) of the judgment of Sir Anthony Mason NPJ).
Sir Anthony Mason NPJ concluded on this point (para 74, 104A-B) that “the legislative history strongly confirms the meaning already placed on s 8(2) and regulation 2”. By that he did not of course refer to the judgment of the Court of Appeal, which was reversed, but to para 49 (99E-H) of his own judgment. Some reference has already been made to that paragraph but it is best to set it out in full:
Viewed in the light of these well-established principles of rating law, the purpose of reg.2 seems to be reasonably clear. It is to overcome the problem that building sites are not rateable tenements for the purposes of the Rating Ordinance. The regulation achieves this purpose by providing that the rateable value of the leased land before any part of it is developed shall be ascertained ‘as if the leased land were a tenement liable for assessment to rates under the Rating Ordinance’. The regulation says nothing about how the rateable value is to be ascertained. That function remains to be dealt with, as s.8(2) of the Rent Ordinance prescribes, in accordance with ss.7 and 7A of the Rating Ordinance. The regulation makes no attempt to displace the operation of these sections. So the preferable meaning to be given to the words quoted above is that they require the rateable value of the leased land to be ascertained on the assumption that it is rateable tenement.
There was also some discussion, in the course of argument, about the contrast between s 2 and s 4 of the Rent Regulation. Section 4 contains the LARV rule as described in para 24 above. Mr Howell suggested that as s 4 undoubtedly looks to the past, it would be odd if s 2 were to be construed as looking to the future in order to charge tax by reference to the completed development. But against that it might be said that what the sections have in common is that neither seeks to levy tax in respect of value added by expenditure on construction works during the development (or redevelopment).
A route map for discussion of the issues
It will be apparent from the preceding sections of this judgment that the Court’s task, with the assistance of counsel’s very thorough written and oral submissions, is to decide questions of some difficulty and great public importance. The essential issue is one of statutory construction, but it is an area of law with a long history and a lot of authority. The best way forward may be to simplify and rearrange the issues set out in para 27 above, as follows:
Does s 2 of the Rent Regulation create a special regime?
What is the content of the rebus principle in relation to active development sites falling within s 2?
How does the yearly tenancy postulated by the rating hypothesis apply in relation to those sites?
Each of these questions (which inevitably overlap to some extent) must be taken in three stages:
Was it decided by this Court in Agrila?
Did the Tribunal make a significant error of law, either by misunderstanding or misapplying the principles and guidance to be found in Agrila, or by getting the law wrong on a point not covered by Agrila?
Did the Court of Appeal correct any part of the Tribunal’s reasoning which was unsound, or reinforce its reasoning when it was sound?
As to what Agrila decided, valuable guidance is available in the judgment of Ribeiro PJ in Hong Kong Electric Co Ltd v Commissioner of Rating and Valuation  4 HKC 509, 521-528, paras 24-58, with which the other members of this Court agreed. As to the Court of Appeal’s decision in this case, it may be noted that although this appeal is in form an appeal from the judgment of the Court of Appeal, counsel’s arguments have largely focused on the decision of the Tribunal. But some references have been made to the judgment of Tang Ag CJHC in the Court of Appeal. In the discussion that follows “CFA” refers to the judgment of Sir Anthony Mason NPJ in Agrila, “LT” to the Tribunal’s decision in this appeal, and “CA” to the judgment of Tang Ag CJHC in the Court of Appeal.
Does section 2 create a special regime?
This question, if it is to be answered shortly, admits of only one answer: No. Sir Anthony Mason NPJ made that clear (CFA para 49, 99E-H, set out in full in para 68 above). It is also spelled out in the answer to point 1, and reinforced by the answer to point 2 (“This point does not arise”), and by the reference to s 7(2) of the Rating Ordinance in the answer to point 4 (the terms of which, as originally formulated, were not satisfactory: CFA para 98, 109C-D).
This is however an area in which any one-word answer is incomplete, and may give rise to misunderstanding. Section 2 altered (for the purposes of Government rent) something which had until then been regarded (CFA para 48, 99B-E) as a fundamental proposition of rating law, without (CFA para 49, 99E-H) saying anything about how the rateable value of a development site was to be ascertained. This Court has held that the alteration cannot be regarded as nugatory. It was therefore necessary to go back to first principles in considering the implications for the ascertainment of rateable value of this novel enactment which requires development sites to be treated as rateable tenements (whereas previously they had been treated as not rateable, because they were regarded as not in actual and/or beneficial occupation).
Sir Anthony Mason NPJ touched on this point, rather inconclusively, in a passage emphasizing the distinction between rateability and rateable value (CFA paras 86 and 87, 106F-107B). Then after a reference (CFA para 89, 107F) to a well-known passage in the judgment of Scott LJ in Robinson Brothers (Brewers) Ltd v Houghton and Chester-le-Street Assessment Committee  2 KB 445, 468-469 as to the need for the value to take into account “every intrinsic quality and every intrinsic circumstance which tends to push the rental value either up or down”, Sir Anthony emphasized the importance, in ascertaining the rateable value of a development site, of recognizing its character as such a site (CFA para 94, 108E-F):
It follows that, in ascertaining the rateable value of the sites, it is permissible to have regard to their character as development sites for that is an intrinsic characteristic of each property. Having regard to that characteristic entails taking account of the likelihood of development taking place and proceeding to completion.
This is an important passage in which the use of the words “character” and “characteristic” is significant, since s 7A(2)(b) of the Rating Ordinance refers to “the mode or character of occupation”. Like a gravel pit or a landfill site in active use, an active development site has an intrinsic characteristic of change. In the case of a development site the direction of change is, in the normal case, towards completion of the development.
The Tribunal recognized the importance of this point. It is worth noting that all the passages in its decision relevant to this appeal come under the heading (immediately after LT para 18), “The implications of the site being a development site”. The point is developed in a passage starting at LT para 26:
In our view, it is almost self-evident that once it is concluded that the site is a development site, there must be a sufficient likelihood of change of use so as to affect the rent that a hypothetical tenant would pay.
The change of use referred to is of course the completion of the development, when (in this case) the new building would be available for letting as over a hundred retail and office units. The discussion of this point continues to LT para 29, a passage that has already been quoted (para 50 above). It is picked up again in LT paras 33 and 52 to 54. These are among the passages that Mr Howell has singled out for criticism.
Tang Ag CJHC (CA para 36) expressed his complete agreement with LT para 54, part of which is as follows:
It is difficult to see how the development potential of a development site can properly be taken into account if the rating hypothesis is to be construed in such a way to confine the expected duration of occupation to the period of construction or a short time afterwards for marketing of the units .... The intrinsic quality in terms of the development potential of the site (which is the actual purpose of occupation in the real world) will not be taken into account. It would become the value of a bare construction site without any development potential. Such an approach is inconsistent with the judgment of the Court of Final Appeal in Agrila and it is also against the principles set out in the other authorities mentioned above [notably Robinson Brothers and Consett Iron].
Tang Ag CJHC went on (CA paras 36 and 38) to answer an objection raised by Mr Howell in relation to CFA para 94 (108E-G). He returned to mode and character of occupation in CA para 60.
On this question as a whole I would conclude that although s 2 of the Rent Regulation certainly has not introduced a new regime displacing ss 7 and 7A of the Rating Ordinance, it has posed a new problem as to how the new deeming provision in s 2 is intended to interact with the rating hypothesis in ss 7 and 7A. The solution to that problem has in my view been correctly identified as starting with the recognition of the intrinsic character of these sites as development sites.
What is the content of the rebus principle in relation to active development sites?
Sir Anthony Mason NPJ referred (CFA paras 86-88, 106F-107D) to Dawkins  2 AC 366, including Lord Wilberforce’s reference (at p 386) to the hypothetical tenant taking into account “not only any immediately actual defects or disadvantages (such as planning restrictions) but disadvantages, or advantages, which he can see coming”, and Lord Pearson’s reference (at p 393) to “a present probability of a future happening, and the present probability affects the present value of the hereditament”. Sir Anthony continued (CFA para 89, 107E-G):
Dawkins is an answer to the main thrust of the respondents’ case that s 7A(2)(a) and the rebus principle require that the valuation must be based on an actual tenement in its existing state. The point is that, although the rebus principle requires the tenement to be valued as in fact it is, the valuer must consider ‘every intrinsic quality and every intrinsic circumstance which tends to push the rental value either up or down’ (Robinson Brothers (Brewers) Ltd v Houghton and Chester-le-Street Assessment Committee  2 KB 445 at 468-469, per Scott LJ.
Sir Anthony Mason NPJ then applied these general principles to a tenement with the intrinsic character of a development site (CFA para 92, 108A-C):
The fact that occupation of a construction site does not enable a hypothetical tenant to make a profit during the construction period does not mean that the property has no rateable value. Occupation of the site may nevertheless be valuable and command a significant rent (London County Council v Erith  AC 562 at 591).
Then after a reference to Consett Iron he continued (CFA paras 93 and 94, 108C-F):
If there is a sufficient likelihood of a change of use of the property as would affect the mind of a hypothetical tenant and alter the rent he would pay for it in its existing state, that is a matter to be taken into account in the valuation. And the current occupier is to be regarded as a party who might become the hypothetical tenant (London County Council v Erith).
It follows that, in ascertaining the rateable value of the sites, it is permissible to have regard to their character as development sites for that is an intrinsic characteristic of each property. Having regard to that characteristic entails taking account of the likelihood of development taking place and proceeding to completion. But this does not mean that the sites should be valued as completed developments.
These are very important paragraphs in the Agrila judgment. They are expressed in general terms, because (as Ribeiro PJ pointed out in Hong Kong Electric, para 29) this Court was in Agrila dealing with issues in principle and in relation to 59 development sites, on some of which (CFA para 13, 92I-93B) development had not advanced beyond the erection of boardings. The above passages emphasise the value of occupation (in rating terms, the value to the hypothetical tenant), this being the statutory intention behind s 7(2) of the Rating Ordinance: see Hong Kong Electric, para 65.
London County Council v Erith  AC 562 was concerned with a tenement which was incapable of ever yielding a profit, but was nevertheless rateable because its occupation was of value to the statutory undertaker in the performance of its functions. The facts are summarized in the headnote:
The London County Council were owners of land and premises consisting of a pumping-station and works, which they occupied and used as a necessary part of the metropolitan sewage system and to enable them to perform statutory duties. So long as the land and premises were used as part of the sewage system they were incapable of yielding a profit and the London County Council were practically the only possible tenants.
There is a single speech by Lord Herschell LC with which the rest of the Appellate Committee of the House of Lords concurred. It makes clear, in a long passage at pp 588-592, that occupation can be of value even though no pecuniary profit is derived from the occupation. There is a citation from the judgment of Fry LJ in R v School Board for London (1886) 17 QBD 738:
The term ‘sterility’ has been introduced into the cases because, as a general rule, a profit is produced; but it does not by any means follow that because there is no profit there is no value.
There is also a reference, at p 593, to what can be recognized as the CB valuation method.
The Tribunal cited CFA paras 92 and 93 (108A-E) at the beginning of its discussion of the implications of the site being a development site (LT paras 20 to 23). But the fact that Sir Anthony Mason NPJ chose to refer (in CFA para 93, 108D) to a change of use (rather than the successful completion of the development) seems to have led the Tribunal to suppose that a change of use would by itself necessitate the coming to an end of the hypothetical tenancy. This was in my view an error of law (but not one which ultimately affects the outcome of the appeal). It was an error because the hypothetical tenancy is to be regarded as continuing indefinitely, unless it is terminated by notice given by the hypothetical landlord. It is true that in the real world the letting of the completed building as more than a hundred separate tenements would lead to a multiplicity of new hypothetical tenancies as part of a valuation exercise undertaken in or about the year 2000. That was foreseeable, but it was irrelevant to the valuation exercise carried out as at 28 June 1997, for reasons already considered (para 64 above).
This error led the Tribunal to the notion (which emerges in a clear form in LT para 37) of the developer continuing to be entitled to occupation (in substance, to enjoyment of profits from the completed development) after the end of the hypothetical yearly tenancy. This tends to distort some of the Tribunal’s reasoning. Nevertheless the Tribunal understood and followed the essential point of what Agrila has to say about how the rebus principle applies to development sites: that the prospect of successful and profitable development is already, during the course of construction, an intrinsic characteristic of a development site. That is clear from LT paras 41 and 46-49.
In those paragraphs the Tribunal used the expression “development potential” on more than one occasion. Mr Howell was critical of this, and Mr Holgate was prepared to accept that the expression is ambiguous. It is an imprecise expression, but its general meaning is reasonably clear. Developers develop sites because of their development potential, which in economic terms is the prospect of profiting from the development. In this case the clearest indication of the site’s development potential is the sum of $760 m that Best Origin paid at the keenly contested Government land auction in December 1996. Paragraphs 28 and 29 of the agreed statement of facts show that this was not a freak result, but a reflection of the strong demand and limited supply on the market at that time. Mr Howell’s submission, which is central to his case, that the Tribunal felt bound to conjure up a demand that could not have existed consistently with the rating hypothesis, depends almost entirely on the yearly nature of the hypothetical tenancy. That submission is considered in the next section of this judgment.
In the Court of Appeal Tang Ag CJHC referred (CA paras 20 and 50) to CFA paras 90 and 92 (107G-I and 108A-C), and to the prospect of continuation of the hypothetical tenancy. In parts of his judgment he seems, as I read it, to have come close to commenting on and perhaps disagreeing with the Tribunal’s view that the developer might remain in occupation after the end of the hypothetical tenancy. In CA para 61 he described the point (correctly in my view) as irrelevant to the ascertainment of the rateable value of the site as a development site.
It may be worth mentioning, but only as a footnote, that in his submissions Mr Holgate insisted that the site should during the construction period be regarded as a bare site, with the building in course of erection on it disregarded. This submission must have been made in order to assist his secondary, or fallback, case as to the severance of the building from the land. It does not assist his primary case, and so far as it is relevant at all I do not think it is correct. The intrinsic character of a development site is to grow, in its physical appearance and in its value. It is not immediately productive of profit, but its long-term growth and productivity is the antithesis of the “sterility” referred to by Fry LJ in R v School Board for London.
On the second question I conclude that in Agrila this Court recognized the importance, in the application of the rating hypothesis, of the intrinsic character of a development site. To answer the question posed by Lord Pearson in Dawkins  2 AC 366, 393, the carrying out of the development was part of (indeed, the heart of) the res. The Tribunal understood this and applied it, though the Tribunal fell into error over the termination of the hypothetical tenancy. The Court of Appeal followed what this Court had said in Agrila, and rightly regarded the Tribunal’s error as irrelevant. Mr Howell’s rather sanguinary submission that the rebus principle has been eviscerated cannot be accepted. But the principle has been adapted to make it apply to the wholly novel situation brought about by s 2 of the Rent Regulation.
How does the hypothetical yearly tenancy apply in relation to development sites?
It is clear that in Agrila this Court was invited to give guidance as to how the yearly tenancy postulated by the rating hypothesis was to be applied by valuers to development sites falling with s 2 of the Rating Regulation. On this appeal counsel referred the Court at some length to the parties’ lists of assumptions and the responses to them filed in Agrila, and to extracts from the transcript of the oral argument in that case on 8 and 9 February 2001. However, in Agrila, this Court (wisely, in my respectful opinion) did not go very far in giving detailed guidance.
It was wise to do so because the appeal in Agrila was confined to preliminary issues of law affecting a large number of development sites in different areas and with differing prospects of construction work being undertaken in the immediate future (CFA paras 13-15, 92I-93D). Although there were apparently some agreed facts in relation to some specimen sites, there had been no Tribunal hearing to determine contested facts. The question posed at point 4 (CFA para 117, 113I-114A) was expressed in general terms that Sir Anthony Mason NPJ considered to be unsatisfactory (CFA para 98, 109C-D).
So far as this Court did deal with the point the relevant part of the judgment is CFA paras 79 to 95 (105B-108H). But this passage is, as the heading before para 79 (105B) indicates, concerned mainly with the likelihood of the development being commenced and completed. In other words, it is concerned mainly with the intrinsic character of a development site, and much of the passage has already been quoted or summarized in the last section of this judgment (paras 78 to 80 above). Sir Anthony Mason NPJ does in this part of his judgment refer to the hypothetical tenancy as being of indefinite duration, and there are references to Humber and China Light & Power (CFA paras 90 and 91, 107G-J). But apart from London County Council v Erith there is mention of the leading cases in which the House of Lords has confronted the problem of tenements whose character is such that in the real world no one would ever take them on a yearly tenancy. Probably these cases were not cited to the Court.
The relevant passage in the judgment ends as follows, in continuation of the second quotation at para 79 above (CFA paras 94 and 95, 108F-109H):
But this does not mean that the sites should be valued as completed developments. Nor does it mean that either of the Commissioner’s methods of valuation or what has been described as ‘the contractor’s method’ of valuation should be adopted. The appropriate mode of valuation, apart from what is prescribed by relevant principles of law, is a matter for the Lands Tribunal to determine. It is not for this Court to express an opinion about valuation or about the appropriateness of any method of valuation.
In these cases much will depend on the estimated duration of a yearly tenancy which the hypothetical tenant might secure. It might be sufficiently long to allow for completion of the relevant development, so that the hypothetical tenancy would extend eventually to such a situation.
I read this as a considered indication that further and more detailed guidance would have been premature. It would have to wait until there was an appeal from the Tribunal after it had heard evidence and legal argument, and reached its own decision based on its special skill and experience in valuation matters. The Tribunal’s decision in the present case, upheld by the Court of Appeal, gives this Court the opportunity to give further guidance.
I have already considered much of the Tribunal’s decision. I concentrate now on the significance of the yearly tenancy hypothesis, without repeating what has already been said about the “change of use” point. LT para 29 states:
Obviously, there are tenements that no tenant would take up a tenancy from year to year in the real world. Lord Pearce gave the examples of sewage works, portion of railway lines. Development sites fall within the same category. In the real world, no developer with his right mind will take up a site for development based on a yearly tenancy. In this sort of situation, the rating hypothesis of yearly tenancy cannot be permitted to operate in such a way to exclude a tenancy for that particular mode or character of occupation. Otherwise, it would be putting the cart before the horse and the rating hypothesis would frustrate instead of facilitating a proper valuation for rating purposes.
That paragraph, and para 33 which develops the same line of thought, were severely criticized by Mr Howell. His argument was that the Tribunal seemed to be under some compulsion to create a demand for a yearly tenancy for which there would be no demand in the real world.
I would accept some of this criticism, but only to a limited extent. I think that the Tribunal, lacking any detailed guidance from Agrila, started with a valuable insight. The Tribunal recognized that development sites, like sewage works and portions of railway lines, are tenements that no developer and no statutory undertaker would contemplate taking on a yearly tenancy. In that type of case the yearly nature of the hypothetical tenancy compelled “rating experts and the Courts to have recourse to hypotheses of a more or less violent character,” as Lord Cave LC put it in Kingston Union v Metropolitan Water Board  AC 331, 338. These methods were the profits (or “R&E”) method and the CB method, both of which are discussed in the speeches of Lord Cave and Lord Atkinson. They are methods under which the possibility of the hypothetical landlord serving notice to quit becomes an irrelevance.
Having set off in the right direction the Tribunal then, in my respectful opinion, took a wrong turning. Instead of following the public utility cases, where the yearly nature of the hypothetical tenancy becomes irrelevant, the Tribunal were tempted into elaborating the hypothesis, a course warned against in Hoare (Valuation Officer) v National Trust and the other cases mentioned in paras 62 and 63 above. One elaboration, the hypothetical tenant’s continued enjoyment after the change of use on completion of the development, has been discussed already. The other (LT paras 44-46) was an extension of the assumption that the hypothetical landlord and the hypothetical tenant “both are reasonable people, the landlord not being extortionate, the tenant not being under pressure”: Lord Denning MR in R v Paddington Valuation Officer ex parte Peachey Property Corp Ltd  1 QB 380, 412.
There is no doubt but that the initial bargaining as to the rent payable under the hypothetical tenancy must be supposed to take place between two parties acting reasonably. But it is an unwarranted extension of this (as Mr Howell rightly submitted) to suppose that the hypothetical landlord would refrain from acting in his own best interests if he could, by serving notice to quit soon after completion of the development, lawfully obtain a great financial advantage. The solution to the puzzle is not to elaborate the hypothesis by supposing that the hypothetical landlord would act contrary to his interest, but to recognize that it goes beyond the statutory purpose and scope of the rating hypothesis.
These are not however fatal defects in the Tribunal’s reasoning. The essential point was that it recognized (LT para 29) the feature that a development site has in common with some tenements owned by public utilities or statutory undertakers: that it is not immediately productive of income or profits, but is nevertheless of real value to the occupier. After straying a little from the right path by over-elaboration of the statutory hypothesis, the Tribunal came back to the right path with another valuable insight (LT para 72) as to the novelty of the situation created by s 2 of the Rent Regulation.
In LT para 78 it is stated to have been common ground that the building constructed on the site should be treated “for these purposes” (which were not precisely defined) as the property of the hypothetical tenant. Mr Howell told the Court that this was a misunderstanding, and that that concession was made only for the limited purpose of assessing the rateable value of the site by the R&E method (which was not the method advocated by the Commissioner, or adopted by the Tribunal). I readily accept that there was some misunderstanding here. The point is in any event irrelevant to the CB method of valuation (which the Commissioner did advocate from the outset, and which the Tribunal accepted). The CB method takes account of the initial expenditure on the acquisition of the site (subject to appropriate adjustments) but does not take into account the cost of constructing the new building. The adoption of the CB method, and its application in this case, are comprehensively covered in the part of the decision prepared by Mr Lo, and are not a separate issue in this appeal.
In the Court of Appeal Tang Ag CJHC, in his summary of Agrila, cited (CA paras 16 and 17) Lord Herschell’s observation in London County Council v Erith  AC 562, 595, that the words of the rating hypothesis:
only provide the means of arriving at what is the annual value of the premises
and the observations of Scott LJ in Townley Mill Company (1919) Ltd v Oldham Assessment Committee  1 KB 585, 643, that
the hypothesis of the imaginary tenant is intended to assist and not hinder that process.
He also cited (CA para 20) observations by Sir Anthony Mason NPJ (CFA paras 90 and 92, 107G-I and 108A-C) as to the indefinite duration of the yearly tenancy and the importance of occupation being of value to the hypothetical tenant, even if he receives no immediate rental income.
Tang Ag CJHC then addressed (CA para 25) Mr Howell’s reliance on the yearly nature of the hypothetical tenancy, noting that Best Origin’s primary case before the Tribunal had been that the only possible tenant would be a building contractor. He went on to refer to passages from the judgments of Scott LJ in Robinson Brothers (Brewers) Ltd v Houghton and Chester-le-Street Assessment Committee  2 KB 445, 475 and Lord Coleridge CJ in Smith v Birmingham (Churchwardens) (1889) 22 QBD 703, 705-706. These passages serve to reinforce the fundamental points that I take from the Tribunal’s decision, that is (i) that as a result of s 2 of the Rent Regulation, development sites present a novel problem, since they are potentially of great value to their developer, but do not produce income during the construction period; and (ii) that the solution to the problem lies in the public utility cases, which are analogous. Tang Ag CJHC also referred with approval (CA para 36) to a passage in the Tribunal’s decision (LT para 54) already cited (para 76 above).
Tang Ag CJHC then referred (CA para 39) to the prospect of profitable development (which was, of course, to be judged on evidence available in 1997). He observed that there was no evidence before the Tribunal that the development might be unprofitable. It is of course necessary to avoid hindsight. But the price paid by Best Origin in 1996 and the agreed facts about the property market at that time (para 84 above) indicate that there was every reason to expect that it would be profitable. CA paras 41 to 46 address the argument that the hypothetical landlord would be either unable or unwilling to obtain the completed building for himself by giving notice terminating the tenancy. I have already explained (para 97 above) why this point, and the further discussion (CA paras 50-58) about ownership of the completed building under the rating hypothesis, are in my view irrelevant.
For these reasons I conclude that the Tribunal and the Court of Appeal were correct in recognizing that s 2 of the Rent Regulation produced a novel situation which called for some development in this area of the law. It is not however a novel or unprincipled development. It extends to development sites, during the period of active construction while they are not producing income, the treatment which has since the 19th century been accorded, regularly and uncontroversially, to the unproductive parts of undertakings owned by municipalities or statutory bodies, such as sewage treatment works, waterworks, gasworks, and so on (there is a comprehensive list of 28 types of tenement in Halsbury’s Laws 5th ed Vol 70 para 166). This extension is of course only for the purposes of assessing Government rent; for rating purposes a development site is still regarded as not being a rateable tenement.
Earlier in this judgment (para 52 above) I thought it unnecessary to multiply citation of authority. But in conclusion it may be useful, in order to make good the point that this is not a novel or unprincipled development, to set out some of the principal authorities which establish (i) that the yearly character of the hypothetical tenancy has been interpreted as indicating an indefinite duration and equivalent to “taking one year with another”; (ii) that in the absence of clarifying legislation, the courts have had no option but to adapt the rating hypothesis to make it workable in relation to some tenements of a special character; and (iii) that the CB method has often been approved by the court as appropriate for tenements which, although of value to their occupiers, were not income-producing (though the detailed application of that method is a matter for valuation experts).
Most of these authorities are decisions of the House of Lords. Several of them were presided over by the Lord Chancellor, indicating the importance that has always been accorded to this branch of the law. Most of them have been mentioned already, but it may be useful to bring them together, in chronological order: R v Coventry Canal Co (1859) 28 LJ (MC) 102, 104 (Lord Campbell CJ); Great Eastern Railway Co v Overseers of Haughley  LR 1 QB 666, 679 (Cockburn CJ); Smith v Churchwardens and Overseers of Birmingham (1889) 22 QBD 703, 705-706 (Lord Coleridge CJ), 707 (Lindley LJ); London County Council v Erith  AC 562, 585, 588-593, 596 (Lord Herschell LC); Kingston Union v Metropolitan Water Board  AC 331, 338-339, 344-345 (Viscount Cave LC), 346-350 (Lord Atkinson); Railway Assessment Authority v Southern Railway Company  AC 266, 273-274 (Viscount Hailsham LC); Robinson Brothers (Brewers) Ltd v Houghton and Chester-le-Street Assessment Committee  2 KB 445, 468-471, 474-475 (Scott LJ, affirmed by the House of Lords  AC 321); and Dawkins (Valuation Officer) v Ash Brothers and Heaton Ltd  2 AC 366, 380 (Lord Guest), 382-384 (Lord Pearce), 386-388 (Lord Wilberforce) and 392-393 (Lord Pearson).
It is unnecessary to consider issue estoppel, or Mr Holgate’s fallback position on hypothetical ownership of the completed building. The striking-out application is no longer a live issue.
I would therefore dismiss Best Origin’s appeal and make an order nisi that the Commissioner’s costs should be taxed (if not agreed) and paid by Best Origin. In the event that either party contends for a different order, that party must lodge written submissions with the Registrar of this Court within 21 days of the handing down of this judgment. Any written submissions in reply should be lodged within 21 days thereafter. Failing any submissions to the contrary, the costs order nisi will become absolute.
Chief Justice Ma
The Court unanimously dismisses the appeal and makes an order for costs in terms of para 105.
 In the official published report the paragraphs of the judgment are not numbered. But most on-line versions have numbered paragraphs and paragraph numbers were used, for convenience, at the hearing. These paragraph numbers will be used in this judgment. The relevant page numbers in the official report are inserted in brackets.
John Howell QC and Nigel Kat, instructed by Woo, Kwan, Lee & Lo, for the appellant.
David Holgate QC and John Litton, instructed by the Department of Justice, for the respondent.
all rights reserved