Independent Candidate Association South Africa NPC v President of the Republic of South Africa and Others (CCT 144/23) [2023] ZACC 41 (4 December 2023)

CONSTITUTIONAL COURT OF SOUTH AFRICA



Case CCT 144/23

In the matter between:


INDEPENDENT CANDIDATE ASSOCIATION SOUTH
AFRICA NPC Applicant

and

PRESIDENT OF THE REPUBLIC OF SOUTH AFRICA First Respondent

SPEAKER OF THE NATIONAL ASSEMBLY Second Respondent

CHAIRPERSON OF THE NATIONAL COUNCIL
OF PROVINCES Third Respondent

MINISTER OF HOME AFFAIRS Fourth Respondent

INDEPENDENT ELECTORAL COMMISSION Fifth Respondent

ALL POLITICAL PARTIES REGISTERED FOR
ELECTIONS FOR THE NATIONAL ASSEMBLY Sixth Respondent



Neutral citation: Independent Candidate Association South Africa NPC v President
of the Republic of South Africa and Others [2023] ZACC 41

Coram: Zondo CJ, Maya DCJ, Kollapen J, Mathopo J, Mhlantla J,
Rogers J, Schippers AJ, Theron J and Van Zyl AJ

Judgment: Mhlantla J (unanimous)

Heard on: 29 August 2023

Decided on: 4 December 2023

2

Summary: Electoral Amendment Act 1 of 2023 — provision limiting seats
that independent candidates can contest in the provincial and
national elections does not infringe any of the fundamental rights
in Chapter 2 of the Constitution

Right to stand for public office — Right to vote — Right to free
and fair elections




ORDER



On application for direct access:
1. Direct access is granted.
2. The application is dismissed.
3. There is no order as to costs.



JUDGMENT




MHLANTLA J (Zondo CJ, Maya DCJ, Kollapen J, Mathopo J, Rogers J,
Schippers AJ, Theron J and Van Zyl AJ concurring):


Introduction
[1] The applicant, the Indep endent Candidate Association South Africa NPC, is a
registered non-profit company that represents and promotes the interests of independent
candidates in the electoral system. The applicant has brought an application for direct
access in terms of wh ich it seeks a declarator that i tem 1 of Sc hedule 1A
(impugned schedule) to the Electoral Act 1 is inconsistent with sections 1(c); 3(2)(a);

1 73 of 1998.
MHLANTLA J
3
9(1); 19(2); 19(3) and 46(1)(d) of the Constitution to the extent that the impugned
schedule provides for 200 seats in the National Assembly to be filled by independent
candidates and candidates from regional lists of political parties (called regional seat s)
and 200 seats to be filled by candidates from national lists of political parties
(called compensatory seats). The applicant contends that Parlia ment acted
unconstitutionally by splitting the seats in the National Assembly into 200/200.
This challenge stems from Parliament’s passing of the Electoral Amendment Act 2
which, amongst others, amended item 1 of the impugned schedule to allow for
independent candidates to contest for seats in the National Assembly.

[2] The primary remedy sought by the applicant is a combination of a striking down
and reading-in, which wil l result in the word “half” in i tem 1(a) of the
impugned schedule being changed with immediate effect to “350” and, similarly, the
word “half” in item 1(b) of the impugned schedule being changed t o “50”. The
reading- in will have the effect of amending i tem 1 of the impugned schedule to apply
to the 2024 elections. In the alternative, the applicant seeks a suspension of the
declaration of invalidity of the impugned schedule for a period of 36 mon ths for
Parliament to address the unconstitutionality and an interim reading-in in item 1 of the
impugned schedule in the manner set out in the preceding paragraph.

[3] The application is opposed by the Speaker of the National Assembly
(National Assembly), th e Chairperson of the National Council of Provinces (NCOP)
and the Minister of Home Affairs (Minister). The President filed a notice of intention
to oppose the application jointly with the Ministe r but did not file an
answering affidavit. The Independent Electoral Commission (Electoral Commission)
abides the decision of the Court and has provided an explanatory affidavit.


2 1 of 2023.
MHLANTLA J
4
[4] In order to understand this case, it is imperative that I commence by setting out
the electoral scheme that operated in South Africa since the advent of our democracy in
1994 and the events that culminated in the amendment of the Electoral Act.

The electoral system since 1994
[5] The Constitution itself does not specify the national and provincial election
system, instead, it outlines the requirements that the elec toral system should comply
with. These requirements include the need for the system to be based on a national
common voters ’ roll, provide for a minimum voting age of 18 years, and result,
in general, in proportional representation. While there are a number of electoral systems
that comply with these requirements, the Constitution leaves the decision on the
preferred system to Parliament. However, this does not grant Parliament unrestricted
freedom. This Court has made it clear that provisions which disenfranchise citizens or
preclude them from standing for office are not permitted.3 At the heart of this case lies
the question of how Parliament chooses to give effect to these requirements.

[6] The number of seats available for contestation in the National Assembly is, in
terms of section 46(1) 4 of the Constitution, capped at 400. Sections 46(1)(d) and
105(1)(d)5 of the Constitution require the adoption of an electoral system that results,

3 See AParty v The Minister for Home Affairs, Moloko v The Minister for Home Affairs [2009] ZACC 4; 2009 (3)
SA 649 (CC); 2009 (6) BCLR 611 (CC) (AParty), Richter v The Minister for Home Affairs (with the Democratic
Alliance Intervening, and with Afriforum as Amici Curiae) [2009] ZACC 3; 2009 (3) SA 615 (CC); 2009 (5)
BCLR 448 (CC) ( Richter), Minister of Home Affairs v National Institute for Crime Prevention and the
Re- Integration of Offenders ( NICRO) [2004] ZACC 10; 2005 (3) SA 280 (CC); 2004 (5) BCLR 445 (CC)
(NICRO), August v Electoral Commission [1999] ZACC 3; 1999 (3) SA 1 (CC); 1999 (4) BCLR 363 (CC)
(August) and New National Party v Government of the Republic of South Africa [1999] ZACC 5; 1999 (3) SA 191
(CC); 1999 (5) BCLR 489 (CC) (New National Party).
4 Section 46(1) provides:
“(1) The National Assembly consists of no fewer than 350 and no more than 400 women
and men elected as members in terms of an electoral system that—
(a) is prescribed by national legislation;
(b) is based on the national common voters roll;
(c) provides for a minimum voting age of 18 years; and
(d) results, in general, in proportional representation.”
5 Section 105(1)(d) provides:
MHLANTLA J
5
in general, in proportional representation for the election of the National Assembly and
Provincial Legislatures. Sections 46(2)6 and 105(2)7 of the Constitution prescribe that
national legislation must provide a formula for determining the number of members in
the National and Provincial Assembly. Pursuant to these provisions, Parliament passed
the Electoral Act which, amongst others, regulates elections of the National Assembl y
and the Provincial Legislatures and provides for related matters.

[7] From 1994 to 2019, the election of representatives in the National Assembly was
based on a two -tier compensatory system, which fille d the 400 seats in the
National Assembly as follows: (a) 200 seats were filled from regional lists submitted
by political parties; and (b) 200 were compensatory seats based on the national lists of
political parties. The nine regions for which there were regional lists corresponded with
the nine provinces. During this period, only political partie s registered with the
Electoral Commission could contest elections. 8 This meant that an adult citizen was
not allowed to contest an election as an individual. Voters had one vote that determined
the regional and national seats. In other words, a voter in any particular region could
only cast one vote on a single ballot paper which listed the contesting political parties,

“(1) A provincial legislature consists of women and men elected as members in terms of an
electoral system that—
. . .
(d) results, in general, in proportional representation.”
6 Section 46(2) provides “[a]n Act of Parliament must provide a formula for determining the number of members
of the National Assembly.”
7 Section 105(2) provides “[a] provincial legislature consists of between 30 and 80 members. The number of
members, which may differ among the provinces, must be determined in terms of a formula prescribed by national
legislation.”
8 The Electoral Commission is established in terms of section 3 of the Electoral Commission Act 51 of 1996,
pursuant to section 190 of the Constitution. Section 190 provides:
“(1) The Electoral Commission must—
(a) manage elections of national, provincial and municipal legislative bodies in
accordance with national legislation;
(b) ensure that those elections are free and fair; and
(c) declare the results of those elections within a period that must be prescribed
by national legislation and that is as short as reasonably possible.
(2) The Electoral Commission has the additional powers and functions prescribed by national
legislation.”
MHLANTLA J
6
but his or her vote for that political party counted once towards the re gional seats and
once towards the compensatory seats.

[8] The purpose of the regional seat s was to ensure that voters in the respective
regions (provinces) were represented in proportion to the population size of their region
in the National Assembly. Once votes were tallied for a particular province, they were
divided by the number of National Assembly seats allocated to that province plus one,
and one was added to the result (ignoring fractions), to determine the quota (the number
of votes) needed per regional seat. The weighted average of the regional quotas in the
2019 national election , in which 17 437 379 people voted, was 83 511 votes per seat.
This method of determining a voting quota – by dividing the total number of votes by
the number of seats ava ilable plus one, and then adding one to the result – is called a
Droop quota.9

[9] The purpose of the national (compensatory) seat was to restore overall
proportionality as between the represented political parties due to potential distortion
created by the r egional system. The total number of votes cast was divided by 400
(being the number of seats in the National Assembly) plus one, and one was added to
the result (ignoring fractions), to determine the Droop quota for a compensatory seat.
In the 2019 national election the quota was 43 485 votes per seat.10 To determine how
many compensatory seats a political party was entitled to, the total number of votes
achieved by the party nationally was divided by the compensatory quota per seat to give
a provisional number of compensatory seats for the party, and the regional seats won
by that party were then deducted from the provisional number of compensatory seats to
arrive at that party’s final compensatory seats.


9 This method is named after the English mathematician, Henry Richmond Droop, who devised it in the 1860s. It
is used widely in election systems around the world. As a mathematical formula, the equation for a Droop quota
(DQ) can be expressed thus: 𝐷𝑄 = (
𝑇𝑜𝑡𝑎𝑙 𝑣𝑜𝑡𝑒𝑠 𝑐𝑎𝑠𝑡
[𝑇𝑜𝑡𝑎𝑙 𝑠𝑒𝑎𝑡𝑠+1] ) + 1.
10 This is about half of the regional quota. The reason for this is that, i n the Droop Formula, the nominator
(total votes cast – 17 437 379) was the same for regional and compensatory seats, whereas the denominator was
different for the regional and compensatory quotas – 201 and 401, respectively.
MHLANTLA J
7
New Nation judgment
[10] Shortly before the 2019 general elections,11 the New Nation Movement NPC and
its associates launched an urgent application for leave to appeal directly to this Court
against a judgment of the Western Cape Division of the High Court, Cape Town.12 The
High Court had dismissed an application to declare certain provisions of the Electoral
Act unconstitutional in so far as they did not allow independent candidates to stand for
elections. This Court held that the application was not urgent and heard it in the
ordinary course.13

[11] On 11 June 2020, this Court, in a judgment referred to as New Nation II,14 held
that the Electoral Act was inconsistent with the Constitution as it made no provision for
independent candidates to contest elections for the N ational Assembly and
Provincial Legislatures. The Court a cknowledged that an electoral system that would
accommodate independent candidates had to comply with certain requirements, one of
which was the section 46(1)(d) requirement that the electoral system must be one that
results, in general, in proportional representation. The choice of an electoral system
which must comply with these objectives and requirements was left for Parliament to
make. The operation of the declaration of unconstitutionality was suspended for
24 months to afford Parlia ment an opportunity to remedy the defect giving rise to the
unconstitutionality.15

[12] After the New Nation II judgment, Parliament was obliged to design an electoral
system that would accommodate both independent candidates and political parties.
That system still had to result in general, in proportional representation as required by

11 The elections were held on 8 May 2019.
12 New Nation Movement PPC v President of the Republic of South Africa [2019] ZAWCHC 43; 2019 (5) SA 533
(WCC) (High Court judgment).
13 New Nation Movement NPC v President of the Republic of South Africa [2019] ZACC 27; 2019 (9) BCLR 1104
(CC) (New Nation I).
14 New Nation Movement NPC v President of the Republic of South Africa [2020] ZACC 11; 2020 ( 6) SA 257
(CC); 2020 (8) BCLR 950 (CC) (New Nation II) at para 4 of the order.
15 Id at para 5 of the order.
MHLANTLA J
8
the Constitution. The revised electoral system effectively had to be in place before the
2024 national and provincial elections.

Process leading to the Electoral Amendment Act
[13] Dr Pakishe Aaron Motsoaledi, the Minister of Home Affairs, filed an answering
affidavit in which he explained the process that led to the amendment of the
Electoral Act. He explained that after the New Nation II judgment, he commissioned a
panel of experts called the Ministerial Advisory Committee (MAC) to produce technical
proposals in response to the judgment. The MAC had to develop policy options on the
electoral system that would cure the defects in the Electoral Act.

[14] After investigation and de liberation on different policy option s, the MAC
proposed two options. There was a 3:4 split between committee members as to which
option was preferable. The first , and majority, option resembled the current local
government electoral system, in that it p roposed combining the first -past-the-post
method and proportional representation system, making it a mixed -member
proportional system. This option would have entailed making extensive changes to the
Electoral Act. The second option entailed the modificat ion of the existing
multi- member electoral system to accommodate independent candidates in the national
and provincial elections without many changes in the legislation, including not
interfering with the constitutionally required general proportionality. Under this option,
the 400 seats in the National Assembly would continue to be divided into two:
200 regional seats and 200 compensatory seats. This option was referred to as the
minimalist approach. The Minister and his team preferred the minimalist approach and
decided to adopt it.

[15] Thereafter, the Minister instructed four independent advocates with expertise in
constitutional and electoral law t o prepare a draft amendment Bill in line with the
chosen “minimalist approach”. In November 2021, a draft Bill was presented to Cabinet
where it was accepted for introduction to Parliament. The Bill was introduced to
MHLANTLA J
9
Parliament and published in the Government Gazette on 31 December 2021 as the
Electoral Amendment Bill, B1-2022 (Electoral Amendment Bill).

[16] An answering affidavit filed on behalf of the Speaker of the National Assembly
and the Chairperson of the NCOP sets out the process followed by Parliament in
amending the Electoral Act. The answering affidavit was deposed to by Mr Mosa Steve
Chabane, the cha irperson of the National Assembly’s Portfolio Committee on
Home Affairs (Portfolio Committee).

[17] On 1 and 2 March 2022 , the Portfolio Committee heard oral submissions from
20 civil society organisations and individuals. The Portfolio Committee thereafter split
into two groups and conducted public hearings across the country from 9 to 23 March
2022. The Electoral Commission also made extensive submissions to the
Portfolio Committee on the Electoral Amendment Bill.

[18] In terms of this Court’s order in New Nation II, Parliament had 24 months to cure
the constitutional defects in the Electoral Act as identified in that judgment. That period
was due to expire on 10 June 2022 . On 26 April 2022, Parliament applied for a
six- month extension, which was granted, thereby shifting the deadline to 10 December
2022.

[19] Between 10 June and 20 October 2022, the Portfolio Committee discussed and
considered various aspects of the Electoral Amendment Bill in no less than 16 meetings
held at least once a week, where the Port folio Committee also received inputs from
various stakeholders and parties.

[20] The Electoral Amendment Bill was passed by the National Assembly on
20 October 2022 and was then referred to the NCOP for its consideration. Before
passing the Electoral Amendmen t Bill, the National Assembly (either through the
Minister or the Portfolio Committee) received at least four memoranda or opinions from
Counsel on various questions raised by Parliament or in response to public comments
MHLANTLA J
10
on various aspects of the Electoral Amendment Bill. Counsel’s opinion specifically
addressed questions on the 200/200 split that are the subject of consideration in this
matter.

[21] The NCOP’s Select Committee on Justice and Security (Select Committee)
called for written submissions from the public upon receiving the Electoral Amendment
Bill from the National Assembly. Between 2 November 2022 and 25 November 2022,
the Select Committee deliberated on the Electoral Amendment Bill and received inputs
from various stakeholders, including the Minister and the Electoral Commission.

[22] Towards the end of November 2022, the Select Committee proposed various
amendments to the Electoral Amendment Bill pursuant to the public engagement
process. As a result, the Electoral Amendment Bill was returned to the
National Assembly for further public engagement on the NCOP’s proposed
amendments as required in terms of standard parliamentary process.

[23] On 29 November 2022, t he NCOP amended and passed the
Electoral Amendment Bill. In December 2022, Parliament sought a further extension
on an urgent basis to finalise the Electoral Amendment Bill. On 9 December 2022, this
Court made an interim order suspending the declaration of invalidity to 31 January
2023, pending a final determination of Parliament’s further extension application. The
further extension application was granted by this Court on 2 0 January 2023, thereby
shifting the deadline to 28 February 2023. 16 Following this further extension, the
National Assembly passed the Electoral Amendment Bill on 23 February 2023 and
referred it to the President for assent.

[24] On 13 April 2023, the President signed the Electoral Amendment Bill and it was
published in the Government Gazette on 17 April 2023, resulting in the

16 This Court’s reasons for the interim and further extension orders were delivered on 20 April 2023: Speaker of
the National Assembly v New Nation Movement NPC [2023] ZACC 12; 2023 (7) BCLR 897 (CC) ( New Nation
III).
MHLANTLA J
11
Electoral Amendment Act. The impugned item 1 in Schedule 1A as amended by the
Electoral Amendment Act provides as follows:

“The seats in the National Assembly are as determined in terms of section 46 of the
Constitution and item 1 of Schedule 3 and are allocated as follows:
(a) Half the seats are filled by independent candidates from lists of candidates of
parties contesting the nine regions and these shall be referred to as regional
seats; and
(b) half the seats are filled by candidates from lists of candidates of parties and
these shall be referred to as compensatory seats.”

Election of the National Assembly under the Electoral Act as amended
[25] The allocation of seats in the National Ass embly in terms of the
Electoral Amendment Act is still divided into equal halves: the first 200 seats referred
to as regional seats and the other 200 as compensatory seats . Independent candidates
and political parties can both contest for the 200 regional seats, but only political parties
can contest for the 200 compensatory seats. The amended Electoral Act therefore
retains the two-tier system while providing an opportunity for independent candidates
to contest elections for the National Assembly as required by New Nation II.

[26] Whereas previously one ballot determined the allocation of regional and
compensatory seats in the National Assembly, voters now receive two ballots – one to
vote for a regional seat and one for a compensatory seat.

[27] The allocation of regional s eats is determined in terms of i tems 4 and 5 of
Schedule 1A to the amended Electoral Act, in terms of which , a s before, the
200 regional seats in the National Assembly are allocated to nine regions, which
coincide with the provincial boundaries. The Electoral Commission determines the
number of seats allocated to each region based on the number of registered vote rs in
each region. Provinces with higher population s of registered voters will thus be
MHLANTLA J
12
allocated more regional seats compared to provinces with smaller population s of
registered voters.

[28] A voter will receive a ballot paper specific to their region, which contains both
independent candidates and political parties. For convenience, I shall refer to this as
the regional ballot. The voter can vote for either a political party or an
independent candidate on the regional ballot. All votes cast in a region will first be used
to determine a provisional allocation of that region’s regional seats to both independent
candidates and political parties. If an independent candidate meets the relevant quota
for one seat, they will be elected to the National Assembly.

[29] The regional quota at the first stage of the computation is the same for both
political parties and independent candidates. The quota of votes needed to obtain a seat
in each region is determined using the Droop F ormula. This provides a provisional
allocation of seats for each political party and independent candidate. Any unallocated
seats are then allocated using the largest remainder method.17

[30] Even if an independent candidate gets sufficient votes to be provisionally
awarded two or more seats, the candidate is only entitled to one seat. This is because,
unlike a political party which has many candidates and may get many seats, that is not
possible for an independent candidate as they are contesting the elections in their
individual capacities and are not affiliated to any party. Once the single seats won by
each of the successful independent candidates are allocated, a revised Droop quota is
calculated to determine the final allocation of regional seats to the political parties, using
the remainder of the regional seats and the number of votes obtained by political parties
in the regional ballot.


17 Item 5(d) of Schedule 1A provides that “the surplus of votes accruing to any party, parties or independent
candidates in respect of the relevant region, competes for the remaining seats in sequence of the highest surplus
of votes”.
MHLANTLA J
13
[31] The amended Electoral Act provides that independent candidates forfeit any seat
or votes they get over and above their one allocated seat. Similarly, if an independent
candidate contests for more than one region and wins seats in more than one region,
they forfeit the additional seat(s) in the other re gion(s). The Minister submits that this
is the inevitable consequence of the choice made by a citizen to contest elections as an
individual instead of doing so through a political party.

[32] Then there are the 200 compensatory seats, the allocation of which is determined
in terms of item 6 of Schedule 1A of the amended Electoral Act. The voter in a region
gets a second ballot paper, listing only the contesting political parties. For convenience,
I shall call this the compensatory ballot. The compensatory b allot does not elect
compensatory representatives directly . The Droop Formula is also used t o allocate
compensatory seats. This quota takes into account all 400 seats in the
National Assembly minus the seats allocated to independent candidates and is based on
the total number of the votes cast for the political parties in both the regional and
compensatory ballots. As before, this exercise will result in a provisional allocation of
compensatory seats to a political party, from which are subtracted the regional seats
won by that party in order to arrive at the party’s final allocation of compensatory seats.
The stated purpose of the compensatory seats is to correct the disproportionality in
representation in the results of the election.

[33] Therefore, the new position after the passing of the Electoral Amendment Act is
that 200 regional seats are filled by independent candidates and candidates from
political parties, and the 200 compensatory seats are filled by the candidates from the
lists of political parties only. The independent candidates are excluded from contesting
for the compensatory seats.

[34] The affidavit deposed to by Mr Michael Louis, the chairman and director of the
applicant, states that the applicant does not object to the principle of a two-tier
compensatory system as contained in the impugned schedule. The applicant also does
not object to the principle that independent candidates can only compete for
MHLANTLA J
14
regional seats whereas political parties can compete for both the regional and
compensatory seats. However, the applicant tak es issue with the fact that
independent candidates can only compete for 200 regional seats, and argues that
independent candidates will, as a result, be required to earn more votes in order to secure
a single seat in the National Assembly compared to political parties.

[35] The contention that candi dates for regional seats will have to earn more votes
than candidates for compensatory seats is based on an analysis by Mr Michael Atkins
(Atkins Report) , whose expertise lies in the fie lds of mathematics and
computer science.18 Briefly, Mr Atkins’ thesis is that, based on data from the 2014 and
2019 elections, independent candidates will require more votes than political parties to
win a seat in the upcoming national elections.

[36] In the 2019 national election,19 in which 17 437 379 people voted, the weighted
average quota for regional seats was 83 511 while the quota for compensatory seats was
43 485. Before the Electoral Amendment Act, this difference in quota was
non- discriminatory, because there was only a single ballot and only political parties
contested elections. The formulas in the impugned schedule for determining the
allocation of regional and compensatory seats were the same for all political parties.

[37] According to Mr Atkins’ analysis, things have changed with the introduction of
two ballots in order to accommodate independent candidates in the regional ballot.
Because independent candidates can only contest for regional seats, they would – based
on the 2019 data – require 83 511 votes to win a seat. The new method for calculating
the compensatory quota still uses, as it did before, the total number of seats in the
National Assembly as the starting point fo r the denominator in the Droop F ormula
(400 plus one), save that one must de duct the number of seats won by

18 Mr Atkins conducted simulations of elections with independent candidates in terms of the Electoral Amendment
Act and analysed the results. The results and analysis formed part of detailed submissions to Parliament in
consultation with civil society organisations. The Atkins Report was submitted in support of the applicant’s
submissions in this Court and was peer reviewed and verified by two independent actuaries.
19 For brevity, my analysis will only focus on the 2019 data, as such, I will not include the 2014 data.
MHLANTLA J
15
independent candidates from the denominator. However, the number used as the
numerator is now the total votes cast for political parties in both the regional and
compensatory ballots. So, if all voters voted only for political parties, the numerator –
based on the 2019 data – would be 34 874 758 (that is, double the number of single
votes cast in the 2019 election).

[38] The compensatory quota will thus be significantly higher than it was in the 2019
election. Indeed, if no independent candidates obtained votes, the compensatory quota
would be very close to the regional quota. But, says Mr Atkins, the difference is that in
relation to the compensatory seats , political parties get the benefit of votes from both
the regional and compensatory ballots. Mr Atkins makes the fundamental assumption
that persons who vote for political parties in the regional ballot will vote for the same
political parties in the compensatory ballot. To illustrate what he contends to be a
discriminatory effect against independent candidates, he assumes total voter support for
independent candidates at two alternative levels, 2% and 8%.

[39] At these assumed levels, and based on the 2019 data, the compensatory quota for
a compensatory seat would be 86 532 and 85 846, respectively. However, on the
assumption that voters will vote for the same political parties in both ballots, these
figures suggest that a party will need the support of only half this number of voters in
order to win a compensatory seat. After certain minor adjustments, the figure of actual
voter support at which Mr Atkins arrives for a political party to win a compensatory
seat is 43 703 or 44 712, which is only just over half the voter support which a contestant
for a regional seat (including an independent candidate) would need to garner.

[40] After running multiple scenarios on varying numbers of regional and
compensatory seats, Mr Atkins concludes that an allocation of 350/50 as between
regional and compensatory seats will result in an app roximation of the actual voter
support required for regional and compensatory seats , respectively, while preserving
proportional representation for political parties and posing minimal risk of a so -called
overhang. I shall later explain the concept of overhang.
MHLANTLA J
16

[41] The applicant therefore seeks an order in terms of which the compensatory seats
would be reduced from 200 to 50. As a result, those competing in the regional tier will
contest for 350 seats instead of 200 seats. This will result in a 350/50 split as opposed
to the 200/200 split contemplated in the Electoral Amendment Act.

Issues
[42] The issues to be determined by this Court are as follows:
(a) Should direct access be granted?
(b) Is the 200/200 split rationally connected to a legitimate governmental
purpose?
(c) Does the 200/200 split give rise to an unjustifiable limitation of
fundamental rights in the Constitution, in particular, the right to vote and
to stand for public office?
(d) If so, is the limitation justifiable?
(e) If not, what is an appropriate remedy?

Direct access
[43] Section 167(6)(a) of the Constitution, read with rule 18 of the Rules of this Court,
provides for direct access to the Constitutional Court. In terms of these provisions, this
Court has discretion whether to grant direct access , but such discreti on will only be
exercised in favour of the applicant if it is in the interests of justice to grant direct access.
As stated in Zondi,20 whether it is in the interests of justice to grant leave for direct
access depends on various factors, including:

“[T]he importance of the constitutional issue raised and the desirability of obtaining an
urgent ruling of this Court on that issue, whether any dispute of fact may arise in the

20 Zondi v MEC for Traditional and Local Government Affairs [2004] ZACC 19; 2005 (3) SA 589 (CC); 2005 (4)
BCLR 347 (CC).
MHLANTLA J
17
case, the possibility of obtaining relief in another court, and time and costs that may be
saved by coming directly to this Court.”21

[44] For direct acce ss to be granted, an applicant must show that exceptional
circumstances exist. This Court is reluctant to be a cou rt of first and last instance as
this deprives parties of the right of appeal and deprives the Court of the benefit of other
courts’ insights.22

[45] The applicant submits that it would be in the interests of justice for this Court to
grant direct access in that: (a) the application raises legal issues of fundamental
importance relating to the fairness of the 2024 elections; (b) the Court has to determine
the validity and constitutionality of the impugned schedule; (c) the challenge has
reasonable prospects of success; and (d) the 2024 elections are imminent, and there is
thus insufficient time to approach the High Court for the relief sought and thereafter
seek confirmation of the order of invalidity in this Court as required by section 167(5)
of the Constitution.

[46] The respondents do not oppose the applicat ion for direct access. They accept
that time is of essence and that it is imperative that the issues raised in the application
be determined as a matter of urgency to allow the Electoral Commission sufficient time
to prepare for the 2024 elections.

[47] I agree with the parties tha t the circumstances of this case are exceptional and
that the issues raised are purely legal in nature. This application concerns constitutional
issues, in particular, whether the impugned schedule violates political rights enshrined
in section 19 of the Constitution. This matter also raises questions of separation of
powers, in particular, the extent to which courts can interfere with the electoral process
chosen by Parliament.

21 Id at para 12.
22 Women’s Legal Trust v President of the Republic of South Africa [2009] ZACC 20; 2009 (6) SA 94 (CC) at
para 27.
MHLANTLA J
18

[48] Further, having regard to the explanatory affidavit filed on behalf of the
Electoral Commission, there will be insufficient time for the matter to be considered by
the High Court and then brought to this Court for final consideration , either on appeal
or confirmation. Finality is required to enable the Electoral Commission to pro ceed
with the preparations for the 2024 general election, which must be held on a date to be
proclaimed by the President of the Republic of South Africa between 22 May 2024 and
14 August 2024.23 Therefore, it is in the interests of justice that direct access be granted.

[49] I now proceed to consider the merits of this application.

Is the 200/200 split rationally connected to a legitimate government purpose?
[50] According to the applicant, the 200/200 split is irrational and inconsistent with
the rule of law and, therefore, contravenes section 1(c) of the Constitution. 24 The
irrationality does not stem from the fact that independent candidates may only compete
for regional seats but, so the applicant argues, from the quota for regional seats being
much lower than t he quota for compensatory seats. As a result, the applicant submits
that the equality of the vote and equal opportunity to be elected are violated. On the
applicant’s version, there is no justification or reason for the split other than for the
improper purpose of undermining the prospects of independent candidates getting
elected.

[51] In determining this question, it is apposite to first remind ourselves of the legal
principles relating to the powers of Parliament and the test for rationality.

23 This is the effect of section 49 of the Constitution, which prescribes a five-year term for the National Assembly
and provides that an election must be held within 90 days of the expiry of that term, on a date to be proclaimed
by the President.
24 Section 1(c) states:
“(1) The Republic of South Africa is one, sovereign, democratic state founded on the following
values:
. . .
(c) Supremacy of the constitution and the rule of law.”
MHLANTLA J
19

[52] It is trite that in order to pass constitutional muster, the exercise of public power
by the Executive and other public functionaries must, at the very least, comply with the
threshold of rationality. Decisions must be rationally related to the purpose for which
they are given. This is an objective enquiry. A court is not allowed to substitute its
own opinion as to what it thinks is appropriate.

[53] In New National Party,25 this Court had to consider a constitutional challenge in
respect of certain provisions of the Electoral Act that prescribed that only bar -coded
identity documents could be used for the purposes of voting and registering for the 1999
national and provincial elections. Yacoob J, writing for the majority, said the following
on the role of Parliament and onus of proof where an infringement of a right has been
alleged:

“It is to be emphasised that it is for Parliament to determine the means by which voters
must identify themselves. This is not the function of a court. But this does not mean
that Parliament is at large in determining the way in which the electoral scheme is to
be structured. There are important safeguards aimed at ensuring appropriate protection
for citizens who desire this foundational right. The first of the constitutional constraints
placed upon Parliament is that there must be a rational relationship between the scheme
which it adopts and the achievement of a legitimate governmental purpose. Parliament
cannot act capriciously or arbitrarily. The absence of a rational connection will result
in the measure being unconstitutional. An objector who challenges the electoral
scheme on these grounds bears the onus of establishing the absence of a legitimate
government purpose, or the absence of a rational relationship between the measure and
that purpose.
A second constraint is that the electoral scheme must not infringe any of the
fundamental rights enshrined in chapter 2 of the Constitution. The onus is once again
on the party who alleges an infringement of the right to establish it.”26


25 New National Party v Government of the Republic of South Africa [1999] ZACC 5; 1999 (3) SA 191 (CC);
1999 (5) BCLR 489 (CC).
26 Id at para 19-20.
MHLANTLA J
20
[54] Chaskalson P reiterated the rationality requirement in Pharmaceutical
Manufacturers27 and held that:

“It is a requirement of the rule of law that the exercise of public power by the Executive
and other functionaries should not be arbitrary. Decisions must be rationally related to
the purpose for which the power is given, otherwise they are in effect arbitrary and
inconsistent with this requirement.
. . .
The question whether a decision is rationally related to the purpose for which the power
was given calls for an objective enquiry.
. . .
What the Constitution requires is that public power vested in the Executive and other
functionaries must be exercised in an objectively rational manner.
. . .
As long as the purpose sought to be achieved by the exercise of public power is within
the authority of the functionary, and as long as the functionary’s decision, viewed
objectively, is rational, a Court cannot interfere with the decision simply because it
disagrees with it, or considers that the power was exercised inappropriately.”28

[55] The wide remit given by the Constitution to Parliament to enact an electoral
system was restated in New National Party as follows:

“The right to vote contemplated by section 19(3) is therefore a right to vote in free and
fair elections in terms of an electoral system prescribed by national legislation which
complies with the aforementioned requirements laid down by the Constitution. The
details of the system are left to Parliament.” 29

[56] In AParty,30 Ngcobo J reiterated the statement in New National Party and said:


27 Pharmaceutical Manufacturers Association of South Africa: In re Ex Parte President of the Republic of South
Africa [2000] ZACC 1; 2000 (2) SA 674 (CC); 2000 (3) BCLR 241 (CC).
28 Id at paras 85-90.
29 New National Party above n 25 at para 14.
30 AParty v Minister of Home Affairs, Moloko v Minister of Home Affairs [2009] ZACC 4; 2009 (3) SA 694 (CC);
2009 (6) BCLR 61 (CC).
MHLANTLA J
21
“Parliament has the constitutional authority and duty to design an electoral scheme to
regulate the exercise of the right to vote. This is apparent from sections 46(1), 105(1),
and 157(5) of the Constitution.” 31

[57] This principle was again restated in New Nation II, where Madlanga J held:

“The pros and cons of this or the other system are best left to Parliament which in terms
of sections 46(1)(a) and 105(1)(a) of the Constitution has the mandate to prescribe an
electoral system.”32

[58] In Albutt,33 Ngcobo CJ said:

“The executive has a wide discretion in selecting the means to achieve its
constitutionally permissible objectives. Courts may not interfere with the means
selected simply because they do n ot like them or because there are other more
appropriate means that could be selected. . . . What must be stressed is that the purpose
of the enquiry is to determine not whether there are other means that could have been
used, but whether the means selected are rationally related to the objective sought to be
achieved.”34

[59] Therefore, as long as the purpose sought is within the authority of the functionary
and as long as the functionary’s decision is rationally connected to that purpose, a court
cannot interfere with the decision simply because it disagrees with it.

[60] In the matter before us, the applicant contends that the 200/200 split means that
the quota of votes a political party must obtain to be allocated compensatory seats is
much lower than the q uota of votes the independent candidate must obtain to be
allocated regional seats, and this infringes on the equality of vote and equal opportunity.
In this regard, the applicant, as stated earlier, relies on the report produced by Mr Atkins.

31 Id at para 6.
32 New Nation II above n 14 at para 15.
33 Albutt v Centre for the Study of Violenc e and Reconciliation [2010] ZACC 4; 2010 (3) SA 293 (CC); 2010 (5)
BCLR 391 (CC).
34 Id at para 51.
MHLANTLA J
22
On Mr Atkins’ model, the quota for re gional seats is much higher than the quota for
compensatory seats – almost double.35

[61] The applicant argues that the 200/200 split as introduced in the
Electoral Amendment Act is arbitrary and therefore substantively and procedurally
irrational in that it fails to meaningfully give effect to section 19(3) of the Constitution
and is not reasonably capable of attaining the desired outcome. The applicant adds that
although the inadequacies in the 200/200 split were highlighted during the public
participation processes, they were not addressed by the Parliamentary Committee.

[62] The respondents submit that the Electoral Amendment Act is rational in that the
two different methods used by the Electoral Amendment Act to fill regional and
compensatory seats legitimately differentiate between independent candidates and
political parties for the express purpose of achieving proportional representation and
avoiding the risk of ov erhang. Overhang occurs where the election formula requires
political parties to be allocated more seats than are actually available in the Legislature.
Put differently, overhang occurs where a party wins more regional seats than the
compensatory or nati onal party vote entitles it to. This, in turn, has implications for
proportional representation. In a Legislature of 400, overhang of one would indicate
that, after applying all the relevant formulas in the legislation , 401 seats have to be
allocated. The rationality of the 200/200 split therefore hinges on whether (a) it results
in proportional representation, in general and (b) avoids the risk of overhang.

Proportional representation in general
[63] Section 46(1)(d) of the Constitution requires Pa rliament to design an
electoral system that results, “in general, in proportional representation”. There are 400
seats available for contestation in the National Assembly which , in terms of the
impugned schedule, are divided in half: 200 regional seats and 200 compensatory seats.

35 See Mr Atkins’ proposed model at page 19 of the Atkins Report. The weighted average for the 200/200 split is
83 511 versus 43 703 with a 91% difference. This is as opposed to the 350/50 split, where the weighted averages
are 48 556 for regional seats versus 44 034 for compensatory seats, with a 10% difference.
MHLANTLA J
23
The stated purpose of the 200 compensatory seats, according to Parliament, is to correct
the disproportionality in representation in the results of the election.

[64] Parliament contends that proportional representation systems are, in genera l,
subject to distortions with the introduction of independent candidates. This is so
because irrespective of how many votes an indep endent candidate receives , they are
restricted to a maximum of one seat. By contrast, political parties can hold the number
of seats proportional to the number of votes received.

[65] According to Parliament and the Minister, the only way to include independent
candidates in a two-tier system is to allow them to retain the seats they win and obtain
proportionality by allocating the remaining seats to the parties that gain representation.
As a result, independent candidates cannot compete for compensatory seats as they are
reserved exclusively for political parties.

[66] The Lynge and Rosen Report, 36 filed on behalf of the Electoral Commission,
concludes that the 200/200 split achieves proportionality. The applicant accepts that
the 200/200 split results, in general, in proportional representation. This is also accepted
in the Atkins Report filed in support of the applicant’s case. The respondents accept
that a 350/50 split would also achieve proportionality in general. The applicant also
acknowledges that reserving compensatory seats for political parties is necessary to
achieve proportionality because the formulas for regional seats have the potential to
distort proportionality, and this potential has increased because independent candidates
can now contest for regional seats but can only hold one seat.

[67] Properly construed, the dispute between the parties is, therefore, not whether it
is irrational for Parliament to distinguish between regional and compensatory seats in

36 Halfdan Lynge and Simon Rosen, “Analysis of the effects of the Electoral Amendment Bill” at 2 .
Dr Halfdan Lynge is a senior lecturer at the Wits School of Governance at the University of the Witwatersrand .
Dr Lynge has published extensively on political rights. Mr Simon Rosen is a data scientist, full -stack developer,
and graduate student in the School of Compute r Science and Applied Mathematics at the University of the
Witwatersrand.
MHLANTLA J
24
the National Assembly. Rather, the dispute concerns the number of compensatory seats
that ought to be reserved for political parties, with the applicant proposing a
350/50 split. The applicant argues that the 200/200 split debases the value of votes for
regional seats. In comparison , the 350/50 split will make the quota for regional and
compensatory seats similar while still retaining overall proportionality and without
giving rise to a significant risk of overhang. Because both the 200/200 and 350/50 split
results i n proportional representation, the question is whether avoiding the risk of
overhang justifies the disparity in votes needed to obtain a compensatory and regional
seat, respectively. The crucial consideration is whether the 200/200 does, in fact, reduce
the value of votes in the regional ballot. I deal with this point later in determining
whether the 200/200 split amounts to an unjustifiable limitation of sections 1(c); 3(2)(a);
9(1); 19(2); 19(3) and 46(1)(d) of the Constitution.

[68] Parliament’s stated obj ective for the 200/200 split is the achievement of
proportional representation as required by the Constitution. Even if the 350/50 split
proposed by the applicant might arguably be fair er and achieve proportionality, an
assessment as to whether the 200/200 split achieves proportional representation in
general is one that should be conducted by Parliament. The case law referred to above
demonstrates the wide latitude given to Parliament to consider the manner in which to
conduct the electoral system. Sections 46(1) and 105(1) expressly leave the choice of
electoral system in Parliament’s hands. Although the Constitution does set out
requirements, there are a number of possible electoral s ystems which would comply
with these requirements. At least on the requirement of achieving proportionality, the
200/200 split chosen by Parliament passes constitutional muster.

Overhang
[69] The second stated objective of the 200/200 split by Parliament is t hat it avoids
the risk of overhang. The Ele ctoral Commission explains that overhang is when more
seats are required to be allocated to restore proportionality as between represented
parties after the allocation of regional (or constituency) votes, than ar e available in the
Legislature. Overhang occurs when a party wins more regional seats than it is overall
MHLANTLA J
25
entitled to when the compensatory tier outcome is calculated. This can happen with
regionally based parties with strong support but much less support in the rest of the
country. Another possibility is when a small party wins a regional seat based on the
largest remainder method but does not meet the threshold for a seat when the
compensatory calculations are done. The consequence of such situations is that seats
for parties cannot be allocated in the National Assembly according to their inter -party
proportional entitlement.

[70] Section 46(1) of the Constitution limits the number of seats in the National
Assembly to 400. The implication of overhang in ou r system is that more seats are
required to be allocated to ensure proportional representation as between the represented
parties after the allocation of regional votes than are available in the Legislature. Put
differently, when overhang occurs, seats in the National Assembly cannot be allocated
according to inter -party proportional entitlement and, as a result, the
Electoral Commission may not be able to declare the results of an election.

[71] Although the applicant accepts that the 200/200 split avoids the risk of overhang,
it contends that the 350/50 split will bring the quota for a regional seat closer to the
quota for a compensatory seat without posing a significant risk of overhang or
disturbing the proportional representation of political parties. To verify that the
350/50 split sufficiently minimises the risk of overhang, Mr Atkins ran a
Monte Carlo simulation37 of elections in terms of the Electoral Amendment Act based
on the 2014 and 2019 election results. The simulation was repeated with 200, 100, 75,
50, 25 and 10 compensatory seats for comparison and with assumed levels of support
for independent candidates ranging from 1% to 10% and assessed the risk of overhang
with reference to “main parties” ,38 “minor parties” 39 and the “largest party” .40

37 A Monte Carlo simulation is a mathematical technique that predicts the outcome of an uncertain future event
based on past data. Raychaudhuri “Introduction to Monte Carlo simulation” (2008) Winter Simulation Conference
91 at 91.
38 Parties that obtain more votes than the quota for a single seat.
39 Parties that obtain fewer votes than the quota for a single seat.
40 The ANC, whose results were included in the “main parties” results.
MHLANTLA J
26
Mr Atkins’ tables show that the risk of overhang almost always reduces as support for
independent candidates increases, so the greatest risk is where in dependent candidates
win only 1% of voter support.

[72] According to the Atkins Report, the 350/50 split proposed by the applicant gives
rise to a foreseeable risk of overhang of up to 15% per election (based on the 2014
election data), limited to one seat. 41 The applicant argues that, while foreseeable, this
risk under the 350/50 split is remote and limited to one se at. On the 200/200 split,
however, there is virtually no risk of overhang, with the only recorded risk found in the
Atkins Report being at 0,03% at 6% support for independent candidates in respect of
one seat, based on the 2014 election data.

[73] With each simulation of elections for regional seats, the 2014 and 2019 election
results were varied in terms of the number of independent candidates contesting each
regional election, as well as the level of support for independent candidates at 1%; 2%;
4%; 6%; 8% and 10%. The number of votes for independent candidates were added to
the total votes for political parties. As a result, the relative reduction in support for
political parties was kept constant across parties.

[74] Once the regional elections were simulated , the votes for political parties were
added together and used to simulate the overall proportional representation of political
parties for the National Assembly. The number of seats available to political parties
is 400 minus the number of seats won by independent candidates. The simulation relies
on two assumptions. First, a vote for a political party in the regional ballot will translate
to a vote for the same political party on the compensatory ballot. Second, voters that
supported an independent ca ndidate in the regional ballot will not vote for one of the
three largest political parties but instead vote for one of the smaller parties. The basis
of these assumptions does not appear from the Atkins Report. Further, the

41 Atkins Report above n 18 at 27.
MHLANTLA J
27
Atkins Report does not tell u s how the risk of overhang would be affected if these
assumptions were changed.

[75] Based on the simulation, t he 350/50 split gives rise to a foreseeable risk of
overhang. The risk of overhang of one seat is at its greatest where independent
candidates obtain 1% support. The risk of overhang reduces as support for independent
candidates’ increases . On the 2019 election data, and based on 1 % support for
independent candidates, the risk of overhang of one seat, while being 0% for the largest
party, is about 2% for the main parties and about 7% for minor parties. On the 2014
election data, and while the risk for the largest party remains at 0%, the risk for main
parties reduces to just under 1% but increases to about 15% for minor parties.

[76] The Atkins Report further shows that parties with strong regional support have
a noticeable risk of a single-seat overhang. The implication is that political parties with
strong regional support may be awarded a seat based on the regional ballot while not
being awarded a s eat in the N ational Assembly based on the final seat allocation for
political parties. In light hereof, the applicant argues that while there is virtually no risk
of overhang on the 200/200 split,42 the risk of overhang under the 350/50 split is remote
and limited to one seat. The applicant also states that, in any event, the Lynge and
Rosen Report which the Electoral Commission relies on accepts that a one seat
deviation is within the bounds of proportional representation.

[77] In considering the question of overhang, the Electoral Commission filed an
explanatory affidavit to respond to the Atkins Report. Mr Phatudi Simon Mamabolo,
the Chief E lectoral O fficer of the Electoral Commission, explained that there is no
mechanism under the current Electoral Act to correct the problem should overhang
occur. Secondly, the applicant has not presented any viable solution to combat the risk
of overhang, or remedy it should overhang occur. Furthermore, overhang poses the risk
that the Electoral Commission would not be able to declare the election should the risk

42 On the 200/200 split, the 2014 election data show a 0.03% risk of overhang in respect of one seat for minor
parties where independent candidates obtain 6% support.
MHLANTLA J
28
materialise and any declaration o f the election results where overhang has occurred
would likely be challenged in court.

[78] The Electoral Commission states that the fact that a one seat deviation still
complies with proportional representation is irrelevant. The Lynge and Rosen Report
did not consider the risk to proportional representation should overhang occur. Rather,
Lynge and Rosen ran a Monte Carlo simulation to determine whether the 200/200 split
under the amended electoral formula would introduce additional disproportionality in
comparison to the electoral formula prior to the amendment. They concluded that it
would not – the deviation from absolute proportionality never exceeds one seat.

[79] Parliament and the Minister submit that, even if the risk is as remote as proffered
by the applicant, the consequences are severe as explained by the
Electoral Commission. To place the risk of overhang in perspective, the Minister refers
to the election of the German Fe deral Parliament, the Bundestag, where overhang
occurred and 138 seats had to be awarded to restore proportionality.

[80] I agree with the respondents that, if overhang were to occur in South Africa under
the current construction , there would be no option to a dd additional seats to the
National Assembly to restore proportional representation, as was done in Germany.
Germany has a comparable population size to that of South Africa. The Bundestag is
comparable to South Africa’s National Assembly and aims to ach ieve proportional
representation as best as it can. Similar to our two -tier system, German voters cast
two ballots – one for a candidate to represent their single member constituency and one
for a political party. Unlike South Africa’s National Assembly, the German Bundestag
is not limited to a specific number of seats. Thus, when overhang occurs, extra seats
are made available to ensure proportional representation. In the 2021 Bundestag
election, which produced overhang, a total of 138 extra seats had to be awarded to
restore proportionality. To put things into perspective, South Africa can only allocate
MHLANTLA J
29
a maximum of 400 seats in the National Assembly. At 736 members, the Bundestag is
now the largest democratically elected parliament.43

[81] We do not have the luxury of pursuing the path taken by the German government
— merely increasing the seats when the need arose.44 The Constitution of South Africa
is clear in this regard and capped the maximum number of seats at 400. Any increase
beyond the 400 seats available for contestation would require a constitutional
amendment and this is something the applicant has failed to take into account.

[82] Besides showing that the risk of overhang is remote on the 350/50 split, the
applicant does not offer a single solution in its pleaded case to combat the risk or how
it should be dealt with should the risk materialise . The fact that the risk is limited to
one seat is immaterial – the issue is that there is a risk that, if it materialises, would have
dire consequences. T he Electoral Act does not provide a mechanism that allows the
Electoral Commission to correct the problem should overhang occur. Furthermore, the
Electoral Commission has explained that, if overhang occurs, it will not be able to
declare that election.

[83] Considering the above, on the 200/200 split there is virtually no risk of overhang.
As a result, Parliament’s second stated objective of the 200/200 spilt – to avoid the risk
of overhang – is achieved. On the other hand, the foreseeability of overhang under the
350/50 split as proposed by the applica nt is potentially destructive to the applicant’s
case. So too is the acknowledgement that the 200/200 split serves a legitimate

43 Witting and Goldberg “Germany passes law to shrink its XXL parliament” DW (21 January 2023, last updated
17 March 2023), available at https://amp.dw.com/en/germany-passes-law-to-shrink-its-xxl-parliament/a-
64471203.
44 Recent legislative reforms will, when enacted, abolish the rules on the overhang which resulted in the continuous
increase in the number of seats in the Bundestag. The reforms will reduce the size of the Bundestag to 630 from
the 2025 general election without renouncing the combination of personal and proportional representation. At the
time of writing this judgment, a constitutional challenge to this amendment is still pending in the German
Constitutional Court. In a recent Joint Opinion, however, the Venice Commission and the Office for Democratic
Institutions and Human Rights stated that the amendments conform to international electoral standards. See
European Commission for Democracy Through Law (Venice Commission) and OSCE Office for Democratic
Institutions and Human Rights (ODIHR) Germany: Joint Opinion on the Amendments to the Electoral Act
(CDL- AD (2023)020, 12 June 2023) at 15, available at https://www.venice.coe.int/webforms/documents/defaul
t.aspx?pdffile=CDL-AD(2023)020-e.
MHLANTLA J
30
governmental purpose – that is, the split ensur es proportionality in results.
Consequently, and as already demonstrated, the 200/200 split passes the rationality test
and cannot be said to be arbitrary.

Possible solutions to the risk of overhang vis-à-vis the applicant’s pleaded case
[84] Before concluding on the issue of overhang, I find myself constrained to address
the applicant’s submission that the method utilised at the municip al level to prevent
overhang offers a potential solution to the risk of overhang in the 350/50 split. Th is
submission was made for the first time during the hearing.

[85] It is well -established that parties in motion proceedings should make out their
case in their founding or answering affidavits and not in their heads of argument or in
the course of the hearing. This is because the right to a fair trial requires that parties
know the case they have to meet, and also ensures that the issues that the courts have to
determine are properly placed before them.45 As such, and in deciding this matter, this
Court cannot take into consideration the submission by the applicant on a potential
solution to the risk of overhang as it is not properly before this Court. The exception to
this position was articulated in My Vote Counts46 where Khampepe J said the following:

“It is, in any event, imperative that a litigant should make out its case in its founding
affidavit, and certainly not belatedly in argument. The exception, of course, is that a
point that has not been raised in the affidavits may only be argued or determined by a
court if it is legal in nature, foreshadowed in the pleaded case and does not cause
prejudice to the other party.”47 (Emphasis added.)

[86] In the present case, this exception does not apply. The proposed solution to the
risk of overhang which the applicant proffe red in the hearing is not legal in nature nor

45 See, for example Swissborough Diamond Mines (Pty) Ltd v Government of the Republic of South Africa 1999
(2) SA 279 (T) at 323F-G.
46 My Vote Counts NPC v Speaker of the National Assembly [2015] ZACC 31; 2016 (1) SA 132 (CC); 2015 (12)
BCLR 1407 (CC).
47 Id at para 177.
MHLANTLA J
31
was it foreshadow ed in the applicant’s pleaded case. The applicant did not offer any
solution to the risk of overhang in its pleaded case. Had the issue been properly raised
by the applicant in its founding a ffidavit, the respondents may well have put up a case
against this proposed solution. The respondents may have offered reasons why the
solution to overhang at municipal level could not work at national level. If this Court
were to consider the proposed s olution to the risk of overhang as proffered by the
applicant in the hearing, that would certainly be prejudicial the respondents as they
would not have had the opportunity to deal with the issue.

[87] While I do not attach weight to the applicant’s proposed s olution to the risk of
overhang, I wish to make the following remarks : although the 200/200 split is rational
in that it avoids the risk of overhang and ensures proportional representation, this should
not be inter preted as an indication that overhang will always be an insurmount able
challenge. The risk of overhang in the Atkins Report on the proposed 350/50 split, as
well as on the 300/100 or 325/75 split, arises for two reasons.

[88] First, like the Electoral Act, the Electoral Amendment Act prescribes that the full
400 seats of the National Assembly must be made available for contestation. This is
the maximum number of seats that the Constitution prescribes. On the proposed
350/50 split, where there is a foreseeable risk of overhang, it is impossible to increase
the number of seats in the National Assembly by the number of seats required to cure
the overhang.

[89] The election of the New Zealand Parliament offers another example of h ow
overhang can be resolved , one which does not inc rease the size of Parliament to the
extent of the German Bundestag. In New Zealand, the Electoral Act 48 introduced a
mixed-member proportional system to elect Members of Parliament , which ordinarily
comprises of 120 seats. Section 192(5) of the New Zealand Electoral Act provides that
if overhang seats are won, the number of seats in Parliament is increased by the number

48 87 of 1993 (New Zealand Electoral Act).
MHLANTLA J
32
of overhang seats until the next election. For example , there were 121 members after
the 2005 and 2011 elections, and 122 following the 2 008 election.49 Of course, our
Constitution does not allow the National Assembly to have more than 400 seats.

[90] However, should the number of seats available for contestation in Parliament be
reduced to a number above 350 but below 400 ,50 it may be possible to allocate
“overhang seats” and increase the number of seats in the National Assembly by the
number of overhang seats (up to 400) until the next election. By way of illustration, if
Parliament were to prescribe that 390 seats (of the 400) were available for contestation
in the National Assembly, and for the relevant election year it happens that there is
overhang of two seats , following the New Zealand model, the Electoral Commission
would be able to allocate the two overhang seats until the nex t election. This would
mean an allocation of 392 seats in the National Assembly until the next election (that
is, the 390 available for contestation plus the two overhang seats), which is still below
the 400 constitutional cap.

[91] Secondly, unlike the Local Government: Municipal Structures Act,51 the
Electoral Amendment Act does not provide a formula to address overhang should it
arise. In local government elections for a metro council or a local council with wards,
each voter has two votes: one fo r the direct election of a ward candidate (who may be
an independent candidate or a political party representative) and one for a political party
to ensure that political parties are proportionally represented in the council. 52
Ward candidates and political parties contest elections for the same number of seats. 53
Item 16 of Schedule 1 to the Municipal Structures Act provides:

49 “New Zealand’s 10th MMP General Election ” New Zealand Parliament (12 October 2023), available at
https://www.parliament.nz/en/get-involved/features/new-zealand-s-10th-mmp-general-election/; and “How are
MPs elected” Electoral Commission New Zealand (2023), available at https://elections.nz/democracy-in-nz/what-
is-new-zealands-system-of-government/how-are-mps-elected/.
50 This would still be in line with section 46(1) of the Constitution, which requires the National Assembly to
consist of no fewer than 350 and no more than 400 members.
51 117 of 1998.
52 Id at item 9 of schedule 1 to the Local Government: Municipal Structures Act.
53 Id at section 21(1).
MHLANTLA J
33

“(1) If, through the election of ward candidates, a party listed on the part of the
ballot paper for parties has obtained a number of seats that is equal to, or greater
than the total number of seats in the council to which it is entitled under item
13, that party must not be allocated any seats from its list of party candidates.
(2) The seats of ward candidates are not affected.
(3) A new quota for a seat must be determined in accordance with the following
formula (fractions to be disregarded)
. . .
(4) (a) The total number of valid votes cast for each party, both on the party
vote and for the ward candidates representing the party, excluding the
party that has excessive number of seats, must be divided by the quota
of votes for a seat. The result is the total number of seats to which
each party is entitled.
(b) If the calculation in paragraph (a) yields a surplus not absorbed by the
seats a warded to a party, that surplus must compete with similar
surpluses accruing to any other party or parties and any undistributed
seat or seats must be awarded to the party or parties concerned in
sequence of the highest surplus.
(c) If the surplus for two or more parties is equal, the seat must be awarded
to the party that received the highest number of valid votes”.54


54 The formula for calculating a new quota of votes for a seat is determined in terms of the following formula set
out in the Municipal Structures Act above n 51 at item 16(3) of Schedule 1:
= ( 𝐴 − 𝐵
𝐶 − [𝐷 + 𝐸]) + 1
“Where
A represents the total number of valid votes cast for all parties, consistin g of those cast
on the party vote and those cast for ward candidates representing parties;
B represents the total number of valid votes cast for the party that has forfeited seats,
both on the party vote and for ward candidates representing the party;
C represents the number of seats in the council;
D represents the number of seats awarded to the forfeiting party; and
E represents the number of independent ward councillors elected in the election. ”
MHLANTLA J
34
[92] I recognise that the absence of a formula to address overhang in the Electoral Act
as amended may well be a result of the f act that there i s no risk of overhang on the
200/200 split. I refer to the Municipal Structures Act to illustrate that the risk of
overhang is not a novel concept in South Africa and that the Legislature has , in the
context of local government elections, provided a mecha nism to deal with it should it
arise.

[93] On the facts and submissions made by the parties in this case, I conclude that the
200/200 split passes the rationality test as it achieves proportional representation and
avoids the risk of overhang , as intended. The next question is whether the electoral
scheme has infringed any fundamental rights in the Constitution.

Does the 200/200 split give rise to an unjustifiable limitation of fundamental rights in
the Constitution?
[94] The applicant contends that the 200/200 split as contained in the Electoral
Amendment Act is unconstitutional as it violates the sections 1(c), 3(2)(a), 9(1), 19(2),
19(3) and 46(1)(d) of the Constitution.

[95] In New National Party, where an infringement of voters’ rights was alleged, this
Court held that the onus is on the party who alleges an infringement of the right to
establish it.55 The applicant therefore bears the onus to establish the rights violations so
alleged. I now proceed to consider the issues raised by the applicant.

Sections 1(c) and 46(1)(d) of the Constitution
[96] The applicant submits that the section 1(c) provision is infringed upon in that the
split is irrational and inconsistent with the rule of law. The applicant also alleges in its
founding affidavit that the split violates section 46(1)(d)56 of the Constitution in that the

55 See, New National Party above n 25 at para 20.
56 Section 46(1)(d) requires the adoption of an electoral system that results, in general, in proportional
representation for the election of the National Assembly.
MHLANTLA J
35
election of the National Assembly in terms of the Electoral Amendment Act would not
result, in general, in proportional representation. This is because, so argues the
applicant, there are bound to be wasted votes and excess votes for independent
candidates which will be allocated to political parties, thereby skewing the outcome of
the elections. The applicant did not persist with this argument in the hearing.

[97] As already shown, t he applicant acknowledges that there is a legitimate
governmental purpose served by the 200/200 split – the split ensures proportionality in
results. The split also avoids the risk of overhang. Consequently , and as already
demonstrated, the 200/200 split passes the rationality test. The section 1(c) challenge
is, therefore, without merit.

[98] Similarly, th e applicant’s section 46(1)(d) challenge to the 200/200 split as
alleged in its founding affidavit should fail. During the hearing, counsel for the
applicant accepted that the 200/200 split achieves proportional representation in general
as contemplated in section 46(1)(d). This was also accepted in the Atkins Report.

Sections 3(2)(a) and 9(1) of the Constitution
[99] The applicant argues tha t the 200/200 split violates section 3(2)(a) 57 of the
Constitution in that a vote for an independent candidate carries less weight than a vote
for a political party, and that this is inconsistent with citizens voting equally for
independent candidates and i ndependents running for public office. This is because,
argues the applicant, the compensatory quota requires only half the number of voters to
support a political party than is required by an independent candidate. In reaching this
conclusion, the Atkins Report rests on the critical assumption that voters who vote for
a political party in the regional ballot will vote for the same political party in the
compensatory ballot. Consequently, a vote for a political party will count twice when
compared to a vote for an independent candidate.


57 Section 3(2)(a) provides that “[a]ll citizens are equally entitled to rights, privileges and benefits of citizenship”.
MHLANTLA J
36
[100] The applicant argues further that the split violates section 9(1) 58 of the
Constitution in that it arbitrarily differentiates between independent candidates and
political parties.

[101] In response, Parliament submits that to the extent there is a rational basis for the
200/200 split, there can be no infringement of sections 1(c) , 3(2)(a) or 9(1) of the
Constitution. For this proposition, Parliament submits that this Court has made it clear
that there can be no infringement of the right to equality even where there is a
differentiation between different categories of persons, provided that there is a rational
connection between the measures taken and the legitimate government purpose of
facilitating the effective exercise of the right to vote.

[102] Parliament argues that, on the evidence, the 200/200 split does not infringe the
right to equality as protected in the Constitution as it is rationally connected to a
legitimate government purpose in that the split (a) ensures a system th at results in
proportional representation in general, and (b) avoids the possible risk of overhang
which threatens constitutional imperatives.

[103] In response to the applicant’s sections 3(2)(a) and 9(1) challenge, the Minister
argues that the applicant’s con tention that a vote for an independent candidate carries
less weight for no reason is fundamentally flawed. The Minister submits that
independent candidates compete for the same quota as political parties in respect of
regional seats, and they do so before the same voters and on the basis of the same ballot
paper.

[104] With regard to the compensatory seats, the Minister argues that the
differentiation between political parties and independent candidates is not a result of a
devaluation of the votes for indepen dent candidates as suggested by the applicant.
Rather, it is because an independent candidate can only hold a single seat regardless of

58 Section 9(1) states that “[e]veryone is equal before the law and has the right to equal protection and benefit of
the law”.
MHLANTLA J
37
the percentage of the vote they get and this distorts proportionality. By contrast,
political parties can fill seats i n proportion to the percentage of votes they get , and the
compensatory seats can thus be used to restore proportionality. The Minister adds that
just as like cases should be treated alike, unlike cases should be treated differently. This
is a general axiom of rational behaviour.

[105] In its explanatory affidavit, the Electoral Commission submits that the assertion
made in the Atkins Report that the number of votes an independent candidate or political
party must obtain in order to gain a seat in the regional tier should be equivalent to the
quota for a compensatory seat is logically flawed.

[106] This submission in the Atkins Report, explains the Electoral Commission,
ignores the fact that South Africa has a two -tier electoral system for the national
election, and as such, different quotas apply to the regional tier and the compensatory
tier. The two tiers are inherently different and serve different purposes: they are not the
same in size of the constituency ; the voter base or voters’ roll ; and the seats in
contention. Consequently, the basic premise underpinning the applicant’s criticism that
the number of votes per seat in the regional tier must be equal or approximately
equivalent to the number of votes required per seat in th e compensatory tier is flawed .
At the regional tier , where independent candidates compete with political parties, the
quota per seat, as already stated, is exactly the same.

[107] The Electoral Commission further explains that, while the quota for a regional
seat is based on one ballot, the compensatory quota is based on the aggregate of two
ballots. The use of two ballots increases the quota for the allocation of compensatory
seats to be broadly similar to those that apply in the regional tier. On projecting th e
2019 election results onto the new electoral system, the Electoral Commission found
that the compensatory quota tier would likely be higher than it would be in six of the
nine regions.

MHLANTLA J
38
[108] The Electoral Commission therefore concludes that the argument that
independent candidates require approximately double the number of votes to gain a seat
in the National Assembly , when compared to p olitical parties, i s misguided. The
Electoral Commission submits that this distorts the true calculation required for political
parties to gain a seat in the National Assembly.

[109] The Atkins Report and the applicant accept that nominally the quota for a
compensatory seat is higher than the quota for a regional seat. However, and on the
basis of the assumption that a vote for a political party counts twice, the Atkins Report
concludes that political parties require half the votes needed by an
independent candidate to gain a seat in the National Assembly. Based on the 2019
election results, the Atkins Report finds that “the likely threshold for [political] parties
to gain a seat in the National Assembly is support from about 44,000 voters” whereas
independent candidates would need support from about 85,000 voters. In its
explanatory affidavit, the Electoral Commission aptly points out that, in concluding that
independent candidates need double the votes required by political parties to gain a seat
in the National Assembly, the Atkins Report assumes, without any basis, that voters
will vote for the same political party on both ballots.

[110] In New National Party, the applicant in that case argued that the requirement of
a bar-coded identity document excluded eligible voters who did not have bar -coded
identity documents and that this breached section 9(1) and 9(2) of the Constitution. In
considering this argument, Yacoob J said the following:

“Before this Court, the appellant advanced an argument based on what was alleged to
be a breach of sections 9(1) and 9(2) of the Constitution. However, it is clear from
what has been said in this judgment that although the documentary requirements in
issue may be said to differentiate between different categories of people, there is a
rational connection between the measure and the legitimate governmental purpose of
MHLANTLA J
39
facilitating the effective exercise of the important right to vote. No discrimination or
unfairness has been established.”59 (Emphasis added.)

[111] In Democratic Party ,60 this Court heard a similar application to that in
New National Party on appeal from the then Transvaal High Court. The relief sought
was an order that the bar -coded identity document requirement for registering and
voting in the 1999 elections was unconstitutional. The applicant argued that although
facially neutral, the documentary requirement for voting constituted indirect
discrimination against discrete vulnerable groups. 61 To support this contention, the
applicant had relied on surveys conducted by the Human Sciences Research Council
(HSRC) and from Opinion 1999: Voter Participation in the 1999 Elections .62 In
assessing the evidence provided by the applicant in support of its contention that the
documentary requirement infringed on the right to equality, Goldstone J for the majority
said the following:

“On the assumption that the opinions expressed in the HSRC and Opinion 99 reports
are correct, there is no evidence as to which category of persons referred to therein
might be among the millions of South Africans who, after the promulgation of the
Electoral Act, applied for and were issued with the necessary documents, and as a result
were able to register on the national common voters ’ roll. In the absence of evidence
showing that the impugned provisions have had the effect suggested by the DP, it
cannot be found that the provisions, on that account, were unconstit utional.”63
(Emphasis added.)

[112] This Court, therefore, rejected the section 9 equality challenge in
Democratic Party and in New National Party on the basis that the applicant failed to

59 New National Party above n 25 at para 48.
60 Democratic Party v Minister of Home Affairs [1999] ZACC 4; 1999 (3) SA 254 (CC); 1999 (6) BCLR 607
(CC).
61 Id at para 11.
62 A survey released on 10 November 1998 , which was conducted jointly by the South African Broadcasting
Corporation, Institute for Democracy in South Africa and Markinor.
63 Democratic Party above n 60 at para 12.
MHLANTLA J
40
provide evidence to support the alleged inequity. This Court in Democratic Party
continued and said the following:

“[I]n any event, it must be accepted that there are very few laws of general application
that will not, directly or indirectly, have the potential to affect different categories of
people in different ways, whet her, for example, by reason of where they live, their
standard of literacy or political beliefs. There is no evidence to show what the impact
of the Electoral Act has in fact been on the various categories of persons referred to by
the [Democratic Party]. Whatever the different impact, if any, might be, it is not
possible to determine whether such impact constitutes unfair discrimination within the
principles endorsed by this Court, unless it is established that such different impact is
caused by the impugned legislation, and is not the result of some other cause.”64

[113] From this Court’s judgments referred to in the precedin g paragraphs, it is clear
that the fact that a law affects different categories of pe ople differently does not prove
a violation of t he right to equality as provided in section 9 of the Constitution. The
person alleging the violation should provide evidence to that effect.

[114] This principle was reiterated by this Court in United Democratic Movement ,65
where the legislation being challenged made it possible for a member of the National
or Provincial Legislature or a municipal council to leave the party to which she or he
belonged without at the same time losing her or his seat (so-called ‘floor crossing’).
The legislation provided that, e xcept for the initial period, at least 10% of a party had
to leave that party for the legislation to apply . In assessing the challenge, this Court
held:

“The fact that a particular system operates to the disadvantage of particular parties does
not mean that it is unconstitutional. For instance, the introduction of a
constituency- based system of elections may operate to the prejudice of smaller parties,
yet it could hardly be suggested that such a system is inconsistent with democracy. If

64 Id at para 12.
65 United Democratic Movement v President of the Republic of South Africa [2002] ZACC 21; 2003 (1) SA 495
(CC); 2002 (11) BCLR 1179 (CC).
MHLANTLA J
41
defection is permissible, the details of the legislation must be left to Parliament, subject
always to the provisions not being inconsistent with the Constitution. The mere fact
that Parliament decides that a threshold of 10% is necessary for defections from a party,
is not in our view inconsistent with the Constitution.”66

[115] In my view , the applicant has failed to discharge the onus of proving that the
model articulated by Parliament infringes on the equal protection provisions provided
for in sections 3(2)(a) and 9(1) of the Constitution as alleged.

[116] Firstly, and as already demonstrated, the applicant’s argument tha t the
200/200 split arbitrarily differentiates between independent candidates and political
parties is without merit. Parliament has articulated and proven that there is a rational
basis for the split, which is to facilitate proportional representation and to avoid the risk
of overhang. In keeping with what this Court said in New National Party, as there is a
proven rational connection between the split and the articulated legitimate
governmental purposes, it cannot be maintained that the split violates sec tion 9(1) of
the Constitution.

[117] Secondly, the proposition that a vote for an independent candidate carries less
weight when compared to a vote for a political party is also without merit. As argued
by the respondents, independent candidates and political parties compete for the same
quota in regional elections and the votes carry exactly the same weight. There is no
differentiation in respect of regional seats.

[118] The applicant accepts the reservation of compensatory se ats for political parties
but rejects the reservation of 200 seats on the basis that independent candidates need
double the votes that political parties need to win a seat since a vote for a political party
counts twice. It is on this basis that the applicant alleges that a vote for an independent
candidate carries less weight when compared to a vote for a political party in violation
of section 3(2)(a).

66 Id at para 47.
MHLANTLA J
42

[119] This argument must also be rejected. As pointed out by the
Electoral Commission, the applicant fails to appreciate that, in a two-tier system, there
are two different quotas that apply in respect of compensatory and regional seats.
Within the different regions, there will also be different quotas. Consequently, the
proposition that the number of votes a n independent candidate or political party must
obtain in order to gain a seat in the regional tier should be equivalent to the quota for a
compensatory seat is logically flawed . Evidently, some difference must be had, the
issue lies in the extent of the difference and whether it serves to exclude independent
candidates disproportionately. This, though, the applicant has failed to establish.

[120] Finally, the applicant’s argument that the Electoral Act as amended devalues a
vote for an independent candidate is based on the assumption that a voter that votes for
a political party will vote for the same party on both the regional and compensatory
ballot, which in turn means that independent candidates need double the vote s needed
by political parties to gain a seat in the National Assembly. Put differently, the applicant
and the Atkins Report, without providing any basis to this Court, assume that vote rs
who vote for a political party will not split their votes across different parties in the two
ballots. By parity of reasoning, vote splitting beyond a de minimis (trivial) level by
voters who vote for political parties would be fatal to the applicant’s right’s violation
challenge.

[121] While the 2024 national elections will be the first election s South Africans will
get to vote using two ballots for the National Assembly, it is not inconceivable that
voters may choose to split their regional and compensatory vote s. There is in fact
evidence of vote splitting by South African voters in the last pr ovincial and national
elections.

MHLANTLA J
43
[122] By way of illustration, the 2019 report of the Mapungubwe Institute for Strategic
Reflection (MISTRA)67 looked at, amongst others, vote splitting in the 2019 provincial
and national elections in South Africa. 68 The report f ound that “ there were no cases
where the party-specific vote totals for the national and provincial votes corresponded
exactly”.69 In this regard, the report said the following:

“The ANC in both Election 2014 and 2019 was the only party that scored more v otes
nationally in all nine provinces than provincially. In addition, in all provinces but the
Eastern Cape, this ANC margin (or ‘vote edge’) grew from 2014 to 2019 (Table 9).
For all of the rest of the top-four parties (DA, EFF and FF+ in 2019) there were diverse
trends. Amongst the 2019 opposition parties that fared better at national than provincial
level, were the FF+ in six of the provinces, the EFF in four, and the DA in three
(KwaZulu-Natal, North West and Limpopo). In the Eastern, Northern and
Western Cape, and in KwaZulu-Natal, the EFF did better nationally than provincially.
In 2014, in the EFF’s first provincial elections, it fared better nationally than
provincially in six of the nine provinces. The DA in contrast fared better at the
provincial than the associated national level in 2019 in six of the provinces – compared
with doing better nationally than provincially in seven provinces in 2014. This
indicates that many of its su pporters endorsed another party. . . at the national level
rather than voting DA at both the provincial and national levels in 2019. The EFF’s
2019 national-level performance was affected similarly at several sites. In five of the
provinces, the EFF recorded fewer votes at the national than the provincial level (the
inverse therefore of the situation outlined regarding 2014, when its national
performance was better): Gauteng, North West, Limpopo, Mpumalanga and the Free
State. Hence, judged by these vote -splits, both DA an d EFF supporters in 2019
bolstered the ANC’s ‘Ramaphosa project.’”70


67 MISTRA is an independent think tank that carries out research on strategic challenges facing South Africa in
the political, economic and social space.
68 The election results for South African elections are available at the Electoral Commission’s website.
“Independent Electoral Commission” Electoral Commission, available at https://www.elections.org.za/pw/.
69 Booysen et al Voting Trends 25 Years Into Democracy: Analysis of South Africa’s 2019 Election
(MISTRA 2019) at 25.
70 Id.
MHLANTLA J
44
[123] Considering the above, it is clear that one cannot discount the possibility of vote
splitting with the introduction of the second ballot. When one considers the 2021 local
government elections, different political party support across the different provinces is
even more stark when compared to the 2019 provincial and national elections. The
Electoral Commission makes this exact point in its explanatory affidavit. The Electoral
Commission explains that “in municipal election where independent candidates stand
for political election as ward candidates, most voters that support independent
candidates also choose to support a political party on the second ballot”. Admittedly,
local governme nt elections are conducted differently to provincial and national
elections, not to mention that voters may well have different considerations when voting
in local government elections compared to national elections . However, the different
levels of suppo rt that political parties receive d in the provinces across the different
elections is nonetheless instructive with regard to vote splitting.

[124] It must be acknowledged that the opportunity for vote splitting is not the purpose
of the recent am endments to the Electoral Act. The possibility of vote splitting is an
unavoidable consequence of Parliament having introduced two ballot papers in order to
accommodate independent candidates to comply with this Court’s ruling in
New Nation II. Nevertheless, because two ballots have been introduced, voters will
now have a tactical opportunity to split their vote between regional and compensatory
ballot (or vote for the same political party on both ballots if they so choose) and one
cannot say with certainty that voters will not use this opportunity to split their vote .
This proposition is even more compelling when regard is had to the fact that some voters
appear already to be voting for different political parties in the national and provincial
elections as evidenced by the 2019 election results.

[125] Notwithstanding the above, I do not think an electoral law could ever be properly
formulated on the basis of an assumption as to how people will vote as the applicant’s
challenge suggests. As already explained, the applicant’s challenge to the 200/200 split
is premised on the assumption of identical voting across the two ballots for political
parties. An electoral law must always assume a complete freedom for the electorate to
MHLANTLA J
45
vote as they please, including splitting their vote where an opportunity to do so exists.
Section 19 of the Constitution guarantees citizens a right to make free political choice.
If, when formulating an electoral law, the Legislat ure had to take into account how
people have voted in th e past, it might have to chang e the law every election cycle as
voters are free to vote as they choose and often vote differently in different elections.

[126] In addition to the above , it must be stressed that the applicant has failed to
provide compelling evidence to support the alleged inequity between independent
candidates and political parties as required by the jurisprudence of this Court. The
applicant has not been able to prove that (a) the split is arbitrary and (b) a vote for an
independent candidate carries less weight. Bearing in mind that the mere difference in
effect of the split is not, in and of itself, evidence of unequal treatment in the manner
protected against in the Constitution.

[127] For these reasons, the section 3(2)(a) challenge also fails.

Section 19(2) and 19(3) of the Constitution
[128] The applicant argues that the 200/200 split violates section 19(2) 71 of the
Constitution in that it undermines the fairness of the outcome of the elections. The
applicant argues further that the split violates the “right to vote” and to “stand for public
office” as provided in section 19(3) 72 of the Constitution. The applicant submits that
section 19(3) entails a right to a vote that has equal weight and the right to stand an
equal chance of being elected. In support of this submission , the applicant relies on

71 Section 19(2) provides that “[e]very citizen has the right to vote in free, fair and regular elections for any
legislative body established in terms of the Constitution.”
72 Section 19(3) provides:
“Every adult citizen has the right —
(a) to vote in elections for any legisl ative body established in terms of the Constitution, and to do
so in secret; and
(b) to stand for public office and, if elected, to hold office.”
MHLANTLA J
46
international law 73 and jurisprudence from Canada, 74 the United States of America 75
and Australia,76 which recognise the principle of one person, one vote . The applicant
also refers to the United Nations Human Rights Committee’s General Comment 25 on
Political Participation in which this principle is reflected as:

“[T]he vote of one elector should be equal to the vote of another. The drawing of
electoral boundaries and the method of allocating votes should not distort the
distribution of voters or discriminate against any group and should not exclude or
restrict unreasonably the right of citizens to choose their representatives freely.”77

[129] Therefore, the applicant submits that the 200/200 split infringes on the right to
equal benefit of the law and the right to stand for public office as votes for
independent candidates weigh less than votes for political parties and makes it more
difficult for an independent candidate to be elected.

[130] Parliament submits that the split does not infringe on the right to free and fair
election, nor does it infrin ge on the “right to vote” and to “stand for public office ”.
Parliament notes that there is no internationally accepted definition of the term “free and
fair election”, a nd that this is ultimately a value judg ement. Parliament submits that
this Court has distilled some elements in Kham78 as being fundamental to the conduct
of free and fair election s, which include, amongst others, that: (a) a person entitled to
vote should be registered to do so, and those not entitled to vote should not be permitted

73 See, for example Article 23(1)(b) of the American Convention of Human Rights (22 November 1969),
Article 5(c) of the International Convention on the Elimination of All Forms of Racial Discrimination
(21 December 1965), Article 13(1) of the African Charter on Human and Peoples’ Rights CAB/LEG/67/3 rev. 5,
21 I.L.M. 58 (1982) , and Article 21 of the Universal Declaration of Human Rights (10 Dece mber 1948) UNGA
Res 217 A(III).
74 See, for example Reference re Prov. Electoral Boundaries (Sask.) [1991] 2 SCR 158.
75 See, for example Gray v Sanders, 372 U.S. 368 (1963) at 379 and Reynolds v Sims, 377 U.S. 533 (1964) at 554-
5.
76 See, for example McGinty v Western Australia [1996] HCA 48; (1996) 186 CLR 140 at paras 8-9.
77 United Nations Human Rights Committee (HRC), CCPR General Comment No. 25: Article 25 (Participation
in Public Affairs and the Right to Vote), The Right to Participate in Public Affairs, Voting Rights and the Right
of Equal Access to Public Service, 12 July 1996, CCPR/C/21/Rev.1/Add.7 at para 21.
78 Kham v Electoral Commission [2015] ZACC 37; 2016 (2) SA 338 (CC); 2016 (2) BCLR 157 (CC).
MHLANTLA J
47
to do so; and (b) the Constitution protects not only the act of voting and the outcome of
elections, but also the right to participate in elections as a candidate and to seek p ublic
office.

[131] Parliament argues that the Electoral Amendment Act gives effect to section 19(2)
and 19(3) in that the Act allows for citizens to stand for political office and to hold
office if elected. The Electoral Amendment Act also allows for citizens to vote in an
election, and they can choose to either vote for an independent candidate (at the regional
tier) or a political party. The requirements and quota for a single regional seat are the
same for both independent candidates and political parties. The compensatory seats are
created and reserved for political parties for reasons already discussed.

[132] The difference in outcome , argues Parliament, emanates from the inherent
difference between political parties and independent candidates . The principle of
one person one vote is not violated by the Electoral Amendment Act. According to
Parliament, the international law and foreign jurisprudence cited by the applicant do not
advance the applicant’s case because those countries do not have a constitutiona l
injunction for an electoral system that results, in general, in proportional representation.
Further, the Australian and Canadian cases that the applicant relies on do not require
strict equality of voting power. Instead, in Reference re Prov. Electoral Boundaries,79
the Canadian case that the applicant relies on, it was recognised that some deviation
from the concept of equal representation can be acceptable if it is justifiable under the
Canadian Charter.

[133] Parliament further argues that, pursuant to this Court’s judgment in
New National Party,80 for the section 19 challenge to be successful, the applicant must
not merely show that Parliament has performed its function in a manner which has
resulted in a denia l of the vote to a substantial number of citizens . Rather, they must

79 Reference re Prov. Electoral Boundaries above n 74.
80 New National Party above n 25.
MHLANTLA J
48
show that the machinery, mechanism or process provided by the
Electoral Amendment Act is not reasonably capable of ensuring the vote for those who
want to vote and take reasonable steps to exercise their right to vote.

[134] The Minister, in response to the section 19 challenge, submits that the applicant
has failed to demonstrate how the freeness and fairness of the election would be
compromised by the 200/200 split as contend ed. This challe nge, argues the Minister,
is inconsistent with the concession made by the applicant and Mr Atkins that the
200/200 split will achieve proportionality. The Minister also states that the argument
that votes for independent candidates count for less is witho ut merit since
independent candidates compete for the same quota as political parties in respect of
regional seats and the difference in compensatory seats has been justified.

[135] Like Parliament, the Minister also states that the jurisprudence of Australia,
Canada and the United States of America does not assist the applicant. According to
the Minister, while the South African Constitution prescribes an electoral system that
must result, in general, in proportional representation, the electoral systems of Australia,
Canada and the United States of America are constituency -based. In
constituency- based systems, the political party or candidate who wins the highest
number of constituencies wins the election regardless of whether they have the highest
level of proportional support. Thus, overall party support is inconsequential. The
Minister submits that this is the exact opposite of what our Constitution envisages and
what it guards against. Further, the case law referred to concerned issues surrounding
whether voters in constituencies were evenly distributed from a geographical
standpoint. According to the Minister, this is irrelevant to the issues in this matter.

[136] As already demonstrated, the Electoral Commission denies that the 200/200 split
causes a vote for an independent candidate to count for less towards the outcome . It
MHLANTLA J
49
argues that this argument is premised on fundamental misconceptions by the applicant
and Mr Atkins as discussed in this judgment.81

[137] For reasons I have stated in the preceding paragraphs, the applicant has failed to
establish that a vote for an independent candidate carries less weight than a vote for a
political party.82 As such, the argument that the 200/200 split infringes the section 19
rights on the basis that a vote for an independent candidate does not have equal weight
and that independent candidates do not stand an equal chance of being elected , is
without merit.

[138] The applic ant fails to get out of the starting block s on this challen ge.
Notwithstanding this shortcoming in the applicant’s case, it is important to consider
how this Court has dealt with section 19(2) and 19(3) challenges in other cases.

[139] I think it apt to remind ourselves of the test for determining a rights violation set
out by this Court in Ex Parte Minister of Safety and Security: In Re S v Walters,83 where,
in assessing whether the impugned provisions limited the right to life, to human dignity
and to bodily integrity as protected by the Constitution, Kriegler J said the following:

“As observed at the outset, the Bill of Rights spells out the fundamental rights to which
everyone is entitled and which the S tate is obliged to respect, protect, promote and
fulfil. An enactment (like section 49) may limit these rights only if – and to the extent
that – the limitation can be justified under section 36 of the Constitution. Otherwise it
has to be declared invalid under section 172(1). This is essentially a two-stage exercise.
First, there is the threshold enquiry aimed at determining whether or not the enactment
in question constitutes a limitation on one or other guaranteed right. This entails
examining (a) the content and scope of the relevant protected right(s) an d (b) the
meaning and effect of the impugned enactment to see whether there is any limitation of
(a) by (b). Subsections (1) and (2) of section 39 of the Constitution give guidance as

81 See [105] to [109].
82 See [117] to [125].
83 Ex Parte Minister of Safety and Security: In Re S v Walters [2002] ZACC 6; 2002 (4) SA 613 (CC); 2002 (7)
BCLR 663 (CC) (Walters).
MHLANTLA J
50
to the interpretation of both the rights and the enactment, essentially requiring them to
be interpreted so as to promote the value system of an open and democratic society
based on human dignity, equality and freedom. If upon such analysis no limitation is
found, that is the end of the matter. The constitutional challenge is dismissed there and
then.”84 (Emphasis added.)

[140] The first leg of the Walters test examines “the content and scope of the relevant
protected right(s)”. In the present case, this refers to the political rights protected under
section 19(2) and 19(3) of t he Constitution. In New National Party , this Court
recognised the importance of the right to political participation, where it stated that
“[t]he importance of the right to vote is self -evident and can never be overstated. . . .
[T]he right is fundamental to a democracy for without it there can be no democracy.”85

[141] In August,86 Sachs J, commenting on the importance of the right to vote, stated
that “[t] he universality of the franchise is important not only for nationhood and
democracy. The vote of each and every citizen is a badge of dignity and of personhood.
Quite literally, it says that everybody counts.”87

[142] The Constitution in various sections provides for the regulation of the exercise
of the right to vote. As stated in New National Party, the effect of these measures on
the effective exercise of the right to vote is the following:

“(a) National, provincial and municipal elections must be held in terms of an
electoral system which must be prescribed by national legislation.
(b) The electoral system must, in general, result in proportional representation.
(c) Elections for the national assembly must be based on t he national common
voters roll.

84 Id para 26.
85 New National Party above n 25 at para 11.
86 August v Electoral Commission [1999] ZACC 3; 1999 (3) SA 1 (CC); 1999 (4) BCLR 363 (CC).
87 Id at para 17.
MHLANTLA J
51
(d) Elections for provincial legislatures and municipal councils must be based on
the province’s segment and the municipality’s segment of the national common
voters roll respectively.”88

[143] Commenting on the inter-relatedness of the rights conferred in section 19(2) and
19(3), this Court in New Nation II said the following:

“The condition [the singular condition that section 19(3) imposes for the exercise of
the right to vote] reveals the inter-relatedness between the right to vote and the right to
free, fair and regular elections which is guaranteed by section 19(2). If the elections
are no t free and fair, there can be no proper exercise of the right to vote and
consequently the content of the right to vote itself would be emasculated. And that
would place at risk the entire democratic project. This illustrates that the right to vote
is vi tal to our democratic order, not only with regard to who gets the honour of
exercising political power, but also in respect of the policies to be adopted in governing
the country.”89

[144] Therefore, the right to vote and the right to free and fair elections are closely
linked. While the right to vote gives content to the right to free and fair elections, the
right to free and fair elections ha s implications for how the right to vote must be
exercised.

[145] In Kham, this Court set aside the results of municipal by-elections held in various
wards in the Tlokwe Local Municipality between August and December 2013 on the
ground that they were not free and fair due to irregularities in voter registration. In a
unanimous judgment, Wallis AJ held that there is no internationally accepted definition
of what constitutes a free and fair election, and that ultimately, whether an election is
free and fair is a value judgement which must always be assessed in context.90 A Court

88 New National Party above n 25 at para 13.
89 New Nation II above n 14 at para 152.
90 Kham above n 78 at para 34.
MHLANTLA J
52
required to make this value judg ement must weigh all the evidence and determine
whether the constitutional requirement that an election be free and fair was satisfied.91

[146] The Court then went on to distil the following elements as fundamental in the
conduct of free and fair elections:

“First, every person who is entitled to vote should, if possible, be registered to do so.
Second, no one who is not entitled to vote should be permitted to do so. Third, insofar
as elections have a territorial component, as is the case with municipal elections where
candidates are in the first instance elected to represent particular wards, the registration
of voters must be undertaken in such a way as to ensure that only voters in that
particular area (ward) are registered and permitted to vote. Fourth, the Constitution
protects not only the act of voting and the outcome of elections, but also the right to
participate in elections as a candidate and to seek public office.”92

[147] The cases cited in the preceding paragraphs give meaning to the “content and
scope” of the section 19 political rights.

[148] The second leg of the Walters test requires of us to examine the “meaning and
effect” of the 200/200 split as provided in the amended Electoral Act to assess whether
there is any limitation to the section 19 political rights as alleged by the applicant . I
have already explained the meaning of the amended Electoral Act, and in particular, the
impugned 200/200 split.93 It is the “effect” of the impugned provisions on the section 19
rights that we are concerned with at this stage.

[149] Before determining the question of the effect of the impugned provisions , it is
worth noting that Walters did not concern section 19 political rights, but rather, at issue
was the constitutional validity of the provisions of section 4 9(1)(b) and 49(2) of the

91 Id at para 90.
92 Id at para 34.
93 See [25] to [33].
MHLANTLA J
53
Criminal Procedure Act 94 which essentially permitted force to be used when carrying
out an arrest. The Walters two-prong test nonetheless applies with authoritative weight
to this case as it is a general approach for determining a rights violation – regardless of
the content of the right(s) alleged to have been violated.

[150] Now returning to the effect of the 200/200 split on the section 19 political rights.
As shown above, this Court in Kham distilled elements fundamental to free and fair
elections as protected under section 19(2) of the Constitution; these elements include
that (a) persons entitled to vote should be registered to do so, and (b) the protection of
the right to participate in elections as a candidate and to seek public office. In my view,
the elements distilled by this Court in Kham are met under the 200/200 split. As
Parliament points out in its submissions, the Electoral Amendment Act entitles every
adult citizen to a vote, a nd they can, in keeping with this Court’s ruling in
New Nation II, vote for either a political party or an independent candidate. The
Electoral Amendment Act also allows citizens to stand for political office and to hold
office if elected.

[151] The contention that a vote for a political party counts double to that of an
independent candidate, which as the applicant argues leads to a limitation of the
section 19 rights, is, as I have already demonstrated, without merit. This argument is
based on the assumption that voters will not split their vote, an assumption which I have
already shown to be unsustainable.

[152] Further, and as stated in New National Party, the effective exercise of the right
to vote as protected under section 19(3) includes, among others, the following: (a) that
elections must be held in terms of an electoral system which must be prescribed by
national legislation; (b) an electoral system that must, in general, result in proportional
representation; (c) that elections for the National A ssembly mu st be based on t he
national common voters ’ roll. The Electoral Amendment Act gives effect to these

94 51 of 1977.
MHLANTLA J
54
principles in that (a) it is national legislation that prescribes an electoral system, (b) the
200/200 split prescribed therein results in proportional repr esentation, and (c) it
provides for a national common voters’ roll for the election of the National A ssembly.
Therefore, it cannot be said that the Electoral Amendment Act limits section 19(3).

[153] It is worth reiterating what this Court said in United Democratic Movement,95
that is, laws of general application will invariably , directly or indirectly, have the
potential to affect different categories of people in different ways , and the fact that a
law affects different categories of people differently is not evidence of a violation of the
Constitution.

[154] While it is quite clear that provisions which disenfranchise citizens or preclude
them from standing for office are unconstitutional, this must be contrasted with
instances where the design of the electora l system (which entails legitimate competing
options available to Parliament) is challenged – that is, the manner in which Parliament
chooses to give substantive content to the right to political participation.96

[155] The applicant was required to show that the measures adopted by Parliament
constitute a limitation of the political rights alleged. For the reasons stated, the applicant
has failed to discharge this onus and as such the 200/200 split passes the second leg of
the Walters test. It follows that the section 19 challenge must also fail.

Foreign jurisprudence and international instruments
[156] Before I conclude, I wish to provide a brief response to the submissions on the
foreign jurisprudence relied on by the applicant. I agree with the respondents tha t it

95 United Democratic Movement above n 65.
96 In August above n 86, the Court declared that prisoners were entitled to vote and directed the Commission and
Minister of Home Affairs and Minister of Correctional Services to make arrangements necessary to enable the
prisoners to register as voters on the national common vote rs’ roll. In NICRO above n 3, the Court struck down
legislative provisions which deprived convicted prisoners serving sentences of imprisonment without the option
of a fine of the right to participate in elections during the period of their imprisonment. In Richter above n 3 and
AParty above n 30, the Court declared invalid legislative provisions which precluded South African citizens
abroad from voting. Most recently, in New Nation II above n 14 this Court declared the Electoral Act
unconstitutional because it completely precluded independent candidates from contesting elections.
MHLANTLA J
55
does not assist the applicant’s case. The South African Constitution prescribes an
electoral system that must result, in general, in proportional representation. The
electoral systems of Australia, Canada and the United States of America , which the
applicant relies on , are constituency -based, where the overall party support is
inconsequential. The Minister correctly submits that this is the exact opposite of what
our Constitution envisages and what it guards against.

[157] We must never forget that the C onstitution represents a decisive break from the
past. The Constitution is intentional in affording all South African citizens the right to
political participation on equal terms. While this means that all South African citizens
now have the right to vote and stand for office, I believe that, in the context of this case,
the more important consideration is the break from geographic voting constituencies
which allowed the National Party to come to power in 1948 with 37% of the total vote
while the United Party had 49%. Throughout its rule, the National Party designed a
system which entrenched their power and excluded most South Africans from
participating in public affairs. The respondents are, therefore, correct that our electoral
systems are incomparable.

Section 36 limitation analysis
[158] In the light of my conclusion that the impugned schedule does not violate any of
the fundamental rights alleged by the applicant, it is not necessary to conduct the
justification analysis under section 36 of the Constitution.

Conclusion
[159] The applicant has not established that the impugned legislation is irrational, nor
that it infringes a provision in the Bill of Rights. Therefore, it has not made out a case
to justify the declaration of constitutional invalidity. Accordingly, the application must
be dismissed.

MHLANTLA J
56
Costs
[160] As established in Biowatch,97 in order to combat the unduly chilling effect of
adverse costs on constitutional litigation, this Court held that the general rule for an
award of costs in constitutional litigation between a private party and the state is that if
the private party is successful, then it should have its costs paid by the state; however,
if the private party is unsuccessful, each party should then pay its own costs. 98 In this
matter, the applicant has not been succ essful, therefore, each party must pay their own
costs.

Order
[161] The following order is made:
1. Direct access is granted.
2. The application is dismissed.
3. There is no order as to costs.


97 Biowatch Trust v Registrar Genetic Resources [2009] ZACC 14; 2009 (6) SA 232 (CC); 2009 (10) BCLR 1014
(CC).
98 Id at para 43.


For the Applicant:




For the Second and Third Respondents:


For the Fourth Respondent:



For the Fifth Respondent:


Counsel for the Amicus Curiae:




T G Madonsela SC, H J De Waal SC ,
T Palmer, Y S Ntloko and C Louis
instructed by Strauss Daly Incorporated.


K Pillay SC and L J Zikalala instructed
by the State Attorney.

S Budlender SC , A Nacerodien and
M De Beer instructed by the State
Attorney.

A Bham SC and J Bleazard instructed by
Moeti Kanyane Incorporated.

M du Plessis SC, P Maharaj-Pillay and
C Kruyer instructed by Webber
Wentzel.