The Constitutionality of Expropriation

Without Compensation
Tembeka Ngcukaitobi*
Michael Bishop†

I INTRODUCTION
Does the Constitution permit expropriation without compensation? If so, in what
circumstances? If compensation is paid, how must the amount be calculated, and what
process must be followed?
The idea of expropriation without compensation is not novel in our law. As far back as
1915, the Appellate Division recognised that Parliament had a right to expropriate without
compensation, although there was a presumption against it. 1 And both the Constitutional
Court2 and the Land Claims Court3 have posited the idea that expropriation without
compensation could be constitutional in certain circumstances. Even conservative academics
have argued that compensation is inefficient, and that it is better to price in the risk of
uncompensated compensation through insurance, and discounts on the market price of
property.4
Moreover, expropriation without compensation is common in our law. Property, including
land, is regularly seized and forfeited to the state, particularly when it is connected to
organised crime.5 And what may often be thought of as expropriation in ordinary terms – the
compulsory transfer of rights from one private party to another – does not trigger a
constitutional duty to pay just and equitable compensation.6
It is true that some authors hold the opposite view. Rautenbach, for example, has written:
“There is no provision in section 25 for expropriation without compensation. It can only be

1*
Member of the Johannesburg Bar.
†
Honorary Research Associate, UCT; Member of the Cape Bar; Counsel, Legal Resources Centre. The authors
would like to acknowledge the research assistance of Eshed Cohen and Michael Tsele in preparing this paper.
Simmer & Jack Pty Mines Ltd v Union Government (Minister of Railways & Harbours) 1915 AD 368 at 398,
holding “It is, of course, within the power of the Legislature to deprive an owner of valuable property without
compensation”. See also Krause v SAR&H 1948 (4) SA 554 (O) at 562-3; and Sandton Town Council v Erf 89
Sandown Extension 2 (Pty) Ltd 1986 (4) SA 576 (W) at 579. On appeal in Sandton Town Council v Erf 89
Sandown Extension 2 (Pty) Ltd 1988 (3) SA 122 (A) it was held that it is within the power of the legislature to
decide whether or not to expropriate property without compensation. However, where there is ambiguity in the
legislation, it should be construed in favour of the subject whose rights of ownership are extinguished or
attenuated by the act of expropriation.
2
First National Bank at para 97, holding “The formulation of property rights and their institutional
framework differ, often widely, from legal system to system. Comparative law cannot, by simplistic
transference, determine the proper approach to our property clause that has its own context, formulation and
history. Yet the comparative perspective does demonstrate at least two important principles. The first is that
there are appropriate circumstances where it is permissible for legislation, in the broader public interest, to
deprive persons of property without payment of compensation. ”
3
Nhlabathi and Others v Fick 2003 (7) BCLR 806 (LCC).
4
R Posner Economic Analysis of Law (6 ed, 2003) at 57.
5
See the Prevention of Organised Crime Act 121 of 1998.
6
Agri South Africa.
 2

inserted by constitutional amendment. The judiciary does not have the power to do so.”7 He
is wrong.
The question under the Constitution is not whether expropriation without compensation is
constitutionally permissible. It is. The question is when it is permissible to expropriate
without compensation. In particular, when is it constitutionally permissible to expropriate
land without compensation in order to advance land reform?
If compensation is to be paid, how much should be paid, and how should the amount be
determined. At present, the government and the courts have generally paid market value for
property. And the law requires that – if there is a dispute about the amount of compensation
owed – it must be determined by courts.
This may impede land reform. It makes the expropriation of land more expensive both
because of the amount that must be paid, and the cost of a court process to determine that
amount.
This is not constitutionally required. The Constitution permits the payment of
compensation that is below, even significantly below, market value. And the Constitution
permits the amount of compensation to be determined through administrative means – as long
as it is subject to court approval of some type. Precisely when expropriation is permitted
without compensation, or with limited compensation, is complicated by the multi-layered
structure of the property clause.
But stripped of technical legal analysis, the questions are fairly simple. If a law is passed
that expressly provides that no compensation, or limited compensation will be paid, it will
pass constitutional muster if: (a) the law is reasonably required to advance the goal of land
reform; (b) the amount of compensation is proportional to the nature and degree of the loss
suffered by the owner; and (c) courts can exercise review jurisdiction over the manner and
the amount of payment of compensation.
In summary, our view is that the Legislature is entitled to pass laws to further land reform
that provide for expropriation without compensation in limited categories where: (a) the land
is not being efficiently used; (b) the owner will suffer only economic harm; (c) there is a
safety valve where expropriation without compensation would suffer particular harsh
consequences; and (d) it does not apply to foreign-owned land.
In addition, the state can legislate to provide for compensation to be administratively
determined at a value significantly below market value, provided that the amount is subject to
some form of judicial review. Whether those laws will survive an inevitable constitutional
attack will depend on the details of the law, particularly the extent to which they are flexible
and subject to judicial oversight. It will also depend on the strength of the evidence the
government can produce to show that paying more compensation, or providing for a more
intensive judicial role, would impede land reform.
The remainder of this article is structured as follows. Part II sets out the meaning of s 25.
Part III summarises the relevant international law. Part IV advises on what types of laws
permitting no or limited compensation will be consistent with the Constitution and
international law.

II THE MEANING OF SECTION 25

This Part sets out the meaning of s 25 of the Constitution. As will become apparent, s 25 has
multiple inter-related parts that need to be considered both separately, and together.

7
I Rautenbach ‘Expropriation and arbitrary deprival of property: five forensic constructions’ (2013) TSAR 743
at 753.
 3

A The Historical Context to section 25

The Courts have repeatedly made it clear that s 25 must be interpreted within the historical
context in which it was enacted. In Shoprite Checkers, Froneman J aptly explained:

“the pre-constitutional conception of property … entailed exclusive individual entitlement.

Put simply, that is largely a history of dispossession of what indigenous people held, and its
transfer to the colonisers in the form of land and other property, protected by an economic
system that ensured the continued deprivation of those benefits on racial and class lines. That
history of division probably explains the concerns both the previously advantaged and
disadvantaged still have. The former fears that they will lose what they have; the latter that
they will not receive what is justly theirs.” 8

Most recently, in Daniels, the Court spent significant time on the history of dispossession,
particularly as it related to farm dwellers and labour tenants. Madlanga J wrote:

“Dispossession of land was central to colonialism and apartheid.  It first took place through
the barrel of the gun and “trickery”. This commenced as soon as white settlement began, with
the Khoi and San people being the first victims. This was followed by “an array of laws”
dating from the early days of colonisation.  The most infamous is the Native Land Act
(subsequently renamed the Black Land Act) (Black Land Act).  Mr Sol Plaatje, one of the
early, notable heroes in the struggle for freedom in South Africa who lived during the time
this Act was passed, says of it, “Awaking on Friday morning June 20, 1913, the South
African native found himself, not actually a slave, but a pariah in the land of his birth”.” 9

The Court’s analysis of history is important because it demonstrates its awareness of the
urgent need for land reform. As the Court held in AgriSA: “We must therefore interpret
section 25 with due regard to the gross inequality in relation to wealth and land distribution
in this country.”10 In addition, it demonstrates that s 25 protects conception of property that is
not a traditional, individualist notion of property rights, but a social understanding of
property. “This brings to the fore the obligation imposed by section 25 not to over-emphasise
private property rights at the expense of the state’s social responsibilities.”11

B Overview of section 25
Section 25 has multiple parts. They must be interpreted together to form a coherent whole.
They must be seen as mutually supportive, not as creating internal conflict.
Section 25 has two general categories of clauses. Sections 25(1), (2) and (3) entrench
negative property rights – the rights not to be arbitrarily deprived of property, and the right
for property not to be expropriated without just and equitable compensation, determined by
agreement or approved by a court.
Next, ss 25(4) to (9) are the positive elements of the right – they “underline the need for
the redress and transformation of the legacy of grossly unequal distribution of land in this
country.”12 They include express and repeated commitments to “land reform” and to

8
Shoprite Checkers (Pty) Limited v Member of the Executive Council for Economic Development,
Environmental Affairs And Tourism, Eastern Cape 2015 (6) SA 125 (CC) at para 34.
9
Daniels v Scribante and Another [2017] ZACC 13; 2017 (4) SA 341 (CC); 2017 (8) BCLR 949 (CC) at para
14.
10
Agri SA at para 61.
11
Agri SA at para 62.
12
Haffejee NO and Others v Ethekwini Municipality and Others 2011 (6) SA 134 (CC) at para 30.
 4

“equitable access” to land and natural resources.13 They also create express rights to secure
tenure,14 and to restitution for dispossession after 1913.15
Accordingly, in the first case to consider the property clause, the Constitutional Court said:

“The purpose of section 25 has to be seen both as protecting existing private property rights
as well as serving the public interest, mainly in the sphere of land reform but not limited
thereto, and also as striking a proportionate balance between these two functions.” 16

This unique structure of s 25 means that the negative elements of the right must be
interpreted in line with the positive elements of the right. This is expressly required by: s
25(4) which states that the public interest includes land reform; s 25(5) which mandates the
state to take measures to promote equitable access to land; and s 25(8) which, as we detail
below, requires the other provisions of s 25 to be interpreted so that they do not impede land
reform.

C Expropriation and Deprivation

The focus of this paper is expropriation. However, it is necessary to be clear about the
difference between an expropriation (requiring compensation under s 25(3)) and a
deprivation (that does not demand compensation may not be arbitrary under s 25(1)). Certain
types of conduct that may ordinarily described as expropriations are not, in constitutional
law, regarded as expropriations that demand just and equitable compensation. As the Court
held in Reflect-All:

“The purpose of the distinction between expropriation and deprivation by regulatory measures
is to enable the state to regulate the use of property for public good without the fear of
incurring liability to owners of property affected in the course of such regulation.” 17

The line between deprivation and expropriation is therefore important to assessing when
the constitutional obligation to compensate arises.
Deprivation means the interference of property that is significant enough to have a legally
relevant impact on the rights of an affected in order for the action to qualify as deprivation. In
Mkontwana,18 the Constitutional Court observed that:

“Whether there has been a deprivation depends on the extent of the interference with or
limitation of use, enjoyment or exploitation . . . . [A]t the very least, substantial interference
or limitation that goes beyond the normal restrictions on property use or enjoyment found in
an open and democratic society would amount to deprivation.19

In Mkontwana, for example, the law at issue prevented the transfer of property until
outstanding rates had been paid. This limited part of the right of ownership – the right to sell
the property – but did not constitute an expropriation because ownership remained with the
owner.

13
Constitution ss 25(4), 25(5) and 25(8).
14
Constitution ss 25(6) and (9).
15
Constitution s 25(7).
16
First National Bank of SA Limited t/a Wesbank v Commissioner for the South African Revenue Services and
Another [2002] ZACC 5; 2002 (4) SA 768 (CC) at para 50.
17
Reflect-All at para 63.
18
Mkontwana v Nelson Mandela Metropolitan Municipality and Another; [2004] ZACC 9; 2005 (1) SA
530 (CC); 2005 (2) BCLR 150 (CC) (Mkontwana) para 32.
19
Ibid para
 5

Expropriations are a sub-category of deprivations. All expropriations are also

deprivations, and must, in theory, comply with s 25(1) and ss 25(2) and (3).20 We explain in
the section titled Structure of Expropriation Analysis why compliance with ss 25(2) and
(3) will also constitute compliance with s 25(1).
Internationally, there is much debate about where the line is crossed from a deprivation to
an expropriation. In our law, it seems that an expropriation differs from other deprivations of
property in three ways: (1) An expropriation constitutes a deprivation of the “core content” of
the property right; (2) When property is expropriated, the right is transferred to the state; and
(3) The transfer to the state is not temporary.21
As the Constitutional Court has explained:

“To prove expropriation, a claimant must establish that the state has acquired the substance or
core content of what it was deprived of. In other words, the rights acquired by the state do not
have to be exactly the same as the rights that were lost 22… There can be no expropriation in
circumstances where deprivation does not result in property being acquired by the state.” 23

This issue was dealt with in Agri South Africa which addressed a challenge to the Mineral
and Petroleum Resource Development Act. The Applicants argued that the MPRDA
constituted an expropriation because it made the state the custodian of all mineral rights, and
gave it the power to determine who could exercise those rights. The Constitutional Court held
that it did not constitute an expropriation, because the mineral rights did not vest in the state
As the concurring judgment in Agri SA explained, this approach to expropriation has
serious consequences: “This construction in effect immunises, by definition, any legislative
transfer of property from existing property holders to others if it is done by the state as
custodian of the country’s resources, from being recognised as expropriation.”24
The application of this principle in the land reform context is obvious. If the state declares
itself the custodian of certain categories of land and transfers that land from the current
owners to others, on a strict application of Agri SA it does not expropriate the land. It is then
not required to comply with ss 25(2) and (3), but only to establish that the deprivation is not
arbitrary in terms of s 25(1). That, for example, is the approach taken by the Land Reform
(Labour Tenants) Act which provides for transfer directly from the landowner to the labour
tenant.25
However, the principle must be tempered in two ways. First, the complete transfer of land
from one party to another may be arbitrary in terms of s 25(1) unless it is accompanied by
some compensation. Still, the manner of determining compensation, and the ultimate amount
will not be subject to the strict standards in ss 25(2) and (3).
Second, the Constitutional Court is aware of the need not to take an overly technical
approach to the concept of “acquisition” by the state. In adjudging whether any particular
legislative measure constitutes an acquisition, the Court will consider both the needs to
protect property rights, and “the constitutional imperative to transform our economy with a
view to opening up access to land and natural resources to previously disadvantaged
people”.26 It is not guaranteed that the form of transfer of property rights adopted by the
MPRDA will always be held to be a deprivation, particularly if the Court believes it is an

20
FNB at paras 57-8.
21
Harksen v Lane NO at para 35.
22
Agri SA v Minister for Minerals and Energy 2013 (4) SA 1 (CC) para 58.
23
Ibid para 59.
24
Ibid at para 105 (our emphasis).
25
In Uys N.O and Another v Msiza and Others [2017] ZASCA 130, the Supreme Court of Appeal treated this as
expropriation.
26
Agri SA at para 63. See also para 64.
 6

improper attempt to expropriate by another name, or if it does not strike the appropriate
balance.

D Section 25(1): Arbitrary Deprivation

As demonstrated in the previous section, certain state conduct that results in the transfer of
land does not constitute an expropriation. It will, however, constitute a deprivation that, in
terms of s 25(1) must: (a) Occur in terms of a law of general application; and (b) not be
arbitrary.
We do not understand there to be any intention to expropriate land other than through
legislation. We therefore focus on the requirement of non-arbitrariness.
This is the part of s 25(1) that has received the most attention of the Constitutional Court.
Four parts of its jurisprudence are relevant.
First, there is both a substantive and a procedural component to the arbitrariness analysis.
A deprivation is arbitrary “when the ‘law’ referred to in section 25(1) does not provide
sufficient reason for the particular deprivation in question or is procedurally unfair.”27
Second, whether a deprivation is ‘arbitrary’ depends on an evaluation of the nature and
extent of the deprivation, compared to the purpose of the deprivation. A deprivation of
ownership of land will generally require a more compelling purpose than deprivation of
lesser rights, in moveable or incorporeal property.28
Third, the extent of the relationship between means and ends occurs on a sliding scale
between mere rationality, and full proportionality. As Nkabinde J explained in Reflect-All:
“In some instances a deprivation will escape arbitrariness if a rational connection between the
means adopted and the ends sought to be achieved is present. In other instances, however, the means
adopted will have to be proportional to the ends in order to justify the deprivation in question.
Marginal deprivations of property will ordinarily not be arbitrary if they are rationally connected to a
legitimate purpose. More severe deprivations will ordinarily have to be shown to be proportionate.” 29
Fourth, a deprivation can be rendered non-arbitrary if some compensation – whether in
money or in kind – is paid.

E Section 25(2): Compensation Determined by a Court

Section 25(2) reads:

“(2) Property may be expropriated only in terms of law of general application—

(a) for a public purpose or in the public interest; and
(b) subject to compensation, the amount of which and the time and manner of
payment of which have either been agreed to by those affected or decided or
approved by a court.”

Section 25(2) sets the three key requirements for an expropriation.

First, it must be in terms of a law of general application. This includes legislation or
regulation. Again, as we do not understand that there is any intention to expropriate outside
of the bounds of a law of general application, we do not consider this requirement any
further.
Second, it must be for a public purpose or in the public interest. Section 25(4) holds
expressly that the “public interest includes the nation’s commitment to land reform, and to
reforms to bring about equitable access to all South Africa’s natural resources”. This paper
is limited to the constitutionality of expropriation with limited or no compensation for the

27
FNB at para 100.
28
Ibid.
29
Reflect-All at para 49.
 7

purpose of land reform. We do not consider the constitutionality of expropriation for other
purposes.
Third, the final s 25(3) requirement is about the process for determining compensation.
Compensation must either be determined by agreement, or be “decided or approved” by a
court. This provision is central to any attempt to expropriate land without compensation or
with limited compensation.
The provision is an aspect of the broader constitutional right enshrined in s 34 of the
Constitution to have “any dispute that can be decided by the application of law decided in a
fair, public hearing before a court”. The Constitutional Court has held: “In a constitutional
democracy founded on the rule of law, disputes between the state and its subjects, and
amongst its subjects themselves, should be adjudicated upon in accordance with law. The
more potentially divisive the conflict is, the more important that it be adjudicated upon in
court.”30 The SADC Tribunal has found that legislation that sought to strip Zimbabwe’s
courts of jurisdiction to determine whether just and equitable compensation had been paid
was contrary to international law.31
A total ouster of the judicial role to “decide or approve” the amount of compensation
would not only be contrary to s 25(2)(b), it would be extremely difficult to justify under s
25(8). Courts are notoriously protective of their role, and will be very unlikely to give up this
specifically-assigned constitutional role.
However, there is considerable flexibility within s 25(2)(b). It requires that the amount of
compensation is “decided or approved” by a court. This entails that the court can either play
an original role (decide), or a reviewing role (approve).
It would be consistent with s 25(2)(b) for the amount of compensation to be determined, in
the first instance, by an executive or administrative body – a state valuer for example – or by
the application of a formula, provided that: (a) The discretion of the administrator is confined
by adequate guidelines contained in legislation or regulation; and (b) there is some form of by
a court to ensure it does not have unduly harsh consequences in specific instances.
The more difficult question is what standard the court should apply. Must the court be
granted the unfettered right to approve the amount of compensation? Or would it be
permissible to limit the court to reviewing whether the initial determination was rational,
reasonable, or some other standard?
There are reasons to believe that a court would accept the second, more limited role. In
expropriation a court plays two primary roles. The first is an optional review role where there
is a challenge to the administrative decision to expropriate. The second is a mandatory role in
determining the amount and time and manner of payment of compensation. In the latter role,
the court is directly involved in the setting of a price. Absent legislative guidance, this
function is less than inviting for the courts, whose institutional make up is more suited to the
determination of rights, rather than the setting of price. Limiting the role of the courts in also
consonant with the language of the Constitution which requires the courts to “decide” or
“approve” the compensation (s 25(2)(b)). Thus, requiring courts to review the primary
decisions of administrators will not violate the language of the Constitution. An analogy can
also be drawn to the Constitutional Court’s attitude to minimum sentencing legislation
adopted in the late 1990s. The law provided minimum sentences, including life sentences, for
various offences. This was contrary to the ordinary principle that courts would impose
sentences based on the individual circumstances of the offender. It also violated the right not
to be treated in a cruel, inhuman and degrading manner because it could permit a sentence
that was grossly disproportional. As the Court explained;
30
Zondi v MEC for Traditional and Local Government Affairs 2005 (3) SA 589 (CC) at para 82. See also
Lesapo v North West Agricultural Bank and Another [1999] ZACC 16; 2000 (1) SA 409 (CC) at paras 16-7.
31
Mike Campbell (Pvt) Ltd and Others v Republic of Zimbabwe [2008] SADCT 2.
 8

“To attempt to justify any period of penal incarceration, let alone imprisonment for life as in
the present case, without inquiring into the proportionality between the offence and the period
of imprisonment, is to ignore, if not to deny, that which lies at the very heart of human
dignity. Human beings are not commodities to which a price can be attached; they are
creatures with inherent and infinite worth; they ought to be treated as ends in themselves,
never merely as means to an end.”32

Nonetheless, the Constitutional Court upheld the law on the basis that it provided a “safety
valve”. It permitted the Court to alter the minimum sentence if there were “substantial and
compelling circumstances” to do so. This phrase was interpreted to ensure that no grossly
disproportionate sentence could be imposed.
A similar principle could be applied to the determination of compensation. It may well be
consistent with s 25(2) for the state to determine the amount of compensation, subject only to
court interference if there are “substantial and compelling” circumstances to intervene.
The justification for such an approach would be to relieve the bottleneck of requiring
every compensation amount to be determined by a court. It would allow for far quicker
expropriation – and therefore far quicker land reform – to: (a) place the duty on the
landowner to approach a court; and (b) to limit the court’s ability to interfere with the
amount. If such a regime is properly structured, we believe it could pass constitutional
muster.

F Section 25(3): Just and Equitable Compensation

Before we address s 25(3) directly, we briefly consider theories about why we pay
compensation. These are important for determining when compensation should be paid, and
how much should be paid. We then deal with the courts’ current approach to determining
just and equitable compensation, and explain why it is wrong and is likely to be overturned
by the Constitutional Court.

1 Theories of Compensation
There are two general philosophical or economic reasons why compensation is required for
expropriation.
First, compensation is paid to prevent inefficient expropriation. The idea is that requiring
the state to pay compensation prevents the state from expropriating land for a purpose that
will be less economically or socially efficient than the current use of the property.
This rationale obviously breaks down where the land is not currently being efficiently
used, or where the purpose of the expropriation is plainly economically or socially preferable
to the current use. Where expropriation with limited or no compensation incentivises more
efficient state conduct, it will be justifiable under this rationale.
Second, compensation is paid to spread the cost of expropriation. If no compensation is
paid, the owner bears the full cost of whatever public purpose is served. That is not only a
cost in money, but a cost in liberty and ability to participate in the political community.
Compensation is fair because it spreads the burden to the entire community who fund the
expropriation through taxes.
The obvious difficulty with this argument in the South African context of expropriation
for land reform must take into account the historical fact of racialised dispossession. Paying
full compensation will require Black people to fund the purchase of land that was taken from
them under colonialism or Apartheid. Fairness in the South African context requires that the
burden is borne primarily by those who benefited from past wrongs.

32
S v Dodo [2001] ZACC 16; 2001 (3) SA 382 (CC); 2001 (5) BCLR 423 (CC) at para 38.
 9

The other note is that sometimes expropriation will not have consequences that
substantially affect liberty or political participation. Where an owner holds large amounts of
property, and the state only seeks to expropriate a small portion, the impact on the owner is
far lower than if the same property was the owner’s sole source of wealth or income.

2 Section 25(3)
The calculation of just and equitable compensation is governed by s 25(3), which reads:

“The amount of the compensation and the time and manner of payment must be just and
equitable, reflecting an equitable balance between the public interest and the interests of those
affected, having regard to all relevant circumstances, including -
(a) the current use of the property;
(b) the history of the acquisition and use of the property;
(c) the market value of the property;
(d) the extent of direct state investment and subsidy in the acquisition and beneficial
capital improvement of the property; and
(e) the purpose of the expropriation.”

The courts have regularly stressed that s 25(3) does not privilege market value above the
other listed (and unlisted) factors. In Du Toit, for example, the Constitutional Court wrote:
“Section 25(3) indeed does not give market value a central role. Viewed in the context of our
social and political history, questions of expropriation and compensation are matters of
acute socio-economic concern and could not have been left to be determined solely by market
forces.”33 The wording of s 25(3) – which expressly calls for an “equitable balance” – make
that conclusion obvious.
However, the recent decision of the Supreme Court of Appeal in Uys NO v Msiza (Msiza)
undermines that principle. Although the SCA too repeated the mantra that market value is
merely one factor, it adopted two methods that, in practice, make market value the primary
determining factor: (a) the two-stage approach; and (b) the prohibition on “double counting”.

3 The Two-Stage Approach

First, according to the Supreme Court of Appeal, 34 courts must adopt a two-stage approach
to determining just and equitable compensation. In step one, the court determines the market
value of the property. Then, in step two, it decides whether that value should be adjusted up
or down in light of the other s 25(3) factors. This approach is generally justified on the basis
that market value is the only factor that can be objectively quantified.
In our view, the two-stage approach is mistaken and, if it is given the opportunity, the
Constitutional Court is likely to reject it. We say this for three reasons.
One, market value is not the only factor that can yield a numerical value. As the Land
Claims Court has held: “it is easy to work out the history of acquisition, and reference may
also be had to historical records to determine the value of land over time.”35 Put differently,
one can determine how much the current owners paid for the property, and then increase it
based on, for example, the CPI. This is the approach used to determine “just and equitable”
compensation for claimants under the Restitution Act where restoration is not possible. 36
33
Du Toit v Minister of Transport [2005] ZACC 9; 2005 (11) BCLR 1053 (CC); 2006 (1) SA 297 (CC) at para
36.
34
Uys v Msiza at paras 11-13.
35
LCC Judgment at para 38: Record Vol 4, p 387.
36
See Florence v Government of the Republic of South Africa [2014] ZACC 22; 2014 (6) SA 456 (CC); 2014
(10) BCLR 1137 (CC); Farjas (Pty) Ltd v Minister of Agriculture and Land Affairs of the Republic of South
Africa and Others, Rainy Days Farms (Pty) Ltd v Minister of Agriculture and Land Affairs of the Republic of
South Africa and Others [2012] ZASCA 173; [2013] 1 All SA 381 (SCA); 2013 (3) SA 263 (SCA).
 10

Moreover, as Harms JA accepted in Helderberg, the amount of state subsidy is also capable
of determination and yields a concrete value.37
Two, the two-stage approach is inconsistent with the text of s 25(3). As appears from this
case, it can result in an improper disregard for equally relevant factors. As Van Wyk has
recently argued:

“The two-step approach that was developed in the courts is not ideal. One must accept that the
goal is not ‘market value’ but ‘just and equitable’ compensation. A contextual determination
of compensation that aims at the "just and equitable" is required. The courts must be mindful
of this.”38

Or as Langa ACJ explained: “In my view, the Constitution expressly insists upon a
different approach – one which makes justice and equity paramount, not as a second level
“review” test but as the test for the calculation of compensation.”39 Psychological evidence
also strongly suggests that adopting the two stage approach will have an “anchoring effect”
on judges that will cause them to reach an ultimate answer that is close to market value than
if they started with a different number.40
Three, it is a fallacy that market value is objectively determinable. While there may be
cases where valuers agree on the market value, often they disagree wildly. That is especially
the case when there are not many comparable sales of the land. As King AJ put it, the task of
a court determining market value “is an Alice in Wonderland world in which the
consideration of principles of valuation and the opinions expressed by experienced property
valuators make the task of the super valuator seemingly ‘curiouser and curiouser’.”41
Four, in cases where market value is in dispute, it will often be extremely difficult to
extricate the other factors listed in s 25(3) from a determination of market value. Assessing
market value requires the court to consider not only the current, actual value of the land, but
also what the value would have been but for the expropriation. That necessarily requires an
assessment that is intimately connected to “the current use of the property” (s 25(3)(a)), “the
history of the acquisition and use of the property” (s 25(3)(b)), and “the purpose of the
expropriation” (s 25(3)(e)).

4 “Double Counting”
In Msiza v Uys, Land Claims Court determined that just and equitable compensation in a
labour tenant claim was R1.5 million, when the market value was R1.8 million. It did so
because of the use of the land, the history of its acquisition, and the fact that the landowners
knew that labour tenants occupied the land when they bought it.
The Supreme Court of Appeal overturned the reduction from market value. It held that the
37
Helderberg (n 42) at para 19.
38
J Van Wyk ‘Compensation for land reform expropriation’ (2017) TSAR 21 at 27. See also J Zimmerman
‘Property on the Line: Is an Expropriation-centered Land Reform Constitutionally Permissible?’ (2005) 122
SALJ 378 at 417 (criticising the adoption of the two-stage approach and the over-emphasis on market value:
“The shared emphasis on market value in the policy and jurisprudence on the new property clause appears to
stem from an extreme wariness of nonquantifiable Constitutional compensation factors and an over reliance on
transnational or 'universal' compensation doctrine. Where the flexible and unusual provisions of s 25 have the
potential to differentiate the South African property clause from its comparative analogs, they are either
ignored or attributed very cautious and conservative content.”)
39
Du Toit (n 11) at para 84 (emphasis in original).
40
WH Gravett ‘The Myth of Rationality: Cognitive Biases and Heuristics in Judicial Decision-Making’ (2017)
134 SALJ 53. See also E Foster ‘Anchoring and the Expert Witness Testimony: Do Countervailing Forces
Offset Anchoring Effects of Expert Witness Testimony?’ (2009-2010) 77 Tennessee LR 623; B Englich ‘Blind
or Biased? Justitia's Susceptibility to Anchoring Effects in the Courtroom Based on Given Numerical
Representations’ (2006) 28 Law & Policy 497.
41
Southern Transvaal Buildings (Pty) Ltd v Johannesburg City Council 1979 (1) SA 949 (W) at 956A.
 11

various factors that the Land Claims Court had considered to justify the reduction “had all
been taken into account [by the valuer] in considering market value.”42 It was therefore
inappropriate to take them into account again in determining what compensation was just and
equitable.
In our view, this approach is mistaken and is likely to be overturned by the Constitutional
Court. There are two related errors in the SCA’s reasoning. One: the Constitution expressly
requires all those factors to be considered, knowing that some of them would obviously be
relevant in the determination of market value. Two: the weight of the various factors must be
weighed by the court when it determines what is just and equitable, not by the valuer when
she determines market value. Even if the same factors are considered, the purpose is
different, as is the entity determining the weight.
Particularly in the context of expropriation for the purposes of land reform, the history of
the property, the purpose of the expropriation should be given particularly heavy weight.
Msiza precludes that possibility.

5 Msiza should be reversed

The consequence of Msiza is that – without intervention – just and equitable compensation
will be market value by another name. There may be small deviations in unusual cases, but
in most cases courts will decide that market value is just and equitable. While Msiza stands,
is virtually inconceivable that a Court will conclude that it could be just and equitable to pay
no, or nominal compensation.
In our view, that position is mistaken. At least where the expropriation is aimed at land
reform, it will generally be consistent with s 25(3) to pay substantially less than market value.
Both courts and academics have regularly assumed that is the case. In some limited
instances, it may be permissible to pay no compensation or only token or nominal
compensation. We return to what those instances might be below.
But as the law stands after Msiza, courts will not apply existing laws permitting
expropriation to allow compensation below market value. Under Msiza, a law permitting
below market value would be difficult to defend on the basis that it grants “just and equitable
compensation”. This is a serious stumbling block for land reform. In our view, there are
three possible mechanisms to address the problem.
The state or civil society could provide the Constitutional Court with an opportunity to
overrule Msiza. Unfortunately, the Department of Rural Development and Land Reform
elected not to appeal the Msiza judgment. Mr Msiza, too, will not lodge an appeal in the
Constitutional Court. If the Constitutional Court does not hear it, the Government or civil
society should find an alternative test case to bring before the Constitutional Court.
Alternatively, Parliament could pass legislation that expressly provides for an alternative
mechanism to calculate just and equitable compensation at below market value (we suggest
some possibilities below). The Government could then defend that legislation on the grounds
that: (a) Msiza was wrongly decided and the legislation is consistent with s 25(3); or (b) if it
limits s 25(3), the limitation is justifiable under s 25(8).
Lastly, Parliament could amend the Constitution to rewrite s 25(3) to allow for
compensation significantly below market value in certain instances.

6 The Correct Approach

There is no single correct approach to calculating just and equitable compensation. As we
have argued above, the Constitution does not require compensation to be determined on a
case-by-case basis by a court. It permits the Legislature and the Executive to pass legislation
determining how that calculation will occur.
42
Uys v Msiza at para 25.
 12

That does not require, in the first instance, a consideration of all the listed factors. It
permits a more blunt approach to determining what is just and equitable, as long as it is
subject to correction by courts where the application has unduly harsh consequences.
A factor that is not mentioned in s 25(3), but that seems particularly relevant is the impact
on the owner. The expropriation of the same property may have very different impacts. If it
is owned by a company and the property constitutes a small portion of the company’s total
wealth, expropriation with no or little compensation will have a limited impact. But
expropriating the same property from a person who has no other meaningful assets will have
a particularly serious impact. The principle is the same as progressive taxation. Those who
own more can be expected to contribute more. Those who own less, should contribute less.

7 Timing of Compensation
Section 25(3) requires not only that the amount of compensation is just and equitable, but
also that “the time and manner of payment must be just and equitable”. aaaaaaaaIn
Haffejee,43 the Constitutional Court held that this does not necessarily mean that
compensation must be determined and paid prior to the expropriation. While it will generally
be just and equitable to determine the compensation before expropriation, there will be
situations where the determination of the amount can occur after expropriation. This is
particularly so given that, even if land is expropriated, the former owners will be occupiers
entitled to the protections against eviction in s 26(3) of the Constitution.
This is an important possibility. One of the stumbling blocks to land reform is that
landowners can drag out a dispute about compensation. While that occurs, the land remains
in the hands of the landowner. In Msiza, for example, there was a delay of fourteen years
after he was awarded the land to the judgment by the SCA.
One way to overcome that obstacle is to provide for a bifurcated process for determining
compensation. The state first determines administratively – by a state valuer, or applying a
formula. Once that amount is determined, it is paid and the property can be expropriated. If
the landowner is dissatisfied with the amount, she is free to approach a court to approve the
amount. But doing so will not prevent the expropriation.
That approach preserves the right of the landowner in terms of s 25(2) to have the
expropriation “approved” by a court, and ensures that the time and manner of the
compensation is just and equitable.

G Section 25(8): Limitation

To the extent that any law contravenes one of the provisions of ss 25(1)-(3), it may be
justified in terms of s 25(8). The provision reads:

“No provision of this section may impede the state from taking legislative and other measures
to achieve land, water and related reform, in order to redress the results of past racial
discrimination, provided that any departure from the provisions of this section is in
accordance with the provisions of section 36(1).”

The import of s 25(8) is accurately summarised by the following passage:

“Section 25(8) clearly envisions categories of expropriated land for which 'market value'
should not factor into the compensation analysis; legislation authorizing expropriations for the
purpose of land redistribution should accordingly be justifiable under s 36 to the extent that it
is procedurally sound and duly takes into account the welfare and livelihood of expropriatees.
Proportionate concern for the individual or family affected by an expropriation is the test of s
43
Haffejee NO and Others v eThekwini Municipality and Others [2011] ZACC 28; 2011 (6) SA 134 (CC); 2011
(12) BCLR 1225 (CC)
 13

36. Proportionality may require that expropriatees be left with more than merely the means to
make a living, but it will not always demand an enquiry into the market value of the land
expropriated. It is far from obvious, for example, that large landholders should receive
compensation at some percentage of market value where they stand to lose only a fraction of
their land wealth. It may very well be 'reasonable and justifiable' under the proportionality
analysis of s 36 to forego calculation of market value where the expropriated family retains
enough land wealth to secure a very comfortable livelihood. It must never be forgotten that
the Constitution is committed to addressing the appalling conditions of poverty under which
the vast majority of the rural population lives. There is only so much wealth to go around in
South Africa; this inescapable fact suggests that some amount of redistribution from rich to
poor, white to black, is inevitable if the promise of the new democratic order is to be
realized.”44

As Zimmerman points out, s 25(8) has an interesting drafting history. It is the product of
compromise – the ANC’s desire to ensure that property rights did not impede land reform,
and the NP’s goal to protect white property rights from unjust expropriation. It was amended
during last minute negotiations to subject land reform to the proportionality review of the
general limitations clause.45 Some academics argue that the strange juxtaposition of the first
and second sentences that resulted from the negotiations render the provision superfluous.46
We disagree. In our view, s 25(8) serves two clear purposes.
Interpretation: It must influence the interpretation of ss 25(1) to (3). It means that
deprivations that are necessary for land reform are less likely to be arbitrary. It means that
limitations on the role of courts in s 25(2)(b) in the land reform context should be more
tolerable. And it should influence the outcome of the s 25(3) analysis, particularly by
emphasising the purpose of the expropriation when land reform is at stake. As noted earlier,
we have relied on this aspect of s 25(8) in our discussions of ss 25(1)-(3) above.
Limitation: If a law is found to limit s 25(2)(b) or s 25(3),47 s 25(8) “should weigh heavily
in on the side of the poor and landless in any proportionality review under s 36.”48 Put
differently, if the criterion in s 25(8) is met, it is a strong indication that the law is
“reasonable and justifiable in an open and democratic society”.
The more difficult question is when an interpretation or application of ss 25(1)-(3) can be
said to “impede” land reform. Impede means to obstruct, hinder, or interfere. We consider
what sort of evidence would need to be put up to make the case that paying more (or any)
compensation would impede land reform.
Lastly, s 25(8) refers to s 36(1) of the Constitution. It provides:

“(1) The rights in the Bill of Rights may be limited only in terms of law of general
application to the extent that the limitation is reasonable and justifiable in an open
and democratic society based on human dignity, equality and freedom, taking into
account all relevant factors, including -
(a) the nature of the right;
(b) the importance of the purpose of the limitation;
(c) the nature and extent of the limitation;
(d) the relation between the limitation and its purpose; and
(e) less restrictive means to achieve the purpose.”

44
Zimmerman.
45
Zimmerman at 413-5.
46
See, for example, Van der Walt Comparative Analysis at 147; Budlender et al Juta’s New Land Law (1998) at
1-73.
47
It is difficult to imagine how s 25(8) could be relied on to conclude that a limitation of s 25(1) – with its
flexible, potentially proportional standard – could still be justified. See, for example, Opperman at paras 73-5.
48
Zimmerman at 416.
 14

The meaning of the provision was accurately summarised by the Court in S v Manamela:

“In essence, the Court must engage in a balancing exercise and arrive at a global judgment on
proportionality and not adhere mechanically to a sequential check-list. As a general rule, the
more serious the impact of the measure on the right, the more persuasive or compelling the
justification must be. Ultimately, the question is one of degree to be assessed in the concrete
legislative and social setting of the measure, paying due regard to the means which are
realistically available in our country at this stage, but without losing sight of the ultimate
values to be protected.”49

In our view, the core issue in determining whether a no or limited compensation regime
meets this proportionality standard is its flexibility. While strict rules make land reform more
efficient, they tend to lead to unjust results in certain cases. Courts will find those measures
unattractive, particularly in this context where the Constitution assigns them a specific role.
The key in designing measures that pass the proportionality requirement will be to balance
efficiency with flexibility. The other key will be to produce evidence to show that less
restrictive means (systems that provided greater or more individualised compensation) would
impede land reform. We return to this below.

H The Structure of Expropriation Analysis

As we make clear below, the basic substantive questions that a court will ask when it
considers whether expropriation without compensation is constitutional are clear. But how
those questions arise within the s 25 analysis is complicated. Because we do not believe this
difference is likely to affect the outcome of any challenge to legislation authorising
expropriation without compensation, we do not seek to answer it conclusively. However, it is
important to be clear about the possible routes to the key substantive questions addressed
below. There are subtle differences that – although they may not determine the outcome –
will determine how the arguments are presented.
There are two possible approaches to the analysis.
First, the approach set out by the Constitutional Court in FNB. On this approach, a court
must always ask the following questions, in the following order:

“(a) Does that which is taken away … amount to “property” for purpose of section 25?
(b) Has there been a deprivation of such property?
(c) If there has, is such deprivation consistent with the provisions of section 25(1)?
(d) If not, is such deprivation justified under section 36 of the Constitution?
(e) If it is, does it amount to expropriation for purpose of section 25(2)?
(f) If so, does the deprivation comply with the requirements of section 25(2)(a) and (b)?
(g) If not, is the expropriation justified under section 36?”

Under this approach, a court must always first consider whether a deprivation (including
an expropriation) is arbitrary within the meaning of s 25(1). Given the nature of the
arbitrariness test, it is very unlikely that ss 25(2) or (3) will alter the outcome. Put differently,
“an expropriation that complies with the justification requirements for expropriation will
also be a deprivation that complies with the requirements for non-arbitrariness, and an
expropriation that does not comply with the justification requirements for expropriations will
also not comply with the requirements for non-arbitrariness.”50 As Roux has pointed out,
this approach “renders all of the other stages of the constitutional property clause inquiry

49
[2000] ZACC 5; 2000 (3) SA 1 (CC); 2000 (5) BCLR 491 (CC) at para 32.
50
Rautenbach at 757.
 15

largely redundant, thus undermining the capacity of these stages to perform a rule-setting
function”.51
The alternative view is that, when the law at stake provides for an expropriation, the Court
should not conduct the s 25(1) analysis, but should instead determine whether the law
complies with ss 25(2) and (3). This is the approach followed in other countries, and places
far greater emphasis on the distinction between deprivation and expropriation.
In our view, the second approach is preferable. The Constitution sets out clear
requirements for lawful expropriation in ss 25(2) and (3). An expropriation that meets those
requirements could never be regarded as arbitrary in terms of s 25(1). There is, therefore, no
purpose in making that far more general inquiry.
However, for the purposes of this paper, it matters not. The key substantive factors that
drive both the arbitrariness analysis under s 25(1), and the expropriation analysis under ss
25(2) and (3) (with both tests applied in light of ss 25(4)-(8)) are likely to be the same.

III INTERNATIONAL LAW

In this Part, we briefly consider the relevant international law. We do so for three reasons.
First, customary international law is law in the Republic. 52 If it limits the ability to
compensate without expropriation, that must be considered when the Legislature acts.
Second, South African has international obligations in terms of multilateral regional treaties,
bilateral investment treaties (BITs) and customary international law. As a matter of
constitutional principle, South Africa is not obliged to comply with those obligations on the
domestic plane. But it should be aware whether any domestic law will breach its
international duties. And third, international law is relevant for interpreting the Bill of
Rights.53
We consider four sources of international law that touch on compensation for
expropriation: Customary international law; Bilateral investment treaties; The South African
Development Community Treaty (SADC Treaty); and The African Charter on Human and
People’s Rights (African Charter).

A Customary International Law

Customary International Law (CIL) recognises states’ right to expropriate property within
their territory, including property belonging to foreign nationals. However, when it comes to
property owned by foreign nationals, CIL imposes certain requirements on such an
expropriation. First, the expropriation must be for a public purpose and must not be
discriminatory. Second, the expropriation must lawful. This second requirement means that
the expropriation must not breach principles of international law, including treaties.54
If an expropriation is unlawful, then the normal principles of state responsibility apply,
and the breaching state may be liable to pay for compensation and damages for its unlawful
51
T Roux ‘The “Arbitrary Deprivation” Vortex: Constitutional Property Law after FNB” in S Woolman & M
Bishop (eds) Constitutional Conversations (2008) 265 at 275.
52
Constitution s 232.
53
Constitution s 39(1)(b).
54
TW Bennet and J Strug Introduction to International Law (2013) at 182; Patrick J. Smith ‘Determining the
Standard Compensation for the Expropriation of Nationalised Assets: Themes for the Future’ 23 Monash U. L.
Rev. 159 (1997) at 159; C. F. Amerasinghe ‘Issues of Compensation for the Taking of Alien Property in the
Light of Recent Cases and Practice The International and Comparative Law Quarterly, Vol. 41, No. 1 (Jan.,
1992), pp. 22-65 at 37-8. L. Yves Fortier & Stephen L. Drymer ‘Indirect Expropriation in the Law of
International Investment: I Know It When I See It, or Caveat Investor’ ICSID Review - Foreign Investment Law
Journal, Volume 19, Issue 2, 1 October 2004, pp 293–327 at 295-6; Kenneth M. Siegel, The International Law
of Compensation for Expropriation and International Debt: A Dangerous Uncertainty, 8 Hastings Int'l & Comp.
L. Rev. 223 (1985) at 234.
 16

expropriation.55 For example, to expropriate property in violation of a treaty will give rise to
the liability of the breaching state vis-à-vis the innocent state (and its foreign nationals who
are owners of that property).
The first important point to note is that CIL imposes no restrictions on states’ rights to
expropriate its own nationals’ property. CIL generally regulates relationships between states,
not between states and their own citizens. However, the fact that CIL places limitations on
the expropriation of foreign-owned property is something that must be considered in
designing any alternative expropriation regime.
With regard to foreign-owned property, the pertinent question is whether compensation is
a further requirement for a lawful expropriation of foreign property under CIL. It is.
Although there have been historical objections to the obligation to pay compensation, the
generally accepted position in CIL is that there is an obligation to pay some compensation.
The debate is about the nature of the compensation that is required. Various developed
states (especially the United States of America) have asserted that CIL requires “prompt,
adequate, and effective” compensation.56 This is known as the Hull Formula, and its
supporters argue that it represents the default, traditional rule in CIL. It translates to
compensation in the form of market value, and placing the owner of the property in the
patrimonial position they would have been in but for the expropriation.
However, the Hull Formula – and the idea of market value compensation – is certainly not
accepted in CIL. The Resolution on Permanent Sovereignty over Natural Resources 1803
(XVII) of 1962 reads:

“Nationalization, expropriation or requisitioning shall be based on grounds or reasons of

public utility, security or the national interest which are recognized as overriding purely
individual or private interests, both domestic and foreign. In such cases the owner shall be
paid appropriate compensation, in accordance with the rules in force in the State taking
such measures in the exercise of its sovereignty and in accordance with international law.”

What amounts to appropriate compensation can vary significantly. This approach

suggests that South Africa would be afforded significant leeway to determine how much
compensation to pay – as long as it pays some compensation.

B BITs
We have not considered the BITs that South Africa has in fact concluded. We note only
that it is common for BITs to include provisions regulating the expropriation of land held by
nationals of the foreign state. The precise provisions of those BITs will determine the nature
of South Africa’s obligations, and whether they limit the legislature’s ability to expropriate
with no or limited compensation.

C African Charter
Article 14 of the African Charter reads: “The right to property shall be guaranteed. It may
only be encroached upon in the interest of public need or in the general interest of the
community and in accordance with the provisions of appropriate laws.” Article 14 is
55
Chorzow Factory Case (Indemnity) (Merits) (1928) PCIJ Reports, Series cA, No 17, 46; Amoco International
Finance Corp v Iran (1987) 15 Iran-USCTR 189, 193; Amerasinghe op cit at 37-8.
56
Heibein op it at 767; C. F. Amerasinghe ‘Issues of Compensation for the Taking of Alien Property in the
Light of Recent Cases and Practice The International and Comparative Law Quarterly, Vol. 41, No. 1 (Jan.,
1992), pp. 22-65 at fn 10; Libyan American Oil Co. Arbitration (1977), (1982) 62 I.L.R. 140 at 207; Bernard
Kishoiyian ‘The Utility of Bilateral Investment Treaties in the Formulation of Customary International Law’ 14
Nw. J. Int'l L. & Bus. 327 (1993) at 328; Brice M. Clagett ‘The Expropriation Issue before the Iran-United
States Claims Tribunal: Is Just Compensation Required by International Law or Not’ 16 Law & Pol'y Int'l Bus.
813 (1984) at 815.
 17

elaborated on in the Pretoria Declaration on Economic, Social and Cultural Rights in Africa
of 2004. The Declaration is not binding, hard law, but is soft law used to interpret the
Charter. Item 5 reads:

“The right to property in article 14 of the Charter relating to land and housing entails among
other things the following:
a. Protection from arbitrary deprivation of property;
b. Equitable and non-discriminatory access, acquisition, ownership, inheritance and
control of land and housing, especially by women;
c. Adequate compensation for public acquisition, nationalisation or
expropriation;
…
e. Equitable redistribution of land through due process of law to redress historical
and gender injustices;
f. Recognition and protection of lands belonging to indigenous communities”

The African Commission on Human’s and People’s Rights has held that the right in art 14
accrues to nationals of a state and to foreign nationals who own property in the state. 57 That
right cannot be violated except “in the interest of the public or in the general interest of the
community” and “in conformity with the provisions of the appropriate laws”; this latter aspect
referring to domestic and international laws.58 This makes the international laws discussed in
the previous section relevant to interpreting art 14.
The limitations to the right to property should be determined in the light of the principle of
proportionality, meaning that interference in the right to property must be “proportional to a
legitimate need, and should represent the least restrictive measure possible”.59 Limitations, to
be lawful, also require “adequate compensation determined by an impartial court of
competent jurisdiction”.60
Accordingly, under the African Charter, South Africa is required to pay “appropriate” or
“adequate” compensation when it expropriates property. Given the express acknowledgment
in the Pretoria Declaration about the need for historical redress, it is unlikely that appropriate
will always demand market value compensation. However, it is unclear whether no
compensation can ever be “appropriate”.
There is one communication of the African Commission that held expropriation without
compensation was inconsistent with the Charter. 61 But that in that instance the law of the
relevant country – the DRC – required compensation. It is far from certain whether the type
of narrow exceptions permitting expropriation without compensation would be inconsistent
with the Charter.
In our view, South Africa is likely to be afforded a margin of appreciation in this area. As
long as it has clear laws and permits judicial intervention to avoid unnecessarily harsh results,
South Africa should be able to defend its laws.

D SADC Treaty
The SADC Treaty contains no express provision regulating expropriation. However, the
SADC Tribunal has held that compensation is a requirement as a matter of SADC law. In

57
Communication 286 /2004 – Dino Noca vs Democratic Republic of the Congo para 128-9.
58
Dino Noca para 144; Centre for Minority Rights Development (Kenya) and Minority Rights Group
International on behalf of Endorois Welfare Council vs. Kenya, Communication no 276/2003, May 2009, par.
211.
59
Endorois Welfare Council para 214; Dino Noca para 145.
60
Dino Noca para 147.
61
Dino Noca.
 18

Campbell v Zimbabwe,62 the Tribunal had to consider Zimbabwe’s laws that permitted
expropriation of land without compensation of agricultural land for resettlement purposes.
The Tribunal held that the laws were contrary to Zimbabwe’s international obligations:

“It is difficult for us to understand the rationale behind excluding compensation for such land,
given the clear legal position in international law. It is the right of the Applicants under
international law to be paid, and the correlative duty of the Respondent to pay, fair
compensation. Moreover, the Respondent cannot rely on its national law, its Constitution, to
avoid an international law obligation to pay compensation.” 63

It is not clear what the source for the supposed “clear legal position” was. As we have
explained above, CIL only requires compensation for the expropriation of foreign-owned
land. The African Charter requires compensation, but it is not clear that the SADC Tribunal
had the authority to enforce that treaty against Zimbabwe. The Tribunal’s ruling has been
criticised on exactly these grounds.64
In our view, the Tribunal’s finding should not be seen as an absolute international
prohibition on expropriation without compensation:
The ruling is legally dubious. Outside the African Charter, there is no international
obligation on South Africa to compensate nationals. The Tribunal has no power to create
new legal obligations on states.
The Zimbabwean law was particularly wide and unlimited. It applied to a wide swathe of
land, without exception, and without any judicial limitations. The types of laws we have
motivated for in this paper are far narrower, and far more likely to withstand international
scrutiny.
It is unclear whether the SADC Tribunal will continue to have jurisdiction to resolve
complaints by individuals. While, in theory, it retains that jurisdiction it is not currently
operational and a recent protocol seeks to remove that jurisdiction; although that protocol has
yet to come into force.65

IV EXPROPRIATION WITH NO OR LIMITED COMPENSATION

In this Part, we consider various options that depart from the current approach of either
“willing buyer willing seller” or “close to market value” that the state currently applies. With
regard to each option, we explain whether it would be consistent with s 25(1)-(3), or s likely
to constitute a limitation of the right in terms of s 25(8).
We stress that it is difficult to predict with certainty how the Constitutional Court will react to
notional pieces of legislation. There are three core issues that will determine whether any of
the proposed measures will survive constitutional attack. First, the evidence that the state is
able to produce that adherence to the current practice will “impede” land reform. Second,
whether the system seeks to balance the nature of the loss with the amount of compensation
based on justifiable factors. Third, the flexibility of the system – courts will be more likely to
uphold a system that does not set absolute rules, but allows deviation in special
circumstances.

62
Campbell.
63
Ibid.
64
Gino J Naldi ‘Mike Campbell (Pvt) Ltd et al v The Republic of Zimbabwe: Zimbabwe’s Land Reform
Programme Held in Breach of the SADC Treaty’ (2009) 53 Journal of African Law 305 at 317-8.
65
See generally Law Society of South Africa and Others v President of the Republic of South Africa and Others
[2018] ZAGPPHC 4.
 19

A Compensation will Impede Land Reform

The key issue to justify any expropriation without compensation – or even with limited
compensation – will be to show that the legislation is a proportionate measure to avoid an
impediment to land reform. This requires an assessment of the following questions. What
qualifies as “land, water and related reform in order to redress the results of past racial
discrimination”? Can the state show that it does not have sufficient funds to pay more
compensation? Why would it impede land reform to allow a court decide on a case-by-case
basis what amount of compensation is justified? Given the failure to efficiently implement
existing measures for land reform, can the state rely on the cost of paying “full”
compensation as a justification?

1 What is land reform?

Section 25(8) does not apply to all laws. It is triggered only by “legislative and other
measures to achieve land, water and related reform, in order to redress the results of past
racial discrimination”. The Constitution recognises three forms of land reform, all three of
which are designed to redress racial discrimination: (a) Promotion of “equitable access” to
land envisaged under ss 25(4) and (5) (land redistribution); (b) Provision of secure tenure
under s 25(6) (tenure reform); and (c) Restitution of land dispossessed after 19 June 1913 in
terms of s 25(7) (land restitution).
Legislation designed to advance any of these three goals will certainly meet the definition
of land reform to redress racial discrimination in s 25(8). Given the breadth of land
redistribution, this will cover a wide range of possible approaches to land reform.
Land reform is often regarded relating primarily to agricultural land. There is no reason
why that should be the case. The idea of “equitable access” to land is not only about the
amount of land that is owned by various races. It also relates to where that land is owned.
Measures that are designed to address apartheid urban geography should also qualify under s
25(8). For example, measures to access land in inner cities to make them available to Black
people who have historically been forced to live on the outskirts of cities could be justified
under s 25(8).
The “water and related reform” is harder to define with precision. It would obviously
include measures to ensure that Black farmers have equitable and sufficient access to water to
make use of their land.

2 Budget
The obvious rebuttal to any attempt to expropriate without compensation is that if the state
wants to expropriate, it should simply re-allocate resources so that it can pay more
compensation. Put differently, the argument will be that paying close-to-market-value
compensation will not impede land reform because the state can afford to pay compensation
at that level.
There is precedent for this type of argument. In Blue Moonlight the Court held that the
City of Johannesburg’s failure to budget to provide alternative accommodation to people
evicted from private residences was no justification for not providing that alternative
accommodation.66 The analogy would be that if the state’s budget cannot be relevant to
determining whether, and to what extent, it should pay compensation under s 25.
There are two problems with this type of argument. First, the issue in Blue Moonlight
concerned only one constitutional obligation – the obligation to provide alternative
accommodation – and the failure to budget for it. Other constitutional obligations were
66
City of Johannesburg Metropolitan Municipality v Blue Moonlight Properties 39 (Pty) Ltd and Another 2012
(2) SA 104 (CC) at para 74 (“it is not good enough for the City to state that it has not budgeted for something, if
it should indeed have planned and budgeted for it in the fulfilment of its obligations.”)
 20

affected only indirectly by reductions in budget to meet those ends. In this context, there are
two constitutional obligations directly at play: the obligation to pay just and equitable
compensation, and the obligation to reform unequal access to land. These competing
obligations are mediated in the ways identified above.
Nonetheless, a proper calculation of the comparative cost of paying compensation is likely
to be extremely persuasive to a court. This analysis would seek to compare the cost of paying
compensation at close-to-market-value, combined with the administrative cost of judicial
determination on a case-by-case basis with the cost of paying limited or no compensation,
and the cost of an alternative administrative means of determining compensation (with a
judicial safety valve). If that is combined with a significant increase in the budget for land
reform, a court will be hesitant to question the state’s decision about how to allocate
resources.
However, if there is no meaningful increase in the resources allocated to land reform, and
no assessment of the consequences of paying greater compensation, a court will be less likely
to accept that the measure meets the purpose of s 25(8) – whether in its role of interpreting
the rights in ss 25(1)-(3), or limiting those rights. The goal for the state would be to
demonstrate that paying compensation according to the current system would “impede” land
reform because it would significantly delay the its ability to transfer land as it waited for
funds to become available.

3 The determination will impede land reform

The prior section considered how the budget necessary to fund the current approach will
impede land reform. But the system might also cause delay. As we noted earlier, both the
Restitution Act and the Land Reform Act require that, if there is no agreement about
compensation, it must be determined by a court. This has several negative consequences for
speedy land reform.
It disincentivises owners from agreeing to compensation, at least while the Msiza approach
prevails. It incentivises the state to pay market value to avoid the delay and cost of litigation.
It incentivises landowners to force litigation in order to either delay the expropriation, or
force a better offer. All of this increases both the time and the cost of land reform.
In our view, the major advantage of a system that provides for no compensation, or
predetermined compensation is to avoid these perverse incentives. It removes the ability for
landowners to delay expropriation or push up the price. And it will avoid delays in
expropriation while any fights about compensation are determined. The state will have to
demonstrate – based on evidence of how the current system has operated – that it impedes
land reform. There are several studies that already provide this type of evidence.

4 The State is Responsible for Delays

A strong likely rebuttal to any reliance on s 25(8) is that it is not the existing legislation that
causes the impediment, but the state’s failure to properly implement it. This is certainly
partially true. The state has been inexcusably remiss in its implementation of both the
Restitution Act and the Labour Tenants Act. It has also failed to enforce the protections
provided in ESTA. If it had properly implemented the existing statutes, and taken full
advantage of all the mechanisms they provide, there is no doubt that substantially more
progress would have been made.
It must also be accepted that reducing compensation and altering the manner in which it is
determined are not, on their own, the answer to all issues regarding land reform. The state
must still put in place the mechanisms to speedily assess contested restitution and labour
tenant claims by significantly increasing the capacity of the Land Claims Court and the
alternative dispute resolution mechanisms those statutes permit. It must also provide the
 21

necessary pre- and post-restitution support to communities who are given land to ensure that
the make effective and equitable use of the land, and ensure that people – particularly in the
former homelands – have secure tenure to their own land and are not subject to the whims of
traditional leaders. Without these and other related reforms and investments land reform will
remain ineffective even with alternative compensation mechanisms.
Nonetheless, the history of neglect and the existence of additional impediment will not, in
our view, be a bar to relying on s 25(8) to justify compensation reforms. The government’s
obligations under ss 25(4) to (7) cannot be affected by its past conduct. This government,
today, must take reasonable measures to advance land reform. Measures cannot become
unreasonable or unjustified because the government, in the past, acted unreasonably. The
government should be able to show that even if it had fully applied the existing measures,
they would still constitute an impediment to land reform for the budget and system reasons
set out above.
The existence of other impediments does not mean that the calculation and determination
of compensation is not an impediment. As long as the legislature does not treat expropriation
with limited or no compensation as the silver bullet for land reform, and demonstrates a
commitment to address the other issues, a court is unlikely to hold that the existence of other
impediments means the government cannot address one of them.
In sum, any defence of limited compensation laws must acknowledge the existence of past
failures and other obstacles. But the government should still be able to justify limited or no
compensation laws.

B No Compensation
The first option is to expropriate property without any compensation. There are three
possible ways this can be achieved under s 25.
First, the legislation could be constructed so that the loss of property is a deprivation, not
an expropriation. This can be done by transferring the land directly from existing landowners
to new landowners. The law will then have to be justified in terms of s 25(1).
Second, the state could argue that, in a very limited set of circumstances, expropriation
without compensation s “just and equitable” under s 25(3).
Third, the state could accept that expropriation without compensation would limit either
s 25(1) or s 25(3), but seek to justify the limitation in terms of s 25(8).
Whichever theoretical route is used to justify expropriation without compensation, the
fundamental questions are likely to be the same. Are the categories narrowly defined to
cover situations where the ordinary justifications for compensation do not apply? Is
expropriation without compensation reasonably necessary to further land reform? Is provision
made for exceptions where expropriation without compensation would cause especially harsh
results?

1 Narrow Categories
Allowing compensation without any compensation is unlikely to be upheld by the Court in
general circumstances. However, if the Legislature identifies specific categories where
legislation
In our view, there are four possible circumstances where property could be expropriated
without compensation: (a) the land is abandoned or unused; (b) the land is held purely for
speculative purposes; (c) the land is under-utilised and owned by public entities; or (d) the
land is actively farmed by labour tenants in the absence of a title deed holder. This applies
only to expropriation without compensation in these instances will be for the purpose of land
reform and not for other public purposes such as the building of roads, dams and so on.
 22

Expropriation without compensation outside the land reform context will be extremely
difficult to justify.
What justifies treating these particular categories of land as not warranting any
compensation? Two issues stand out. One: the common theme for all the categories is that
there is no emotional connection to the land in any of those cases. The owner will suffer, at
worst, pure economic loss. In some situations there will be little or no loss at all. Two: the
land is not being used productively. The justification for land reform is both to redress
historic wrongs, but also to ensure that access to land is “equitable”. Allowing land to be
unutilised, while others are landless – even if it is not subject to a specific restitution or
labour tenant claim – does not promote “equitable” access to land. It may be possible to
identify other situations that meet these two criteria – only economic meaning, and unutilised
– that would also justify expropriation without compensation.
Whichever of the three technical options is used – s 25(1), s 25(3) or s 25(8) – these two
factors are the key bases to justify expropriation without compensation.
Section 25(1): when assessing whether the deprivation is arbitrary, the nature and extent of
the loss is compared to the reason for it. Ordinarily the complete deprivation of land requires
strong justification. But in these instances the land only has economic meaning to the owner.
It is, in some ways, the equivalent of taxation rather than expropriation. The justification is
strong because: (a) it is to meet the purposes in ss 25(4) to (8); and (b) the land is not
currently serving a useful purpose.
Section 25(3): if the payment of zero compensation is defended as just and equitable under
s 25(3) the “current use” of the property, and the “purpose of the expropriation” strongly
support reducing compensation from market value.
Section 25(8): even if a court concludes that zero compensation cannot be justified under s
25(3), the legislation can be defended under s 25(8). The key issue here will be the showing
that compensation with compensation would impede land reform.

2 Exceptions
Even though the four categories are identified to be situations where expropriation without
compensation will not be unduly harsh on the current landowners, there are likely to be
situations where it will have a disproportionate impact to expropriate without any
compensation. For example, the land may be abandoned, but there may be good reasons for
the present abandonment, and a reasonable intention to use the land in the future. Or the land
may be held for speculative purposes, but it may be the only investment of the landowner.
Expropriation without compensation may leave the landowner destitute.
The way to deal with these possibilities is, as foreshadowed earlier, to allow landowners to
approach a court to contest whether their land falls in the specified category and argue that,
even though they are in the category, there are substantial and compelling circumstances that
justify a departure from the default rule of zero compensation.
In order not to defeat the purpose of not impeding land reform, it will be important that
this opportunity to approach a court is a burden that rests on the landowner and does not
prevent the expropriation from continuing while the challenge is determined.

C Limited Compensation
In our view, expropriation without any compensation will only be justified in the limited
situations identified above (and other similar situations). It is doubtful whether expropriating
that land alone will be sufficient to meet the demands for land reform. What may provide a
far more effective way to achieve land reform is to combine expropriation without
compensation in limited circumstances with limited compensation in other circumstances.
The limits relate to both the determination and the calculation of compensation.
 23

1 Calculation
There are a variety of possible ways in which compensation can be determined other than by
placing competing valuer reports before a court. Some options include:
o Determination by a state valuer with clear guidelines for how to weigh competing
concerns. There could also be an internal appeal mechanism that would be
cheaper and more efficient than a court. This avoids the difficulty of competing
valuation reports. But it will still be an expensive exercise to set up the
bureaucracy of a state valuer.
o Compensation could be determined by market value reduced by a particular
percentage. For example, 50% of market value. Of course, this still relies on a
determination of market value. This option
o An alternative value could be used, such as the municipal valuation, or an altered
version of the municipal valuation. The advantage of this approach is that the
municipal valuation will already be determined.
o Or it could be calculated with reference to the original purchase price adjusted
according to inflation. The Constitutional Court has accepted this is an acceptable
method to determine just and equitable compensation for restitution claimants
where restoration is not possible.
o The government could determine a formula that considers multiple factors –
municipal valuation, previous purchase price, current use and others – to
determine the compensation payable.
o A flat rate could be payed per hectare, which is predetermined for different areas
and different types of land.
We do not express a view about which of these options would best serve the state’s goals.
Whatever method might be adopted would have to balance competing concerns: efficiency
and cost of calculation, the amounts the method will generate, the flexibility, and the
likelihood that it could be defended against constitutional attack. The ability to defend any
particular method of calculation will also depend on the extent to which the initial calculation
can be altered by judicial intervention.

2 Judicial Role
Requiring courts to make the initial and final determination of compensation is inefficient.
Excluding courts altogether will almost certainly be unconstitutional. There are multiple
ways to conceive a judicial role between those two extremes.
There are three questions. First, how does the court become involved? The calculation
could be subject to automatic review by the Court, or to optional review by the owner or the
government. The legislation could also set the time periods within which reviews must be
brought and determined.
Second, on whom does the onus lie? Must the state justify the amount, or must the owner
put up evidence to show that the amount is not justifiable?
Third, what standard will the court apply? The court could review on ordinary
administrative grounds. Or it could be limited to a “substantial and compelling” type
standard. The legislation could also specify the types of substantive considerations that the
Court should take into account – the prejudice to the owner, the deviation from market value
and so on.
These are difficult systemic questions. But even the most restrictive conceptualisation –
review only by the owner, who bears the onus to show substantial and compelling
circumstances – will likely pass constitutional muster if the evidentiary basis for s 25(8) is
met.
 24

V CONCLUSION
Is it constitutionally permissible to expropriate land or property without compensation? That
is the question we are asked to answer in this paper. The answer is yes. But only in limited
circumstances, and subject to procedural and judicial safeguards.
The Constitution permits expropriation without compensation in one of three ways. First,
as a non-arbitrary deprivation in terms of s 25(1). If the state does not acquire property, but
transfers it from one private party to another, it does not expropriate property and is not
obliged to pay compensation. Second, as just and equitable compensation in terms of s 25(3).
The courts’ current interpretation of “just and equitable compensation” would likely exclude
this possibility. But in our view, the Constitutional Court is likely to uphold an interpretation
that permits no or little compensation to be paid if the purpose is to advance land reform.
Third, as a limitation of the rights in ss 25(1) or (3) as permitted by s 25(8). Section 25(8)
makes it clear that the other parts of s 25 should not “impede land reform”. If the
government can show that paying compensation would impede land reform, and paying no
compensation is proportional in the circumstances, it will be permissible under s 25(8) to
either expropriate with limited or no compensation.
We are unable to find any rational basis for the argument that expropriation without
compensation should apply as a default position to all circumstances. Expropriation of
property is used in two ways under the Constitution. First, as a measure to enable the state to
acquire property for public purposes, and second as a measure to facilitate access to property
in the broader public interest, particularly by persons who previously were denied access to
property by state action. In the first category (public purposes) expropriations tend to affect
everybody regardless of background. Many black persons are compelled to give up their
property to enable the state to fulfil public goals. No rational argument exists why those
people should not be compensated. However, in the second category (of land expropriated in
the broader public interest) different considerations apply. Property may have been acquired
through racial discrimination, and may be required to achieve the constitutional goals of
equality and access to land.
In our view, there is no clear line or closed list of situations where expropriation without
compensation will be constitutionally permissible. The key constitutional standard is
proportionality between the loss of the owner, and the public benefit of the expropriation.
Because the Constitution demands land reform, expropriations for that purpose may be
justifiable with no or limited compensation. Expropriation is more likely to be justified
where it has the following characteristics: (a) the loss suffered by the owner is purely
economic; and (b) the land is not currently being used productively by the owner.
The precise categories may be subject to debate, but the following situations are likely to
justify expropriation without compensation: (a) abandoned land; (b) hopelessly indebted land;
(c) land held purely for speculative purposes; (d) unutilised land held by state entities; (e)
land obtained through criminal activity67; and (f) land occupied by labour tenants.68
Whether these or similar categories of land are chosen, to survive constitutional scrutiny
legislation permitting expropriation without compensation must ensure that: (a) the categories
will have to be carefully defined; (b) expropriation without compensation will not have
unduly harsh consequences for the owner; and (c) the expropriation without compensation
will have to be subject to the procedural limits set out above.

67
We have not considered the consequences for the application of legislation such as Prevention of Organised
Crime Act 121 of 1998.
68
This is a slight exception as the justification for expropriation without compensation is also that the labour
tenants have, through their (and their forefathers’) labour, already compensated the owner.
 25

Under international law, it is generally unlawful to expropriate foreign-owned property

without compensation. However, it is permissible to pay significantly below market value.
Legislation may have to exclude foreign-held land from being expropriated without any
compensation.
The Constitution requires that the amount of compensation to be paid must be
“determined or approved” by a court. It will be unworkable to require courts to determine in
the first instance whether expropriation without compensation is just and equitable in each
instance. It is therefore necessary for legislation to establish an administrative mechanism to
determine whether land falls into a category subject to expropriation without compensation.
That could be determined by executive officials, or by an independent administrative body.
The administrator would determine whether the land falls within the categories specified
in legislation for expropriation without compensation, and whether there are substantial and
compelling circumstances to pay some compensation. The procedure before the
administrator should comply with the ordinary standards for administrative decision-making.
In order to survive constitutional attach, that decision must be subject to judicial oversight.
However, the duty to approach the court can be placed on the owner. Moreover, the bringing
of the review need not delay the expropriation. The Court can be limited to determining
whether the administrator acted consistently with the Promotion of Administrative Justice
Act. That would include the power to intervene if it believed substantial and compelling
circumstances justifying some compensation existed.
We have focused on the possibility of expropriation without any compensation. While
that is possible under the Constitution, the instances in which it is permissible are limited. On
its own, we do not believe this will likely to fulfil the state’s obligations with regard to land
reform.
The more effective mechanism is to use legislation to reverse the current approach that
requires just and equitable compensation to be determined by agreement or by courts, and
permits little (if any) deviation from market value. An efficient administrative system for
determining compensation at significantly below market value, and created a limited judicial
role, has far more potential to realise the existing constitutional possibilities for land reform.