Administrative Law in South Africa

Cora Hoexter

Chapter 1 – Introduction to Administrative Law

1.1 THE NATURE AND SCOPE OF ADMINISTRATIVE LAW (BACKGROUND)

Administrative law regulates the legal relations of public authorities or bodies exercising public powers
or public functions. It has come about as a result of separation of powers where the judiciary regulates
the exercise of public power by the other branches of government [Pharmaceutical Manufacturers
Association]. The question is thus what constitutes a public function.

(a) Public powers and functions

In South African, as in English law, the focus is more on the nature of the power being exercised than on
the entity which does the exercising. The two judgments which deal with this issue are AAA Investments
and Calibre Clinical Consultants. In England, USA and Canada the idea of public powers have been
associated with governmental activities. However in SA the ‘governmental’ inquiry has not always been
regarded as necessary [AAA]. In CCC Nugent JA insisted that the governmental inquiry was relevant.
State involvement, or lack of autonomy, may imply a public power. Public funding could also be a
significant factor. The governmental test must not be applied too literally or narrowly as the idea of
‘public’ as a wider meaning. There are many other characteristics which may contribute to something
being public such as it being in public interest which may go beyond governmental interest. The source
of the power is also relevant and how that power is exercised with regards to its effect on the public. It
is a broad and flexible approach, not formalistic and mechanical. There is no one single test of universal
application [CCC].

(b) The focus of administrative law

1
There is an overlap with constitutional law but it differs in that administrative law has its emphasis on
one particular branch of the state system, the public administration. The SCA has described it as ‘the
conduct of the bureaucracy….’ But the term bureaucracy is rather lose and often pejorative. The more
technical term would be ‘public administration’ which is generally understood to mean the organs and
functionaries of the executive branch of the state that are concerned with the day-to-day business of
implementing law and administering policies. All government departments are covered as well as quasi
non-governmental organizations e.g. Telkom, Eskom, SABC etc.

The term ‘administrative action’ has acquired a specific and technical meaning as the constitutional
rights to administrative justice are attached. It is not necessarily something performed by members of
the public administration as it is the function not the functionary which is relevant.

Administrative law has an intimate relationship with the interpretation of legislation as legislation is
often the most prolific source of administrative power. Many cases thus begin with an analysis of the
empowering provisions.

There is a close relationship with political science and the study of public administration. The main
distinction being that administrative law focuses on the legal aspects of these studies. However there is
an overlap as both are concerned with controlling and improving public administration.

(c) Aspects of administrative law

There are general and particular categories of administrative law. Particular administrative law deals
with rules in specialized areas such as social welfare. General administrative law deals with rules and
principles common to most or all kinds of administrative action. It could be described as the ‘regulation
of regulation’.

Administrative law is broader than judicial review even though the two are often equated.
Administrative law is concerned with both judicial and non-judicial safeguards against poor decision
making. Judicial focuses on the diagnosis of problems while administrative law has a more positive side
and includes empowerment of administrators, facilitation of administration and encouraging good
decision making. Jy

2
 (d) The domain of administrative law

It covers a vast area because it can be relevant to any action involving public powers or functions. It
permeates virtually every facet of the legal system. There has been a growth of the administrative state
and so the public administration has become overburdened and unwieldy.

1.3 THE RELATIONSHIP OF CONSTITUTIONAL AND ADMINISTRATIVE LAW

(BACKGROUND)

The constitution inevitably has a fundamental bearing on the nature of the administrative law system.
The shift to constitutional democracy has had a profound effect on South Africa’s administrative law.

Constitutional law is concerned with the establishment and structuring of a system of government and
the separation of powers. It provides a framework for legislation and policy. Administrative law on the
other hand is concerned with the daily business of government: implementing and administering of
enacted policy and the exercise of delegated powers.

Constitutional law is “the state at rest” while administrative law is “the state in motion”. Constitutional
law is primarily concerned with structure while administrative law is primarily concerned with function
but they both deal with both aspects.

(a) Administration and the separation of powers

The functions performed by the administration straddle all three organs of the state and so the
distinctions have become increasingly blurred. In the case of SARFU the court confirmed that the
administration is that part of the government concerned with the implementation of legislation. The
implementation of legislation is a broad concept which covers virtually any method of making legislation
or giving it practical force. Administrators may actually make a greater number of laws than the
legislature itself. Although this may seem to infringe on the doctrine of separation of powers, this issue
is inevitable as the Constitutional Court itself has conceded. In the New Clicks Case Chaskalson CJ said
that ‘the making of delegated legislation by the executive is an essential part of public administration’.

The legislature is overburdened with the duty to make legislation and as such only produces the bare
minimum required. Further the legislature is not familiar with the day-to-day administration and may
not be able to foresee all the practical difficulties which may arise. Thus the administration is often

3
empowered to make the detail rules needed for effective regulation. The administration may be given
powers to make delegated legislation and also to flesh out statutes by making additional policy. It has
become difficult to make a distinction between administration and policy-making. In many cases the
administrators cannot give effect to the relevant legislation without making policy.

Just as law making is not the sole function of the legislature, so dispute resolution is not the sole
function of the judiciary. Administrative tribunals may play a role in adjudicating disputes in various
fields.

(b) The overlap of constitutional and administrative law

In the early years of our democracy it was understood that a matter seeking enforcement of section 33
was as much a constitutional as administrative law. It was also understood that the work traditionally
done by the old administration was now being done by the Constitution.

While a previous SCA decision held that judicial review under the Constitution was entirely different to
that under the common law, the CC in PMA held that there is one system of law which derives its force
from the Constitution and is subject to constitutional control. The court further took the view that
judicial review has always been a constitutional matter. There is only a narrow band of non-
constitutional administrative law matters.

(c) The role of the common law

The Constitution has not done away with common law administrative law rather section 33 and other
relevant provisions must be seen as incorporating and expanding on the established principles of the
common law in line with section 39(2).

In PMA the CC held that the common law will continue to inform the content of administrative law. This
is obviously not applicable where it is incompatible with the Constitution. In Bato Star Fishing O’Regan
held that the relevance of the common law must be worked out on a case-by-case basis. In practice it
has proved to be very useful to the courts.

1.4 SOURCES OF ADMINISTRATIVE POWER

There is a basic requirement that all power is legally authorized. There are various possible sources of
that power. Administrative power must be conferred by law. The sources of administrative powers also

4
define the boundaries of what administrators may do. PAJA recognizes some sources of administrative
law in its definition of an ‘empowering provision’ but it does not recognize them all.

(a) The Constitution

It is the ultimate source of administrative power or authority. The Constitution is the most important
source of power even though legislation is the most prolific. All organs of state are bound by the
Constitution and it takes precedence over any other law. Any law inconsistent with the Constitution is
invalid regardless whether that inconsistency is procedural or substantive.

Courts control public power under the Constitution which defines their role, powers and the constraints.
One of the most fundamental constraints is that the source of public power must be consistent with the
Constitution particularly in the case of legislation.

Authority for administrative action may also come directly from the Constitution [Kyalami Ridge
Environmental Association].

(b) Original and delegated legislation

Legislation is the ideal modern source of administrative power as it is capable of being enacted,
amended and repealed expeditiously and is readily accessible and knowable. Almost all administrative
power has a legislative source.

Government delegates authority to the administration by means of original or delegated legislation.

Original legislation is enacted by Parliament or a legislature in accordance with powers conferred by the
Constitution. Typically it authorizes an administrator to make rules, award licenses, charge fees ior
appoint or dismiss staff.

Delegated legislation is made under the authority of original legislation. An example would be a
proclamation by the President. The regulations should be consistent with the original legislation. There
are also Constitutional constraints on the ability to delegate.

Administrators also rely on standards or quasi-legislation which are less official. For example a circular or
manual which governs how an administrator acts but is not necessarily published as rules. Standards can
be extremely helpful [Sasol Oil] but can also be viewed with suspicion when they are not published or
readily accessible. They are further not always subject to section 101(3) of the Constitution which

5
requires them to be accessible to the public. Administrative lawyers believe they should be available to
those affected by them.

Standards are only really regarded as a source of administrative authority where a statute appears to
anticipate their creation and where they do not interfere with statutory or common law rights. The
definition of empowering provision in PAJA may be read to include standards such as policy statements,
directives and guidelines. Their validity however must still be judged in relation to their effect in
particular cases and the terms of the relevant enabling legislation.

(c) Prerogative powers

This was a non-statutory source recognized before 1994. It was inherited form English law. In SA the
most important prerogatives were the power to issue passports, appoint a commission of inquiry and
conclude state contracts. They became increasingly reviewable towards the end of the pre-democratic
era but they did not survive the enactment of the interim and final Constitution.

Section 84(1) confers on the president ‘the powers…necessary to perform the functions of Head of State
and head of the national executive”. Section 84(2) lists activities for which the president is responsible.
Many of these resemble prerogative powers. The nature of the executive powers were considered in the
case of Hugo and held that they flowed from the Constitution and so were not dependent on legislative
enactment. There are thus no powers derived from the royal prerogative other than those in the
Constitution.

The president is acting as an executive organ of the state regardless of whether he is acting as head of
state or as head of the executive. The Bill of Rights binds all organs of states and so the powers of the
president are subject to review in the same way as all other Constitutional powers.

However, not all the powers are subject to review as ‘administrative action’. In the SARFU case the
decision to appoint a commission of inquiry was conferred on the president in his position as head of
state and was related to policy rather than the implementation and thus did not bear the hallmark of
administrative action. Powers of a strong political nature are also likely not to be subject to review as
administrative. In any event, almost all of the powers are excluded from the definition of administrative
action in section 1 of PAJA. While the powers are still reviewable other than under section 33, the courts
view them with respect because of their highly political character.

6
 (d) The common law

Often administrators perform activities for which there is no specific authority. This void may be filled
with general constitutional duties as well as by ‘non-prerogative’ common law powers of the state
[Kylami]. In this case the court held that the government had all the rights of a private land owner but
within the constraints of the Constitution and applicable legislation.

Where the state’s ability to contract is absence from any legislation it can be seen as an incident of its
common law powers. Agreements entered into by administrators are a further source of administrative
powers as recognized by PAJA. Again the common law powers must be consistent with applicable
legislation and the Constitution. In the pre-1994 case of Dilokong Chrome Mines the court held that
‘purely beneficial dispositions’ are a source of administrative power which seems to derive authority
from the state’s common law powers. The common law seems the most likely contender as the other
options would be estoppels or no authority at all.

(e) African customary law

Pre-1994 there was a question as to whether or not custom and African customary law were legitimate
sources of administrative power while those exercising customary powers were clearly subject to
administrative law.

Customary law derives from practice considered obligatory by the community and the chiefs and
headmen exercise considerable power over the people resident in their jurisdiction. This power derives
from African customary which is acknowledged as an empowering provision in PAJA.

The Constitution also acknowledges customary law in a number of places and like any other law,
customary law is subject to the constraints of the Constitution particularly in relation to section 39(3)
and section 211(1).

Like the common law it may be become entwined with legislation to form part of the amalgam of SA
law. There is no doubt it is a source of administrative power. However, as it is an unwritten source an
administrator relying on customary law may have difficulty proving the existence and extent of power.

7
 (f) Estoppel

Important principles of administrative law as well as the balancing of public and private law are at stake
when determining whether or not estoppel can be a source of administrative power. If A makes a
representation of fact to B who relies on that representation as the truth to B’s detriment, A can be
estopped from denying the truth of that representation.

Administrators often make representations on which people rely to their detriment. Cases have shown
that the person relying on the representation may suffer harm if the administrator is not held to the
representation.

The question is whether or not an administrator can acquire powers by estoppel which are not found in
legislation or the Constitution where the decision was ultra vires. This would have the effect of ratifying
a decision which the administrator was not lawfully allowed to make and thus it conflicts with the
principle that all decisions and actions need lawful authority. Public authorities should not use powers
which they do not have. Waiving legal requirements would overlook the principle of public policy.

In the case of Contractprops 25 the court refused to apply the doctrine of estoppel saying that contracts
concluded ultra vires cannot be allowed to stand as if they were intra vires.

Estoppel may also violate the principle that public bodies should not fetter their free discretion in future
by making promises in advance as to how they will exercise that discretion.

Courts have, however, applied the doctrine of estoppel to public authorities but only when the decision
was not made ultra vires or where it does not result in the waiver of statutory duties. Twenty years ago,
estoppel might operate where a policy rather than legal provision was at issue. Further it may be used
where it would only result in the waiver of an internal requirement rather than legal formality or where
a legal duty has been violated was directory not mandatory.

This cautious view was challenged in Peter Klein Investments where Boruchowitz J held that there would
be no immunity from estoppel where it would be of minimal benefit to the plaintiff and cause great
hardship and injustice to the defendant. The court drew on an English decision which suggested that
individual and public interests should be balanced.

However in the case of RPM Bricks the SCA described this as fallacious and held that estoppel cannot be
used to give effect to that which is not permitted by law and where a transaction is unlawful the

8
consequence cannot vary from case to case. Ponnan JA drew a distinction between acts in excess of
legal powers and irregular or informal exercise of an existing power. Estoppel could be relied on in the
second category of acts where all internal formalities had been complied with but never in the first
category.

An enrichment action or delictual damages are still available to the plaintiff. The plaintiff could also
argue that the representation constitutes a decision and so the administrator was discharging an official
function and could not revoke, withdraw or revisit the decision.

1.5 TYPES OF ADMINISTRATIVE POWER (BACKGROUND)

The interpretation of legislation is inextricable linked with administrative power thus the basic
categories of administrative power are relevant. These categories are not explicit in the Constitution or
PAJA but are frequently referred to in case law. There is a basic distinction between powers and duties
and powers which are express or implied, discretionary or mechanical, and mandatory or directory.

(a) Powers and duties

Powers enable things to be done while duties require that they be done. A measure of discretion is
implied in a power but not in a duty. The language used in a legislative provision can express whether or
not it confers a power or a duty.

The distinction is not always complete though. Powers may often be coupled with a duty in a legislative
provision. There may be a power to do something with a duty to exercise that power in a certain way.
Public powers always accompany a duty of some kind. At common law this duty is to act in the public
interest. Chapter 2 of the Constitution places numerous duties on holders of public powers and ‘organs
of state’ as well as non organ of states by way of the principle of legality and the rule of law.

(b) Express and implied powers

Generally express powers are needed for the actions and decisions of administrators while implied
powers may be ancillary and exist as a necessary or reasonable consequence of express powers. What is
reasonably incidental must be considered as impliedly authorized. For instance it is reasonable
necessary to suspend municipal councilors to investigate corruption.

9
There is a proposal to create a new Interpretation of Legislation Act which includes the fact that a power
conferred by legislation to perform an act must be read as impliedly conferring all powers ‘necessary for’
or ‘incidental to’ the performance of the act.

The factors to be considered include:

(i) The language of the legislation

Language is crucial to discover intention. Peremptory language may encourage a court to read implied
provisions in.

(ii) The breadth of express power

Where an express provision is broad and discretionary, an implied power is more likely to be found and
vice versa.

(iii) The context of the provision

The Act as a whole and its purposes may indicate whether further powers are incidental. for instances
where the main power cannot be achieved without it or it is necessary to the proper functioning of the
administrative body. A term cannot be implied where it contradicts express terms.

(iv) The nature of the action

Where action is coercive, oppressive or likely to have far-reaching effects, implied authorization is
unlikely to be found. The power to regulate does not necessarily include the power to prohibit.

(c) Discretionary and mechanical powers

Discretionary powers are recognized by permissive words such as ‘may’ or ‘shall’. There is an element of
choice for the holder. Traditionally there is suspicion of discretion and lawyers prefer neutrality of legal
rules. Modern legal systems acknowledge the need for discretion to complement rules as it allows for
flexibility and the fixing of gaps in rules. It allows for general rules to be applied to a specific
circumstance in a fair manner.

Mechanical powers involve little or no choice for the holder and purely mechanical powers fall more into
the nature of duties.

10
Discretion may be wide but it is never unfettered as one must act wisely and after due reflection. The
Constitution requires constraints on broad discretionary powers in order to minimize the danger of
violation of rights and so those who are affected by broad discretionary powers will know when they can
seek relief.

(d) Mandatory and directory provisions

The question on whether or not non-compliance with a statutory provision leads to invalidity is
answered by looking at whether or not it is a mandatory or directory provision. Failure to comply with a
mandatory provision will inevitably result in invalidity while failure to comply with a directory provision
will not.

The court will use the following factors in interpreting formalities. None of them are decisive or
necessarily entirely reliable.

(i) The use of peremptory language

Words such as ‘shall’ or ‘must’ create mandatory provisions while permissive language such as ‘may’
does not. However this is a general guide as the circumstances may negate that construction.

(ii) The use of the negative form

Negative language is more likely to be treated as peremptory while positive would be directory
particularly where it is vague and there is no sanction.

(iii) The purpose of the provision

Where legislation allows for the acquisition of a right rather than the infringement of an existing one it is
more likely to be imperative.

(iv) The context, history and importance of the provision

The provision must be interpreted in light of the act as a whole. If a provision is mere technicality then
strict compliance is not needed and it must be asked whether the legislature would have intended
invalidity to result or if strict application would lead to fraud or injustice.

(v) The nature of the action

11
Impact and importance may indicate intention. Urgency may lead to a provision being interpreted as
merely directory.

Chapter 3 – Judicial Review

3.1APPEAL AND REVIEW (BACKGROUND)

(a) The distinction between appeal and review

They are both ways of reconsidering a decision. Appeal relates to the merits of the case while review is
concerned with how a decision was arrived at. A decision can be reviewed even where the court may
have come to the same conclusion. This distinction is fundamental in the theory of South African law. In
practice however the focus of an administrative review often falls on the decision rather than the
process. It may be impossible to separate the merits from the rest of the matter. The courts cannot look
at reasonableness and rationality without looking at the merits of a case. Only in a case of the most
technical nature can the courts avoid the scrutiny of the merits.

However the courts continue uphold the decision because of the doctrine of separation of powers as
commenting on the merits of administrative action would be usurping the powers of the executive. It is
only really when scrutiny of the merits is translated into intervention that separation of powers has
been affected.

(b) Types of appeal in South African law

There are two main types of appeal: appeals from lower courts to higher courts and administrative
appeals. Generally they lie from one administrative agency to another but may lie to specialized courts
or even ordinary courts of law.

(c) Types of review in South African law

A century ago only three forms of review existed: reviews of inferior courts, common law review of
administrative authority and a wider form of statutory review. This list has been expanded as a result of

12
modern developments including the Constitution. There is now automatic review which occurs from
lower courts even without an application. There is also Constitutional review. Common law
administrative review has largely been constitutionalised by section 33 and placed on a statutory footing
by PAJA. The forms of review are now as follows:

(1) Review of the proceedings of inferior courts

High Courts may be asked to review decisions of inferior courts on grounds set out in section 24 of the
Supreme Court Act which are absence of jurisdiction, bias or corruption, gross irregularity and admission
of inadmissible evidence.

(2) Automatic review

Certain statutes make provision for reviews to be automatically statutorily triggered rather than by an
aggrieved individual.

(3) Judicial review in the constitutional sense

Usually means the power of courts to declare any legislation or state conduct unconstitutional. Review
in administrative law is now largely constitutional review.

(4) Judicial review in the administrative law sense

Refers to the power of courts to scrutinize and set aside administrative decisions or rules on certain
grounds. This used to be a common law power but is now governed by section 33 and PAJA. It is now a
statutory matter. Where PAJA is not applicable, it will be governed by the Constitution (public powers)
or common law (private powers).

(5) Special statutory review

It is sometimes wider than an ordinary review and so is more akin to an appeal but may also be
narrower with courts confined to particular grounds of review or remedies. The extent of the power will
depend on the statutory provision concerned.

13
The first two types of review are not of much relevance to administrative law although there may be
some overlap. Generally the term judicial review in administrative law means type (4) and it is now
largely constitutionalised and is used in the alternative with type (5) review under PAJA.

3.2ADMINISTRATIVE LAW REVIEW: PAST AND PRESENT

(a) Judicial review before and after 1994

In the pre-democratic era review took place largely in terms of the common law. The justification for the
courts interfering in administrative decision was the ultra vires doctrine which is an important English
Law Doctrine. In essence, the legislature confers powers on administrators and sets statutory
boundaries of those powers. Courts must ensure that the intention of the legislature is carried out and
that administrators act intra vires, within the boundaries of the powers. Courts may set aside powers
exercised ultra vires. However this doctrine does not explain the enforcement of non-statutory
requirements such as the common law. Common law review is now only one of five pathways available
to review. It has not been easy for courts to sort out the proliferation and hierarchy of pathways.

The courts first concern was to determine the relationship between the common law and the
Constitution more generally. In the early years the effects of the Constitution were not fully appreciated
and so there was a willingness to draw a sharp distinction between ‘constitutional’ and ‘non-
constitutional’ matters. Courts insisted on not reaching a constitutional matter unless it was relevant for
adjudication of the case. Courts thus initially maintained that administrative law review was completely
different under the Constitution and under the common law.

This view was rejected in Pharmaceutical Manufacturers Association (PMA) which placed the
Constitution at the center of administrative law. The court held that there is only one system of law
governed by the Constitution. The court noted that what would have been ultra vires at common law in
now invalid according to the constitutional doctrine of legality. Judicial review of public power is
inevitably a constitutional matter. Common law review is essentially confined to the realm of private
power.

14
Today the main reason for courts to interfere in administrative matters is the right to administrative
justice found in section 33. The legislative intention of the ultra vires doctrine is no longer necessary.
They can now rely on the binding effect of the Bill of Rights and on the constitutional duty to uphold
fundamental rights. The vires in question are now those of the Constitution and not merely the
legislature. The doctrine now explains how courts control administrative action which falls outside the
operation of law and the Constitution.

(b) Section 33, PAJA and special statutory review

The situation has been further complicated by the coming into force of PAJA which is mandated to give
effect to section 33. PAJA provides the most immediate justification for judicial review. PAJA does not
amend or replace section 33 but rather gives effect to it. The Act contains a list of grounds of review
familiar at common law.

Applicants must bring cases for review under PAJA where possible. In the case of Bato Star v Minister of
Environmental Affairs O’Regan confirmed that “the cause of action for judicial review of administrative
action now ordinarily arises from the PAJ, not from the common law as in the past”. The common law
informs the provisions of PAJA and the Constitution, and derives its force from the latter. The extent to
which the common law remains relevant to judicial review will need to be decided on a case-by-case
basis.

The displacement of the common law may be regarded as an aspect of the constitutional principle of
subsidiarity. This principle limits the availability of direct constitutional review under section 33. Where
legislation has been enacted to give effect to a right, an applicant should rely on the legislation or
challenge it as being inconsistent with the Constitution. Direct review is only available where the
constitutionality of PAJA is in question or where legislation is challenged for infringing section 33 of the
Constitution.

Another aspect related to subsidiarity is that general norms should be resorted to only when specific
norms have run out. Specific provisions which focus on constitutional principles must be applied before
the general norm is invoked. Further there is the principle of avoidance which requires recourse to the
legislation or common law before the Constitution.

15
Before applying PAJA and section 33 it must be determined whether it is in fact a review of
administrative action. The definition given by the Constitutional Court is not the same as that which is
found in PAJA which creates a discrepancy. The elaborate definition in PAJA excludes some actions
which were reviewable before 1994.

These actions may be still be reviewable through special statutory review where the relevant enabling
legislation makes provision for it. It operates to the exclusion of default review under PAJA owing to it
not being administrative action. Where an action is reviewable through PAJA or another statutory
provision, presumably the applicant will have a choice of how to proceed.

In the case of Sidumo the court held that the Labour Relations Act was suffused with the standard of
reasonableness found in section 33 of the Constitution and since the action described in the Act would
qualify as administrative action under section 33, the LRA was specialized legislation alongside PAJA.

In that situation the grounds listed in the specific statute (the LRA) are rendered almost meaningless
unless they are more extensive than the grounds found in section 33. The grounds would need to be
suffused with the standards of section 33 and so full scale administrative review is applied irrespective
of the limits of the grounds themselves.

The question arises of whether or not the suffusion is limited to the labour context. It may be difficult for
the court to justify withholding the benefits of section 33 but if the court does not restrict the suffusion
then actions specifically excluded under PAJA may qualify under section 33 and an inconsistent situation
could arise. An example would be acts which can specifically be reviewed under PAIA but not PAJA.

(c) The principle of legality

Non-administrative action is catered for in the Constitution and by the principle of legality which is an
aspect of the rule of law. Administrative conduct is subject first to PAJA then section 33 and only in the
last instance to the rule of law.

Pre-1994 the principle of legality and the requirements were inferred from what it had been held
administrators may not do. The requirements were expressed negatively in the common law grounds of
review. Ultra vires were the negative side and legality the positive.

In the constitutional era legality may still refer to the same requirements which are the grounds of
review applicable to administrative action. Now the concept of legality is more explicit than at common

16
law. Section 33 requires administrators to act lawfully, reasonable, procedurally fairly and to give
reasons in certain circumstances. The detailed content of legality has been written down in the grounds
of review in section 6 of PAJA.

Legality goes beyond administrative action and refers to the constitutional principle that governs all
public power. It is an aspect of the rule of law which is the fundamental idea that all public power is only
legitimate where it is lawful. Working out the detailed content is a continuing process embarked on by
the Constitutional Court in a series of non-administrative action cases.

In the case of Fedsure the court identified legality as an aspect of the rule of law. A body exercising
public power had to act within the powers lawfully conferred on it. In SARFU the principle requires the
holder of public power to act in good faith and no misconstrue his powers. In PMA it was determined
that the exercise of public power should not be arbitrary or irrational. In Allbut it was held that
procedural fairness was a requirement of rationality and in Competition Commission of SA v Telkom
procedural fairness was brought in through the requirement of lawfulness. Procedural fairness can also
be acknowledges as a requirement on its own right.

The principle of legality is a safety net that gives the courts some control over action which not
administrative but still involves the use of public power. The principle allows the government to take
control while the court maintains a supervisory role. The review in terms of legality is more limited than
that in terms of PAJA or section 33 however it is already far reaching and there is considerable overlap.
Rationality which is a component of reasonableness is demanded by legality and this in turn may require
procedural fairness. There is judicial support for the proposition that the principle of legality demands
the giving of reasons. It is an evolving concept in our jurisprudence.

A reason for the development of the principle is the narrowness of the definition set by PAJA. It is more
simple, general and inclusive than PAJA. Our courts seem willing to rely on the principle and concept of
the rule of law.

(d) Judicial review and private entities

In accordance with PMA all four pathways thus far are concerned with public power and are thus
governed by the Constitution. However it is also possible under the Constitution for judicial review of a
private body. What matters is the nature of the function not the functionary.

17
 (i) Public powers and functions

Section 8(1) makes the Constitution binding on all organs of state. These are defined to include
departments of state or administration in different spheres of government (except courts) and any other
functionary exercising a public power or function in terms of legislation. This has huge implications for
private bodies exercising public functions.

A powerful example is the case AAA Investments where the Micro Finance Regulatory Council was
found to be exercising a public function in terms of a constitution or legislation in regulating the micro-
lending sector.

The Constitution states that the exercise of public power is always subject to constitutional control and
the rule of law and specifically the principle of legality. Even if the power is not being exercised in terms
of a constitution or legislation, the power will still be subject to the principle of legality.

PAJA’s narrow definition of administrative action is liberal when it comes to private actors. It is not
limited to conduct in terms of legislation but extends to include a law, common law, customary law,
agreement, instrument or other document.

(ii) Private powers and functions

Before 1994 the courts regularly reviewed the decisions of private bodies in exercise of powers derived
from the contract between them and members. Mostly this was in disciplinary matters.

Today these matters are reviewable and an explanation may lie in section 8(2) of the Constitution which
contemplates the direct horizontal application of the Bill of Rights in the private sphere. It has been
suggested that the control of private power could be subject to section 33 but the rights in section 33
are naturally associated with the activities of the state. The rights are confined to administrative power
which is public more than private power. Currie and De Waal suggest that the nature of the right to just
administrative action is precluded from applying directly to private conduct.

However it may apply indirectly by virtue of section 39(2) which arguably provides a constitutional
justification for the review of private power.

(iii) Justifying the review of private power

18
Review in terms of private administrative law would seem to require more than not being objectionable
to the Bill of Rights. The pre-1994 ultra vires doctrine focused on legislative intent and had difficulty in
explaining non-statutory review. English judges have called it a “fairy tale” and legislative intention does
not come into it at all. The courts simply impose their own ideas about good governance on the
administration in the form of lawfulness, reasonableness and fairness. However the doctrine is still
upheld by the courts.

There is a modified ultra vires doctrine or ‘legislative intent model’ which finds that the legislative
intention by reasoning that Parliament intends administrators to conform to lawfulness, reasonableness
and fairness and that it intends the courts to apply these principles.

There is also the common law model which accepts that principles applied in court are a product of
judicial creation. Judges set standards for administrators through the common law. If the sovereign
British Parliament does not like these standards, it can make this clear in legislation. This avoids the
problem that the ultra vires doctrine has of not being able to explain the review of non-statutory power.

South Africa avoids the justification almost entirely except where it comes to reviews of private power.
Section 39(2) provides for something similar to the common law model. It makes no reference to the
‘nature’ of rights and duties and can be seen as authority to set public law standards for private
administrators especially in cases of coercion or inequality. The power being yielded then mimics the
power controlled by the rights to administrative justice. Thus judicial review is fully in line with the spirit
of the Constitution.

(e) A proliferation of pathways – and avoidance of the PAJA

The courts have had a problem with the different pathways and have had a tendency to treat them as
free alternatives. In the New Clicks case the majority saw no need to indicate which form of review was
being relied on in its review of non-administrative action on various grounds. The court seemed to
assume that the choice of pathway made no difference.

This approach did not meet with the approval of the Constitutional Court on further appeal. Chaskalson
CJ confirmed that PAJA is a codification of section 33 and a litigant cannot avoid the provisions of PAJA
by relying on section 33, the Constitution or common law. This would negate the purpose of effecting
legislation to codify section 33. It could create a parallel system of law which would be inconsistent with
the purpose of section 33 which is ‘to establish a coherent and overarching system for the review of all

19
administrative action’. Ngcobo J in a different case too highlighted that there is only one system of law
grounded in the Constitution. The provisions of PAJA must be read in line with section 33.

The implications of the above for the principle of legality are directed particularly at parallel systems of
administrative law growing out of section 33 or the common law; and the message of these warnings is
that the PAJA must not be bypassed but must be applied where applicable. The principle of legality
however poses a greater and more immediate challenge to the PAJA. It is appealing because it is general
and flexible. It may also be appealing to use in order to bypass certain procedural requirements in PAJA.

Avoidance of PAJA seems not to have been addressed in case law but the principle is the same as that in
New Clicks: use PAJA where it is applicable and do not use the principle of legality in order to bypass it.
Nevertheless a growing number of cases are not relying on PAJA. The Allbutt case is an example of
deliberate avoidance of the PAJA. The court not only bypassed PAJA but asserted that this was a better
way of proceeding. The fact that difficult questions will arise does not seem a sufficient reason for
avoiding PAJA. If the approach taken by the court was taken by more courts then PAJA would disappear.
Every case would simply be decided on its own unique features and what the principle of legality had to
say to them.

The principle of legality remains an essential safeguard for action which does not qualify as
administrative action by doing at least some of the work that administrative action would do. To the
extent that there is action which is not administrative, there will always be the principle of legality. It is
particularly necessary at the moment because of the narrow and complicated definition so until such
time as PAJA is brought in line with section 33 by means of judicial interpretation, the principle of
legality will no doubt continue to flourish,

Chapter 4 – Administrative Action

4.2 ADMINISTRATIVE ACTION UNDER THE CONSTITUTION (BACKGROUND)

The constitutional meaning of administrative action is strongly informed by separation of powers. In the
first few years the Constitutional Court had several opportunities to fence off action that is associated
with the political process and to distinguish administrative action from legislative, executive and judicial
action. Other categories of administrative action will be discussed below.

(a) Administrative action in general

20
In SARFU the Constitutional Court admitted that it was difficult to decide what administrative action is
and it would have to be done on a case by case basis. The following are relevant considerations: the
source of power, nature of power, its subject matter, whether it is a public duty, and how closely it is
related to policy matters (not administrative) or the implementation of legislation. The court pointed out
that membership in the executive does not necessarily identify administrative action. What matters is
the function not the functionary and whether the task itself is administrative action. It is thus possible
for judicial officers or legislators to exercise administrative functions.

The diagnoses may either be difficult or relatively uncomplicated.

(b) Executive action and ‘policy’ decisions

In SARFU the Constitutional Court decided that the President’s decision to appoint a commission of
inquiry into the administration of rugby was a power conferred on the President as head of the state
rather than head of the executive and was akin to a prerogative power. It was thus executive rather than
administrative action. The court observed that the tasks of formulating policy and initiating legislation
are constitutional responsibilities of the Executive branch and cannot be construed as administrative
action. Implementing legislation is however typically administrative.

In Ed-U-College O’Regan explained the difference between policy formulation in the broad political
sense and the narrow administrative sense. Policy formulated by the Executive outside of a legislative
framework generally involves a political decision and will not constitute administrative action. Policy
may also be formulated by a member of the Executive implementing legislation which is narrower and
generally administrative. She pointed out that it is possible for action to be administrative even with
political implications.

It has been noted that few administrative acts are devoid of underlying policy and administrative action
is often the implementation of policy that has to be given legal effect [Grey’s Marine]. Judicial scrutiny
cannot merely be avoided by attaching the epithet of “political”.

(c) Legislative action and ‘political judgment’

In PMA ‘political judgment’ featured to bring the action closer to legislative than administrative action.
In Fedsure the court held that legislative bodies exercising original, deliberative lawmaking powers are

21
not engaged in administrative action. It is possible for law made by administrative functionaries but not
legislative bodies to classify as administrative action.

The court has looked at whether the action is achieved by means of a specific statute or from some
other source such as a discretionary power [Ed-U-College].

The implementation of legislation is an important factor in determining whether or not something is

administrative action. In PMA however even though it clearly dealt with legislative implementation, the
court held that it was not administrative action as it was done prematurely and so the president acted
beyond the scope of his powers. It required a political judgment as to when to bring the statute into
force and so the action lay closer to the legislative process.

The court seemed to rely on the distinction between the mechanical implementation of legislation and
the antecedent decision to bring the statute into force. However the trouble is that the proclamation is
the only evidence of the antecedent decision and so they should really be treated as a single transaction
when they are so closely related. In fact only the most mechanical of tasks will not require decisions
about timing and content.

It is significant that no court has been able to decide whether delegated legislation amounts
administrative action under the Constitution or PAJA. The question was left open in Eisenberg. A few
years later in New Clicks Chaskalson CJ wrote in favour of it being regarded as administrative action but
the court was unable to agree.

(d) Judicial Action

In Nel v Le Roux the Constitutional Court found that the procedure for the summary sentencing of
someone refusing to give evidence was judicial rather than administrative. It further doubted whether
an investigation to gain factual information could be regarded as administrative.

There are more examples of cases where the court found action to be judicial in nature rather than
administrative (see pages 182-3 of Hoexter).

(e) Procurement and licensing

Ordinarily this action is administrative under the Constitution. In Umfolozi the SCA held that the award
of a state tender was administrative action. It was reasoned that the steps preceding the contract were

22
administrative decisions and that public money was being spent by a public body in the public interest.
This reasoning is also extended to adjudicating tenders.

In Logbro the court rejected an argument that tender conditions gave rise to contractual rights outside
the realm of administrative law.

Mining rights are analogous to a tender and thus administrative. Refusal of application and renewal of
firearm licenses are also administrative action.

(f) Public and private power

Administrative action under the Constitution is restricted to powers and functions of a public nature.
There are two issues which need to be interrogated by the courts.

(i) Powers exercised by public bodies in a private law setting

In Cape Metropolitan Council the court held that cancellation of a contract was not administrative
action as it involved the use of common law contractual power rather than public power as it took place
on grounds of fraud rather than legislation. There was also equality of bargaining powers and there was
no stronger position by virtue of a public power.

This conclusion is not easily defended. The court purported to apply the factors from SARFU but those
factors pointed to public power and thus administrative action. There was legislation which made
provision for termination of an outsourcing agreement on grounds including fraud.

The reasoning in Cape Metro was characterized somewhat differently in Logbro. Cameron JA rejected
the argument that a tender contract could be exercised without regard for administrative justice. He
pointed out that Cape Metro distinguished between tender and employment. The question of whether
a public body is required to act fairly when exercising a contract was left open. The administrator in
Logbro was undoubtedly acting from a position of superiority evidenced by it dictating the tender
conditions and so the decision to call for new tenders was administrative action even if sourced in
contract. More recently the SCA took the view that once a tender had been awarded it was a contractual
not administrative relationship.

In MV Snow Crystal it was held that it is difficult to draw the line between public duties and contractual
obligations.

23
The SCA has displayed more consistency in cases dealing with property than contract. In Bullock the
court emphasized an element of public interest in the granting of a servitude over state owned land and
that it must be administrative action. In Grey’s Marine, a case decided under PAJA, the court did not
differ from that reasoning.

The Constitutional Court has generally not permitted the contractual features of a case to blind it to the
relevance of public power. In Masetha the court cautioned that the president’s powers are subject to
constitutional constraints even if there is a contractual element. In Joseph the court held that Eskom’s
relationship with tenants was governed by more than merely contract but also administrative law.

(ii) Private bodies exercising public powers

A purely institutional test would suggest private bodies are incapable of administrative action but this is
not so. Private bodies can exist which exercise public powers typical of governments. These are
becoming popular where activities are contracted out. PAJA explicitly includes such actors and does so
without insisting their actions be taken in terms of legislation. It does not define what gives a power its
‘public’ character. The courts are answering this incrementally.

Some private bodies gave no direct statutory powers but act coercively and are integrated into the
system of statutory regulation. Their decisions have sometimes been reviewable at common law. In
Datafin the English court looked beyond the source of the power and acknowledged the nature. The SA
version Dawnlaan Beleggings reviewed the decision of the JSE, a non-statutory body under a statutory
duty to act in public interest.

In Cronje the court found that the UCB was a voluntary association wholly unconnected from the State.
Its origin was in contract, powers were contractual, it was privately funded and had no contractual
official responsibility for cricket. A private body under a statutory duty to act in the public interest is
likely to be found to be capable of exercising public powers. This was not the case of the UCB. This tends
to ignore the monopoly power of the body and that it performs the equivalent of governmental
functions. In contrast in Coetzee the court held that the National Soccer League performs a public
function in the public interest and anyone wanting to play professional soccer is subject to their rules.

In AAA Investments the High Court drew on Datafin and Dawnlaan Beleggings and took into account
the coercive nature of moneylending rules and purported to be of general application. It was not merely
a situation of voluntary submission as the rules applied to everyone and part of governmental regulation

24
o microloans. The SCA took a different view and held that the rules apply only to those who are willing
to submit. The judgment says little about what characterizes public or private power.

The views of the SCA were rejected by the CC. Yacoob J noted that as the council was performing
functions in terms of legislation it was an organ of state and subject to the Bill of Rights. There was also
absolute control of the Minister over its functioning. O’Regan offered the following criteria in
determining whether rules are public in character: ‘whether they apply generally to the public or a
section of the public, whether they are coercive in character and effect, whether they are related to a
clear legislative framework and purpose’.

(g) Procedural and formal action

In Smith the SCA held that the launching of legal proceedings was procedural not administrative. This
reasoning is open to doubt as administrators may be suing in public interest and in pursuit of statutory
obligations to recover money.

In Gamevest the SCA held that the receipt of a land claim was formal not administrative which is also
open to doubt as it suggests mechanical acts are incapable of being administrative. It does however
seem to accord with PAJA’s requirement of a ‘decision’.

(h) Miscellaneous

Various acts constitute administrative action in terms of PAJA. See pages 193-4.

4.3 ADMINISTRATIVE ACTION IN TERMS OF PAJA

Section 1 of the PAJA defines administrative action as:

… any decision taken, or any failure to take a decision, by—

(a) organ of state, when—

(i) exercising a power in terms of the Constitution or a provincial constitution; or

(ii) exercising a public power or performing a public function in terms of any legislation; or

25
(b) a natural or juristic person, other than an organ of state, when exercising a public power or
performing a public function in terms of an empowering provision, which adversely affects the rights of
any person and which has a direct, external legal effect…”

The Act then lists nine specific exclusions. The first five seem to be an attempt to capture the approach
and anticipate the attitude of the Constitutional Court in cases such as Nel, Fedsure and SARFU. The next
four are a mixed bag.

The definition in the Act is extremely narrow and highly convoluted. In Grey’s Marine it was described as
‘cumbersome’ and pointed out that it does not attribute meaning so much as to limit it with
qualifications. It is a mix of South African, German and Australian law.

The narrow definition is achieved by exclusions as well as a number of separate conditions that must be
fulfilled. A host of questions must be answered. The requirements are a mix of substantive elements and
the effects of the action on the applicants. This finds no support from the construction of section 33.

It is convoluted as the terms of the definition are within themselves defined. This means each word in
the definition must be read in the light of its own definition.

Notwithstanding all this complexity there have been some relatively simple and unsurprising diagnoses
of administrative action under the PAJA.

The adjudication of tenders and related actions have been taken for granted under PAJA even though
they have long since qualified as administrative action under section 33. Recently a more positive
diagnosis has however been made [Simunye].

The definition can be divided into seven main elements. They do overlap to some extent and whether or
not something amounts to administrative action still depends on the facts of each case.

(a) A decision

The definition of decision was taken from Australia’s Administrative Decisions Act. For the purposes of
PAJA decision means:

…any decision of an administrative nature made, proposed to be made, or required to be made, as the
case may be, under an empowering provision, including a decision relating

26
(a) making, suspending, revoking or refusing to make an order, award or determination;

(b) giving, suspending, revoking or refusing to give a certificate, direction, approval, consent or
permission;

(c) issuing, suspending, revoking or refusing to issue a license, authority or other instrument;

(d) imposing a condition or restriction;

(e) making a declaration, demand or requirement;

(f) retaining, or refusing to deliver up, an article; or

(g) doing or refusing to do any other act or thing of an administrative nature, and a reference to a failure
to take a decision must be construed accordingly.

Two additional elements are introduced here – ‘empowering provision’ and ‘of an administrative
nature’. They will be discussed separately.

The definition of ‘decision’ is detailed and surprisingly inclusive. It encompasses decisions which are
proposed to be made, omissions and failure to take a decision. In Sibiya Wallis J found that the definition
was broad enough to encompass a situation where an identity document was issued but not delivered.
He warned against fragmenting what ought to be a single action into parts which may and may not be
administrative action.

However not every official action will qualify as a decision. Certain clerical acts will not amount to
administrative action for instance the signing of a declaration.

The Australian origin of the definition raises some questions as to whether or not administrative action
includes administrative lawmaking and conduct preparatory to the making of the decision. PAJA omitted
these from the definition which suggests an intention to rule these stages out. This is reinforced by the
additional requirement of adverse and direct effects on rights. PAJA however refers to decisions
‘proposed to be made’ which counter the suggestion that preparatory stages are excluded. In Offit
Enterprises the SCA made it clear that PAJA does not encompass future administrative action which may
not happen.

It remains unclear as to whether or not lawmaking amounts to a decision under the PAJA. The high court
in New Clicks the majority found that rule making could not be regarded as a decision while the

27
dissenting judge held that this would be unconstitutional. The Constitutional Court failed to clear up the
controversy. Chaskalson CJ supported the view that rule making fell into the definition of decision. The
majority however disagreed with this.

In McDonald it was noted that it is still an open question. In Cable City the SCA agreed that regulation
making was administrative action and cited New Clicks as reasoning. In Hospital Association it was
stated that the majority in New Clicks had found that it did constitute administrative action.

Administrative rulemaking and preliminary action pose a problem. It has been noted by commentators
and Chaskalson CJ in New Clicks that although the jurisprudence of PAJA was inspired by Australia and
Germany it would be wrong to transplant provisions from other countries into our constitutional
framework and so courts are advised to interpret ‘decision’ broadly to avoid conflict with section 33.

(i) Of an administrative nature

The purpose of this phrase is not entirely clear but may simply emphasise the point made in the
exclusions that executive, judicial and legislative action are not administrative action. It has also been
held in Competition Commission v Telkom to contrast action of an administrative nature to that of an
investigative nature.

The phrase could also be interpreted as it is in Australia to exclude administrative functions which are
not purely administrative such as judicial but this interpretation should be resisted as the classification
of administrative functions has been discredited in our law.

In the New Clicks case the phrase was held to bring regulation making within the scope of a ‘decision’
and it was held that regulations are indeed subject to review under PAJA.

In Sokhela it was held that there must be a positive finding of ‘administrative nature’ and the fact that
action does not fall under the exclusions does not automatically render it administrative action.

(iii) Made under an empowering provision

It is clear from the definition of a ‘decision’ as well as the definition of administrative action relating to a
natural or juristic person that there must be an empowering provision for an action to be reviewable.
The definition of empowering provision is broad and it makes it clear that private actors may be subject
to judicial review under PAJA. This is different to the Australian narrow definition of an ‘enactment’.

28
In Marais v Democratic Alliance it was held that ultra vires actions could not be reviewable as they
lacked an empowering provision but adopting this view would restrict reviewability to lawful
administrative action only and this would be awkward.

(b) By an organ of state (or a natural or juristic person)

Organs of state are capable of performing administrative action in terms of the Constitution, provincial
constitution (any power) or legislation (in cases of public powers and public functions).

This seems to entirely exclude the exercise of private powers by private bodies which could only be
reviewed in terms of common law. There is room for the argument made in Datafin that some private
regulatory bodies exercise public functions particularly where the private body performs a function
which would have to be performed by an organ of state if that body did not exist.

(c) Exercising a public power or performing a public function

There has been a conservative approach taken by the courts in some cases such as Marais and
Pennington. Both of these judgments ignore the public role played by some entities such as the political
parties and medical aid schemes with which these cases were concerned.

A broader approach was taken in Van Zyl where it was held that the National Council of Provinces had a
‘strong public component’. Also in Max v Independent Democrats which concerned a political party. The
Sokhela case also showed a broader approach.

In Tirfu it was held that ‘significant public interest’ is relevant which is contrary to the Marais case. In
Calibre Clinical Consultants it was held that no support is found in SA or abroad for an approach which
says that a body is exercising a ‘public function’ only because of public interest.

In National Horseracing Authority public interest was a feature discussed by Wallis J in determining
public power but ultimately the question was left open in this case.

In Calibre Clinical Consultants concerned an action taken by a bargaining council. In this case it was held
not to be public power as it ‘was not woven into a system of government control’ or ‘a privatization of
the business of government itself’. The statutory basis was merely because of existence in terms of
company legislation. The council was not spending public money but money collected from members

29
and the industry. The council could have appointed a service provider without tenders or proposals.
Section 217 of the Constitution did not apply.

The Constitutional Court’s decision in AAA Investments shows that a different approach could have
been taken. More emphasis could have been made by the court on section 28 of the Labour Relations
Act which authorizes bargaining councils to act in such a way and so could be seen as integrating the
council into a system of statutory regulation. The extension of a collective agreement could be seen as
an element of coercion and funds collected were compulsory not voluntary. Although the council may
not have been required to comply with section 217 of the Constitution, administrators do many things
that they are not strictly required to do which my still entail public power as was seen in Simunye
Developers.

It was argued in Airports Company South Africa that since the decision in Calibre Clinical Consultants the
essential enquiry was ‘whether the said conduct arises from the exercise of a governmental function or
not’. This was rejected as being too restrictive and it was held that all relevant factors must be looked at
together.

(j) Public-sector employment relationships: Public or private?

One view is that employment relations of public entities should be governed by section 23 of the
Constitution and associated legislation to the exclusion of section 33 and PAJA. The other view is that
the exercise of public power attracts the protection of administrative law as well as labour law
regardless of the context.

In the SAPU case the first approach was taken and a distinction was drawn between inherently public
activities and internal employment relations premised upon a contractual relationship of trust and good
faith where parties or on equal terms. Due to the extension of labour rights to public-sector employees
post-1994 there was no longer a reason to use administrative law to advance labour rights.

The second approach can be seen in the POPCRU case where it was held that power to dismiss vested
with a public functionary who was required to exercise it in the public interest. Factors such as the
statutory basis of the power and public character of the department weighed heavily on the decision. It
was pointed out that individuals should have more legal protection than less.

30
The SCA could decide the debate in Chirwa but a divided court offered three different answers which
complicated the problem. Jafta JA concurred with SAPA, Cameron JA and Mpati DP concurred with
POPCRU while Conradie JA held that the legislature did not want labour matters to be litigated in other
courts.

On appeal the majority of the Constitutional Court concurred with Conradie JA. Skweyiya J held that the
labour court had exclusive jurisdiction to prevent public sector employees being afforded more
protection and the development of a dual system of law. Ngcobo J endorsed the jurisdictional solution
but held that the dismissal did entail the exercise of public power but was not administrative action as it
was contractual. Langa CJ dissented on the jurisdictional solution and further found it was not an
exercise of public power.

The decision is problematic and mutually destructive as the finding on jurisdiction leaves no room for a
finding on administrative action.

The problem with the jurisdictional solution was finding a legal (not policy) based justification as the
High Court has jurisdiction in labour cases. Further, the appellant had framed the dispute as one of
administrative action not labour since her initial claim in the CCMA but the court insisted it was a labour
dispute based on the initial claim in the CCMA.

This approach came about as a result of a compartmentalized characterization of section 23 and section
33 of the Constitution which was rejected a few weeks earlier in Sidumo. The SCA pointed out in
Makhanya that the law does not exist in distinct boxes and is interrelated.

(ii) The (new) requirement of ‘public impact’

Due to the difficulties in the case of Chirwa the court abandoned the jurisdictional solution in Gcaba and
again focused on the interdependence of rights such as those in section 23 and section 33. The court
used administrative law to promote the policy aims from Chirwa.

The court held that the problems of the parallel systems of law etc would be avoided if the conduct
were not administrative action and so made that finding but with little analysis. Part of the decision
relied on the decision in Chirwa that the dismissal did not amount to administrative action but unlike
this case, there was no legislation which governed the dismissal and so the relevance of the conclusion
reached in Chirwa is not clear. It is also unclear whether or not the court started with section 33 or the

31
PAJA and the court does not mention which factor on the SARFU list is unsatisfied or which element in
the PAJA is not met.

The court also held that it was essentially a labour related matter which had few or no direct
consequences for other citizens. This element of ‘public impact’ is not set out in the PAJA and seems to
have been introduced by the courts. The element of public impact was often used in cases involving
private bodies exercising a public power [Dawnlaan Beleggings] [AAA] however this case is different in
that the action was clearly an exercise of public power and so using the element of public impact to
justify it not being administrative action seems perverse. The new requirement thus raises questions
such as whether it only applies to labour related matters or is it a new ingredient in administrative
action? It was pointed out in POPCRU that existing administrative action may only affect the individual
concerned.

Notably, the court did not mention the requirement of ‘external’ legal effect which was given similar
meaning to public impact in SAPU and which in Ntshange was held to be satisfied because it affected aat
least the appellant.

Ultimately Gcaba raises more questions and causes complexity and confusion. The idea that something
is either a labour matter or an administrative matter is seriously mistaken. Litigants are entitled to the
full protection of all and any applicable rights, particularly section 33 which is a general right that covers
areas governed by more specific rights.

The court’s reasoning that labour related matters are not administrative action may not discourage
applicants from approaching the High Court in terms of the principle of legality and so this problem of
parallel systems and duplicate jurisdiction is not entirely eliminated.

(d) In terms of any legislation (or in terms of an empowering provision)

Organs of state must be exercising public power in terms of the Constitution, provincial constitution or
be a public power or performing a public function ‘in terms of any legislation’. The requirement is easily
satisfied by the prevalence of legislation or delegated legislation.

It may be argued that action taken in terms of a contract is not taken in terms of any legislation. In Cape
Metro the court reasoned that the cancelling of an outsourcing contract on the basis of fraud was based

32
on common law contractual grounds and not legislation even though legislation provided for such a
cancelation.

The case was decided under section 33 and the court ruled that there was no use of public power at all
which reasoning could also be used under the PAJA. There are thus two approached.

A similar view to the first approach was taken in Chirwa where the majority judgment focused on the
absence of the implementation of legislation as laid down as a pointer in SARFU. The dissenting
judgment engaged directly with PAJA and it was held that in the absence of a legislative provision
governing the dismissal it did not amount to administrative action and was contractual in nature.

The phrase ‘exercising a public power…any legislation’ also appears in PAIA and in cases surrounding this
act, a liberal broad approach was taken to interpret the phrase which could include functions performed
in accordance with legislation.

(e) That adversely affects rights

This is one of the most important aspects of the definition. It was held in Walele that the definition in
section 1 can be read to encompass not only rights but also legitimate expectations.

(i) Determination or deprivation?

The verb ‘affect’ can be interpreted to mean either determining rights or taking away or depriving of
rights. In terms of the determination theory the limiting effect is negligible however the deprivation
theory creates an unacceptably high threshold for admission to administrative action. It would exclude
applicants who have not yet been given rights as they are in the process of being granted.

Although PAJA is not clear, the determination theory ought to be accepted on the basis of section 33(1)
of the Constitution which does not depend on the affected right when being applied unlike section 33(2)
(right to reasons) so the deprivation theory would render the Act unconstitutional as it would not give
effect to section 33(1). Fortunately the courts seem to be moving towards the determination theory.

(ii) Determination and the role of variability

Some may argue that the determination theory is too broad. Lawfulness already applies across the
board but how can administrators apply reasonableness and fairness to all administrative conduct? The

33
idea of variability may be able to supply boundaries to this. It allows courts to vary the content of
administrative justice according to the circumstances. This approach is determination-plus-variability.

Variability is well accepted in procedural fairness where flexibility is crucial and fairness must be tailored
to suit the circumstances of the case. The question should not be ‘when can we withhold fairness’ but
‘what fairness does this case require’ which limits the content of fairness not the application. This can be
applied to lawfulness and reasonableness too.

In Joseph the court supported the notion of determination-plus-variability when it said, of fairness, that
administrative efficiency informs the content of the duties not the scope of application of administrative
law.

(iii) Interpreting ‘rights’ to include administrative justice rights

Another method to mitigate the deprivation approach is to construe ‘rights’ as including the right to
administrative action as was done in Transnet in context of the right to reasons. One could thus always
argue that the adversely affected right is the right to administrative justice [Goodman Brothers].

However this is circular reasoning and one feels that the rights to be affected must surely lie elsewhere.
This would make the requirement of ‘adversely affects…rights’ redundant as it would always be
satisfied.

(iv) The ‘capacity to affect legal rights’

The reasoning in Goodman Brothers has not been used in relation to various rights called for by the
statute in other sections. The meaning of ‘rights’ may still be softened. In Bullock it was suggested that
rights should not be restricted to those enforceable in a court of law.

In Grey’s Marine Nugent JA said that ‘rights’ was not intended to be taken literally and just serves to
emphasise that administrative action impacts directly and immediately on individuals. This is close to the
determination theory. Further the adverb ‘adversely’ was seemingly disregarded.

This judgment was relied on in Kiva in terms of rights to fair labour practices. Further in Dunn and
Wessels decisions not to appoint were held to be administrative action in terms of legal rights. In
Steenkamp, Moseneke DCJ referred to a direct effect on ‘legal rights or interests’ but this was a brief
description which does not carry much weight. In Harvey a generous application of the approach was
used in contrast to the more narrow approach taken in Competition Commission of SA v Telkom.

34
 (v) Whose rights?

Section 1 of the PAJA indicates a broad approach must be taken and so someone’s rights must be
affected, not specifically the applicant’s. the SCA has been content with a decision where the lessee’s
rights (not the applicant’s) were affected and it was beneficial not even adverse [Grey’s Marine].

(f) That has direct, external, legal effect

This was a last minute borrowing from German law.

(i) The meaning of ‘direct, external legal effect’ in German law

The term ‘legal effect’ entails determinations and deprivations. A ‘direct’ effect refers to immediacy or
finality. This expresses that only final decisions be subject to judicial scrutiny. Thus all preparatory
decisions are in principle not reviewable.

The ‘external’ effect means that it must not be a purely internal matter of departmental administration
or organization. A decision regarding performance of duties would not be administrative action while a
dismissal would be.

This is all well and good in German law but poses more of a problem in South African law and it is not
clear why parliament chose to include the phrase.

(ii) South African meanings

If the German meanings of the words were to be followed by our courts it would not solve the
determination/deprivation puzzle and the application of ‘external’ did not pose a problem at common
law.

The allusion to finality and ripeness from ‘direct’ does seem to add something. In common law, decisions
must be ripe before review so as not to waste time with a decision which may yet change. It could be
interpreted further to mean that preliminary decisions are not reviewable however the jurisprudence
from the courts have recognised the significant effects that preliminary decisions can have and so
‘direct’ should not be read to make preliminary decisions entirely unreviewable although limits may still
be placed.

35
The requirement of ‘decision’ seems to accord with the idea of finality but then the Act does talk of
decisions ‘made, proposed to be made...case may be” and the definition relating to ‘doing or refusing to
do…’ seems wide and vague enough to cover preliminary decisions in the same way as administrative
lawmaking. The idea of final may also run contrary to the wide scope of the Act itself and section 3 and
section 4 set out procedures to be followed before decisions are made.

Not much reliance must be placed on additions made hurriedly and at a late stage or those transplanted
from foreign law as they may produce different results from the countries from which they were taken
[New Clicks]. The courts will still need to work out the South African meanings.

Some progress has been made by the courts though not consistent. In Joseph it was held that the
requirement was satisfied whenever rights are materially and adversely affected for section 3 of the
PAJA. In Gcaba it was not mentioned at all. In Viking Pony Pumps it was fatal but this case emphasized
that an administrative action enquiry depends on the facts of the specific case.

(aa) Direct effect

There is pre-PAJA support for decisions needing to be final and the reasoning was used in a number of
cases.

Since the coming into operation of PAJA, the requirement of finality has been referred to in case law. In
Grey’s Marine it was held to emphasise that administrative action impacts directly and immediately on
people. In Oosthuizen a liberal approach was taken and it was held to be sufficient if there was capacity
to affect legal rights and the administrative action can be directed towards or aimed at direct
consequences. Further, preliminary decisions were held to be reviewable particularly where laying the
foundation for a possible decision.

The liberal approach cannot be pushed too far and in Viking Pony Pumps although the Grey’s Marine
position was accepted, it was held that an investigation which did not determine culpability could not
adversely affect rights with a direct and external legal effect.

(bb) External effect

There are allusions to a requirement of external effect in the pre-PAJA case law such as Goodman
Brothers where weight was given to the internal nature of the action.

36
The PAJA requirement has been dealt with more explicitly in cases such as Metcalfe, Botha and Sokhela.
It has featured in several public-sector employment cases. In SAPU a decision to change the shift system
was held to be an internal matter of departmental organization while conversely in Nxele it was held
that ‘external’ should not be taken literally so as to exclude actions that affect members of the public
body. This is in line with the general approach in Grey’s Marine. The court was surprisingly silent on the
element in Gcaba but gave weight to public impact which was given a similar meaning to that of
‘external’ in SAPU to which the court referred to with approval.

(cc) Legal effect

This requirement is easy to satisfy as it overlaps with ‘adversely affects rights’ so it can be inferred
where that is satisfied [Joseph]. Skweyiya J refused counsel’s invitation to use it to narrow the scope of
administrative action.

4.4 EVALUATING THE DEFINITION IN THE PAJA

PAJA’s definition of administrative action is far more circumscribed and narrow than that crafted by the
courts. The question is thus, does the Act’s narrow definition matter?

(a) Arguments for reducing the realm of administrative action in the PAJA

It is not longer necessary for an extended role of administrative law in the democratic era and it is
understandable that the courts have sought to cut the ambit down to size where section 33 is concerned
but PAJA seems to have gone too far.

The narrow definition may be as a result of section 33(c) which may be interpreted to be a constitutional
mandate to reduce the burden on the administration by narrowing section 33 in legislation.

Efficiency is an attractive goal but must not be valued above correctness. It is only one of the values
governing the administration in section 195 of the Constitution. When balancing efficiency with fairness,
administrative justice may be deprived of its vitality. In terms of PAJA, Currie points out that ‘efficient
administration’ can be read ‘downwards’ to reduce the burden and increase cost-effectiveness or
‘upwards’ to require accountable and responsive decision making. Therefore the constitutional mandate
may not be read to mean the legislature should reduce the legal burden. In Joseph it was held that
administrative efficiency should inform the content not the scope of administrative action.

37
It seems that the definition in the PAJA aims to prevent a flood of legal actions that can hamper the
efforts of government to deliver and threaten the transformation process. The next question is thus,
does it succeed in preventing the flood?

(b) The negative effects of the PAJA’s definition

The Act aims to exclude much of what had been reviewable at common law however the narrow
definition fails to stop the flood of litigation because non-administrative action is reviewable in ways
other than the PAJA. Other rights may be applicable as well as special statutory review or common law
review of private powers. The most significant however is the principle of legality which is regarded as
an essential safety net and turned to often. Judges have observed that it makes virtually the same
demands as regular administrative law.

Courts are always able to control the abuse of power even with the narrow definition in the Act. An
argument could thus be made that the definition of administrative action is practically insignificant
because as long as there is a way to control the power, the source does not matter. Therefore people
could have the free choice of which pathway to proceed by. However, PAJA is unavoidable and must be
used, even without the assistance of a legal education. Users of the PAJA may be more flummoxed than
guided by the definition making it inaccessible, even to lawyers and the courts. It is not acceptable that
something so fundamental is so uncertain. The administrative action enquiry is an unfortunate drain on
resources and there should be greater focus on more substantive issues. It puts the focus of inquiry in
the wrong place and so cases may fizzle out after the definition stage. It uses the formalistic use of the
administrative action inquiry to solve substantive problems.

One of the most significant difficulties is the disparity between the definition in the PAJA and section 33.
Phrases were imported from foreign law that do not exist in relation to section 33. The purpose of the
Act is to give effect to section 33 and thus the tension must be resolved by construing PAJA consistently
with the section 33 meaning [Grey’s Marine] [New Clicks] [Sidumo].

38
Chapter 5 – Lawfulness

AUTHORITY

5.2 THE REQUIREMENT OF AUTHORITY

Section 6(2)(a)(i) of PAJA gives effect to section 33(1) by allowing review of an action taken by an
administrator who ‘was not authorized to do so by any empowering provision’. Similarly section 6(2)( f)(i)
allows for review of an action ‘not authorized by an empowering provision’. This reflects the common
law and the principle of legality. There are a number of ways administrators may fall foul of these
provisions.

(a) Appointment and constitution of administrators

Where there doubt of an identity of an administrator it is usually ascertainable from the legislative
context (‘the officer’, ‘the tribunal’, ‘the board’). Whatever their identity they must be properly
appointed, qualified and constituted when taking administrative action.

Many cases observe this fundamental principle such as Simelane where the absence of a commissioner
meant that the commission was not properly constituted. In JSC v Premier, Western Cape the absence of
the Premier of the province invalidated proceedings.

39
Generally, unless a quorum is specified, action must be taken by all members of the body or it would
defeat the purpose of selecting the members. This may be sacrificed for practical necessity such as
urgent decisions.

Regarding voting, unanimity is required by default unless the legislation opts for a majority or other
basis. In JSC v Premier, Western Cape this was held to mean a majority of all members not simply those
present.

(b) Exceeding authority

The majority of cases relate to administrators who went beyond the powers conferred on them
[Kwakwa] [Watchenuka] [Affordable Medicines Trust] [Hoerskool Ermelo].

Sometimes administrators make decisions for which there is no authority at all. A typical example is
Rangani where an administrator revokes benefits to determine who a proper beneficiary is. The
motivation is understandable but the method is condemned.

Authority may reside in a specific piece of legislation or may be a more general constitutional duty such
as in Kyalami Ridge. This case also showed a common law power as a result of being an owner of land. In
Bullock it was held that ownership (or contractual) rights are always to be asserted within the
framework of the Constitution and any relevant legislation.

(c) Existence of authority elsewhere

The question arises if it is valid for action to be taken under the wrong legislative provision where
authority actually exists elsewhere. In Pinnacle Point it was held that this would validate action. In Latib
the court took the same approach providing there was no direction in the statute requiring the section
to be mentioned. However in Pinnacle Point it was noted that this does not apply where an
administrator acted deliberately in terms of a particular provision.

The challenge is between paying regard to the right to lawfulness and faulting administrators for purely
technical or administrative errors. The liberal approach cannot be pushed to far as an administrator
must fulfill the requirements in respect of which power was actually exercised.

In Sedgefield Ratepayers’ it was held that if two powers could be invoked, one restricted and one not,
and the administrator uses the restricted one without the restriction then he cannot take refuge in the

40
existence of the non-restricted power. In Harris the fact that a minister had ‘consciously opted’ for a
specific provision did not open up for him the ability to rely on the correct provision. Authority
elsewhere can only be used where failure to expressly or accurately invoke the source is immaterial. This
seems to be in line with section 6(2)(f)(i).

5.3 DELEGATION OF AUTHORITY

Section 6(2)(a)(i) indicates that the power given to one administrator must not be exercised by another
unathorised administrator or body. Whether it is a usurping of powers or an abdication it remains
unlawful and there are a number of ways in which this is done.

(a) Delegation of power by a legislature

When original legislators confer authority on administrators they are said to delegate power.
Delegation, unlike assignment, is revocable and less than complete. The delegator retains control in
terms of the purpose and the terms regulating the delegation.

Delegation of powers to adjudicate or make delegated legislation undermines separation of powers and
so there are constitutional limits on the ability to delegate powers.

In the pre-democratic era Parliament could delegate as much power as it liked however the approach by
the Constitutional Court has been different. Where the legislature purported to delegate power to
amend an Act of Parliament it was held that this would subvert the ‘manner and form’ provisions in the
Constitution and gave away too much of the legislature’s lawmaking responsibility [Executive Council,
Western Cape Legislature].

In the case of Justice Alliance the president wished to extend the term of the Chief Justice and this was
held to go to the core of judicial independence and the separation of powers.

Any broad discretionary power may fall foul of the Constitution. In Dawood parliament had given Home
Affairs wide powers to grant or extend residence permits but without guidelines. It was held that the
legislature must limit the risk of an unconstitutional exercise of discretionary power and it would not be
enough to challenge the exercise of the power later. The legislature must provide specific guidance as to
when limitation of rights would be justifiable.

41
There are exceptions to this when the factors relevant in decision making are clear and indisputable,
when the administrator possesses the necessary expertise or where the factors are so numerous that it
would be impossible to identify them all.

This is a fairly rigorous principle, after all the exercise of any administrative power may violate
fundamental rights. The view that whenever the legislature confers a wide administrative power it must
provide, or mandate, appropriate guidelines to the limit the risk is fortified in the case of Armbruster.
Even though the power fell into the third exception above, it was held that the appropriate authority
should formulate some guidelines to encourage consistent and correct application of the power.

(b) Subdelegation by administrators

Often it is practically impossible for an administrator to exercise a power personally. Original legislators
may stipulate that their delagatees may subdelegate. This is the position in terms of section 238 of the
Constitution. The section imposes a fundamental limitation on the activity in that it must be ‘consistent’
with the relevant enabling legislation in that it was authorized, expressly or impliedly by that legislation.
The position is the same under the PAJA.

(i) Express authority to subdelegate

Many statutes contain a delegation clause which expressly allows for delegation and may prohibit
subdelegation of certain powers.

(ii) Implied authority to subdelegate

The question often arises whether the power to subdelegate can be implied in legislation. There are
strong arguments for and against. On the one hand it is practically necessary but on the other hand the
legislature has good reasons for entrusting certain tasks to specific individuals. There is a rebuttable
common law presumption that delegatus delegare non potest which means ‘one on whom power is
delegated cannot himself further delegate that power’. The willingness of the court to approve
subdelegation will depend on various factors discussed below.

A question arises where express powers of subdelegation exist whether further implied powers can then
be found. In Teltron it was held that they could not and this was upheld by Du Plessis J in AAA however
O’Regan J resisted this logic and held that where it is impractical for a person to do something alone, the
power must be construed to permit any subdelegation to an appropriate subdelegatee.

42
 (aa) The nature of the power

‘Judicial’ and ‘quasi-judicial’ powers are traditionally not subdelegable in our law. However these terms
have not always been interpreted consistently. ‘Judicial’ often means a decision based on rules which
has little room for discretion but has been interpreted in terms of subdelegation to describe a highly
discretionary power. A mechanical power is more easily subdelegable.

Subjectively phrased powers indicate the widest discretion so should not be subdelegated [Shidiack].

Powers with far-reaching consequences are not as appropriate for subdelegation as those with less
impact. This was relevant in New Clicks where a regulation to subdelegate the power to determine a
pricing system was struck down as it was a crucial decision. Further, it was held to be a “broad subjective
discretion”.

(bb) The extent to which the power is transferred

Subdelegation is more likely to be approved where discretionary power is only partly transferred so the
original delagatee retains sufficient control or provides sufficient guidelines. The Constitutional Court
has held that the distinction between ‘decentralisation’ and ‘deconcentration’ must be approached with
caution as the starting point must be the Constitution.

(cc) The importance of the subdelegatee

Where a subdelegatee appears to have been chosen for special abilities or qualifications, it is likely that
subdelegation will be disallowed. In Niewoudt inspectors could not subdelegate to an advocate as it was
assumed that those inspectors had been chosen for their particular abilities and qualifications. This was
also relevant in New Clicks as the minister had been appointed due to her special expertise.

(dd) Practical Necessity

The intention of the legislature may need to bow down where it is impossible for a delegate to exercise
power personally. The case of Shidiack may have a different outcome today as ne minister cannot be
expected to mark every language test. The court in Carltona accepted that powers given to ministers are
normally exercised by responsible officials under authority of the minister even without formal
subdelegation of authority. This principle is overly-optmistic in South Africa as it relies on the ministerial
effectiveness to secure accountability and has the potential to swallow the rule against delegation.

43
Practical necessity may also allow for two bodies who have been delegated the same task to
subdelegate to only one of those bodies.

(c) Unlawful dictation and referral

These are more covert forms of delegation and tend to be less obvious and more difficult to prove. They
cannot prevent, or be intended to prevent, a minister from consulting other officials before taking a
decision. The point is that the final decision must be taken by the empowered administrator and not
merely be rubber stamped.

Unlawful dictation is known as ‘taking directions’ is where a decision is in fact made at the dictation of
an unauthorised administrator. It is provided for in section 6(2)(e)(iv) of the PAJA. In Tantoush an
Interpol official pressured a refugee officer to refuse an asylum application.

Unlawful referral is known as ‘passing the buck’ is where a decision is referred to an unauthorised
administrator. While PAJA does not specifically provide for passing the buck, it is covered by the catch-
all ground in section 6(2)(a)(ii). In Scenematic Fourteen the Deputy Director-General relied on advisory
committee to evaluate applications etc and this was allowed as he was legitimately relying on expert
knowledge after weighing up other factors.

In Walele it was seen that it is difficult to determine where legitimate reliance ends and unlawful
referral begins. The degree of involvement expected of the decision maker is relevant. The case dealt
with the extent to which an ultimate decision maker could rely on the recommendation of a Building
Control Officer. The majority held that the decision maker must make the decision himself and must be
satisfied that all of the relevant factors have been complied with and that none of the disqualified
factors have been considered. He must do it himself. However the minority said that the decision maker
should be allowed to give considerable weight to the Building Control Officer’s recommendation
because he was a skilled expert.

JURISDICTION

Administrators are required to remain within the bounds of their powers and not misconstrue them in
terms of legality or PAJA. Judicial review enables courts to decide whether or not this has been done in
other words whether or not they have strayed beyond their jurisdiction. This involves the interpretation
of legislation but there are opposing views as to whether the court or administrator should have the

44
final view as to how to interpret the boundaries of the jurisdiction. This arises in relation to both law and
fact which are two distinct concepts.

5.5 ERROR OF LAW

A wrong or mistaken interpretation of legislation is an error of law. Section 6(2)(d) of the PAJA and the
principle of legality recognise this as a ground of review. Judges are experts in law so should be able to
quash any interpretation they regard as incorrect. On the other hand, correct interpretations are not
any more clear to judges than administrators. Further, judges should not review whether an
administrator was substantively right or wrong but whether the conclusion was arrived at in an
acceptable manner. To understand how a court can review a question of substantive correctness, our
history must be looked at.

(a) Jurisdictional and non-jurisdictional errors

For years, our courts took the view that they could review an error of law which prevented an
administrator from appreciating the nature of the powers and thus prevented the proper exercise of
discretion. However, all errors of law will lead to this outcome and so to tell the difference between a
reviewable and non-reviewable mistake of law, the English approach was adopted.

A distinction was drawn between a jurisdictional error made in determining the limit and extent of
power and a non-jurisdictional error made in the course of deciding a matter which it had jurisdiction to
decide. Where an error was jurisdictional, a court could intervene as the administrator was acting
unlawfully in terms of a decision they had no power to make. However a non-jurisdictional error could
not be reviews as courts would not interfere with the merits of a decision.

This reasoning applied particularly to review of lower courts not administrative tribunals. The reasoning
was taken up by the AD in the case of Doyle v Shenker as well as Chesterfield House where the court
refused to review decisions taken within jurisdiction (or ‘on the merits’).

A different view was taken in Goldfields where a decision by a magistrate was reviewable because he
has misconceived the whole nature of the enquiry or of the duties in connection therewith and so did
not deal with the matter as contemplated by the legislation. Other earlier cases held that if a wrong

45
decision was reached as a result of misinterpreting the law it could be set aside by common law review
proceedings.

(b) A new approach

The traditional distinction is arbitrary and it is difficult to see why any error of law would not prevent a
decision maker from properly considering the matter. This began to be conceded in English law, at least
in relation to decisions of administrative tribunals. In Pearlman Lord Denning MR held that the
distinction should be discarded all together. In Re Racal Communications it was held that the distinction
had effectively been abolished. Legislation could allow administrators to decide on questions of law but
‘this required clear words’.

In South Africa, in Hira v Booysen the court held that the traditional distinction was by no means a clear
one and it was difficult to reconcile cases going one way with those going the other.

The Chief Justice in Theron offered the following guidelines:

 Where powers are of a purely judicial nature, a court will be reluctant to find that a tribunal has
exclusive jurisdiction to decide on the meaning of a legislative provision.
 The materiality of the error must be looked at in determining if the error renders the decision
invalid.
 Where power is discretionary, the approach to establishing the intention of the legislature may
be different.

(c) Error of law in the constitutional era

Any legal system which upholds the distinction between appeal and review will experience problems in
terms of review for an error of law and why a court can pronounce on the substantial correctness of an
administrator’s interpretation.

The problem is exacerbated when there is uncertainty as to the justification for judicial review. In Hira
the court linked review jurisdiction to the intention of the legislature. The Constitution provides a much
broader link in giving rights to lawful and reasonable administrative action. A court can review any error
of law which materially affects the outcome. This is given effect to in PAJA under section 6(2)(d).

46
Courts must not disregard the reasoning in Hira as the legislature’s interpretation of the Constitution
must be given some respect. The courts power can only be exercised with cooperation of other
branches of the state.

PAJA has been applied in a number of cases such as Mikro, Tantoush and Sithole.

5.6 MISTAKE OF FACT

(a) Jurisdictional and non-jurisdictional facts

The distinction again follows from English law and generally only the first category appears in case law.

Jurisdictional facts refer to preconditions which must exist before the exercise of the power and
procedures to be followed, or formalities to be observed when exercising the power. They are
jurisdictional as the exercise of the power depends on their existence. If the facts are not present or
observed then the exercise of power will be unlawful otherwise administrators can arrogate powers to
themselves or inflate their own jurisdiction.

There are hundreds of cases showing the operation of jurisdictional facts in South Africa. See for
example Paola v Jeeva NO. PAJA allows for review in the case of procedural or substantive jurisdictional
mistake of fact in section 6(2)(b).

Non-jurisdictional facts are those on which jurisdiction does not depend. This would be a mistake within
the jurisdiction of the administrator and would not be reviewable in common law. PAJA seems to follow
this by omitting to list such a mistake of fact but the SCA reformed the law to include it nevertheless.

(b) Procedural jurisdictional facts

Legislation frequently imposes procedural requirements on administrators and applicants. Failure to

comply does not necessarily lead to invalidity. The court must decide if it is mandatory or directory
hence PAJA refers to a ‘mandatory’ procedure. At common law, only failure to comply with a mandatory
provision leads to invalidity.

In reality it is not so clear cut. Mandatory provisions need not be strictly complied with but adequate or
substantial compliance may be sufficient which may be why PAJA references a ‘material’ procedure.

47
Directory provisions cannot simply be ignored and there could be a mandamus or an order directing the
administrator in terms of PAJA.

Further, the factors for deciding if a provision is mandatory or directory provide an uncertain guide.
Courts often interpret peremptory language to be directory and the existence of a sanction may in fact
mean that the legislature intended for the sanction to be used in place of invalidity [Estate Van Rhyn].

A court may find mandatory and directory provisions to be closely linked in a provision. The court in
Maharaj held that the test should be whether compliance is necessary to achieve the purpose. This has
been followed in other cases and shows a trend to move from the legalistic to the substantive. Debates
about mandatory and directory provisions must come second to the intention of the legislature and the
purpose.

The Constitutional Court also adopted the purposive approach in ACDP v Electoral Commission and held
that the question is ‘whether what the applicant did constituted compliance with the statutory
provisions in light of their purpose’.

In Phoenix Cash & Carry an administrator read a list of evidence that a tenderer may rely on as being a
list of peremptory clauses and so this was read as excluding reliance on any other forms of evidence.

In Millennium Waste a tender was disqualified for a failure to sign a ‘declaration of interest form’. The
court warned not to apply such conditions mechanically and held that the administrator had made an
error of law in disqualifying the tender for an innocent omission. The administrator was empowered to
condone non-compliance. Our law permits condonation of non-compliance with peremptory
requirements where it is not incompatible with public interest. In this case it was in line with section 217
of the Constitution.

(c) Substantive jurisdictional facts

The jurisdiction of an administrator may depend on a particular state of affairs. There is an analogy to a
policeman who is entitled to arrest where there is ‘reason to believe’ that an offence has been
committed. The question is how the court decides if the prerequisite was met.

48
There is a view that the police officer must decide on the jurisdiction and the court must accept it.
Another view is that the courts must establish if there were grounds for arrest because administrators
can make mistakes. There are also compromise views.

Before 1994 the courts adopted a compromise between objective jurisdictional facts, facts or state of
affairs that had to exist objectively and subjective jurisdictional facts. A court could make an enquiry on
the objective facts and could only interfere with subjective facts where it could be shown that the
administrator acted mala fide, with ulterior motive or that he failed to apply his mind or disregarded
express provisions of the statute.

(i) Substantive jurisdictional facts at common law

To determine whether a jurisdictional fact is substantive or procedural, legislative interpretation is

required. Substantive facts are generally implied by language such as ‘if satisfied that’ which confer wide
unfettered discretion. Subjectively phrased clauses act to subjectively ouster the jurisdiction of the
court. Courts were reluctant to interfere where the legislation vested discretion solely in an
administrator.

In 1989 in Van der Westhuizen v UDF Hefer JA broke down the distinction between wide discretionary
powers and subjective discretionary powers and made it unnecessary to use subjective language to
confer subjective discretion.

Language is not the only factor that shapes a decision and a result may depend more on the judicial
ideology of the court hearing a case.

(aa) ‘Reason to believe’

The subjective reputation of this clause probably stems from the resemblance to the English phrase
‘reasonable cause to believe’ used in Liveridge. The majority held that a court cannot question grounds
for belief and subjective satisfaction is sufficient. Only one judge pointed out that having reason to
believe was different to thinking one has reason to believe.

In South Africa, the question of whether this is objective or subjective was laid to rest in Hurley. It was
held that objective grounds must be shown that give rise to or form the basis of the belief.

49
 (bb) ‘Is satisfied’

In Katofa Rabie CJ took the view that these words conferred subjective discretion. A person merely has
to show that they were satisfied and then the onus is on the other party to show a failure to apply the
mind or that the decision was in bad faith. The real question is not where the onus lies but what is
required to discharge it.

(cc) The requirement of an ‘opinion’

In Dempsey it was held that the prerequisite of an opinion is a subjective jurisdictional fact and the view
on onus was similar to in Katofa. However this view was changed two years later and it was held that
the person must hold an opinion that something is necessary to in order to achieve the purpose and so
merely asserting opinion was not sufficient.

(dd) Cases unrelated to ‘security’

All the above cases refer to security matters such as arrest and detention. Outside the context of
security subjectively phrased clauses were treated differently pre-democracy. The courts adopted an
objective approach to the subjective discretion and the mere say-so of an official was not sufficient [La
Grange].

(ii) Substantive jurisdictional facts and the Constitution

In recent years jurisdictional facts have been judicially considered in a range of contexts including arrest
and detention. Since the decision in Hurley the courts have treated subjectively phrased clauses more
objectively and this is evident in recent case law. Even more subjective clauses such as ‘is satisfied’ were
acknowledged for not entitling unfettered discretion pre-1994. They were however used by the
legislature to minimise the scope of judicial review.

This has become less effective in the Constitutional era. Courts must be able to satisfy themselves that
administrative action was lawful and reasonable in terms of section 33 and the principle of legality.
Reasonableness means a Hurley type approach can be used regardless of the wording. This was
confirmed by the court in Walele where it was held that there must be reasonable grounds for the
subjective opinion relied on. This makes all jurisdictional facts objectively justiciable. However subjective

50
wording may still show the legislature’s intent for judicial deference in some cases. The demise of
parliamentary supremacy does not allow courts to ignore the wish of the legislature; in fact it should be
taken more seriously now. There is a need for a balance to be struck between heeding the voice of the
legislature and protecting rights.

(d) Non-jurisdictional facts

At common law courts regarded facts within the jurisdiction as non-reviewable. In Pepcor it was held
that where jurisdiction is not in issue then courts cannot review solely on the basis of a material mistake
of fact. The courts reasoned that they had no power to enquire into the correctness of conclusions
arrived at by administrators properly vested with the discretion to do so.

Mistakes of facts could be reviewed on the grounds set out in Shidiack being bad faith, failure to apply
the mind, ulterior motive or breach of an express provision. This approach is illustrated in De Freitas
where the court held that where a functionary has the power to decide and applies his mind then the
discretionary decision cannot be set aside.

All of this changed in the Pepcor case where section 33 and the principle of legality were held to
demand recognition of material mistake of fact as a ground of review. Public interest and other factors
such as prejudice to parties must play a role in reviewability and these must be balanced. PAJA does not
expressly provide for a material error of fact but the SCA held that section 6(2)(e)(iii) could be extended
to include mistake of fact and is the most appropriate section for this.

The court pointed out the dangers of this approach. It held that this must not blur or eliminate the
distinction between appeal and review. A court cannot interfere where a functionary could decide what
facts were relevant and if they existed however it is difficult to see how this would not be a jurisdictional
error which is reviewable.

Error of fact is an emerging ground in a number of countries under English law. in South Africa it is now
firmly established under the principle of legality and PAJA.

ABUSE OF DISCRETION

5.7 ULTERIOR PURPOSE OR MOTIVE

51
In terms of section 6(2)(e)(ii) of PAJA administrative action may be reviewed if taken for an ulterior
purpose or motive. This is also covered by the principle of legality [Sex Worker Education and Advocacy
Task Force].

There is a difference between ‘motive’ and ‘purpose’ is important to understand. Common law has not
always treated motive as relevant but action taken for ulterior purposes could always be set aside for
being unlawful.

Powers must always be exercised in the public interest and not for personal advantage. The purpose of a
statute is binding on an administrator acting in pursuance to it. The purpose may be express or
determined through statutory interpretation. Even where powers are taken for praiseworthy purposes
they will be unlawful if this was not the authorized purpose [Van Eck].

There are also cases were the intentions were not praiseworthy [Rikhoto]. A notorious example is
University of Cape Town v Ministers of Education and Culture where education ministries had imposed
certain conditions on universities before they could receive subsidies including preventing boycott
action etc. They required that any transgressions be reported to the government. This was held to be in
order to achieve the purposes relating to law and order which was not the purpose contemplated by the
Universities Act. The court makes it clear that ulterior motive does not necessarily mean a sinister
motive. Purpose would likely be a more appropriate word here as motive seems to suggest a hidden
agenda. PAJA does however cover both situations and there is overlap with section 6(2)(e)(ii).

5.8 BAD FAITH (MALA FIDES)

PAJA permits review of actions taken in ‘bad faith’. This refers to fraud or dishonesty: consciousness that
it is unlawful. This seems to be the reference to motive as distinct from purpose.

In the pre-democratic era bad faith was used in a loose sense to mean dishonesty or mere stupidity. It
was treated similarly to ‘failure to apply the mind’. Some writers even denied that dishonesty was a
ground of review as these sorts of cases would be decided on some other ground and state of mind
alone could not render action invalid. Early case law supports this opinion as motive became irrelevant
once a legitimate purpose was determined [Broadway Mansions] [Whiterock Farms]. Towards the end
of the Apartheid era courts seemed prepared to confront motive [Waks v Jacobs] [Hart v Van Niekerk].

52
There are moral and practical reasons for dismissing the proposition that dishonesty is irrelevant. It is
however often accompanied by other grounds of review but should remain a separate ground. It is
relevant in the award of damages and in persuading the courts to substitute their judgment for that of
the administrator. Dishonesty as a ground is in line with ideals of transparency and accountability.
Conduct must be conscionable particularly as parties are not equally powerful.

PAJA certainly acknowledges dishonesty. Dishonesty could be covered by review of action taken for
ulterior motive, action not authorised by an empowering provision or the catch-all of ‘otherwise
unconstitutional or unlawful’.

5.9 FAILURE TO APPLY THE MIND

This phrase in the PAJA or common law could mean almost anything. There is little case law to clarify,
probably because it overlaps with other grounds of review. In JSE v Witwatersrand Nigel it was held that
it could be shown where a decision was arrived at arbitrarily, capriciously or in bad faith in unwarranted
adherence to a fixed principle or to further an ulterior purpose; or where an administrator failed to take
consideration of relevant considerations or ignored relevant ones. PAJA treats it as a broad ground in
terms of the component parts from the case and some of its own components discussed below.

(a) Failure to decide or to consider

At common law a power was generally accompanied by a duty to exercise the power. Administrators
cannot simply decline to act or decide. PAJA perpetuates this by defining administrative action to
include ‘any failure to take a decision’ and secondly by recognising it as a ground of review. It also
includes a decision being made without proper consideration such as in Littlewood where failure to
consider was held to be failure to exercise the discretion conferred on him.

(b) Failure to act within a reasonable time

PAJA creates a ground of review which always existed but was not well known in section 6(2)(g) and
section 6(3) which read together provide for a ground of unreasonable delay. Section 6(3) provides for
where an administrator then fails to take a decision. Special remedies are found in section 8(2). It was

53
already accepted in case law that a court could compel action in a reasonable time where statute
requires this [McGregor].

In terms of section 6(3) where a period is prescribed then action for review can be taken within the
period and where there is no prescribed period then there must be ‘unreasonable delay’. In Sibiya it was
held that applicants must convey the reasons to prove why they contend that there has been undue
delay.

The ground will not apply in the absence of a duty to take a decision and not for indecisiveness in
planning on future policy issues [Coega].

(c) Relevant and irrelevant considerations

Section 6(2)(e)(iii) of the PAJA provides for review of this nature. It need not be a complete failure to
consider but may be the giving of incorrect weight to various considerations [Bangtoo Bros].

The legislature decides what is relevant. It cannot give wide discretionary powers without some sort of
guidelines [Dawood]. When the legislature is silent then the courts must decide on what is relevant and
this is highly subjective. A typical example is Anchor Publishing were judges disagreed as to which factors
were given the correct weight when deciding whether or not to ban the publishing of ‘glamour’
magazines.

This power comes dangerously close to turning judges into administrators which raises questions about
judges competence to make decisions relating to technical matters. Courts have therefore been
reluctant to exercise this power or have diluted it.

During the 1980’s courts were denied the power to determine relevant and irrelevant considerations,
particularly in security matters [Dempsey]. However in non-security related matters this was not the
case as was shown in Witwatersrand Nigel and this is the position to follow now particularly as it has
been codified in the PAJA.

There are now many cases surrounding this ground of review, notably Eskom Holdings Ltd v New
Reclamation Group.

54
 (d) Fettering

Administrators may not act in ways which will prevent their discretionary powers from being exercised
the manner envisaged. The discretion cannot be unduly limited by its holder. This is similar to acting
under dictation which also prevents full exercise of discretion. However here the decision maker is
making his limits by adhering to rigid policies or promises of how he will act.

Fettering is not specifically mentioned in the PAJA but it is so well known at common law that it could be
regarded as being covered by the catch all ground of ‘otherwise unconstitutional or unlawful’.

(i) Fettering by rigidity

Administrators must develop and apply policies, precedents and guidelines in line with empowering
legislation. But is policies of previous decisions determine new ones then the benefits of discretion in
the particular circumstances are lost. A balance must be found.

Guidelines can be of enormous existence particularly in complex decisions with many factors.
Precedents too can save time. In Britten v Pope it was held that reference to precedent or policy may
indeed be necessary providing an administrator is alert to features which justify a departure. This was
held to still be the case ninety years later in Kemp. It was held policy and precedent must not be
considered binding with the result of no discretion being exercised at all.

In Norman Anstey & Co an application to grant a tea-room license was not considered on its own merits
but rather with rigid application of policy. Moralettasentrum is another example of rigid application of
policy as a hard-and-fast rule.

Fishing quota cases often involve fettering where there are many factors to be taken into account. The
factors may be wide ranging and in tension with each other. There are also large numbers of
applications to be processed. Courts have generally been sympathetic to the use of criteria for
consistency sake. In Bato Star the allegation of rigidity was dismissed in favour of fairness and
consistency. In Scenematic Fourteen it was seen as ‘objective, rational and practical’.

(ii) Fettering by contract

55
This is a theory problem as there is a clash between the power of contracts to bind parties and the
principle that public bodies should not fetter their freedom of discretion. At common law contracts
which purport to fetter future statutory powers may be unenforceable.

The principle from Rederiaktiebolaget ‘Amphirite’ that a government cannot fetter action which must be
determined by the needs of the community has been upheld in a few South African cases. In reality
however, government entered into thousands of these contracts. A test to distinguish between
governmental and commercial contracts has ultimately failed. A better test is that contracts which are
incompatible with the power which has been conferred are void. The incompatibility test is English but
used by the South African courts along with a number of factors to balance public and private interests.

A difficulty comes in compensating the other party of a now void contract. It is argued that normal
contractual remedies are not sufficient and so a special remedy should be read into the PAJA under
section 8(1)(a)(ii) which allows for any order which is just and equitable.

(iii) Promises and assurances

It is dangerous to rely on promises as to future actions. Courts have generally been unwilling to hold
administrators to their representation by means of estoppel. Estoppel may subvert the requirement of
authority as they could arrogate powers to themselves. There is however procedural protection in the
form of legitimate expectations which requires an administrator not to disappoint where a person has a
reasonable expectation of future conduct.

In English law this doctrine protected expectations substantively well which offers freedom from the
conceptual language of estoppel. However English courts take the problem of fettering less seriously.

(e) Arbitrary and capricious decision making

This is provided for in section 6(2)(e)(vi) of the PAJA but does not define these terms. At common law it
is said to mean irrational or senseless without foundation for the apparent purpose. In case law it is
seldom the sole ground of review as it overlaps with others such as bad faith and irrationality.

56
 Chapter 6 -Reasonableness

6.3 REASONABLENESS IN SECTION 33(1) AND THE PAJA

(a) The elements of ‘reasonable’ administrative action

No defined meaning in law however there are recognised elements, rationality and proportionality.

(i) Rationality

Decisions must be based on information/evidence with reasons (objectively capable to meet the purpose
which the power was given) i.e. a reasonable person would have made the decision. There must be a
‘rational objective basis’ [Carephone v Marcus] which is reflected in the PAJA.

S6(2)(f)(ii): (replaced S24(d) interim constitution) decisions must be rationally connected to:
(aa) purpose for which it is taken
(bb) purpose of the empowering provision
(cc) information before the administrator
(dd) reasons given by the administrator.

S6 extends S24(d) [Trinity Broadcasting][Total Computer Services v Municipal Manager, Potchefstroom

Local Municipality] [National Horseracing Authority of SA v Naidoo].

57
S6 is thorough however it requires no perfect rationality only a rational connection.

Can rationality be seen covered within reasonableness? When viewed together they may become the
same enquiry [Calibre Clinical Consultants v National Bargaining Council for the Road Freight Industry].
But they should be distinct as the PAJA separates them and reasonableness goes beyond rationality.

(ii) Proportionality

This entails balancing adverse and beneficial effects of an action, and the possible need for less dramatic/
oppressive steps to be taken to achieve a goal. Essential elements are balance, necessity and suitability
which usually refers to lawful means. Do not use a sledge hammer to crack a nut.

It is integral in SA today and comes from German law in SALRC draft bill [Kruse v Johnson][Roman v
Williams].

It is reflected in S6(2)(h) of PAJA which is a ground dealing with unreasonable effects however not
proportionality itself. S6(2)(h) doesn’t further define reasonableness but rather it is circular linking the
unreasonable action to the unreasonable actor. The ground is similar to the test in Wednesbury which
states a decision is ‘so unreasonable that no reasonable authority could ever have come to it’. It really
serves as a last resort safety net as some other ground is likely to be present in cases of such egregious
unreasonableness.

(b) The Interpretation of section 6(2)(h) of the PAJA

‘Lawful, reasonable & procedurally fair administrative action’ in terms of S33(1) makes it clear that there
is no need for the requirement of gross unreasonableness. The standard is incompatible with the Bill of
Rights [Reynolds].

Dangers of reviewing reasonableness:

(a) It is burdensome on courts

58
(b) It allows judges to substitute their own policy preferences for the administrators. In a democracy
respect must be given to the other functionaries and their respective functions (balance judicial
supervision and the others mandate).

‘Reasonable’ provides a diverse range of possible choices and so does not have to be a perfect or correct
one in courts eyes. To require less would allow for capricious decisions and more would allow a court to
substitute their own opinion. This is reflected in socio-economic rights [Government of RSA v Grootboom].

The Act does not define at what level unreasonable must exist. It could be argued that a decision must be
so utterly and completely unreasonable that no reasonable person could be expected to have come to
that conclusion [Wednesbury] i.e the standard of ‘perversity [Trinity Broadcasting v Independent
Communications Authority SA]. Arguments have been made that this will be too strict and will lack
application.

It could be argued it simply means unreasonable which means no reasonable person/authority would
have taken the decision. This does not mean that the decision was grossly unreasonable.

In Bato Star Fishing v Min of Environmental Affairs it was held that S6(2)(h) must comply with S33
Constitution so one need only show that a reasonable person/authority would not have made the same
choice depending on the circumstances and not and exaggerated version of it.

Certain factors (see pg249 of Hoexter) must be taken into account in order to frame reasonableness,
confirm its variability and set out a basis for the enquiry. The factors give scope to proportionality, by
inviting it either explicitly or indirectly, which would otherwise be missing from section 6 of the PAJA.

In Ehrlich v Min of Correctional Services S6(2)(h) was divided into 3 types of unreasonable decisions
(i) Characterised by a defect in decision-making process (irrationality)
(ii) Violating the principles governing public power (statutory/constitutional imperatives)
(iii) Unnecessarily disproportionate/onerous in effect.

(c) The distinction between review and appeal

59
Courts must not usurp on administrative functions, they must only review and not slide into an appeal
situation [Bato Star Fishing][Trinity Broadcasting].

Can review and appeal be distinguished in these cases? Yes, only scrutiny of the merits may overlap. It is
impossible to judge if a decision is within reason or ‘defensible’ without reviewing the information or
which factor was given more weight. Only in narrow technical cases will there be no scrutiny for example
failing to comply with a mandatory formality.

Courts inevitably look at the merits of the decision in order to determine rationality, review must
encompass the decision itself. Review can be seen as a process and as a remedy. In the process of review
judicial scrutiny is harmless provided it isn’t overzealous. Review for reasonableness doesn’t demand
perfection (or the courts idea of perfection) so it cannot be used as an opportunity for judges to use the
merits for judges to substitute their opinion for that of the administrator [Carephone].

6.4 REASONABLENESS AND RESPECT

Judicial restraint and deference is needed in reasonableness as judges are placed out of their
constitutional & institutional competence.

Restraint must be given by the court where complex and technical issues must be weighed up, where the
court has no expertise [Logbro Properties v Bedderson][Minister of Environmental Affairs and Tourism v
Phambili Fisherie][Bato Star Fishing v Environmental Affairs].

Two points made about deference by O’Regan J:

(1) Deference/respect is part of public law and is not extraneous to the law. It derives from the
constitutional principle of the separation of powers.

(2) Deference doesn’t prevent setting aside unreasonable decisions completely, a trump card of review. A
court must not merely rubber stamp a decision it does not understand.

60
Variability in reasonableness exists, it is context specific [Bato Star][New Clicks]. Bato Star factors give
room for variation and an equilibrium must be struck between all of them.

6.5 REASONABLENESS AND THE RULE OF LAW

The principle of legality is increasingly utilized in administrative law, as it is general and to all public
power. The principle and rule of law may encompass many broader concepts within the rule of law.

Original legislation cannot be vague and must be clear and accessible. This rule against vagueness is
traditionally part of reasonableness in administrative actions [President of RSA v Hugo][Dawood v Min of
Home Affairs][Investigating Directorate: Serious Economic Offenses v Hyndai Motor Distributors].

Delegated legislation also relies on the rule of law due to the uncertainty about delegated legislation. It
must be clear and accessible so that the public does not have to rely on lawyers to interpret legislation
[Affordable Medicines Trust v Minister of Health][Kruger v President of RSA].

The rule of law holds that rationality is a minimum requirement for the exercise of public power
[Pharmaceutical Manufacturers Association of SA: ex parte President RSA]. There must be a relation
between exercise of public power & its purpose. This may have been extended to mimic a S6(2)( f)(ii)
enquiry [SA Predator Breeders Association v Minister of Environmental Affairs]. Objectively all acts must
have a valid nexus between means and end sought to be achieved.

Proportionality has not yet found its way in judgement as a part of legality; Sachs suggests in a minority
it should [New Clicks].

Rationality may require procedural fairness where appropriate [Albutt v Centre for the Study of Violence
and Reconciliation]. Giving reasons may soon also fall under procedural fairness [Wessels v Minister for
Justice and Constitutional Development].

61
Chapter 7 – Procedural Fairness

AUDI ALTEREM PARTEM

7.1 INTRODUCTION

Procedural fairness and natural justice are interchangeable concepts. Procedural fairness is based on
two principles which have been developed by PAJA and case law:

1) Audi alteram partem: the right to be heard by both sides

2) Nemo Iudex in sua casus: there should be no bias on the part of the presiding officer.

PAJA distinguishes between the right to be heard before and after the administrative action is taken.
Section 3 of PAJA relates to audi alterum partem concerning 1 person: in other words it relates to the
rights of an individual person affected by a decision to be heard. Section 4 of PAJA relates to audi
alteram partem concerning the public: it thus related to the right of the public to be heard where a
decision is heard that affects them.

Procedural fairness is sub-divided into two separate grounds of review in PAJA which deal with different
things. As discussed above on ground relates to bias and the other is a general ground of procedural
unfairness encompassing the right to be heard within it:

 Section 6(2)(a)(iii) states that: a court or tribunal has the power to judicially review an
administrative action if the administrator who took it was biased or reasonably suspected of
bias.
 Section 6(2)(c) states that: a court or tribunal has the power to judicially review an
administrative action if the action was procedurally unfair.

Having said that, S6 provides the ground of review however S3 and S4 of PAJA give content to the
standards of procedural fairness, which if not met allow for the grounds of review to kick in.S33 (1) of

62
the Constitution (the right to just administrative action): provides that everyone has the right to
administrative action that is lawful, reasonable and procedurally fair.

7.2 SECTION 3 AND THE VARIABLE CONTENT OF FAIRNESS

(a) Section 3 in outline

S3 of PAJA deals with ‘Procedurally fair administrative action affecting persons’ and provides that:

(1) Administrative action which materially and adversely affects the rights or legitimate
expectation of any person must be procedurally fair.
(2) (a) A fair administrative procedure depends on the circumstances of each case.

(2)(b) In order to give effect to the right to procedurally fair administrative action, an
administrator subject to subsection (4) must give a person referred to in subsection(1): (These
are the minimum requirements for fairness):

These first two relate to pre-administrative action, in other words it relates to the standard of
procedural fairness that must be met before administrative action is taken:

(i) Adequate notice of the nature and purpose of the proposed administrative action
(ii) A reasonable opportunity to make representations

These last three relate to post-administrative action, in other words it related to the requirements for
procedural fairness after administrative action has been taken

(iii) A clear statement of the administrative action

(iv) Adequate notice of any right of review or internal appeal; and
(v) Adequate notice of the right to request reasons in terms of S5.

In order to give effect to the right to procedurally fair administrative action, an administrator may, in his
or her or its discretion also give a person referred to in subsection (1) an opportunity to: These are the
discretionary requirements for fairness:

(a) Obtain assistance and, in serious or complex cases , legal representation;

(b) Present and dispute information and arguments; and

63
 (c) Appear in person

SS4 and SS5 allow for exceptions to the guideline of procedural fairness mentioned above.

(3) (a) If it is reasonable and justifiable in the circumstances, an administrator may depart from
any of the requirements referred to in subsection (2).

(4)(b) In determining whether a departure as contemplated in paragraph (a) is reasonable and

justifiable, an administrator must take into account all the relevant factors, including:

 The objects of the empowering provision

 The nature and purpose of, and the need to take the administrative action
 The likely effect of administrative action
 The urgency of the administrative action or the urgency of the matter; and
 The need to promote an efficient administration and good governance

(4) Where an administrator is empowered by an empowering provision to follow a procedure

which is fair but different from the provisions of subsection (2), the administrator may act in
accordance with that different procedure.

S4 of PAJA deals with administrative action affecting the public (which is not relevant for the purposes of
the exam and will not be discussed any further).

A discussion on the subsections of S3 is required to understand how it gives content to fairness.

“S3(2)(a) a fair administrative procedure depends on the circumstances of each case”. In determining
whether the fairness requirement has been, it truly depends on the circumstances of the case. The
content of fairness varies amongst cases [Metro Projects v Klerksdorp].

(b) Section 3(2): The minimum requirements

“S3(2)(b) In order to give effect to the right to procedurally fair administrative action, an administrator
subject to subsection (4) must give a person referred to in subsection(1):(These are the minimum
requirements for fairness):

These first two relate to pre-administrative action, in other words it relates to the standard of
procedural fairness that must be met before administrative action is taken:

64
 (i) Adequate notice of the nature and purpose of the proposed administrative action
(ii) A reasonable opportunity to make representations

These last three relate to post-administrative action, in other words it related to the requirements for
procedural fairness after administrative action has been taken.

(iii) A clear statement of the administrative action

(iv) Adequate notice of any right of review or internal appeal, where applicable ; and
(v) Adequate notice of the right to request reasons in terms of S5”

In the Joseph v City of Johannesburg case it was held that not all 5 of the guidelines in S3(2)(b) need to
be taken into account in determining whether or not there has been procedural fairness. This is because
S3(2)(a) limits the application of the minimum requirement of procedural fairness to the individual
circumstances of the case. Thus during the process of administrative action, the administrator must
abide by these 5 minimum standards, but when under judicial review, the courts have the discretion as
to which standards they want to enforce in determining whether procedural fairness is present. This is
known as the variability of fairness.

Thus the absence of a requirement does not automatically invalidate the action or render it procedurally
unfair. Accordingly the grounds in s3(2)(a) point towards but cannot be equated with procedural
fairness.

The requirements of fairness stated in PAJA must be read in conjunction with any empowering provision
which also stipulates its own requirements relating to fairness.

(i) Adequate notice of the nature and purpose of the proposed administrative action

The word ‘adequate’ is not defined. However it has been interpreted to mean that sufficient information
should be given to the affected party so that are enabled to exercises their rights. This should be looked
at in relation to the nature and purpose of the administrative action.

In Cape Metro it was noted that the following circumstances could vary the standard of fairness
required in giving adequate notice as well as the nature of the administrative action:

 Practical considerations
 Financial considerations

65
  The cost implications
 The position of the complainant

The Permanent Secretary, Department of Welfare Eastern Cape case dealt with government wanting to
review disability grants. They put this proposal in a general notices such as in pamphlets, radio-
broadcasts and printed notices. The court found that individual notice should have been given in
individual pay-packets. It was deemed that adequate notice was not given and thus the administrative
action was procedurally unfair.

Mahamba dealt with the eviction of occupiers of a particular land. When such administrative action is
taken, the affected persons need to be given adequate notice and an opportunity to be heard. The
problem was that the affected persons were illiterate and did not speak English. Here it was held that
adequate notice would have comprised giving out the normal notice as well as giving notice through a
loud speaker in a language that the inhabitants understood and at a time where they were likely to hear
the notice.

At common law the time and place to make representations should be included in such notice as well as
well as the nature and purpose of the notice.

In the case of Popcru, employees were given 48 hours-notice of disciplinary proceedings. The court
found that this was unfair as the information provided was not sufficient to know and meet the charge

In the case of Nkomo illiterate workers were given 48hours to make written representations regarding
their proposed dismissals. The court found that this requirement was not met and thus the
administrative action was not procedurally fair. This links closely to the idea that notice must be given
timeously so as to enable the persons to exercise their rights.

The giving of such notice is not required where to give notice would defeat the purpose for which the
power was given.

(ii) A reasonable opportunity to make representations

The purpose of making these representations is to give the interested party an opportunity to affect the
discretion of the decision-maker.

66
The ‘hearing’ by which representations by the affected person is made does not have to be an oral one
as per the PAJA or the common law. The form of the hearing is at the discretion of the administrator and
this he has the discretion to grant an oral hearing. Generally written representations are more cost and
time effective. However concessions will be made where the persons affected are illiterate or do not
speak English, it would be appropriate for the administrator to grant an oral hearing in a language that
they understand.

The type of information that should be given in order to satisfy this section is not specified, however it is
implied to mean that the information should be sufficient to be used for:

 Making a proper representation in response

 Understanding the reasons underlying why the administrator is making the decision (and
not the reasons for the decision itself)
 To inform the affected person of the substance of the case

Relevant here are tender cases: this is because tendering generally has its own rules. Issues generally
arise because reports are filled out by referees and it is debated whether those reports should be made
available to the tenderer if they are adverse. The courts have been unclear in answering this question:

 In Thabo Mogudi Security the court said that at the very least, policy considerations would have
to be disclosed so as to allow the affected person to try and satisfy the policy considerations.
 Also generally the affected person should be given an opportunity to respond to any adverse
information which materially affects the administrator’s decision.

It also needs to be established who must hear the representations. The case of Earthlife related to a
challenge of a decision to construct a nuclear re-actor. Here it was decided that the decision maker
himself needn’t hear the representation himself and may rely on the opinion of an expert. The reason
being that the representations were of a highly technical and scientific nature. The administrator can
rely on the conclusions of another so long as he does not defer to the opinion of the expert/another
without considering the matter himself and making his own decision at the end of the day. Reliance on
an expert is not ideal but it would be impractical not to make such provision else the administration
would be inundated. It is especially acceptable where legislation allows it and the credibility of witnesses
relies on it

67
The affected person must be aware that the purpose of a specific occasion is to make representations
which could affect the outcome of the decision.

(iii) A clear statement of the administrative action

When giving a clear statement of administrative action, this can but does not have to include the giving
of reasons. This includes what the decision was, when it was taken, by whom it was taken and on what
legal and factual basis the decision was made.

It must include the factual and legal basis for the administrative action but the administrator need not
have to state how the legal and factual basis was applied. The factual and legal basis for the decision
may to some extent aid in determining the reason for the decision.

(iv) Adequate notice of any right of review or internal appeal, where applicable

Primary or original legislation can make provision for an internal review or internal appeal for
administrative action, and the person who is affected by such administrative action must be made aware
of such internal appeal or review. This is because all internal remedies must be exhausted before a court
can be approached to review administrative action.

However it should be noted that the administrator is not under an obligation to inform a person who is
affected by the administrative action of the right to review administrative action through the courts as
this would place an overly onerous duty on the administrator.

(v) Adequate notice of the right to request reasons in terms of S5 of the PAJA

At common law there was no right to request reasons for administrative action unless a contractual or
legislative duty existed to do so. S33(2) of the constitution states that: ‘Everyone whose rights have been
adversely affected by administrative action has the right to be given written reasons’

S3 (2)(b)(v) of the PAJA triggers S5 of the PAJA. S5 of the PAJA relates to the right to request reasons (if
not already given)after such request the administrator is compelled to provide such reasons within 90
days unless he is exempt from giving reasons in terms of the PAJA itself.

S5 of PAJA only poses an obligation to give reasons where rights have been materially and adversely
affected by administrative action. The giving of reasons help to prevent the abuse of discretion on the
part of the administrator.

68
A situation may occur where reasons have already been given within the clear statement of
administrative action. In such instances there will be no right to request reasons as they have already
been provided.

(c) Section 3(3): The discretionary ingredients

(ii) Legal representation and other assistance

For example in certain instances you have the right to have family support or a translator present.
Whether or not to allow legal representation is generally at the discretion of the administrator.The
administrator should take into account the complexity of the case and the seriousness of its
consequences.

It may be possible that statute may require that an individual be allowed legal representation in certain
instances.

In the Hamata Case it was held that where the presence of legal representation is warranted, it is a
constitutional imperative that it is allowed, despite the administrator’s discretion. For this reason
administrators generally err on the side of caution and allow legal representation. Any blanket rule
prohibiting legal representation regardless of the circumstances is unlawful.

It should be noted that the restriction placed on having legal representation exists because the
introduction of legal representation is costly and time consuming.

(ii) An opportunity to present and dispute information and arguments

There is a mandatory right to make representations, however the right to dispute and challenge
information is discretionary and thus depends on whether fairness in the circumstances warrants it.

In the case of Earthlife the decision to construct a nuclear reactor was challenged. Eskom appointed it
consultants who made representations. Due to the materiality of these representations it was held that
there exists a right to rebut the representations made by other parties. Thus the parties were allowed to
make representations again.

Accordingly, the right to reply or rebut information after representations have been made depends on
the seriousness of the case and the materiality of the new information. Where new submissions are mad
based on old facts, there will generally be no right to reply or rebut.

69
(iii) Personal appearance

The right to appear in person generally implies an oral hearing in which both sides must be present
through-out. It is important to bear in mind that this is not a court hearing and thus the proceedings can
be conducted in any way that is fair given the circumstances.

Thus there is no absolute right at common law or in the PAJA to appear in person. It will ultimately boil
down to the administrator’s discretion given the circumstances. A very serious issue may demand a
formal oral hearing.

(d) S3 (4): Departure from the requirements in section 3(2)

S3(4)(a): If it is reasonable and justifiable in the circumstances, an administrator may depart from any of
the requirements referred to in subsection (2).

S3(4)(b) In determining whether a departure as contemplated in paragraph (a) is reasonable and

justifiable, an administrator must take into account all the relevant factors, including:

 The objects of the empowering provision

 The nature and purpose of, and the need to take the administrative action
 The likely effect of administrative action
 The urgency of the administrative action or the urgency of the matter; and
 The need to promote an efficient administration and good governance”

According to Joseph one can depart form the procedures listed in S3(2)(b) by invoking S3(4). However
even if S3(4) is not invoked the procedure will not necessarily be unfair as the courts still have the
discretion to allow non-compliance with those procedures and still have the administrative action be
procedurally fair as per S3(2)(a).

Individual consideration and individual departure is more favourable than blanket exemptions, and
permission to vary

(e) S3 (5): ‘Fair but different’

“Where an administrator is empowered by an empowering provision to follow a procedure which is fair

but different from the provisions of subsection (2), the administrator may act in accordance with that
different procedure.”

70
If a specific procedure is stated in another empowering provision, the administrator will have a choice
between observing the procedures either in PAJA or the alternate provision. This choice is however
qualified in that if the alternate provision offers less protection to the person affected by the
administrative action then the following of that procedure will be unfair.

Nortjie: as subsequent hearing is generally as advantageous as a hearing beforehand and thus a hearing
after the decision has been made are generally the exception and not the rule. However a hearing after
the decision has been taken may be fair so long as it does not result in prejudice.

(f) Non-compliance with section 3(2)

At common law non-compliance with the minimum/mandatory requirements results in invalidity of the
administrative action. Currently the position is still debated.

According to Klaaren and Currie, if the first mandatory requirements are not complied with then this will
result in invalidity because these are aspects of procedural fairness, however, where the last 3
mandatory requirements are not complied with, this will not necessarily result in invalidity as these are
merely statutory formalities. Failure to comply with the last 3 can be challenged instead under
lawfulness and the failure to comply with a mandatory provision.

It should be noted that as discussed above as per the decision in Joseph, a court may condone non-
compliance with minimum requirements, where however it does not, this will likely result in invalidity.
This is a strict approach in which variability of fairness does not play an imperative role.

According to Hoexter, an all or nothing approach is unsatisfactory. Thus whether or non-compliance will
result in invalidity will depend purely on the what is fair within the circumstances of the specific case.

A pragmatic approach requires looking at the effect of the failure to comply with the minimum
requirements on the outcome of the case.

7.3 S3 (1) AND THE APPLICATION OF FAIRNESS

(a) The pre-democratic era and the classification of functions

The audi alteram rule applied unless statue precluded it. Here one was only entitled to fairness where
existing rights were affected. Therefore where it came to administrative action affecting future rights, or
interests, the affected individual was not entitled to procedural fairness. Furthermore the existing rights

71
which were guaranteed procedural fairness when affected by administrative action were usually limited
only to private law rights and not public rights.

The right to be heard only applied to quasi-judicial decisions (involving rights to property and liberty)
and not legislative or purely administrative decisions (relating to mere applicants). This approach was
obviously the result of applying the deprivation theory.

Natural justice was applied in an all or nothing manner which leads to several injustices.

(c) The development of the legitimate expectation doctrine

The position before Traub was such that the focus was on a distinction between quasi-judicial functions
and judicial functions. Accordingly the courts could only perform a judicial function by making
determinations as to existing rights. Traub noted that this artificial distinction results in unfairness.

This case changed the position in the pre-democratic era by introducing the concept of legitimate
expectation.

The facts of Traub are as follows: this case dealt with doctors in a provincial hospital. New doctors went
through the stipulated hospital practice and were recommended for particular positions. It was past
practice for many years that those recommended would be appointed to the positions. However these
doctors were not appointed to those positions because they signed a public letter which was critical of
the public administration. The court found that the doctors did in facthave a legitimate expectation of
being appointed. Because we are dealing with review, the courts cannot award a substantive remedy.
Therefore instead of the court ordering that the doctors be appointed, instead they ordered that the
appointments be set aside and that the process start again. The court went on further to say that where
past practice is departed from; the persons who will be affected by such deviation have a right to be
hard before such departure.

The court stated that there is a need balance the protection for individuals against unfair decisions
against allowing judicial interference. The seriousness of the consequence can be determinative of
whether legitimate expectation exists.

Under this approach, one was entitled to procedural fairness where his legitimate expectation was
affected by the administrative action.

72
The legitimate expectation of the affected person relates to a legitimate expectation of a particular
outcome or merely the legitimate expectation to be heard before a decision is taken.

There are two circumstances in which a legitimate expectation can exist:

1) There was an express promise of a particular outcome made to the affected person by (or
on behalf of) someone who has the authority to make such promise.
2) Past regular practice (that is not an isolated incident) which the affected person could
reasonably have expected to continue.

It was held that there are 3 types of procedural legitimate expectation:

a) The expectation to be hard

b) The expectation of prior consultation
c) The expectation to make representations

There are 3 instances in which rights are adversely affected:

a) An existing right has been taken away

b) There was a failure to grant a right which a person is entitled to
c) Legitimate expectation

There must be fairness in the procedure and not the outcome and the principles of natural justice must
be observed.

The SARFU case held that a legitimate expectation depends in the circumstances of the case. However
the qualification is that the legitimate expectation must exist objectively and not just in the mind of
affected person. Furthermore it equated the question of whether a legitimate expectation of a hearing
exists with whether the duty to act fairly would require a hearing in those circumstances.

In the Phillips Case it was held the objective test should not be applied rigidly. The court went on further
to list 4 elements of a legitimate expectation:

1) There must be an objectively existent reasonable expectation

2) The reasonable expectation must be induced by the decision maker
3) The representation by the decision maker must be in the form of either:
 A clear promise

73
  A regular past practice

The decision maker must have been competent to make such representation lawfully. In other words
the decision maker must not be acting ultra-vires when making such promise/conducting such past
practice

The case of Duncan addresses the issue of what is ‘reasonable’. The Duncan Case related to fishing
quotas. Duncan already had a medium term licence to line-fish however he now wanted a long term
licence. One of the requirements for a long term licence was that the person should have the necessary
vessel. Duncan’s vessel did not meet their requirements as it was unstable and thus he was not granted
a long term license. Duncan alleged that he was given an undertaking by the administrator that this
requirement would not be enforced strictly and that he thus had a legitimate expectation of being
granted a long term license. The court found that Duncan’s expectation existed only subjectively and
thus there was no existing legitimate expectation. Furthermore it held that unlike in English law, South
African law does not allow enforcement of a substantive legitimate expectation. Mandatory
requirements were complied with in this case.

The case of Cluade Neon addressed the meaning of express promise. Here an official of the Germiston
municipal council told a potential applicant for a tender that he would inform him as to the date that
applications for tenders were due. The official failed to do so and the potential applicant never
submitted a tender. The court held that the potential applicant had a legitimate expectation that he
would be informed when applications for tenders opened. The award of the tender was set aside and
the process started again.

Premier Mpumamalanga decided that legitimate interests and rights were worthy of procedural fairness
but not mere interests.

The case of Batleur Books addressed the issue of past practice. The court said that there is no fixed
period of time in which something becomes a practice as it depends on the circumstances of the case.
However it must be determined whether the incident is isolated or whether there have been several
similar events. Therefore we don’t just look at the period of time but also, how many times the event
occurred. We look to see if the past practice gave rise to future practice. In this case there was a
purchasing system for books which had been in place for 2 years.

74
The case of Public Service Association dealt with the issue of variation of past practice. In this case
magistrates (who were white men) wanted to be promoted to the Regional Magistrates Court. The
Magistrates Court Act stipulated that a magistrate in the District Magistrates court could apply for a
position in the Regional Magistrates Court after serving certain amount of years. However at this stage a
policy change in the form of Affirmative Action occurred. The new point system basically excluded white
men from being appointed as Regional Court magistrates. The court held that the magistrates had a
legitimate expectation to be heard before the policy change was implemented.

The problem lies in that the courts have sometimes applied the legitimate expectation doctrine where
there is no express promise or past practice. For example in Nortjeeit was held that prisoners had a
legitimate expectation to be heard before being transferred to a maximum security part of the prison.
The court’s rationale was that the prisoners would suffer prejudice if no legitimate expectation to be
hard existed because they would experience a decrease in privileges. Hoexter criticises this decision in
that no grounds existed for their finding of a legitimate expectation. Similarly in SARFU it was held that
whether or not a hearing is required depends merely on whether or not fairness demands it. However
where conduct does not amount to administrative action, there is generally no obligation to act
procedurally fairly.

(e) Section 3(1) of the PAJA

There are three requirements for procedural fairness under PAJA to be triggered:

1) The exists a right/ legitimate expectation

2) Which is adversely affected (by the administrative action)
3) And is materially affected (by the administrative action)

Materiality and adverse effect are seen in conjunction by the courts. Therefore the presence of one
implies the other. This is because, according to Kiva and Joseph, rights cannot be adversely affected if
the effect is not also material.

Originally the test was that the adverse effect cannot be trivial which was a very easy test to satisfy.
Later the test changed to that the adverse effect must be significant which became more difficult to
satisfy.

(i) Rights, determination and deprivation

75
The next issue is that of rights. Relevant here are the cases of Walele and Joseph in that they are both
very extreme on the opposite ends of the spectrum as to what constitutes a right.

Walele’s interpretation of what constitutes a right is criticised for being narrow and taking the definition
back to the pre-democratic era. Here it was held that in order for a right to be adversely and materially
affected, there must be a pre-existing vested right which is affected. The background of this case is that
W was denied a hearing before his neighbour’s building plans were approved despite the fact that the
approval affected his peculiar interest (his property would become devalued). The court reasoned that
his property rights were not affected by the approval nor could he prove that the approval would reduce
the value of his property. This seems to indicate support for the deprivation theory and a pre-
democratic quasi-judicial approach. Furthermore this case is an example of an isolated incident as
opposed to an official practice. The court also noted that for a legitimate expectation, there must exist
the capacity to affect legal rights. It also noted that natural justice must be observed when fulfilling a
legislative function. However here there was no legitimate expectation requiring it.

(ii) A public-law right

Joseph moved away from the concept of vested rights/private law rights. In this case the electricity of a
building was cut off. The tenants had paid their landlord but the landlord failed to pay the municipality.
Despite the argument by the municipality that they had no contractual relationship with the tenants, the
court held the tenant’s public law right was affected. This right derived from the ‘Municipal Systems Act’
giving the municipality as a branch of government) a general duty to provide core socio-economic
essentials. The tenants should have been afforded a hearing before their public law right was adversely
affected. Accordingly it is possible for a statutory and general constitutional duty to trigger procedural
fairness. Thus the tenants had a right to be heard before their electricity was cut off. The court noted
that procedural fairness exists in order to allow people to influence decisions which affect them. The
purpose of which is openness, transparency and accountability.

Hoexter criticises this decision for being too broad and having a very flimsy foundational base. This
flimsy foundation makes it difficult for public law rights to be developed in the future. She also criticises
the court for relying on statute and not the Bill of Rights. Furthermore she argues that the constitutional
court needs to limit the scope of this case as it throws the scope of what constitutes a right too wide.

As a side note: affording procedural fairness to interests is precluded from S3 and S4. And S4 precludes
legitimate expectations as well.

76
 (iii) The crucial role of variability

Hoexter argues that we should not be trying to impose additional hurdles to get to procedural fairness.
Instead, we should have a universal application of procedural fairness with a variable content. Fairness
should depend on the circumstances of the case and not whether it fits into certain boxes such as
materiality.

This approach is advantageous in that it is more practical and philosophical and less conceptual. Thus
was we do not apply fairness in an all or nothing manner. We need not try to compartmentalise rights
and expectations.

PAJA seems to mix a philosophy of variability within a procedural framework. The mandatory and
discretional requirements in PAJA should be viewed as giving content to fairness and not restricting it.

7.5FAIRNESS UNDER THE PRINCIPLE OF LEGALITY

Legality comes into play where there is the exercise of public power which does not amount to
administrative action. Legality was an aspect of the rule of law until the democratic era where it has now
become a founding value of the constitution. It requires lawful and rational conduct which encompasses
fairness and the obligation not to act arbitrarily.

The exercise of public power must be exercised:

a) As per the principle of legality itself:

 Rationally
 In good faith
 In accordance with the principle of legality

b) As per S237 of the constitution:

 Diligently
 Without delay

The relevant case here is that of Albutt. In this case the courts blankly ignored PAJA and went straight
onto legality and is thus criticised on this basis. Despite this error, the court made some useful findings.
On the facts, the court found that before granting parole to political offenders, their victims should be
heard. This will ensure that there is a rational purpose for the pardon and the purpose for which the

77
power to pardon was given. It found that the principle of legality was broad enough to encompass the
right to a hearing and thus it can be concluded that legality incorporates procedural fairness.
Furthermore there needs to be a rational connection between the exercise of power and objectives of
the pardoning regime, there would be no rational objective if the perpetrator’s motives weren’t
established and relevant facts established by listening to both sides.

In Albutt procedural fairness was needed to fulfil rationality, it is likely that in all instances the giving of
reasons and importing procedural fairness will enhance rationality.

WHAT TRIGGERS FAIRNESS UNDER THE CONSTITUTION

Under the constitution, so long as it is administrative action, the affected person is entitled to fairness.
Everyone has the right to administrative action which procedurally fair; this can only be limited generally
and not individually.

PROCEDURAL FAIRNESS: SOME CENTRAL THEMES

7.6 LEGITIMATE EXPECTATIONS

(a) Legitimate expectation in the Constitutional era

Legitimate expectations can derive from:

 An express assurance
 A settled practice
 Established policy: still need to consider white men for appointment despite affirmative
action policy because they had a legitimate expectation of employment

In some limited circumstances substantive remedies have been awarded however procedural remedies
still seems to be the norm. For example in Premier Mpumalanga: a promise was made that a book
bursary scheme would continue for another year unless reasonable notice of termination is given. The
scheme was terminated with retro-effect and the court ordered a date at which the bursaries were
payable.

Despite legitimate expectation being legislated into PAJA, the old cases are still used to inform the
interpretation of PAJA.

78
A legitimate expectation cannot arise from confusion or misinterpretation.

SARFU and Nortjie seem to be moving towards a loose approach of legitimate expectation which merely
entails that someone be heard as a matter of procedural fairness if they will be prejudiced otherwise.

(b) Procedural and substantive legitimate expectation

It is unsettled whether once can only have a substantive legitimate expectation or a procedural one too.
It appears that one can have both a legitimate expectation of a specific substantive outcome and a
legitimate expectation of procedural fairness itself.

NEMO IUDEX IN SUA CAUSA

7.9 THE RULE AGAINST BIAS

As mentioned above, procedural fairness is divided into audi alteram partem “the right to be heard” and
Nemo Iudex in sua causa “bias”. Most of the focus up until this point was on audi alteram partem. We
will now discuss the second element of procedural fairness, namely nemo iudex in sua causa.

Nemo iudex in sua causa generally means that you will not have an interest in the matter. In other
words, it is the protection offered from bias. It is a ground of review in both the common law and in
PAJA.

Bias equates to impartiality which would then influence the proceedings. The reason for protection
against bias can be found in the maxim: “justice must be done and be seen to be done”

Even in the presence of bias, the right decision may be reached. However, in order to promote faith and
transparency in our administration, both the decision and the decision making process should appear
unbiased. Furthermore, the absence of bias increases the likelihood that a fair decision will be reached.

The most common context in which bias plays a prominent role is in judicial proceedings and quasi-
judicial proceedings.

 Judicial proceedings equate to actual court proceedings, such as litigation and criminal
trials
 Quasi-judicial proceedings are those which resemble actual court proceedings but which
are not, for example disciplinary hearings.

79
This ground of review may easily overlap with other grounds of review such as failure to apply the mind.

(a) The test for bias

Bias as a ground of review originates from common law. This is in line with the idea that PAJA is a
codification of the common law.

Under the common law and originating from the case of City of Suburban Transport two tests were
developed, each listing an instance in which administrative action could be reviewed, namely:

1) ‘The real likelihood of bias test’ which is very similar to actual bias and thus more difficult to
prove
2) ‘The reasonable suspicion of bias test’ required an individual merely to prove an appearance
of partiality as opposed to its actual existence.

Over time an uncertainty developed as to which test should be applied. In BTR Industries the court
resolved this uncertainty. It discarded the ‘real likelihood of bias test’ in favour of the ‘reasonable
suspicion of bias test’. The court however left some issues unresolved.

These issues were resolved in the State v Roberts Case. Here the court summarised the test for the
presence of reasonable suspicion of bias into 4 elements, namely:

1) There is a suspicion that the judicial officer might be biased

2) This suspicion must be of a reasonable person in the position of the accused or the litigant
3) The suspicion is based on reasonable grounds
4) The suspicion is something which a reasonable person would have.

The test mentioned above is essentially a reasonableness test. There is a need for double
reasonableness in the sense that that the person who alleges bias must be reasonable and the suspicion
itself must also be reasonable. The reason for the great focus on ‘reasonableness’ was because the
courts were hesitant to make a finding of bias due its negative implications of misconduct and bad
motive.

S6(2)(a)(iii) of the PAJA states that: a court or tribunal has the power to judicially review an
administrative action if the administrator who took it was biased or reasonably suspected of bias.

80
Accordingly, S6(2)(a)(iii) distinguishes between two instances on which administrative action can be
reviewed on the basis of bias:

a) Where there is actual bias

b) Where there is reasonable suspicion of bias

Actual bias is a subjective element resting on the state of mind of the decision maker himself and is thus
more difficult to prove than reasonable suspicion of bias. Whereas reasonable suspicion of bias is an
objective enquiry which entails an enquiry into whether there is a reasonable possibility of bias.

The focus hereon will be on reasonable suspicion of bias.

(b) Sources of bias

Pre-constitutional cases were divided into 4 sources of bias:

a) Personal bias
b) Financial Bias
c) Bias relating to prejudice over the subject matter
d) Institutional/ structural bias

It is important to note that the element of reasonableness is very important as one cannot merely rely
on this ground for any type of subjective suspicion

(i) Financial interest

This relates to a pecuniary interest. Whether or not the pecuniary interest is significant (big) or slight
(small) will impact on whether or not there is a reasonable suspicion of bias.

In the Bernert case the applicant had a cause of action against ABSA. The applicant requested that the
judge recuse himself because he had shares in ABSA. The court stated that the idea could not be
automatically dismissed. In other words, the court must firstly consider whether the pecuniary interest
is significant or slight. And secondly whether there is a realistic possibility that this interest will affect the
outcome of the case. The courts generally find that a small pecuniary interest suffices unless the interest
is ridiculously slight or remote. However these small interests should still be disclosed.

81
In Rose Taxis a Transportation board was formed for the purposes of granting certain exemption
certificates to car-hire service companies. The problem here was that the Chairman of the board was the
director of D taxi company. D was Rose Taxi’s competin terms of rs. Rose Taxi objected to the idea that
the director of their competin terms of r would play a part in deciding whether they would be granted
an exemption.

The case of Bam-Mugwanyaa member of the tender board failed to recuse herself or disclose the fact
that she was also a member of a close corporation that stood to benefit from the tender board’s
decision.

(ii) Personal interest

Personal bias arises from the presence of a family, friendship, enmity or other similar relationship
between people.

An example can be found in the SARFU Case where the applicant requested that several of the judges
recuse themselves. This was because these judges all enjoyed personal and professional relationships
with Nelson Mandela (the respondent). These relationships varied from them sharing membership to
the same political party, to one of the judges having been part of Mandela’s defence counsel during
apartheid. The problem with recusal is that it is up to the judge himself to exercise the discretion as to
whether or not he should recuse himself. Here the judges elected not to recuse themselves. This was
because on their opinion, their personal relationships with the respondent would not impact their
impartiality and independence. Accordingly they found that when judged against the standard of a
reasonable man, there was no reasonable suspicion of bias.

In Liebenberg a board made a decision to grant X a licence. The problem was that X’s brother (the
mayor) sat on that very board when the decision was made. The court noted that regardless of whether
the mayor had an influence on the final decision or not, the mere presence of the mayor sufficed to
raise a reasonable suspicion of bias. This relates to the principle that “justice must be done and be seen
to be done”. The presence of the mayor raises issues as to the whether fair procedure was followed
when granting X his license. The decision was thus set aside.

The case of Bula further exemplifies personal bias. Here Y laid a complaint against X (the head of a
provincial department). X was found to be guilty of misconduct and was dismissed. The problem being
that Y’s brother in law (the Premier of the province) assigned a delegate who then had the authority to

82
dismiss X. The court went on further to say that a delegate when making a decision that will affect the
delegator will almost always raise a suspicion of bias.

(iii) Bias on the subject matter

‘Prejudice’ and ‘bias’ are not interchangeable terms. Prejudice refers to a pre-conceived idea.
Accordingly, bias is much broader.

Proving prejudice regarding the subject matter is difficult. It requires showing that the decision maker
has pre-judged the issue or expressly aligned himself to one side of the argument.

The case of Hamata noted that there is a difference prejudice and having a personal view. It went on to
say that human beings innately hold personal views. Thus a line needs to be drawn between that of
having a disposition or attitude and that of having pre-judged the issue. In essence, an administrator
must approach the decision making process with an open mind to the evidence laid before him and any
exceptions to his dispositions and attitude.

Bias to the subject matter was found in Mofongosi where the same members of the original disciplinary
hearing sat on the appeal of the disciplinary hearing. In essence they were hearing an appeal against
their own decision which creates a reasonable suspicion of bias.

In S v Collier the accused was coloured and the magistrate hearing his criminal trial was white. On this
basis the accused requested that the judge recuse himself. It was held that recusal on this basis would
undermine the idea of diversity as a cornerstone of South African Society. Also it would be absurd to say
that a coloured magistrate could never administer justice fairly in a matter involving a white accused and
vice versa.

(iv) ‘Official’ or ‘institutional’ bias

This relates to a situation where the decision maker is closely aligned to an institution and must then
decide whether someone has transgressed rule against the institution.

For example a principal of a school who hears a charge of damage to property belonging to the school.

In Monnig where supporters of the end conscription campaign chaired on the board deciding whether
members of the defence should be found guilty of spying to discredit the end-conscription campaign.

83
The patriotism of these supporters and their string allegiance to the South African Defence Force was
enough to raise a reasonable suspicion of bias. This was accordingly an example of societal bias.

84