ANALYSIS

Are T’sos Really Trusts?



Malcolm Merry*

Traditional Chinese forms of communal landholding, t’so and t’ong, are

habitually referred to by judges and authors as trusts. This article investigates
the basis for this description and examines whether and to what extent the
description is legally accurate. It concludes that there is considerable doubt as
to whether they really are trusts.

Introduction

The Chinese customary phenomena of collective ownership, t’so and

t’ong, by which much land in Hong Kong’s New Territories has long been
held, defy analysis in terms of the common law system. They have been
variously described as institutions, organisations, associations, entities,
endowments, settlements, funds and trusts. Those nouns are often qualified
by a word indicative of their familial nature (clan, lineage, ancestral or
hereditary) or of their purpose (worship, ritual, cult or charitable). These
descriptions are at best approximate, for the phenomena are emanations
of Ming and Tsing China.1 Their preservation in modern common-
law Hong Kong is a matter of irony if not perversity and is the result of
historical happenstance.
The most frequent noun applied to t’sos and collective landholding
t’ongs is the trust. The use of the language of trusts in this connection is
so pervasive that it has become axiomatic that these ancient institutions
are legally trusts. So when Deputy High Court Judge Lam (as then he
was) essayed a summary of the law relating to t’sos in June 2002, he
stated confidently that “it has been established that the concept of trust
is applicable to a Tso with the managers as trustees and the members as

*
Department of Professional Legal Education, Faculty of Law, University of Hong Kong. Whilst
retaining full responsibility for the text, the author gratefully acknowledges the assistance of
Prof Anthony Dicks QC and Dr James Hayes each of whom kindly commented on drafts of this
article.
1
Endowments for the worship of ancestors by t’so and t’ong became popular after edicts of
successive Tsing emperors in the 18 century encouraged the practice, although some existed
before then (information from Prof Anthony Dicks). In 1904, at least a quarter of the land in
the New Territories was estimated to be held by t’so or t’ong: Hayes, The Great Difference, Hong
Kong’s New Territories and Its People 1898-2004 (HKUP, 2006), p 182, n 17.

Binder for HKLJ.indb 669 12/17/2012 4:11:47 PM

 670 Malcolm Merry (2012) HKLJ

beneficiaries”.2 A similar, if more dogmatic, statement, made by Chu J

(as then she was) three months later, is contained in another summary.3
Likewise, when Hong Kong land law texts deal with t’so and t’ong,
the authors have no hesitation in labelling them Chinese customary
trusts.4 The practice has spread to historical and sociological accounts
of customary land tenure. So Professor Hugh Baker calls t’so and t’ong
ancestral trusts and Doctor James Hayes calls them customary trusts.5
Yet, how can these descriptions be legally accurate? The trust is a
concept of the common-law system. T’so and t’ong are concepts of
customary law. In their summaries both judges recognised this apparent
contradiction but drew no conclusions from it. Deputy Judge Lam was
content simply to note that the imposition of a trust upon a t’so “is an
application of concepts of English law to a Chinese customary institution”,
as if this was perfectly natural. He went on to say, however, that “by
reason of local circumstances” certain concepts in English trust law, such
as the rule against perpetuities, had been held to be inapplicable, without
observing that this was inconsistent with the idea that a t’so is a trust.
Chu J merely noted what Deputy Judge Lam had said. No alarm bell rang
that might have led the judges to trace the origin of the suggestion that
a t’so is a trust and to question its authenticity.6
It will be the contention of this article that whilst applying the
description “trust” to t’so and t’ong might seem convenient, it is at best an
approximation and is dubious legally. The assertion, which has become
an assumption, that these customary institutions are trusts is built on a
flimsy foundation which has never been tested at appellate level.
Most of the cases in which a customary land institution has been
described as a trust concern a t’so rather than a t’ong of the communal
landholding variety, so in what follows “t’so” will be used for simplicity
to indicate the customary institution. The description “t’ong” occurs
in respect of organisations which have no connection with the New
Territories or with customary collective landholding, whereas “t’so” is
always a description for clan and family land ownership named after

2
Leung Kuen Fai v Tang Kwong Yu (or U) T’ong or Tang Kwong Yu Tso [2002] 2 HKLRD 705
at 717.
3
In Tang Kam Wah v Tang Ming Yat (unrep., HCA 10141/1998, [2002] HKEC 1443) para 68.
4
Lee and Goo Hong Kong Land Law (3rd edn, 2009), pp 488–493; Wilkinson and Sihombing,
Hong Kong Conveyancing Vol. 1, Law and Practice I [31] and [46]; Nield, Hong Kong Land Law
(2nd edn), Ch 8.
5
H.R.D. Baker, A Chinese Lineage Village (Cass, London, 1968), Ch 4; in Ch 7 Baker calls them
“kinship-ritual trusts”; Hayes, The Great Difference, n 1 above, Ch 3; in the new introduction
to the reissue of The Hong Kong Region 1850-1911 (Hong Kong University Press, 2012), Hayes
refers to them as “corporate lineage trusts” (p xiv).
6
T’sos have been accepted to be trusts in subsequent cases, eg: Wong Kam Pok Tso v 李強 (unrep.,
DCMP 2197 2006, [2010] HKEC 469); Mak Lai Chuen v Lau Kar Yau [2009] 3 HKC 217.

Binder for HKLJ.indb 670 12/17/2012 4:11:47 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 671

a common ancestor. However, there is no distinction between the t’so

and the communal landholding ancestral t’ong which is material to the
question of whether they are trusts.7

Analogies with Trusts

A t’so certainly exhibits similarities to a trust. The founders of a t’so put

land into it, reminiscent of a settlor placing land upon trust. The income
from the land is used for the benefit of descendants of the founders
through the male line, those descendants resembling beneficiaries
under a trust. The income is often also used for religious, educational
and welfare purposes, which call to mind charitable trusts.8 T’sos have
managers, analogous to trustees. The land is usually registered in their
names as well as that of the t’so.
The appellation has gone beyond convenient approximation,
however. The courts in Hong Kong seem to regard t’so as settlement
trusts.9 Judges and some experts have taken to calling and treating the
managers as trustees in law and calling and treating the descendants as
beneficiaries at law. Judges have imposed trustee duties upon the former
for the protection of the latter.
This treatment is justified where the founder has made a declaration
of trust and explicitly appointed managers as trustees when dedicating
land to the t’so. In that event, the founder has chosen to adopt a
mechanism from the common law system so as to create a settlement of
the land upon trustees for the benefit of the members of the t’so. Such
a declaration is, however, rare and could have occurred only where the
t’so was set up during the 20th century, after the common law came to
the New Territories with British administration, and then only in the
event that the founder engaged a solicitor to set up the trust.10 The great
majority of t’sos are of greater vintage, having been founded under Tsing
rule or earlier and being in place when the British took over. Although
those t’sos were established entirely under Chinese customary law, yet the
judicial practice has been to treat them as trusts.

7
In the author’s experience, the colloquial expression “t’so t’ong”, embracing both types of
institution, is used by rural inhabitants.
8
But t’so and t’ong cannot be charitable in the legal sense since they do not benefit the public at
large.
9
As, for instance, Deputy High Court Judge Tang (as then he was) expressly did in Kan Fat Tat v
Kan Yin Tat [1987] HKLR 516.
10
An instance of such a settlement is that in Chu Tak Hing v Chu Chan Cheung Kiu [1968] HKLR 542.
Mak Lai Chuen v Lau Kar Yau [2009] 3 HKC 217, DC may be another example.

Binder for HKLJ.indb 671 12/17/2012 4:11:47 PM

 672 Malcolm Merry (2012) HKLJ

Tang v Tang

The source of this practice is a judgment of Mills-Owens J in 1970,

Tang Kai Chung v Tang Chik Shang11 (hereinafter Tang) in which the
institution of the t’so was first subjected to extended judicial analysis.
The analysis was performed with the benefit of expert evidence as to
Chinese law and custom, for the judge (in common with all members
of the Hong Kong judiciary at the time) professed no special knowledge
of these traditional institutions. The case was a dispute between
members of a t’so of the extensive Tang clan from the northwestern
New Territories: some wanted land held by the t’so in Demarcation
District 121 near Yuen Long to be partitioned, others did not. The
court decided that the law of partition did not apply because the t’so
land was intended to be held perpetually and not to be alienated. In the
course of a 42-page judgment, however, the judge described the nature
of a t’so and examined its relationship with modern common law and
statute. Amongst other matters he held that a t’so is a trust and the
managers are trustees who are subject to the provisions of the Trustee
Ordinance.
Early in the judgment, Mills-Owens J set out his understanding of
what a t’so was. “Speaking generally,” he said:

“A tso may be shortly described as an ancient Chinese institution of ancestral

land-holding whereby land derived from a common ancestor is enjoyed
by his male descendants for the time being for their lifetimes and so from
generation to generation indefinitely.”

The descendant automatically becomes entitled at birth to the interest in

the land for life, the judge explained, and on death that interest merges
with the interests of the surviving members of the t’so, automatically
enlarging their interests.12
The judge’s description was made with the benefit of having heard
expert evidence as to Chinese customary law. It will be noted that the
interest of the male descendant as described is an interest in the land,
not in the t’so, and that there is no mention of the interest being merely
beneficial. Indeed, in defining a t’so, Mills-Owens J does not mention the
managers, the putative trustees and legal owners of the land, at all. This
is not surprising because in giving this short definition the judge was

11
[1970] HKLR 276.
12
Tang at pp 279–280. The qualifications that Mills-Owens J was speaking generally and was
attempting a short description should not be overlooked.

Binder for HKLJ.indb 672 12/17/2012 4:11:47 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 673

confining himself to customary law and Chinese customary law has no

conception equivalent to a trust.
As described by the judge and by experts in Chinese customary law,
the interest of a member of the t’so is a joint interest with other current
living members. Entitlement to that interest arises automatically at
birth and is not derived through some other person. The interest is not
heritable; in that respect the interest is more like joint tenancy13 with
its right of survivorship than tenancy-in-common which passes into the
estate of a deceased co-owner. Unity of possession, unity of interest and
unity of title are also present as with joint tenancy but not the fourth
unity, that of time. According to Mills-Owens J’s account, the interest of
each member vests at birth and ends at death, occasions which will occur
at different times for each member.
The character for t’so (祖) simply implies the male descendants
of the person whose name precedes that character.14 The method of
customary landholding under a t’so defies categorisation in common
law terms. But what is clear is that it is a form of co-ownership and
that each member (living male descendant) receives a direct interest
in the land. It is therefore on the face of matters perplexing that later
in his judgment Mills-Owens J declares the t’so to be a trust with the
managers (registered with the District Office) as trustees, for that
would mean that the members would have merely an equitable interest
as beneficiaries and that the sole legal owners of the land, who should
be registered at the Land Registry as such, would be the managers and
not the t’so.
What impelled Mills-Owens J to that declaration? The judge relied
on the terms of the Crown lease of 1905 under which the land had been
granted (or perhaps more accurately, re-granted) to the t’so by the new
administration in the name of King Edward VII for a term of 75 years
with right of renewal for a further 24 years less three days.15 As is common
with rural New Territories land, the lease took the form of a block grant,
that is to say one deed of lease covering all the lots in an entire district.
This had been used in order to simplify the process of making mass grants
following the acquisition of the New Territories in 1898, since there were
many thousands of lots for which Crown leases had to be issued and the
administration was anxious to levy Crown rent on the owners as soon as

13
Or the defunct method of joint ownership, coparcenary.
14
As recorded on the Land Register and the Crown lease and before that on the Ching land
register.
15
Thus, if the option were exercised, taking the term to 27 June 1997, ie shortly before expiry of
the United Kingdom’s holding of the New Territories.

Binder for HKLJ.indb 673 12/17/2012 4:11:47 PM

 674 Malcolm Merry (2012) HKLJ

possible so as to meet the costs of running the “new territory”.16 The body
of the lease was printed and in general terms, applicable to all the lots in
the district. Details of the individual lots which were the subject of the
grant were given in a schedule at the end of the lease. The schedule was
in common form with columns for the number of the lot, its size, its use,
its address, the term, the name of the owner (ie grantee) in Chinese and
in English, the Crown rent and so on. The entries under the columns
were written in manuscript. Where the grantee of a particular lot was a
customary institution, the practice was to name against the lot number
the t’so or t’ong and follow this with the names of the managers, all in
Chinese. In the next column was a translation (or, more accurately, a
transliteration) of the Chinese names into English.17 In the case of t’sos
and t’ongs, there would be a transliteration of the name of the t’so or
t’ong followed by the name in English of each manager. In the schedule
in question, as no doubt in many others, this name was followed by
“(trustee)”.
On the basis of this parenthetical word, Mills-Owens J concluded that
the grant had been made to the managers as trustees in law.18 In other
words, the trust had been created not by a settlement but by the Crown
grant.

Crown Lease Schedule

To the mind of this writer, there are substantial grounds to doubt the
correctness of this conclusion. First, it overlooks what is manifest from
the contents of the schedule, that the name of the grantees in English
followed by “(trustee)” is a transliteration or translation of the Chinese
entry in the immediately adjacent column. Therefore, the entry in
English has no force additional to that of the Chinese characters. Only
if the Chinese entry contained the equivalent of “trustee” could that
word have effect to impose a trust. Unfortunately, the Chinese entry is
not reproduced in the judgment but it seems highly unlikely that any
indication of trusteeship would have been given there. The description
favoured for the managers of a t’so was the Cantonese “sze lei”, meaning

16
The Hong Kong Government had been instructed by the Colonial Office in London that the
New Territory (as it was initially called) was not to place a financial strain on the colony:
Hayes, The Great Difference, n 1 above, p 30 and p 193, n 3.
17
Mills-Owens J refers to it as a translation but the column in schedules seen by the author is
headed “transliteration” which is a more accurate description of the contents of the column.
This is also the experience of Dr James Hayes (personal communication).
18
Tang at p 304.

Binder for HKLJ.indb 674 12/17/2012 4:11:47 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 675

nominee, manager or administrator. In the 19th century, there were no

Chinese characters for a trustee. The modern Chinese term (信託),
“shun tolk” in Cantonese, is a 20th century commercial innovation
believed to be derived from the Chinese name adopted by western trust
companies. This was not used in connection with managers of t’so. In
any event, the concept of a manager is different from that of a trustee.
This is so in English and is apparent also in Chinese traditional usage
from the custom that sze lei would not act without the blessing of all
the members of the t’so.
The presence of parentheses around “trustee” may also be significant.
They suggest that the description is incidental and may not have been
used in the Chinese but is an explanation added in English. It is as if
whoever filled in the columns by hand was concerned to indicate that
each of the three names was not an owner as such but was a representative
of the t’so. In other words, the choice of “trustee” was not made in order to
reflect a status or impose a trust but to emphasise that the grant was made
to the t’so itself and to explain why three individual names follow the
name of the t’so in the owner column. The composers of the details in the
schedule may have felt that the description “manager” was insufficiently
legalistic.
Who were those composers? The entry itself would probably have
been the work of a clerk since there were 477 block Crown leases with
accompanying schedules created between 1905 and 1910;19 these
would have recorded the details of hundreds of thousands of lots.
The clerk would have been acting on instructions from more senior
staff who in turn would have based the entries on the findings of the
surveyors. Each block lease indenture and each schedule was checked
and signed off by an assistant land officer who was responsible for the
demarcation district to which the lease referred and who was also a
member of the Land Court.20 This court was an informal tribunal so
the members did not need to have legal qualifications: in fact the
officers were cadet administrators, so they were well-educated civil
servants. The officer’s signature indicated that the lease had been
“examined and found correct” before it was signed by the Governor
on behalf of the Crown. Even so, many mistakes in the leases were
subsequently uncovered.21 It seems unlikely that the checking would
have involved an investigation as to whether the “trustees” really

19
Information from Mr Roger Nissim, former District Lands Officer and Adjunct Professor at
the Department of Real Estate and Construction of the Faculty of Architecture, University of
Hong Kong.
20
Information from Dr James Hayes. The Land Court is considered further below.
21
Hayes, The Great Difference, n 1 above, p 39 and p 199, n 53.

Binder for HKLJ.indb 675 12/17/2012 4:11:47 PM

 676 Malcolm Merry (2012) HKLJ

were trustees in law or a conscious decision to give them that status.

Rather, the word was probably a standard translation of the characters
for “sze lei”.
In any event, the schedule contains no words of grant: it is merely a
record of details pursuant to the grant in the body of the lease. That body
declared its purpose to be to grant a lease “to the several persons, clans,
families, and ‘t’ongs’ whose names are set out in the Schedule hereunder
written”.22 The operative words stated that the King “doth hereby grant
and demise unto each Lessee ALL that piece or parcel of ground situate,
lying and being in Survey District No. 121 in the NT in the Colony
of Hong Kong set out and described in the Schedule hereto opposite
to the name of such Lessee … according to the lot number set out in
the Schedule hereto opposite to the name of such Lessee”.23 There is
no mention here of status or capacity except that of lessee. There is no
indication that the grant is in any way qualified by the words in the
schedule. The grantee named in the schedule is the t’so which, having no
personality separate from its members, consists of its members.
Since the grant was plainly intended to be to the t’so (the living male
descendants of the ancestor after whom the t’so is named) rather than to
the individual managers, the intention cannot have been to constitute
the managers as trustees. Trustees are by definition legal owners of the
land which they hold on trust so, had the three names been truly trustees,
the owner column in the schedule should have begun with them rather
than the t’so, perhaps adding that they were holding for the benefit or on
behalf of the t’so or its members. But this treatment would have been in
defiance of the words quoted above from the start of the lease by which
it is made apparent that grantees could include customary institutions:
clans, families and t’ongs.

Contemporaneous Legislation

A grant to the institution rather than to the managers as trustees

would have been entirely consistent with the nature of t’so and with
the contemporaneous legislative treatment of them. In 1905 not only
were block grants made to owners of New Territories land, including
t’so and t’ong, but an ordinance was passed dealing with customary

22
Tang, p 289. The wording reflects that used in contemporaneous legislation, now s 15 of the
New Territories Ordinance, Cap 97.
23
Tang, p 290. There is an obvious typographical error in the report in that it says “demise unto
each Lease” rather than Lessee.

Binder for HKLJ.indb 676 12/17/2012 4:11:47 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 677

law and with managers of traditional institutions.24 That legislation

provided, in what is now s 13 of the New Territories Ordinance,25 that
the courts may apply customary law: this was interpreted in Tang as
being obligatory.
Another section of the 1905 legislation, now s 15 of the New
Territories Ordinance, made convoluted provision regarding managers.
Where land was held in the name of a “clan, family or tong”, a manager
was to be appointed by the clan, family or t’ong to represent it.26 The
appointment was to be reported to the District Office and, if proved,
was to be approved by that office which would register the name on
a register of managers. The manager would then have “full power to
dispose of or in any way deal with the land as if he were the sole owner
thereof”, subject to the consent of the District Office to the transaction.
The manager was also to be personally liable for payment of rent and
charges and for observance of the covenants and conditions in the
lease. Furthermore, every instrument regarding the land executed or
signed by the registered manager in the presence of and attested by the
District Officer was to be effectual for all purposes as if executed and
signed by all the members.
These provisions are inconsistent with a manager being a trustee.
First, the section expressly refers to the land being held in the name of
the clan, family or t’ong, not as being held by the managers as trustees for
them. Second, the section deals with the appointment of a manager, not
a “manager and trustee”; indeed it does not mention trust or trustee at
all. If the intention had been to impose upon managers the obligations
of a trustee, what would have been simpler than to say so? The section
was interfering with the customary law by bestowing upon the manager
powers of sole ownership. If the government’s intention also had been to
interfere with custom by imposing trustee duties upon the manager, one
would have expected that to be in the legislation as well.
Instead, the section refers to the members of the family or t’ong
appointing a manager “to represent it”, in other words to act as an agent.
If the manager had been intended to be a trustee, he would not have
been a mere representative but the legal owner of the land. Elsewhere
in the ordinance there is a reference to trustees, so the absence of the

24
New Territories Land Ordinance 1905.
25
Cap 97.
26
Clan and family land refers to t’so and to t’ong in which a family has chosen to append the
character for t’ong after the proper or informal (nick) name of an ancestor. The separate
mention of t’ong was probably to ensure that s 15 covers non-family common landholding
organisations such as temples, monasteries, nunneries and charities.

Binder for HKLJ.indb 677 12/17/2012 4:11:47 PM

 678 Malcolm Merry (2012) HKLJ

use of that word in s 15 is unlikely to have been an oversight.27 The

section institutionalised and adapted the existing practice in t’so and
t’ong of appointing a “sze lei”, or rather a number of “sze lei”.28 This is a
Cantonese term which translates as manager or representative but not
as trustee.
However, the best point that emerges from s 15 against the manager
being a trustee arises from the express effect of the registration of the
manager: he was to have full power to dispose of and deal with the land
as if he were the sole owner thereof. If managers were trustees, they
would not need this power and status because they would already have
them. It is axiomatic, indeed elementary, that trustees of land are legal
owners of the land. It would make no sense for the legislature to bestow
upon managers the power of disposal of the land as if they were sole
owners unless the legislators, and the drafters of the legislation in the
government, were of the view that the managers were not already the
sole legal owners of the land and therefore not trustees.29
The evident purpose of s 15 is to facilitate dealings in the land. As
Briggs J pointed out in 1966, but for the statutory appointment of the
registered manager as sole representative, the land could be conveyed
only if the documents were executed by all the members of the t’so, a
cumbersome and inconvenient, and often impractical, requirement.30 It
might be added that it was not just conveyance of the land that would
have been difficult. More importantly (since t’so land is in principle
inalienable so sale would have been rare), the mortgaging of the land
would have been hindered, as might the collection of government
rent, rates and other charges and the enforcement of covenants in
the government grant.31 None of these difficulties would have been
encountered if the managers were also trustees.
Section 15 had an additional purpose: the protection of members of
the t’so or t’ong from misfeasance by managers by requiring not only that
managers be approved but also that any transaction be carried out with
the approval and witnessing of the District Office. This, too, would have

27
Section 18 of Cap 97. This argument was considered but rejected by Chu J in Tang Kam Wah v
Tang Ming Yat, n 3 above, para 61, for reasons which are not convincing.
28
The usual arrangement was that each fong or branch (or sub-branch) of the clan would appoint
one of the managers.
29
Most managers are members of the t’so or t’ong and therefore a joint owner in the capacity of a
member but they are not the sole owners. Occasionally a manager is not a member, eg where
she is the widow of a deceased manager: being female, she cannot be a member.
30
Lai Chi Kok Amusement Park (No. 2) v Tsang Tin-Sun [1966] HKLR 124 at 130.
31
The efficient collection of Crown rent would have been a priority for the new administration
in 1905 because of the Colonial Office direction in 1898 that the New Territories were not to
be a burden on the finances of the colony: Hayes, The Great Difference, n 1 above, p 193, n 3.

Binder for HKLJ.indb 678 12/17/2012 4:11:48 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 679

been unnecessary had the managers been trustees, for the duties imposed
in equity upon trustees would have provided protection to the members
as beneficiaries.

Events of 1898–1905

It is not just the legislation at the beginning of the 20th century concerning
customary landholding that points against an intention to impose a
trust upon managers, the history of events at that time surrounding the
acquisition of “the Hong Kong Extension” is also inconsistent with such
an intention.
The new governor of Hong Kong in 1898 was Sir Henry Blake, a man
of genial character with wide experience as a colonial administrator.32
Blake’s experience told him that it was important to reassure the local
population that the change in administration would have little affect on
their daily lives, particularly in respect of land and money. Accordingly
when Britain took possession of “the new territory” in April 1899,
Blake issued a proclamation in Chinese pursuant to the New Territories
Order-in-Council by which the extension had been constitutionally
incorporated into the colony of Hong Kong. The proclamation told
inhabitants of the newly acquired area that their “commercial and landed
interests will be safeguarded and … usages and good customs will not in
any way be interfered with”.33 This was not legislation but it showed the
government’s intention and policy towards land in the New Territories.
In order to emphasise the change in authority and to set a sound
basis for the collection of revenue in the form of rent, Crown leases
were to be issued to existing landholders and a reliable rent roll drawn
up. To achieve this, and to ensure the protection of established landed
interests, the new administrators had first to find out who owned what
land. The land registry of the Chinese government was found to be
inaccurate, so a complete survey of the 365 square miles of the New
Territories had to be undertaken. Hong Kong lacked the expertise to
perform this enormous task, so land surveyors were brought in from
the Survey of India.34 The great land survey, which began in 1899,
took much longer to conduct than anticipated owing to adverse

32
Frank Welsh, A History of Hong Kong (Harper Collins, 1997), p 330.
33
A translation of the proclamation is set out in the judgment of Roberts CJ in Winfat Enterprises
(HK) Co Ltd v Attorney-General [1984] HKLR 32.
34
The Survey of India, the land survey department of the government of India, still exists and is
the oldest extant department of that government.

Binder for HKLJ.indb 679 12/17/2012 4:11:48 PM

 680 Malcolm Merry (2012) HKLJ

conditions (heat, ill health, mountainous terrain, language difficulties

and uncooperative population) and was not completed until 1904.35
There were disputes about ownership, so a Land Court was established
in 1900 to adjudicate upon claims.36 Titles were to be granted once the
claims had been confirmed. Entitlements were judged in accordance
with Chinese law and custom. This meant that members of the court
had first to ascertain the relevant law, which delayed adjudication but
showed that the new administration was serious in its promise to honour
local custom.37
The Land Court having completed its work by early 1905, the New
Territories Land Ordinance was passed that year.38 This gave the Land
Officer and the courts power to enforce Chinese custom and customary
right. As has been seen, it also provided for registration of managers of t’so
and t’ong and gave them full power to deal with the land subject to the
supervision of the District Offices. With the necessary information for
the schedules and the legislation in place, block Crown leases could be
issued with the term of their grant beginning as from 1898. The purpose
of the block leases was to re-grant the land to the established owners in
an administratively convenient way, not to alter the identity or status of
those owners.39
If the block leases of land to customary institutions had been
intended to convert their managers from mere representatives into
trustees, that would have been contrary to the terms of the proclamation
which declared that landed interests were to be safeguarded and
customs were not to be interfered with. In that event, members of
t’so and t’ong would have been demoted from legal owners of the land
to mere beneficiaries, and managers would have been promoted from
their customary role as mere representatives of the owners to legal
owners of the land.
Accordingly, the background to the grant of the block leases suggests
that there was no intent to convert t’so and t’ong into trusts through

35
Hayes, The Great Difference, n 1 above (p 32) suggests that the survey was completed in 1903.
36
Land Court (New Territories) Ordinance 1900.
37
Wesley-Smith, Unequal Treaty (Oxford University Press, Hong Kong, 1980), pp 94–97.
38
Mills-Owens J says in Tang that the court finished its work in 1904 and Hayes, The Great
Difference, n 1 above (p 36) is to the same effect, but Wesley-Smith in Unequal Treaty gives
1905; the latter seems more likely since the court could hardly conclude all adjudication until
the survey had been completed.
39
However, because in Imperial Chinese practice there could be an owner of the topsoil (or skin
of the land) and an owner of the subsoil (or bones of the land), this did necessitate choosing
between the two: generally, the former was chosen and granted a Crown lease; the latter was
(somewhat dubiously) regarded as having a rentcharge only.

Binder for HKLJ.indb 680 12/17/2012 4:11:48 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 681

the medium of the government grant. Ironically, Mills-Owens J set out

much of this background (but not the proclamation) in his judgment
in Tang yet did not draw any lessons of interpretation from it. That
may be because in 1970 the approach to the construction of contracts
and deeds was more literal than now. During subsequent decades, a
contextual, rather than textual, approach has found favour so that the
courts now look at the words used in their setting. This includes the
factual and legal background as well as the practical objects which the
words were intended to achieve.40

Opportunity for Re-consideration?

In one respect at least, Mills-Owens J’s interpretation of the schedule

to the Crown lease has certainly proved incorrect. In summarising the
schedule’s contents, the judge asserts that user of the lot is limited to the
use stated under the heading “description of the lot”: so a lot which is
described as “padi” (as the land in question before him was) could not be
used for any purpose other than rice-growing. The view that the use stated
in the schedule was prescriptive (or restrictive) rather than descriptive
was however exploded in 1982 when the Court of Appeal decided in
Attorney-General v Melhado Investment Ltd41 that the description was just
that and did no more than aid identification of the lot. Yet until then the
view that the description of use was also a restriction on use had been
held with a tenacity similar to the view that t’sos are trusts.
Is there an opportunity for a similar re-consideration of the trust view
of t’sos? The view has been repeated so often, by so many judges, that
it has taken on the appearance of a matter of precedent, even though
(the author hopes) it has been demonstrated to lack a basis in principle.
Nevertheless, the view has been expressed mainly at first instance. Where
the Court of Appeal has accepted it, the view has not been in issue.
Therefore, in Wong Shing Chau v To Kwok Keung the point was
conceded and not argued.42 No expert evidence as to the existence of
trusts in Chinese custom had been presented at trial. Le Pichon JA
summarised the propositions as to customary institutions made by Deputy
Judge Lam in Leung Kuen Fai and stated that they were not challenged.

40
Jumbo King Ltd v Faithful Properties Ltd [1999] 3 HKLRD 757 at 773 (CFA). There is also much
more published research about New Territories land available now than there was in 1970.
41
[1983] HKLR 422.
42
(unrep., CACV 20/2008, [2008] HKEC 969) paras 9 and 12.

Binder for HKLJ.indb 681 12/17/2012 4:11:48 PM

 682 Malcolm Merry (2012) HKLJ

In Tang Yau Yee Tong v Tang Mou Tso, another dispute between
members of the litigious Tang clan, Liu JA observed that in relation
to t’sos, manager, trustee and sze lei were synonymous.43 As a simple
statement as to how these terms are used in fact, this observation is
accurate: they tend to be employed interchangeably in the context
of t’so and t’ong to mean the same thing. Liu JA seems to intend to
convey no more than that. But if he had meant to make a finding that
managers or sze lei are trustees (as one later judge in yet another Tang
clan dispute appears to have thought),44 the observation would have
been incidental and unnecessary for the decision, so would have been
obiter dictum.
Therefore, it is submitted that the question of whether t’sos are trusts
remains open at appellate level. One might anticipate, however, that the
Court of Appeal would be reluctant to disturb an assertion which has
attracted widespread approval at lower judicial levels and which leads to
convenient results. By calling t’so trusts, the courts give themselves the
comfort of a familiar concept and access to a body of familiar rules which
can be employed for the protection of members of these institutions.
Yet that protection is not really necessary, for customary and statute
law already provides it. Any attempt by managers to sell t’so land for their
personal benefit is rendered impossible by the customary requirement
that there be unanimous clan support for the sale and by the provisions of
s 15 of the New Territories Ordinance which impose formal requirements
upon disposals of the land. Likewise, any dealing in the land such as
a letting is regulated by the section and by the customary requirement
that all managers consent to the dealing. The handling of rent or other
income from the land, including distribution of compensation for land
resumed by the authorities for public purposes, would also be the subject
of unanimous decision by the managers under customary practice.
It might be said that the overlaying of a common-law institution upon
a customary one is of little consequence and can do only good: the trust
simply reinforces the dictates of custom. This is generally so where a dispute
is between members or managers of the t’so, as most litigation concerning
them is. However, where the dispute affects the rights of others, the effect
of imposing a trust on the t’so is not so benign. This is illustrated by Leung
Kuen Fai, the case in which Deputy Judge Lam reviewed the principles
applicable to customary institutions. The plaintiff there had occupied t’so
land without permission for a very long time and was claiming title by
adverse possession. Had the owner been an individual, the plaintiff would

43
[1996] 2 HKLR 212; the judge had been counsel for the successful defendant in Tang.
44
Chu J in Tang Kam Wah v Tang Ming Yat, n 3 above.

Binder for HKLJ.indb 682 12/17/2012 4:11:48 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 683

have been able to repulse an action for possession and the owner’s title
would have been extinguished. However, the normal limitation period
was held not to apply because the t’so was also a trust and its members
were beneficiaries. This meant that the limitation period was not the
normal 12 years from accrual of the owner’s right of action but was 6 years
from the attaining of the age of majority (the time of accrual of a cause
of action for possession) by any beneficiary. Since new members of the
t’so were continually being born, the limitation period was continually
being extended. All the defendant had to show was that a child who was
a member of the t’so had been born within the past 24 years (the statutory
age of majority of 18 years plus 6 years’ limitation period) to show that the
statutory bar had not descended.45 Thus the imposition of a trust gives t’sos
a great privilege: as the judge observed, they are effectively immune from
the law of limitation of actions.

Limits of the Trust

Even if the trust analysis based on the parenthetical translation in the

schedule to the Crown lease were to be upheld after full consideration
by the appellate courts, it would apply only to those lots for which the
word “trustee” had been used in the schedule to describe the managers.
The practice of those who composed and completed the schedules may
not have been consistent. The author has come across one block lease
schedule in which only Chinese descriptions appear to have been used.
If the clerks and their seniors at the Land Office followed the habit
of their counterparts at the Land Registry, they would have used the
label “manager” as often as they used “trustee” and apparently without
discrimination. Then there are Crown leases which have been lost and
those that have schedules that are illegible. There would be no basis
upon which to conclude that a trust had been imposed by the grant in
such cases.

An Alternative Basis?

If there is no trust by grant, what of the possibility that t’so and t’ong are
settlement trusts? By and large, since the decision in Tang judges have
not asked themselves about the nature of the trust, being content merely

45
This assumes that the statutory rather than the customary age of majority applies. Chinese
custom was that a person attains majority at about the age of 14 (source: Prof Anthony Dicks).

Binder for HKLJ.indb 683 12/17/2012 4:11:48 PM

 684 Malcolm Merry (2012) HKLJ

to assert that a trust exists. An exception was Kan Fai Tat v Kan Yin Tat46
in which 18 lots of land at Tseung Pak Long near Sheung Shui had been
purchased and registered in the name of a t’so in and between 1935 and
1949. The then Deputy High Court Judge Tang said:

“I believe that in the event of a conveyance to a Tso the law will presume
an intention to create a trust over the properties, in favour of members of
the Tso from time to time, subject to such limitation or conditions as may be
imposed on such properties by Chinese custom or customary rights affecting
such land.”

This, then, is a different basis from that postulated by Mills-Owens J in

Tang. There could have been no trust imposed by government grant in
Kan because the founding of the t’so occurred after the block grants of
1905. Nor, apparently, was it a case of the founder creating an express
trust with the aid of legal advisers as had happened with at least one
other 20th-century t’so. Instead, the deputy judge’s suggestion was that a
trust was implied by law.
The judge in Kan cited no authority and gave no reasons for his
assertion that the law presumes of intention by founders of a t’so to
create a trust. Presumably, the presumption was thought to arise from the
circumstance that the founders had put the land into the t’so by registering
it in the name of the t’so. But, as has been seen, the effect of doing that
was, according to Chinese customary law, to constitute members of the
t’so joint owners of the land rather than mere beneficiaries, and not to
constitute the managers as the owners, as they would have been had
they been trustees. It is difficult to see how a court would be justified in
presuming any intention from the placing of land into the ownership
of a t’so except an intention to create a t’so and thereby make all living
male descendants joint legal owners of the land. To presume an intention
to create a trust with the managers as trustees and legal owners and the
living male descendants merely beneficiaries would be inconsistent with
that intention.
If, by saying that the putative trust was subject to such limitations
imposed on the properties by Chinese custom or customary rights, the
learned deputy judge meant to recognise the paramountcy of those rights
in the question of whether there was a trust, he was in effect contradicting
himself in the course of one sentence, for custom limits the manager’s
role to something less than that of a trustee and allows the members

46
[1987] HKLR 516; hereinafter Kan.

Binder for HKLJ.indb 684 12/17/2012 4:11:48 PM

 Vol 42 Part 3 Are T’sos Really Trusts? 685

rights exceeding those of a beneficiary. But perhaps the qualification

“on such properties” was the deputy judge’s way of confining the effect
of customary law to the land as distinct from the landholding: in other
words, custom dictates only the terms of the trust so far as they relate to
the land and does not affect the existence of the trust.
If the latter was what was meant, it is respectfully suggested that
the distinction between the trust and its terms is unjustified and that
any limitation of the effect of customary law to the terms of the trust
is unsustainable. It is true that in Tang Mills-Owens J explained that
the terms of the trust were the rules of customary law but this was
an attempt to accommodate the customary law within the framework
of the trust which he had deduced to arise from the schedule to the
Crown lease. As it is hoped has been demonstrated, that deduction
was not only dubious but was in plain contradiction of the nature of
a t’so. In finding that there was a trust, the judge in Tang was failing
to obey the injunction in s 13 of the New Territories Ordinance to
apply customary law—ironically, because he himself had found that its
application was mandatory. Similarly, in finding that there was a trust
in Kan, the deputy judge was also failing to obey that injunction, in
effect ousting the custom and amending the nature of a t’so by imposing
a common-law construct upon it.
The basis of the imposition of a trust in Kan is as debateable as that
in Tang. The founders of the t’so were said to have intended a trust,
presumably a trust of the sort which arises when two persons have
a common purpose with regard to property. Perpetual purpose trusts
are not known to the common law system, unless they are charitable
trusts. Although the purposes of t’so or t’ong can include religious and
educational purposes, they are not charitable because the beneficiaries
of the trust are not the general public but a limited class, certain male
persons from a clan who also happen to be members of the t’so or t’ong.
Proponents of the trust analysis would no doubt counter that it is here
that custom trumps equity so as to allow a perpetual settlement: this is
what Deputy Judge Lam seemed to imply in his summary in Leung Kuen
Fai. But why at this point and not earlier? The very fact that proponents
have to qualify the effect of their trust suggests that the trust analysis is
flawed in the first place.
It was in any event straining matters to presume an intention by
the founders of a t’so to impose a trust upon the managers of their
t’so. Such a presumption was not necessary to give effect to the t’so.
The founders were operating within the confines of Chinese custom.
It is doubtful that they would have known what a trust and its
consequences were. They expressly intended to create a t’so, not a

Binder for HKLJ.indb 685 12/17/2012 4:11:48 PM

 686 Malcolm Merry (2012) HKLJ

trust. The processes of conveyance of the land into and registration

of it in the name of the t’so alone realised that express intention. The
presumption or implication that the founders intended to create a
trust is contrary to the evidence of their actions. Its effect is partially
to nullify those actions.

Conclusion

The description of t’sos as trusts is based on repeated assertion but little

analysis and very questionable reasoning. The description appears to
be motivated by a desire of common lawyers to equate these customary
institutions with another institution which is familiar to them and
thereby to protect the members of those institutions. Two attempts have
been made to explain the basis for the assertion. The first, by Mills-Owens
J in Tang, relies upon the contents of the schedule to the block Crown
lease to show the imposition of an express trust by grant. The second, by
Tang Dep-J in Kan, simply states a presumed intention by founders of a
t’so to settle the land on trust. Neither explanation is convincing.
T’sos are not trusts by government grant because that would be
contrary to government policy at the time of the block grants, contrary to
legislation contemporaneous with those grants, contrary to a reasonable
interpretation of the terms of the grant and, not least, inconsistent with
the contents of the customary law which gives t’sos their very existence.
T’sos are not purpose trusts created by common intention because they
are not charitable in nature and their creators have no intention to create
a trust. The customary role of a manager is less than that of a trustee. The
customary interest of a member of a t’so in the land is greater than that
of a beneficiary under a trust. Indeed, the nature of a trust is in conflict
with the nature of t’sos.
T’sos are inventions of Chinese rural culture and reflect the values
of that culture: reverence for ancestors, importance of the land to
identity, solidarity of family and clan, and male dominance. They have
captivated administrators, historians, anthropologists, lawyers and judges
who have come into contact with them. For generations they have been
part of the social fabric of that part of San On county which became
the New Territories. That fabric and those values have however been
worn down with the passage of time. Education, prosperity, building and
infrastructural development, urbanisation, modern travel and improved
communications have all had their effect to undermine these customary
institutions. Let us not add to the list by insisting that t’sos are trusts.

Binder for HKLJ.indb 686 12/17/2012 4:11:48 PM