Elections, Constitutionalism and Political Stability in South Africa


The history of South Africa’s electoral politics dates back to the mid-19th century. Historical familiarity with electoral procedures and the development of a strong party system helps to strengthen modern electoral politics. General elections under universal suffrage have been held three times, in 1994, 1999 and 2004. Good electoral administration has included a strong commitment to conflict mediation procedures, using the skills generated through trade union-based collective bargaining in earlier decades. Electoral reform to enhance parliamentary accountability constitutes the main policy challenge in this sphere today.

Entrenchment of Racial Segregation

South Africa was first colonised by the Dutch East Indies Company in 1652. During the Napoleonic Wars the British assumed control of the Cape Colony. An exodus of Dutch-speaking pastoralists beyond the borders of the colony led to the establishment of independent settler republics. Within these republics, a creolised form of Dutch, Afrikaans, became the dominant language. A second British colony, Natal, was established on the East coast in 1830. From 1863 Indian indentured labourers were recruited to work on Natal sugar African kingdoms were gradually subjugated in a series of wars which lasted through the nineteenth century but in several instances the terms of conquest allowed considerable autonomy to the better organised African authorities, including the Zulu monarchy and the less centralised Transkeian kingdoms.

In 1910 the Union of South Africa was constituted as a self-governing Dominion from four British colonies, Cape, Natal, the Orange River Colony and the Transvaal. The Orange River and Transvaal Colonies were formerly self-governing Afrikaner republics which had lost their autonomy during the Anglo-Boer war. Each of these territories had a history of electoral politics before Union but in the case of the Republics the franchise was completely racially exclusive, limited to white settlers. In the Cape and Natal the franchise was qualified by property and education requirements. From 1853, when representative government began, the voters roll in the Cape incorporated Africans and Coloureds, people of mixed racial descent, and by 1910 their numbers represented 15 per cent of the electorate. In Natal, though, in 1909 only six Africans could vote and the Asian franchise was limited to those who had qualified by 1893. The South Africa Act which established the Union ensured that these arrangements would continue subject to any change a subsequent Union parliament might make.

In 1930 white women were enfranchised without qualifications and in the following year the vote was extended to all white men, though the qualifications in the Cape and Natal remained for black voters. African voters were removed from the common roll in 1936 and in the Cape qualified Africans instead could vote for three white ‘Native Representatives’. In addition, nation-wide a complicated system of indirect elections was established to send four (white) native senators to the upper house. At the same time, Africans could elect 12 African representatives to an advisory Native Representative Council. These arrangements persisted until their abolition in 1959. Between then and 1994 the only elected bodies which represented Africans were those which functioned within the boundaries of the ethnic homeland system.

By 1994 ten separate homeland governments existed, each with their own arrangements for elected representation. The first elected homeland ‘Territorial Assembly’ was instituted in the Transkei in 1963. Of its 109 members only 45 were elected, the rest held their seats as ex officio chiefs. Meanwhile, 50 000 Coloureds were removed from the common roll for parliamentary elections in 1957 after enlargements of the Senate and the Supreme Court enabled the government to remove the entrenched status of Coloured voting rights. Instead, Coloureds were accorded voting rights for special white representatives; by 1963 Coloured voter registration had declined by 80 per cent to under 10 000. Indians and Coloureds were removed from the common municipal rolls in Natal and the Cape in 1964 and 1968 respectively. They were re-enfranchised in 1984 through the ‘Tricameral’ constitution, which established a separate House of Representatives for Coloured ‘own affairs’ and a corresponding House of Delegates for Indians.

The franchises for all these segregated institutions were qualified by assigned ethnic or racial status, but within such limitations they were open to all adults. Generally, all South African elections were conducted within single member constituencies in which the winner was the candidate with the largest share of the votes. Racially segregated elections for African, Indian, and Coloured voters did not usually attract enthusiastic participation.

Despite the endurance of a ‘dominant party’ political regime between 1948 and 1994 amongst white South Africans, their more or less racially exclusive electoral politics featured vigorous competition between parties and fairly high levels of voter commitment. Through much of its rule the National Party benefited from constituency demarcations that enhanced the political significance of certain geographical groups of voters; effectively some votes were worth five times as much as others. Voter interest in elections was sustained by the degree of ideological hostility that existed between parties and the strong sense in which until the 1980s political divisions amongst whites corresponded with ethnic and other kinds of social distinction. This was changing by the late 1980s, though. In 1987, for example, half the English speaking population voted for the (traditionally Afrikaner) National Party.

Within this institutional framework a strong party system developed. The National Party could trace its origins to 1913 when it was formed as the heir to Afrikaner republican traditions. It found its social base among small farmers and white sharecroppers as well as a growing Afrikaans-speaking proletariat. Before the Second World War, South Africa was mostly governed by a succession of political organisations that embodied a social coalition between larger landowners and big business. Of these the most enduring was the South African Party (later the United Party), which was in power between 1910 and 1924 and subsequently between 1933 and 1948. In and out of office, Afrikaner nationalists encouraged a process of local manufacturing based initially on the establishment by parastatal steel and electricity companies. Racially restrictive labour laws protected white workers’ jobs and status.

The Second World War accelerated industrial growth and urban expansion. This process considerably enlarged the number of black industrial workers and magnified the potential threat to white livelihoods they represented through competition as a much cheaper source of labour. These wartime social developments helped to boost the popularity of Afrikaner nationalism which from the 1930s had become an advocate of much more stringent and codified forms of racial segregation. In 1948, the National Party was elected into government. It would retain power for the next 46 years.

Until the mid-1970s, National Party administrations were committed to the implementation of Apartheid, an initially vaguely defined programme of white racial supremacy. As it became systemised its measures would include strict social segregation, a harshly repressive labour regime, including tight ‘influx’ controls on the geographical movement of black labour, and the confinement of African land rights to the native reserves which had been established through earlier legislation in 1913 and 1936. During the 1960s, nearly three million people were forcibly removed from white farms and black peasant free-holdings and despatched to bleak settlements in the reserves. These were now dignified with the status of ethnic homelands. From 1959 efforts were made within the homelands to promote local political authorities which in four cases, Transkei, Ciskei, Bophuthatswana and Venda, would eventually acquire independent national status.

Within the arena of white parliamentary politics, the National Party administration was confronted by a demoralised opposition initially led by the United Party. The first parliamentary organisation to offer white voters any kind of principled opposition to Apartheid was the Progressive Party formed in 1959. Under various different names, the Progressives slowly expanded their electoral support, and by the late 1980s they attracted the backing of approximately a quarter of the white electorate. A series of reforms, though, beginning with the official recognition of black trade unions in 1979 and the abolition of ‘influx control’ in 1986, had narrowed the distance between the government and its liberal parliamentary opponents. Its more important adversaries were outside the boundaries of institutional politics, leading a popular insurrection in the factories and black townships.

Political Challenge to Racial Segregation

The most important organisation that represented black South African aspirations was the African National Congress (ANC), established in 1912 in protest against the impending restrictive land legislation. Its formation was prompted partly by the relative success the Natal and Transvaal Indian Congresses under the leadership of Mohandas Gandhi had enjoyed in mobilising civil disobedience campaigns against racially discriminatory legislation. Until the 1950s though, the ANC only occasionally employed militant tactics, relying instead on efforts to influence government policy through the Cape franchise and white liberal intermediaries. From 1921 a Communist Party of South Africa represented an alternative source of popular political inspiration, and by 1928 the Party had a mainly black member-ship, though it remained very small. During the 1930s and 1940s, both Communists and the ANC contested segregated elections for native representation as well as the local township Advisory Boards. The Communist Party enjoyed more success in municipal politics. In the Cape, coloured voters returned several communist councillors to office during the 1940s. During this period Communists, Gandhists in the Indian Congresses and the ANC leadership worked increasingly closely with each other, and, under pressure from a militant ‘African nationalist’ ANC Youth League, began to embrace a more militant range of tactics.

The ANC’s radicalisation was hastened by the official suppression of the Communist Party in 1950, whose African members thereafter concentrated on strengthening the ANC. At this stage the ANC confined its membership to Africans though it headed a multiracial ‘Congress Alliance’ which included the Indian Congresses, a white (and mainly communist) Congress of Democrats and a non-racial South African Congress of Trade Unions. A decade of strikes, demonstrations, civil disobedience, and boycotts culminated in 1960 in the government banning the ANC and its more radical offshoot, the Pan-Africanist Congress. In 1961 both organisations, the ANC in conjunction with the Communist Party, began planning armed insurgencies. Within four years most of their leadership was either in prison or in exile. In exile the ANC established headquarters first in Dar es Salaam and later in Lusaka; in 1969 whites, Indians and coloured were admitted to membership of the exile Organisation and from 1985 ‘non-Africans’ were elected to its national executive.

During the 1960s another social and political force emerged, namely the black trade union movement. The South African economy experienced structural transformation financed by a large influx of foreign capital, which brought about important changes in the relations of production. These changes resulted in a shortage of semi-skilled and skilled labour unable to be met by Whites or immigrants. This allowed for large numbers of black labour to enter the labour market. But the racially based labour practices created deep resentment amongst black workers and confirmed a growing sense that the political and economic systems were profoundly unfair and unjust. These conditions provided a platform on which Black workers were able to build industrial unions. ‘By concentrating large numbers of workers in production, the material conditions for a new kind of shop floor based trade unions had been created by the early 1970s. The January and February 1973 strikes in Durban highlighted these changes, when an estimated 100 000 workers broke the decade of industrial peace by taking to the streets to demand higher wages’ (Webster 1978:178). Between 1973 and 1976 those involved in disputes ranged annually between 30 000 and 100 000 (Webster 1978:179).

In 1976 a schoolchildren’s revolt against compulsory instruction in Afrikaans spread across the country, helping to create the opportunities and the following which enabled the ANC to rebuild its clandestine networks and initiate a modest programme of urban guerrilla warfare. Rather than become involved in direct political activity, the independent black trade union movement embarked on building strong decentralised structures to survive attacks by the repressive state apparatus. The democratic shop floor structures were predicated on ‘principles of worker control, accountability and mandating of worker representatives, and to developing working class leadership in the factories’ (Webster 1978:179).

The trade union movement was fast becoming a powerful force on the shop floor. It practised a policy of engagement with the state and employers, predicated on strong democratic shop floor structures and practices. At the same time as union membership was on the increase powerful social movements were developing and emerging in the townships and schools, such as civic organisations and youth and student congresses. By the mid-1980s these new community-based movements in civil society began to challenge the state directly, both at central and local level, supplying the ingredients for the most massive and sustained social insurrection ever experienced by any South African government.

The social crisis deepened, as did the various strategies of the pro-democratic forces in South Africa; i.e. the political independence and strategic flexibility of the trade union movement, in contrast to the political and military struggle then waged by the ANC and its internal allies (Adler & Webster 1995:81). The trade union movement had to re-evaluate their strategy for fear of being marginalised and isolated. They joined with the social movements and participated actively in general strikes and political activities from the mid-1980s. In 1985 the Congress of South African Trade Unions (COSATU) was formed ‘signalling a strategic compromise in which the integrity of the industrial unions was acknowledged while the new federation committed itself to participation in national democratic struggle under the leadership of the ANC’ (Adler & Webster 1995:82). This resulted in 1986 in mass action that included strike action by the trade unions, township rebellion and massive civil disobedience campaigns orchestrated by the Mass Democratic Movement (MDM) – reconstituted as such after the United Democratic Front had been banned. This rebellion’s impact was magnified by an international movement in favour of economic sanctions against South Africa. By the mid-1980s discreet talks had begun between government representatives and both the imprisoned and exiled ANC leadership.

In 1990, after most whites had voted in favour of the National Party reform proposals, F.W. de Klerk repealed the bans on the ANC and other prohibited organisations. He ordered the release from their life sentences of the ANC’s leaders, including the guerrilla leader Nelson Mandela, first imprisoned in 1962. Mandela was elected as the ANC’s president one year later. In this role he helped lead a multi-party negotiation process which resulted in the adoption at the beginning of 1994 of a transitional constitution under which would be held South Africa’s first fully democratic elections. This constitution reflected a widespread recognition of the merits of compromise amongst the most significant political actors.

The resulting elections would be held under a system of proportional representation and all parties that achieved a ten per cent share in the vote would be entitled to cabinet positions. Nine elected regional governments would share power with the central administration and allow smaller parties the possibility of executive control. A final constitution would be negotiated after the elections by both houses of parliament serving as a Constituent Assembly. This second constitution would have to conform to many of the essential principles contained in a liberal Bill of Rights. To ensure such continuity and to safeguard constitutional rights more generally, the transitional legislation established a permanent Constitutional Court.

State and Constitutionalism

Before 1994, South Africa was governed through the conventions of parliamentary sovereignty. The South Africa Act of 1910 entrenched the status of two official languages, English and Afrikaans, as well as the non-racial common roll in the Cape, but the protection afforded to both could be overcome by a two-thirds vote in parliament. In 1956 the National Party administration enlarged the upper house of parliament, the Senate, to achieve the majority it needed to remove coloured voters from the Cape roll. From 1961 South Africa had a written constitution but this could be altered in its entirety, except for the provisions relating to official languages, by a simple parliamentary majority.

The adoption of a transitional constitution for the period 1994-1999 and the subsequent enactment of a more permanent constitution in 1996 marked South Africa’s transition to a system in which the constitution has supreme authority. It can be amended by parliament but changes in certain of its most important provisions need a 75 per cent parliamentary majority. All other sections are protected by a two-thirds majority requirement for their amendment. To date there have been only eight minor amendments concerning technicalities of public administration, supported by all parties in parliament. A Constitutional Court can test the constitutional validity of any legislation. The state president appoints the courts’ members for 12-year terms selecting the judges from a list of nominations submitted by the Judicial Service Commission, a body itself partly appointed by the executive, partly by parliament and partly by the legal profession. Chapter 9 of the 1996 Constitution established or enhanced the independent status of a number of other ‘state institutions supporting constitutional democracy’ including a Public Protector, an Auditor General, an Electoral Commission and Commissions for human rights, language protection, and gender equality. Though the 1996 Constitution calls for a separation of powers between the legislature, executive and the judiciary, in fact the executive is accountable to parliament and the President exercises certain judicial powers, including the appointment of judges and rights of pardon over criminals. In 2002 President Thabo Mbeki exercised this latter right with his pardon of thirty prisoners who had been refused amnesty by the Truth and Reconciliation Commission.

Since 1994, the South African state has assumed what constitutional lawyers have termed a ‘quasi-federal’ character inspired mainly by the German model. In the division of powers between central government and the nine elected regional or provincial administrations the national government has the power to override laws passed by the regional legislatures, except in a few minor fields. The provincial administrations have very limited authority to raise taxes, primarily through vehicle licensing and levies on gambling. and depend chiefly on funds allocated by central government. Central authority can recover such funds if they are misused. Central government can also assume control over any section of provincial administration if it fails to perform effectively.

Between 1994 and 1999, the transitional constitution compelled the executive to be composed through a coalition of party representatives with posts being distributed roughly proportionately to those parties that achieved more than five per cent of the votes in the national elections. The National Party withdrew from the Government of National Unity in 1996, partly because of its failure to impose its will on the drafting of the 1996 Constitution. It wanted the maintenance of the transitional power sharing provisions so that minority parties would keep their entitlements to cabinet seats. Regional executives also composed themselves through prescribed coalitions during this period. The constitutional requirement for power sharing did not feature in the 1996 constitution that determined cabinet appointments after 1999. The transitional constitution also protected the jobs of all existing civil servants during the first post-1994 administration. In 1999 a second chamber of the parliament in Cape Town was constituted, the National Council of the Provinces, replacing a senate whose members had been nominated by the National Assembly. Its 90 members are made up of nine equal delegations representing each region, drawn from the regional legislatures. The Council was intended to review legislation with regional implications and over such matters can exercise veto powers.

South African governments are formed by the President, a member of the National Assembly who is elected during its first meeting after an election. The combination in the Presidency of the roles of head of state with head of government dates back to 1994; before then, state presidents were titular and ceremonial figures who replaced in their constitutional function the British Queen’s representative, the Governor General when South African became a republic in 1961. Up to 1984, prime ministers led governments. When elected, the president vacates his or her parliamentary seat and appoints a cabinet of ministers with particular portfolios as well as a deputy president. Since 1994, the president can serve only two five-year terms. South Africa’s first president after universal suffrage elections in 1994 was Nelson Mandela, who served one term, declining on grounds of age to serve another. His successor was Thabo Mbeki. Before their appointments both were elected (though secret ballots) as the presidents of the African National Congress at delegate conferences.

Since Thabo Mbeki’s accession, the President’s office has accumulated enormous functions and powers. Mbeki was one of the two deputy presidents during the Mandela government (the other was F.W. de Klerk until 1996). Even as Mandela’s deputy Mbeki was largely responsible for the day-to-day management of the administration as well as providing the main authority in policy formation. At the end of 1998, Mbeki’s office was merged with Mandela’s for the launch of a ‘super-presidency’. After the 1999 election, the new office was to be supported by more than 300 staff, its own cabinet minister and three director-generals as well as a budget of R70 million. This was to enable it to perform its role as ‘a powerful coordinating structure for government policy and action’. A Cabinet Office with a Secretariat within the Presidency is responsible for this co-ordinating function. The intention was that the office would exercise considerably more control over ministries to ensure a much more ‘integrated’ (a favoured term in modern South African public administration) allocation of resources. The Minister without portfolio within the presidency controls government communications. All ministries are expected to subject policy documents and draft legislation to the Presidency’s Policy Co-ordination and Advisory Services Unit. This trend towards executive power makes the South African system, as in the case of France, a hybrid between parliamentary and presidential government.

Ministers remain members of parliament and are accountable to parliament through question time sessions as well as through their obligations to various parliamentary portfolio and standing committees. As in Britain, therefore, the heads of the various civil service departments are usually professional politicians rather than administrative specialists or qualified experts. The President may also appoint deputy ministers. Under the terms of the 1996 constitution, there are no restrictions on the president’s power to appoint and dismiss ministers, except that on appointment they should normally be members of parliament. The president may appoint to his cabinet two people from outside parliament, a provision intended to facilitate the inclusion of technocrats. In practice, if the President wanted to appoint additional people from outside parliament, as the party leader he can easily create a parliamentary vacancy and appoint that person to it. As will be explained below, in general elections parties compete for parliamentary seats through offering single lists of candidates and consequently, depending on the internal rules of the political parties concerned, party leaderships can decide who should be nominated for public office and whether they should be replaced between elections. For this reason, the executive’s power over parliament is very much greater than in most systems in which executives are accountable to parliament. In theory, South African Presidents and their executives can lose office through votes of no confidence, but given the de facto power over parliamentary office holding exercised by the President, an ANC backbencher revolt would be extremely unlikely. This point was underlined when a deputy minister, General Bantu Holomisa, a former dictator of the Transkei, was dismissed from office in 1998 (and expelled from the ANC) after accusing a member of cabinet of accepting a bribe from a casino proprietor. Holomisa proceeded to found a new party, the United Democratic Movement (UDM), which enjoyed a measure of success against the ANC in the vicinity of the former Transkei in the 1999 election.

During four decades of the National Party’s rule, the state buttressed its powers through manipulating judicial appointments in high courts, granting the police extensive powers through security legislation and building in the South African Defence Force a formidable military capacity. Between 1979 and 1989, during the time of P.W. Botha’s leadership, military commanders acquired considerable political influence through their membership of the State Security Council, a cabinet committee that during this period became the key site of government decision making. During the 1990s, though, much of the state’s coercive capacity eroded.

In certain respects this diminution of executive authority has been positive. To the government’s credit, the senior levels of the judiciary are much freer of executive influence than they were ten years ago. All judges are appointed through a constitutionally sanctioned process that limits executive discretion and ruling party influence. Though senior judicial appointments since 1994 have included several lawyers who before their appointment to the bench were well known ANC supporters, court judgements demonstrate plenty of evidence of robust judicial independence, despite complaints by cabinet ministers about the judiciary usurping their role as policymakers. This independence has been especially obvious with respect to the Constitutional Court. The court demonstrated its independence decisively in 1996 when it refused certification to a draft of the final constitution because it reduced the powers of regional governments. In two key judgements the Court has ruled on how the government should allocate public resources, in respect of its constitutional obligation to supply shelter for groups of forcibly evicted Cape Town squatters and, most recently, in the provision of anti-retro-viral medication for HIV-AIDS patients.

Public respect for legal institutions and constitutional rights, though, requires more than the autonomy and integrity represented by the Constitutional Court and the other senior levels of judicial decision making, commendable as these features of South African justice are. For most citizens, courts are inaccessible and inefficient. Magistrate’s courts especially are understaffed, under-financed, corrupt and demoralised. As a consequence of huge caseloads legal proceedings are extremely slow. Nearly half of South Africa’s prisoners are people awaiting trial. Several thousand cases a year do not reach trial because criminal syndicates bribe court officials to lose or destroy dockets. The continuing evident popular preference for informal modes of retributive ‘street justice’ is testimony to a public view of the courts as time consuming, expensive and ineffectual. Meanwhile, a comparably inefficient public administration continues to frustrate the ‘delivery’ of constitutional entitlements, court judgements notwithstanding.

The Nuts and Bolts of the Electoral Process

In contrast to the shortcomings of its public service, South African electoral administration and politics and practice present rather bright prospects for the country’s future democratic consolidation. When asked in a survey about the efficacy and importance of voting and elections, 62 per cent of South Africans believe that ‘the way you vote could make things better in the future’. Similarly 64 per cent agree that elections in general (rather than their own vote) make a difference, with 73 per cent of citizens agreeing that ‘Our government was voted into power by accepted procedure’ (Friedman 2001:8).

One sardonic observer noted in 1958 that ‘South African elections have a deceptive air of reality about them’ (Farquharson 1958:134) and still today all the main parties conduct their campaigning in a manner to suggest that victory is almost certainly within their grasp. They use opinion polling, market research, image consultants, advertising agencies and many of the other techniques of modern electioneering, increasingly gearing their approach to a population that relies more and more on television for its understanding of public affairs. Television political advertising is prohibited, but party events are calculated to project compelling visual images to news cameras and voters are presented with choices that focus heavily around the presidential attributes of the party leader. Despite the provision of public funding for parties already with parliamentary representation, the main contestants depend heavily on unofficial sources of finance, the ANC still in 1999 collecting much of its R160 million election treasury from friendly foreign governments, while the Democratic Party (DP) and the New National Party (NNP) drew their support mainly from local business.

Despite their generally predictable outcome, South African elections are contested energetically and seriously. In 1994, in certain parts of the country, organisations that were still influenced by their experience as military formations, continued to defend their territory, forcibly excluding their rivals from their ‘home’ locations. By the 2000 elections, such instances of ‘no-go zones’ (there were 165 during the 1994 elections) had become exceptional and even in Black townships in small rural centres – previously the neighbourhoods in which activist intolerance was most pronounced – canvassers from both the Democratic Alliance (DA) and the ANC could visit the same streets on the same days.

Arguably, the electoral system promotes the formation of socially inclusive political parties and civil electioneering. The types of party system which develop, in particular the number and the relative sizes of political parties in parliament, are heavily influenced by the choice of electoral system. Electoral systems can also impact on the way parties campaign and the way political elites behave, thus helping to determine the broader political climate (Reynolds & Reilly 1997:8). It was these considerations that were taken into account at the World Trade Centre in Kempton Park, Johannesburg, in the pre-1994 all-party negotiations to map out the future political dispensation of the country.

In South Africa elections are organised through national list proportional representation. 200 national assembly members are selected from lists submitted by provincial party organisations and 200 are drawn from the parties’ national lists. Effectively the country serves as a single constituency.

Seat allocation is through the Droop quota. The absence of a representation threshold and the employment of the largest remainder method for the allocation of surplus seats ensure that very small parties enjoy parliamentary representation. Early drafts of the electoral law put the threshold for parliamentary representation at five per cent of the national vote, but, in a concession to the smaller parties, the African National Congress and the National Party agreed in early 1994 to drop the ‘mandatory’ threshold. The provincial governments are elected on the same basis, at the same time, on a separate ballot paper. Parties can choose to contest either a single region or several regions, or the national assembly and regions. The parties pay a deposit of R10 000 for each regional list and R90 000 for the national assembly list when they register with the Independent Electoral Commission (IEC). Parties have discretion over the principles they employ to compose their lists.

In theory the size of the national assembly should be determined on the basis of one seat for every 100 000 of the population, with a minimum of 350 seats and a maximum of 400. Provincial seats in the national assembly should reflect their proportional share of the population and these are decided after political party representations by the IEC. Provincial legislatures have between 30 and 60 seats, again on a principle of one seat for every 100 000 people. Elections must be held within 90 days of dissolution of the national assembly and provincial legislatures and are proclaimed by the President and the Provincial Premiers. Municipal elections, while conforming to the constitutional requirement for proportional representation, combine ward councillors and list representatives in a mixed member proportional system.

Party list proportional representation, is often criticised for offering opportunities for extremist parties to obtain office – sometimes as power brokers in shaky coalitions (as in Israel). In South Africa’s case, the double-ballot, List PR electoral system contains strong incentives for moderation given the spatial distribution of the electorate. Every vote a party receives counts in their seat allocation and hence all parties are encouraged to seek votes outside their core support or base areas. In doing so, they will attempt to project their programs in terms which make them generally appealing. The system is less likely to promote ethnic bloc voting than a geographically defined form of representation.

Party leaders can include on their lists people who might not win popular support in electoral contests focused around individual candidates: members of racial minorities or women, for example. The drawback is that parliamentarians hold their seats at the behest of party leadership and they are not directly accountable to voters. In the South African case this has produced a parliament which tends to be deferential towards the executive. Electoral experts and observers maintain that the choice of a List PR system for the first democratic elections in South Africa in 1994 was critical to creating an environment of inclusiveness and reconciliation which has made post-apartheid South Africa an example of hope and political stability to the rest of Africa (Reynolds & Reilly 1997:70).

Since 1994 the IEC has managed South Africa’s elections. The IEC is responsible for administering national, provincial and municipal elections as well as evaluating their integrity and certifying their results. The present IEC was established after the adoption of the 1996 constitution with the appointment of five commissioners, one of whom must be a judge. A broadly representative all-party committee within the national assembly nominates the Commissioners. None of them are allowed, at their time of appointment, to have overt political affiliations. The IEC has administered two national elections since its establishment and a municipal election in 2000.

In the 1999 general election, the IEC’s arrangements included an elaborate registration process in which strenuous efforts were made to register as many people as possible. This was not the case in the 1994 elections when there was no voter registration requirement – voters were required to present a prescribed identity document, of which there was a wide range. Prior to the 1999 elections legislation imposed on prospective voters quite exacting identity documentation requirements – many voters had to acquire new documents – but these requirements did not particularly discriminate in their effects against likely supporters of any particular parties – and they were not of the IEC’s making. Registration was conducted over a six-month period and during this time the IEC’s administrative proficiency clearly improved.

In general, voting itself in 1999 and 2000 was well organised, though long queues on polling days suggested that the IEC’s faith in satellite technology to demarcate roughly even electoral districts was at times misplaced. Even so, despite the long waiting times, voters’ verdicts were favourable. An exit poll based on responses from 11 140 voters at 214 randomly selected polling stations found that only three per cent of respondents felt that the procedures during balloting and vote counting were not free and fair. Parties did not query the outcome of the 1999 election – they too seemed to accept the results as equitable. In national elections voter participation is high – 90 per cent of eligible voters in 1994 and 80 per cent of the registered electorate in 1999. Voter turnout in municipal elections is much lower though, an apparent reflection of public disenchantment with the performance of local councillors, or as has been argued by other commentators, also because this seems to be a world-wide phenomenon.

The public legitimacy of the electoral arrangements over which the IEC presides is evident in the absence of any serious disputes over their results. Acceptance by losing parties of electoral outcomes is partly attributable to the effectiveness of a conflict management system instituted in 1994 and elaborated in successive polls. The functioning of this system is described in the next section.

Conflict Management

There are intense debates and theories about the nature and outcomes of conflict management processes in South Africa. Some analysts argue that the various forums where trade unions and management sought to reach recognition and agreements about shop floor issues in the late 1970s and the 1980s represented a form of conflict management. Others argue that for the emerging black trade unions in the 1970s to have engaged company management in these forums was to allow for co-optation and compromise. A third view was that unions engaged with company management, and later the state, as a strategic response to prevent a unilateral restructuring of labour relations in circumstances in which the state was imposing excessive control. Whatever the merits of such arguments, it is very likely that the experience of institutionalised collective bargaining helped trade unionists to acquire negotiation skills and that such skills were subsequently transferred to community organisations and political organisations.

Arguing in a similar vein to the critics of trade union participation in collective bargaining forums, in 1990 there were powerful advocates of the view that any negotiations with the apartheid state by the banned ANC and its internal allies, namely the MDM and civic organisations such as SANCO (South African National Civic Organisation), had the potential to undermine the course of the national democratic struggle. But the conclusion that ultimately prevailed was that it was important to enter into negotiations with the De Klerk government to ensure that any transitional arrangements that were agreed upon represented the interests of all groups in South Africa. The fear was that if the apartheid government was allowed to shape economic and political policy alone, it would not only ignore the interests of the MDM, but would also create new realities which would be very difficult to undo and which would then dictate the course of the transition (Friedman & Shaw 2001:192).

The economy had been in recession since 1982, resulting in retrenchments and mass unemployment. The crisis was met with unprecedented levels of mobilisation and resistance from both labour and communities over political and economic issues – the highest strike levels were experienced in the history of South Africa, and rent strikes, bus boycotts, school boycotts, consumer boycotts, anti-Constitution campaigns and stay-aways became a permanent feature of the political terrain from 1984 onwards. At the same time the ANC in exile stepped up the armed struggle. This period of resistance created new alliances internal to South Africa, where labour, community organisations and student organisations joined together. Massive stay-aways became the order of the day, despite the continual harassment and repression by the apartheid state (Webster 1978:185). The first few months of 1984 constituted a period of intense urban conflict, leading eventually to the declaration of the State of Emergency on the 22nd July 1984.

It was during this period of upheaval and conflict in South Africa that a group of lawyers and academics established the first independent mediation service (IMSSA). The service provided independent mediation and arbitration services in the industrial relations arena. In the early 1990s IMSSA extended its services to include community mediation as communities were experiencing an upsurge of political conflict, especially in the East Rand townships in the ‘old’ Transvaal and KwaZulu-Natal. Skilled mediators were trained in each of the provinces to deal with the conflicts.

The decision to convene the pre-1994 multi-party negotiations to negotiate transitional arrangements from apartheid rule to a democratic dispensation resulted in a consensus amongst political elites on the legitimacy of the democratic system that could only bode well for political stability in the country (Friedman 2001:6). These negotiations played an important conflict management and conflict prevention role, even although the process was not fully inclusive as several groups decided not to participate. These included the governments of Bophuthatswana and the Ciskei, and parties at both ends of the political spectrum including the Inkatha Freedom Party (IFP), Conservative Party (CP), the Azanian Peoples Organisation (AZAPO) and the Afrikaner Weerstandsbeweging (AWB).

The outcome of the negotiations represented a series of compromises, some of them short-term, others more likely to endure. These compromises may be fundamental to the long-term sustainability and stability of the country. We would not have reached the consensus we did as a nation without these negotiations, which put in place the necessary mechanism to usher in transitional arrangements, such as an electoral system that is inclusive, that represents all stakeholders, is transparent in its management and in the final analysis produces legitimate and credible results, acceptable to all, and important for this discussion, a National Peace Accord.

By January 1993, 11 regional peace committees and 54 local peace committees had been formed throughout the country with skilled, trained mediators to deal with disputes and conflicts that arose in the communities. These committees had multi-party representation, including political organisations, civic organisations, local authorities, churches, business, trade unions and members of the security forces. Their role was to intervene to avert incidents of political violence and to promote the prospects of peace in the areas over which they had jurisdiction. Peace committees were able to anticipate and avert incidents of violence by opening lines of communication between key parties and, in numerous instances such as the march on KwaMadala Hostel by residents of Boipatong following the massacre in that community, on-site mediation by peace committee officials averted imminent confrontation. The peace committees also operated as forums of accountability. The police, for example, were called upon to account for progress in investigations into crimes of political violence. The committees further co-ordinated humanitarian relief efforts for residents driven from their homes because of the political violence. Negotiation and facilitation skills training were also offered at both regional and local levels to deal with conflicts and disputes which arose from political party contestation, civic groupings with political affiliations, student grievances, the white right’s insecurity about change and so forth.

The experiences of negotiation, consensus building and conflict management processes from the 1970s onwards provided the necessary experience for these processes to be included into the legislative framework for the 1994 elections. The significance of this election as well as the way the managing body, the Independent Electoral Commission, was structured and formulated reflects the context of the transitional process of which they were part. The Independent Electoral Commission Act, No 150 of 1993, stipulated that the Commission was to consist of an Election Administration Directorate, a Monitoring Directorate and an Adjudication Secretariat, each operating independently of the others. The task of the Monitoring Directorate was basically to monitor the performance of the Election Administration Directorate, a rather novel idea for one arm of an organisation to monitor and police another. The Monitoring Directorate established a panel of mediators and the Adjudication Secretariat a panel of adjudicators whose role was to investigate, and when appropriate, to mediate election related disputes. If unsuccessful, the disputes and/or complaints were referred to the Adjudication Secretariat. These internal conflict management processes ensured that assistance was on hand if and when disputes and conflicts threatened to destabilise the pre-, during and/or post-election periods. The panels were a great success and were hailed by all parties as having been vital to the successful outcome of the elections.

In addition the Electoral Act favoured the involvement of political parties in all aspects of the election process. The Party Liaison Committee structure provided a space for parties to share their concerns about what was happening on the ground with the electoral authority and each other. This complemented the work that had already been done by the Peace Committees in managing political conflicts and disputes that erupted in various provinces. These structures provided avenues of communication between the political parties, the IEC and the Transitional Executive Council (TEC) respectively. They played a crucial conflict management and conflict resolution role and in the final instance secured the electoral process.

In the months preceding polling in 1999, the IEC established similar structures with a wide network of local and provincial conflict resolution and conflict management structures, drawing upon prominent local personalities to sit on these committees. The Party Liaison Committees were once again resurrected and supplied forums in which local instances of violations of the electoral code of conduct could be addressed. These were once again very important in terms of ensuring that the integrity of the electoral process was upheld. A significant indication of their success was the absence of any litigation following the election as well as the acceptance of the results by all the parties that contested the poll.

Similar arrangements were instituted during the 2000 local government poll, and once again, despite the impressive diversity of contesting parties that included a large number of local associations, the Commission recorded no serious objections to any of the procedures or outcomes. This was largely because of the role played by the IEC’s conflict management panels that altogether responded to (a notably modest) 193 complaints, in most cases to the recorded satisfaction of the parties concerned (Conflict Management Programme 2001:23-24). South Africa has had a long history and association with trying to build consensus around the ‘rules of the political game’ be it at trade union level in the 1970s, or with students and scholars in the 1980s. The dividends of this experience are evident in its relative success in comparison to most new democracies in addressing electoral conflict since 1994.

Electoral System Reform Imperatives

During colonial and republican administrations, South Africa used the British First-Past-The-Post (FPTP) system for electing its public representatives for parliament. This system was in place until the 1993 Interim Constitution and the elections of 1994. After these elections the election reform debate once again surfaced as the country still had to move towards its permanent constitution that, in principle, left the door open for some adjustments to, amongst other things, the electoral system. To this end the Constitutional Assembly convened its Theme Committee Two to debate electoral options for consideration by the lawmakers (Faure 2003:1). Though it became evident after the Arniston Bosberaad (indaba/meeting) in 1996 that the 1994 PR list system would be retained for the 1999 elections, speculation as to what type of electoral system South Africa should change to for the 2004 election gained momentum once again.

The imperative to revisit the model used for the 1994 and 1999 elections was partly constitutional. The provisions of the 1996 Constitution relating to an electoral system did not extend beyond the 1999 elections. The Constitution requires that an electoral system should be introduced through the enactment of national legislation. Thus at present there is no electoral system that prescribes for the conduct of national and provincial elections.

Against the background of ethnic and regional divisions of the apartheid era, PR offered the promise of coalitions and multiparty representation of diverse interests, and forced parties to campaign beyond narrow ethnic or regional bases of support if they hoped to win a majority of seats. It is also the best system in the world for improving the representation of women. South Africa now has 30 per cent representation of women in Parliament and a similar share in local councils. But in addition to the electoral system, it is largely as a result of the ANC’s voluntary commitment to 30 per cent women’s representation on its party lists that South Africa now ranks in the top ten countries in the world for women in parliament.

By the same token, there are disadvantages associated with the South African experience of the current electoral system. For example, the closed party list system may bestow excessive power on party elites – especially where weak intra-party democracy exists. There are also issues of accountability, informed by the argument that members elected through PR have maintained very weak ties to a fixed territorial constituency (even with the assignment of informal constituencies). This is not an automatic flaw; other formal consultative arrangements may contribute to enhancing accountability. However, in a closed list system, as in variants of PR, it is the party rather than the member who ‘owns’ the seat, and this system enhances party power (and independents are unable to compete for provincial or national seats). A related problem is the fact that voters do not know who is going to represent them, whereas some might prefer to associate a face, as well as a party, with their vote. Moreover, just as the PR system may encourage coalitions, so it may enable fragmentation of the party system.

On the basis of these theoretical and practical issues, it seemed timely for South Africans to reflect on the country’s experience with the current electoral system. South Africa had now held two sets of national and provincial elections in 1994 and 1999, as well as municipal elections in 1995/96 and 2000. The next national and provincial elections are scheduled for 2004. However, there are additional dynamics of a political and institutional nature that make the timing of such a review even more imperative.

Since early 2001 there have been signs from several quarters that an electoral system review is to be launched. First, the IEC has undertaken a wide-ranging review of its procedures and regulations, with an eye to improving its conduct of elections. Second, the Minister of Home Affairs has indicated that his ministry would initiate a review of the constitutional provisions related to the electoral system, particularly the prohibition on Members of Parliament (MPs) crossing the floor to join another party. Third, a special task team reporting to the Office of the President, and to be led by Dr. Frederik van Zyl Slabbert, was formed, and its brief was to review the current electoral system.

This review demanded that a complicated technical topic involving highly motivated interest from political parties and voters alike should be handled with a deft concern for the consolidation of South Africa’s democracy. The ultimate test of the contribution of any electoral system review should not be the degree to which partisan interests feel they benefit, but the degree to which the rules of the electoral system are structured to encourage the maximum possible participation by political parties and voters alike, so that their interests are represented. Of course individual parties are always inclined to calculate under which system they feel they would most benefit and it is not desirable to seek to exclude such normative considerations from the debate. But it was very important for this process to remain non-partisan. Additional issues are considered in any electoral system reform process. For example, there are administrative and financial considerations for the IEC. Elections are expensive; the IEC budget for the 1999 elections was close to R1 billion, with political parties spending a combined estimate of R300-R500 million.

Whichever electoral system Cabinet adopts finally, it will impact on the demarcation process, the voters’ roll and the drawing of electoral districts. The electoral schedule will also be adversely affected if the debate is left too late, and more importantly, any changes in the electoral system must be communicated to voters. The municipal elections of 2000 were very complicated with some voters casting three ballot papers. Possible reform could limit voters to two ballot papers in a more consistent electoral system combining PR and FPTP.

Electoral system reforms may also impact on the formation of South Africa’s party system. Independents and smaller parties captured less than five per cent of the combined vote in both sets of local elections – the majority of the electorate continue to vote for a relatively small number of parties, with a strong majority voting for the ANC, and there was relatively little ‘vote-splitting’ by voters with ward and PR ballots. South Africans therefore display high degrees of party loyalty, which is reinforced by the electoral system.

Moreover, those who argue that the introduction of more constituency representation would improve the relationship between voters and their representatives must be balanced against the fact that voter turnout in South Africa’s two sets of municipal elections was approximately 50 per cent despite the fact that these are the only elections in which voters may choose constituency representatives. But the fact that public opinion data has shown that many South Africans do not feel that MPs (elected through PR) do a good job of staying in contact with them shows that there are no easy answers.

The Electoral Task Team (ETT) identified a set of core values against which electoral options should be evaluated: fairness, inclusiveness, transparency and accountability. It was the last of these considerations that proved most contentious both in the team’s closed sessions and in the public discussions it sponsored. Team members agreed that the retention of proportionate representation should be an essential principle of any future electoral design, notwithstanding a 1992 Constitutional Court opinion (on proposed floor crossing legislation) that this was not a core principle in the Constitution. Common sense suggests that in South Africa, the coincidence between traditional communal divisions of race and ethnicity with party political affiliations makes it desirable to ensure that the most significant parties are represented in parliament in accordance with the respective significance of their support. Two-thirds of the team in a majority report recommended the adoption of a mixed member proportional system in which half of parliamentary representation would be through 69 three to seven-member constituencies. The remaining representatives would be selected from national lists and would be allocated to parties in such a way as to compensate those parties who would be under-represented through the results of the constituency elections. This would be more complicated than the present national list system but the same principles were used in the 2000 South African local government elections and they seemed to be understood and favoured by the electorate. The new system, the ETT argued, should only be legislated after the 2004 elections, for it would require extensive administrative preparations.

A minority of the team argued for the retention of the present system, a position also supported by the ANC whose leaders maintain that a mixed member proportionality (MMP) system would probably work against the more popular parties adopting women or members of racial minorities as candidates. The ANC would win most constituency elections and hence its share of the list membership would be small. Maintaining its gender equality quota and its racial diversity in parliament would require it to nominate substantial numbers of women and minority candidates in the constituency polls.

Conclusion: Key Policy Challenges

These proposed reforms represent key policy challenges. In 1996 three million fewer people voted than in 1994. In both local government elections, 1995-1996 and 2000, less than half the electorate participated in the polls. Survey evidence suggests that South Africans are less likely than the citizens of other countries in the region to make contact with their elected representatives and such exercises also elicit high levels of citizen distrust of politicians (Mattes, Davids & Africa 2000:35-38,71). The present electoral system provides few incentives for parliamentarians and members of provincial legislatures to make contact with citizens between elections; the voluntary system through which parties assign their representatives to various constituency offices works at best unevenly. Public cynicism about parliament is accentuated by a system that permits party leaders to ‘re-deploy’ or dismiss parliamentarians at will. The resulting insecurity amongst parliamentarians who hold their positions at the behest of party leadership weakens any predispositions by backbenchers to exercise oversight with respect to the executive.

South African elections are well administered and contested fairly. They benefit from a diversity of well-established parties, several of which have substantial popular followings. Their integrity and legitimacy is protected by a constitutional order to which the ruling party accords appropriate deference. They result in institutions that are impressively inclusive in their membership but that remain very remote from ordinary citizens. A more personalised system of representation might well enhance the public accountability of legislatures and government more generally.


  1. Adler & Webster 2000. Trade Unions and Democratization in South Africa, 1985-1997. Johannesburg: Witwatersrand University Press.
  2. Conflict Management Programme 2001. Report: South African Local Government Elections. Johannesburg: Electoral Institute of Southern Africa.
  3. Farquharson, R. 1959. South Africa, 1958, in Butler, D.,Williams, P., Harrison, M., Pelcynski, Z., Chubb, B., & Farquharson, R., Elections Abroad. London: Macmillan.
  4. Faure, M. 2003. Report of the Electoral Task Team. Presentation at Electoral Task Team Conference, January.
  5. Friedman, S. 2001. A Partial Peace: Democracy and Conflict in Post-Apartheid South Africa. Johannesburg: Centre for Policy Studies.
  6. Friedman, S. & Shaw, M. 2001. Trade Unions and Democratisation in South Africa, 1985-1997.
  7. Lodge, T. 1999. Consolidating Democracy. Johannesburg: Witwatersrand University Press.
  8. Mattes, R., Davids, Y.D. & Africa, C. 2000. Views of Democracy in South Africa and the Region: Trends and Comparisons, in IDASA, Southern Africa Development Barometer. Cape Town: IDASA. October.
  9. Reynolds, A. & Reilly, B. 1997. The International IDEA Handbook of Electoral System Design. Stockholm: Institute for Democracy and Electoral Assistance.
  10. Webster, E. 1978. The Rise of Social-Movement Unionism: The Two Faces of the Black Trade Union Movement in South Africa, in Frankel, Pines & Swilling, State Resistance and Change in South Africa.