Urban Forum

, Volume 18, Issue 4, pp 247–264

Dilemmas of Representation in Post-apartheid Durban


DOI: 10.1007/s12132-007-9019-0

Cite this article as:
Low, M., Ballard, R. & Maharaj, B. Urban Forum (2007) 18: 247. doi:10.1007/s12132-007-9019-0


This paper aims to provide a sketch of the ways in which ‘formal’ institutions of democratic representation worked in practice in Durban/eThekwini in the 2000-2004 period. In so doing, it assesses how the representation of eThekwini’s citizens functioned at both the ward and metropolitan levels. After outlining the formation of the new metropolitan political arena, we consider the relationships amongst political parties at Metro and Ward levels, and, in particular, explore some contextually specific forms of democratic practice which emerged through the interaction of proportional representation and ward representation over time. The election of councillors as such does not resolve a series of dilemmas concerning how to institutionalise democratic representation within a racially diverse, spatially divided, and rapidly changing metropolitan area. Electoral-representative aspects of a urban democratisation are not a straightforward ‘formal’ framework alongside which other processes of democratic discussion, mobilisation and contestation can be easily situated.


Local democracy Durban Post-apartheid 


In the decade leading up to the 2000 elections, local government in South Africa went from a state of permanent crisis and collapse, which few would have described as meaningfully democratic, to a relatively stable and broadly legitimate set of democratic institutions. Against this background, this paper aims to provide a sketch of the ways in which ‘formal’ institutions of democratic representation worked in practice in Durban/eThekwini1 in the 2000–2004 period. It addresses some of the ways in which formal democracy was put into practice in eThekwini during these crucial years for South African urban governance. Our research assesses how the representation of eThekwini’s citizens functioned at both the ward and metropolitan levels. We consider the relationships amongst political parties and, in particular, some contextually specific forms of democratic practice that emerged.

Elections as decision-making devices and as key participatory rituals signalling the emergence of a new South Africa have certainly been necessary conditions for establishing urban democracy. However, the premise of the research underlying all the papers collected in this issue is that these central institutions of ‘formal’ liberal democracy are not usefully isolated from other (sometimes more informal) modes of representation, participation and mobilisation if urban democratisation is to be adequately described and evaluated (see, for example, Saward 2003; Young 2000). Importantly, in the case of Durban/eThekwini, the election of councillors as such does not resolve a series of dilemmas concerning how to institutionalise democratic representation within a racially diverse, spatially divided and rapidly changing metropolitan area.

In outlining the post-2000 political landscape in Durban/eThekwini, we deal with four themes. First, we provide a brief sketch of the reforms of the 1990s in which the fragmented apartheid city was drawn together into a consolidated metropolitan unit. Secondly, we discuss the changing balance of power within the eThekwini council, outlining the electoral performance of the key parties – the African National Congress (ANC), Democratic Alliance (DA), New National Party (NNP), Inkatha Freedom Party (IFP) and the Minority Front (MF). Although the major party, the ANC required alliances and cooperative arrangements with other parties such as the MF, NNP and IFP to secure its political dominance. Thirdly, we move to ward level. It is difficult to convey the diverse and complex political landscape that exists at the local level in South Africa. This diversity cannot, of course, be separated from the history of the apartheid city itself. Ward level research revealed complicated political arrangements at the local level. Some wards are predictably dominated by parties because they have electorates with a relatively homogenous political identity (e.g. former white suburbs or former black townships). Others are more heterogeneous as a result of changes in residential patterns or the way new ward demarcations attempted to combine different types of settlement. Finally, we examine some of the emergent and informal forms of local representation, including an interesting development in which councillors elected on party lists have been deployed to particular wards to represent parties that had not won that ward. We conclude by emphasising the importance, in evaluating urban democratisation, of processes of inter-party negotiation and the practices of representation ‘on the ground’. Together, these inform our sense that the formal-electoral aspects of democratic change, although important, need to be considered alongside other emergent practices of representation, some of which are discussed in the other papers in this issue.

Reforming Local Government

The challenges inherited by the post-apartheid government in relation to local government were vast, and their resolution was seen to be fundamental in addressing the injustices of the past. Apartheid urban planning was characterised by racially fragmented and discontinuous land use and settlement patterns, haphazard, dysfunctional and inefficient spatial ordering, land use mismatches, low level population density and the concentration of the poor in relatively high density areas on the peripheries and the rich in the core intermediate urban areas (Hindson et al. 1992). Only white people elected political representatives until reforms in the 1980s, and the introduction of Black Local Authorities in this decade hardly represented democratic advance. These were described contemptuously as “fictitious councils” by one ANC representative in our research (Buthelezi, 22–04–2004)2. At the time of political transition, the challenges for local government were described by Bennington and Hartley as: (1) democratising local government to make it representative of all people, (2) establishing political alliances across diverse interests, (3) creating mechanisms for delivering services to those previously excluded and (4) restructuring the municipal staff to achieve racial and gender balance (1994: 7–8). Two years later, the new Constitution of South Africa (1996) confirmed the key role that the local government would play in transformation, development and the achievement of social justice (Republic of South Africa 1996; van Donk 1998).

The restructuring of local government in South Africa has been described as internationally unique (Swilling et al. 1995: 16). Its core imperative was to achieve consolidation and integration. As anti-apartheid organisations had been recognised in the 1980s, urban inequalities had been created and reinforced through a fragmented system of local government based on race (Swilling et al. 1991: 176). The Local Government Transition Act of 1994 provided a framework for the disbanding of race-based municipalities, the scrapping of apartheid laws relating to local government and the establishment of transitional councils (Cameron 2006: 77; van Donk and Pieterse 2006: 112). At the core of this process were two rounds of local government boundary redefinition. In preparation for the first post-apartheid local government elections in 1995/1996, about 1,260 black and white local authorities were merged to form 843 municipalities (Giraut and Maharaj 2003). Cities were organised into two tier administrative structures with Transitional Metropolitan Councils covering cities, which is comprised of smaller “substructres” or Metropolitan Local Councils (Cameron 1999). The first democratic local government elections in KwaZulu-Natal took place on 26 June 1996.3 In spite of a history of political conflict (Coleman 1995, Morris and Hindson 1992), the relatively peaceful local elections marked the beginning of a new era for local government.

Transitional Local Councils stood until the second post-apartheid local government elections in 2000.4 In preparation for these elections, a second round of demarcation took place, which resulted in a further consolidation to 284 municipalities (Cameron 2006: 81). In terms of the Municipal Demarcation Act (1998), an independent Board was established to determine the criteria and procedures for the delimitation of municipal boundaries (Sutcliffe 2002). According to the Board’s criteria, the existing metropolitan boundaries in eThekwini were not cohesive or integrated and were unsustainable economically.

The demarcation of the present boundary has not helped to resolve all of the problems associated with service delivery and infrastructure in areas outside the urban core, especially in informally settled areas. Poverty, inadequate housing and infrastructure, land tenure, unemployment, crime, lack of skills and environmental problems continue to manifest themselves in these areas, which remain functionally linked to the Metro core.5

Metropolitan areas were reorganised into single tier “unicities”. The reforms of the 1990s consolidated the developmental function of local government (Republic of South Africa 1998, p. 15). Local authorities are exhorted to focus on achieving developmental outcomes, such as the provision of basic infrastructure and services, the creation of integrated cities and liveable environments, the encouragement of local economic development initiatives and the empowerment of communities. The transitional or interim system of local government came to an end after the democratic municipal elections of 5 December 2000.

Balance of Power: Alliances, Opposition and Floor Crossing

The eThekwini council looks and works much like formal political arrangements in other contexts around the world where elected representatives function as members of parties and determine policy on the basis of party policy. The dominant party, in this case the ANC, largely determines policy and resource allocation. Chief whips of parties ensure coordinated voting and discipline of councillors (Buthelezi, 22–04–2004). Caucusing takes place regularly and is oriented to developing positions ahead of upcoming council meetings. Discussion in the ANC caucus is said to be consensus-based and there is seldom a vote. Nonetheless, the chief whip of the ANC reported extensive debate on matters such as the extension of free basic services, the assistance of those in outer areas and rate rebates, along with attempts by individual councillors to attempt to secure development for their wards. “[Within the ANC caucus] there is a lot of debate, there is so much debate to the extent that at times if you are unable to agree as a caucus you have to refer the matter to the senior leadership of the party” (Buthelezi, 22–04–2004).

Whereas most municipalities in South Africa have executive mayors, local authorities in KwaZulu-Natal are run by executive committees or ‘Excos’ comprised of key councillors from the various political parties in numbers proportional to the composition of the overall council. Representatives of the IFP suggested that differences in political style resulted in the ANC favouring an executive mayoral system, whereas the IFP favoured the Exco model (Ndlela and Mnguni, 20–04–2004). Whilst it was in charge of the KwaZulu-Natal province, the IFP established legislation which stipulated that municipalities were to have executive committees. This presented a problem for the ANC in eThekwini. Whilst it won almost half of the seats, it did not have a majority and therefore could theoretically be opposed at Exco meetings. There were several strategies adopted by the ANC to surmount this problem. One was through the appointment of Mike Sutcliffe as city manager. A local academic, previously an ANC member of the provincial legislature and chairperson of the national Municipal Demarcation Board, he was seen as a strong city manager with substantial political backing who would be able to implement the party line. Senior city officials spoke anonymously to the Mail and Guardian (11–07–2004: 8) as follows:

He made it quite explicit that we were to consider ourselves accountable to [the ANC Regional Executive Committee]….It was clear that Mike himself was not going to be accountable to Obed Mlaba (eThekwini’s ANC mayor), or even really to the Metro executive committee. He was deployed by the ANC when the city began to lose political and administrative direction.

One non-ANC Exco representative argued that this key political appointment had transformed policy decision making in noticeable ways:

Since Dr Sutcliffe has come in and there is now a sense that there is political control of the administration, everything is being decided and debated behind the scenes.… Exco…used to meet twice a week…, in some cases we would meet all day on a Tuesday and a Thursday and we used to thrash things out, debate things thoroughly. Yesterday we had Exco, we were finished in 20 minutes…only things that they have discussed thoroughly and internally amongst [the lead party] come to EXCO and what happens at Exco meetings is we are politely allowed an opportunity to air our views, the ANC do not say anything and it is just voted on and they go ahead. (Interview, 07–2003).

Secondly, alliances with other parties were used to ensure that the opposition would not win a vote. In particular, the ANC had a formal alliance with the MF. The MF had been founded to act as a “voice for minority groups” and in particular “consolidate the Indian support” (Reddy, 05–08–2003).6 Although the MF was a small party, its ability to offer enough seats to deliver a majority on Exco gave it considerable leverage. The MF representative on the executive committee stated:

The DP had more seats than us you know…. Without key positions all you would be doing is making a noise in chambers of parliament, that’s all. But when you have the key positions you actually make delivery take place on the ground and that is what happened…. We used [our seats] as a bargaining tool and…through that we were able to secure millions of Rands worth of development for our constituencies. (Reddy, 05–08–2003)

The second largest bloc of seats in eThekwini in 2000, one quarter, went to the DA, a national alliance primarily between the Democratic Party (DP) and the NNP formed in 1999 (Table 1). The DP’s roots were in the white liberal parties of the apartheid era, whist the NNP was an attempt to re-brand the party that had implemented apartheid itself. The NNP was the weaker party in eThekwini and only held 13 of the DA’s 53 seats during the 2000 election. The DA’s support base was traditionally white, and its ability to make inroads into other groups has been more successful amongst those formerly classified as coloured and Indian than African (Habib and Naidu 2004). The DA claimed a vocal oppositional role, with the media used to particularly good effect. The eThekwini chief whip of the DA said that the ANC did not receive the DA’s attempts to function as an opposition party well and suggested that there was a general belief that they had to keep quiet and cooperate because they “lost” at the elections (Ploos Van Amstel, 29–07–2003). Members of the DA defended their political style:

I have said to the ANC ‘don’t knock the DA…be grateful that we exist because if we didn’t you would get up a one party state and you would no longer be a democracy. The thing you fought for, for so many years, since 1912 your whole charter was aimed at a democracy. But in all of that you need different political parties to be active and you need a strong opposition.’ They don’t understand the concept of opposition, their idea is if you are not for us you are our enemies…they will not vote for anything that we propose no matter how good it is. (Interview, 08–2003)

Table 1

5 December 2000 election results for eThekwini (Adapted from IEC 2001)


Ward seats

PR list seats

Total seats

% Total seats won

African National Congress





Democratic Alliance





Inkatha Freedom Party





Minority Front





African Christian Democratic Party





Action Independent Peoples Party





Ethekwini Ecopeace





National United People’s Organisation





Pan Africanist Congress of Azania















The involvement of the NNP in the DA at a national level was relatively short lived. The NNP opted out of the arrangement in 2001 as a result of its minority status within the alliance (Southall 2004: 7). It chose instead to ally itself with the ruling ANC. The ANC then sought to make it legally possible for politicians to change parties without losing their seats. The requisite legislative amendments were steered through parliament, and this created a floor crossing window period in 2002 for local representatives to change parties. As a result, the ANC did not have to wait until the next local government elections in 2005 for NNP representatives to join their ranks. It also allowed the ANC to make a play for power in provinces, which had been dominated by other parties (Koetzé 2007: 77). Although only 5% of the local councillors changed sides across the country, these defections resulted in alternations in the balances of power of a number of councils (Faull 2004). The new ANC/NNP alliance gained control of about 20 councils, whilst the DA lost control of 16 (Koetzé 2007: 81). In eThekwini, the ANC increased its seats from 95 to 105, whilst the DA shrank from 53 to 35 (Faull 2004). Now in the majority, the ANC no longer required the MF to form a majority at the executive committee, and this and subsequent floor-crossing episodes constitute a third mechanism through which the party has consolidated its position. The ANC’s majority remained marginal, though, and a broadly cooperative arrangement between the ANC and MF remained in place. However, the MF did feel that its ability to leverage resources for its constituencies was depleted after the 2002 floor crossing (Reddy, 05–08–2003). The former MF Exco representative said the “the ANC were loathe to give me projects in my area because those projects they knew would become MF campaign tools, so they would be destroying themselves in the process” (Reddy, 20–04–2004).

The IFP is a prominent party in the province, historically linked to the government of KwaZulu under apartheid. Its political identity was defined as Zulu and was characterised by support of traditional leadership. Since the democratic transition, the IFP has not been successful in large towns and cities in the province (Hart 2002: 240). In eThekwini, it only held 17.5% of the seats after the 2000 election. At a metropolitan level, the historically antagonistic relationship between the ANC and IFP gave way to a relatively cooperative political environment. DA representatives observed that the IFP was relatively uninvolved for an opposition and would often rather abstain from council votes than support counter positions (Ploos Van Amstel, 29–07–2003). An IFP representative said that, although they were in opposition to the ANC, their objective was not simply to “oppose anything” associated with the ANC (Mile, 05–08–2003). There was a pragmatic reason for this, not unlike that recognised by the MF. Representatives felt that it was almost impossible for the IFP to facilitate delivery on its own and secure development in IFP-held wards (Interview, 04–2004). They reported hearing an ANC minister say “if you do not vote for the ANC – you won’t get services” (Interview, 04–2004).

As well as attempting to strike the right balance between opposition to and co-operation with the ruling party, the IFP in eThekwini found itself in a working relationship with the DA as a result of decisions taken by their national leaders. The “Coalition for Change” between the IFP and DA arose after the NNP’s departure from the DA in 2001. This partnership proved difficult to implement at a local level because of differences in both political style and party constituencies. An IFP councillor stated:

It is so difficult sometimes to work with the DA…. The IFP and ANC are sharing the constituents, the need of our members is also the need of the ANC members … [A]lthough we have got working relations with DA, we end up working with the ANC. (Mile, 05–08–2003)

Therefore, despite the official link between the IFP and DA from 2001, the ANC deputy mayor was able to say that the ANC continued to enjoy a good working relationship with the IFP and that its link with the DA was not strong in practice. On certain matters, the “IFP voted with the ANC against their alliance partners” (Naidoo, 09–08–2003). Its racially defined constituency was seen to be important: “being a black political party it has to take into account its constituency and that is why I think the difference between the DA and the IFP, they don’t have that kind of common group for an alliance” (Naidoo, 09–08–2003). The difficulties the largest other parties have had in sustaining co-operation then have been a fourth mechanism through which the ANC has been able to maintain its dominance in the absence of a large majority or an ANC executive mayor.

The first years of the eThekwini unicity were therefore characterised by the ANC’s bypassing the absence of a strong (and partisan) elected Mayor through a key administrative appointment and a series of ambiguous and fluid alliances that were struck nationally and resulted in a number of changes to the local balance of power. Unable to secure an outright majority in the 2000 election, the ANC nonetheless secured levels of cooperation from the MF, IFP and later NNP. All of these saw cooperation with the ANC as strategically important for securing development in their constituencies. The DA adhered to an “oppositional” role, but its legitimacy in doing this was not universally recognised. For their part, ordinary ANC councillors we interviewed repeatedly conveyed a sense that they were now occupying a moment of historical destiny and that the ANC’s past gave it a unique degree of political legitimacy.

You know with the ANC it is the leading party, it is the only party that has a vision, I don’t mean that the other parties don’t have a vision, but the ANC are strong … We have been fighting for this democracy for a long time, then we are here today…. Other parties came because… things are easy [and] there are too many opportunists (Mthembu, 03–08–2006)

The narrower ethnic and racial bases of the other parties, whether the result of party strategies or of the perceptions of the electorate and society, have therefore helped to legitimise the dominance achieved by the ANC in the city, despite its relatively marginal electoral victories (Habib and Naidu 2004).

Councillors and their Wards After 2000

The evolution of local government and the local party system in the Metro region as well as the changing balance of power both between parties and between the different elements of local government form the institutional and political context in assessing how the representation of eThekwini’s citizens takes place. We consider how effective the formal politics of councillors and parties is as a vehicle (obviously not the only vehicle) for democracy in the city. We are also interested in the extent to which the evolving forms of democracy in eThekwini are resulting in new, locally distinctive patterns of political representation, communication or accountability.

There are two kinds of local government councillors. Half the councillors are elected on a constituency basis as representatives of wards, of which there are 100 in eThekwini. Other councillors are elected according to a system of proportional representation (PR) where seats are allocated to parties proportionally according to votes they receive. There are two separate votes recorded on the ballot: one for a person to be the councillor of their ward and one for a party without reference to the individuals who might become councillors.

Party representatives said that ward and PR candidates were nominated by the grassroots. During the nomination period before the 2000 elections, local party structures had yet to re-align themselves with the new political demarcations. However, processes for nomination were put in place. For example, the DA representative said that candidates were nominated by a party ward committee (Ploos Van Amstel, 29–07–2003). For the ANC, “the nominations or the elections of the ANC are very transparent. We call a meeting of the branch all the people, they pull together [and are] told they have to [nominate] a candidate to stand as a councillor” (Buthelezi, 22–04–2004). According to one PR councillor, it was necessary to be nominated by five different wards to be considered for the PR list (Nxumalo, 09–09–2003). Another said:

You don’t apply but people nominate you, it’s the people who send you. I was elected in the meeting.… In a branch meeting then I was elected we were many of us about hundred people I think because we had to fill in a list, you don’t apply. [You qualify because] you’re hard working.…a list committee, party list committee [decides who is on the list]. (Mthembu, 06–08–2003)

Despite the rhetoric, there was evidence, as we shall see, of unpopular candidates being imposed on wards and party hierarchies fast tracking candidates’ access to positions on the PR list. Whilst the electorate has the option of rejecting ward candidates by refusing to vote for them, they cannot do the same for list candidates. Cameron (2003) has raised the concern that, as elsewhere in the world, the PR system concentrates power in the hands of party leaders.

We now review some characteristics of selected case study wards7 (see Fig. 1), highlighting some key aspects of the first Unicity elections in 2000. Ward 64 consists of Yellowwood Park, Montclair, Seaview and several other predominantly lower middle class and working class white constituents. Over the last 15 years or so, it has experienced marked social and ethnic transformation, with some areas, particularly those constructed under apartheid for low-income white people, becoming quite multi-racial in character. Yellowwood Park as such is particularly interesting, having been effectively an independent suburb with its own finances, run by a health board, until the late 1990s. Despite changes, the ward still projects a quite self-contained white identity, after a struggle over its demarcation before the 2000 elections and returned Ward and PR majorities for the DA.
Fig. 1

Selected Wards and Suburbs within Durban

Ward 21, comprising both mainly of white middle-class New Germany and part of the former township of Claremont, is, on the other hand, a good example of the pre-2000 Municipal Demarcation Board’s attempts to generate racially-heterogeneous wards. It is more or less bifurcated by a highway, which formed an apartheid era buffer zone. In the 1990s, an area of informal settlement developed along the highway, but its inhabitants have now been re-settled, ironically restoring the gap between the two social worlds of the ward. Middle-class New Germany has, like Ward 64, experienced some demographic change in recent years, with increasing African and Indian populations, but remains a very different place, socially and morphologically, from Claremont across the highway. The portion of Claremont in Ward 21 – Claremont as a whole stretches across several wards – is a dense settlement of mainly African people with many unimproved roads, a large hostel complex, a Civic centre, and a commercial centre with many informal traders as well as shops. Claremont has an unusual position as an apartheid era township where residents enjoyed property tenure and is relatively close to central Durban. As a result, it has had an ongoing experience of enormous pressure on the housing front, with each successful attempt at housing informal settlers followed by an influx of new residents from other townships and rural areas.

This Ward is one of the few in eThekwini that received some attention nationally during the 2000 elections (see Lodge et al. 2002). In theory, it should be an ANC ward, and the ANC carried the PR vote. However, in the Ward election, because of some disaffection locally over the ANC candidate, an independent candidate siphoned off enough votes from the ANC to allow the DA to win the ward. The independent candidate was subsequently hired by the DA ward councillor, who was the head of the DA Caucus in the city. Interpretations of what happened naturally vary between the parties. The DA highlight what they see as an attempt by the ANC Regional structure to impose an unsuitable candidate in the ward. The ANC draw attention to the subsequent close links between the independent candidate and the DA ward councillor to suggest that the opposition party may have engineered the whole process. There are probably several dimensions to an explanation, but the ANC’s inability to mobilise sufficient support for its ward candidate in an area it should have won was obviously embarrassing, especially so given the high profile position of the winner.

Ward 71 is a predominantly Indian ward consisting of parts of Shallcross and Chatsworth centre, both formerly Indian townships. A hilly landscape with several commercial strips, a civic centre and a socio-demographic profile ranging from middle-class to informal sector poverty, it elected a MF ward councillor in 2000. The main area of African informal settlement in the ward, Bottlebrush, creates the most striking socio-spatial division, but there are important socio-economic, linguistic, religious and other differences that can become politically salient. In 2000, the MF carried both the Ward and PR elections. This is an area that has been characterised by a competitive multi-partyism. In this and neighbouring wards, the parties are inclined to behave very strategically to try and block outcomes that are not in their interests. Therefore, in 2000, the DA did not field a candidate in the Ward election in Ward 71, maximising the chances of the ANC candidate (the leader of a key local ANC-aligned civic organisation) being defeated. Similarly, in neighbouring Ward 73, where the resignation of the Minority Front Ward Councillor in 2004 led to an April by-election, the ANC decided not to run a candidate, fearing that competition between the ANC and the MF might allow the DA to take the Ward (Reddy, 20–04–2004)8.

Wards 80, 82 and 88 are in the very large former African township of Umlazi, south of both the former City of Durban and the former Indian townships of Chatsworth and Shallcross. Umlazi was established in 1962 as a residential zone for African people who were relocated in terms of the Group Areas Act (1950) from areas such as Cato Manor (Maasdorp and Humphrey 1975). Wards 80 and 82 are overwhelmingly ANC in electoral terms, although they do contain some IFP voters in both informal and formal areas. For example, Ward 88 was carried by the IFP in both PR and Ward elections in 2000. Along with informal settlements, the IFP presence in the Ward is associated with the existence of hostels for migrant workers. The analysis given us by key ANC figures in Ward 88 suggests, as in the case of Ward 21, that the outcome here may have been influenced by ANC Regional decisions about Ward candidates. In this case, the Region selected a candidate who was viewed as insufficiently respectable by the local community, resulting in many ANC supporters allegedly staying away from the polls on election day (Buthelezi, 22–04–2004; NgCobo, 26–04–2004).

The social and racial composition of electorates, levels of employment and affluence, and the types of problems typically faced by councillors varies greatly across the selected wards. Levels of party competition and the relationships between parties vary across the wards as well, with attitudes ranging from the largely co-operative to the suspicious or hostile evident in different places. However, in addition to these dynamics, one of our main findings has been that the first-past-the-post system of ward councillor election does not lead to a situation in which the local-level political representation is decided in a clear-cut fashion. In the following section, we discuss a dimension of the representation of communities by councillors that has emerged in eThekwini since 2000. This has involved councillors elected on party lists being assigned to specific wards to perform constituency work alongside the elected ward councillors. There are several reasons for this developing practice, but the results include a growing lack of distinction between councillors representing localised communities and those representing the municipality as a whole, and a complex, and to some extent hidden, geography to the politics of representation in the city.

Democratic Representation in Practice: The Deployment of PR Councillors

As explained above, the eThekwini council, like other municipalities in South Africa, officially operates a mixed electoral system combining at-large proportional representation and the local representation of wards. These municipal arrangements constitute an anomaly in a political system that otherwise operates completely on the basis of proportional representation. The importance of proportional representation in establishing the make-up of the council is partly a reflection of transitional concerns about the representation of minorities in a system dominated by a political majority. It was also informed by concerns, particularly within the ANC, about the dangers of too much local representation within a racially divided urban system and the possibility of the development of local clientelistic politics, with attendant risks both in terms of creating different geographical power bases within the party and undermining attempts to create a municipality-wide focus on longer-term development issues. Proportional representation was, furthermore, an opportunity to address the perceived difficulty of electing women as ward councillors. Aside from the innovation concerning floor-crossing between elections, which has been quite significant in its impacts on city politics, but which is in no sense a purely local phenomenon, the composition of the council and the roles of different sets of councillors seem fairly conventional.

However, in practice, political representation is complicated; it is not fully determined or exhausted by the system as formally constituted. Since the 2000 elections, the ANC has “deployed” extra councillors (those elected on PR lists) to work in wards, generally those where elected ward councillors come from other parties. The chief whip and many others we asked were unsure whether this system was being employed in other cities (Buthelezi, 22–04–2004). Other parties with spare capacity of this kind have followed suit but to a much more limited degree.

There are various reasons for this deployment of PR councillors to wards. The political system, as formally constituted, places a huge burden on ward councillors, the only locally elected politicians in South Africa. Unsurprisingly, their roles are constantly tested and expanded by their constituents, as they are called on to assist in a very wide range of tasks, from mediating with non-municipal services and bureaucracies, to dealing with neighbourhood disputes, witchcraft accusations, harassment cases, burst pipes and the difficulties individual households face in terms of services and resources. It is hard for them, as technically part-time council workers – there are only 10 full time councillors in the system, although most of the ward councillors we talked to worked more than full time jobs on part-time salaries – to effectively combine all this with their more formal local representational duties (Interview, 06–08–2003). On the other hand, politicians we interviewed from all the main parties underlined the problems PR councillors faced in finding a role within the local system, implying that they had not sufficient activity to keep them occupied and that there was an undesirable status gap opening between them and their locally elected ward counterparts.

Ward councillors are therefore sympathetic to the idea that they could do with some assistance in their work, especially where minority communities are concerned. They also, perhaps partly to highlight their own special status as individually elected local representatives, emphasise the unsatisfactorily light workloads and roles of the PR councillors before deployment. However, some incipient conflicts were apparent, with some assigned PR councillors representing themselves as more officially connected with the wards than they actually are. For example, one DA councillor was clearly put out by an assigned PR councillor setting up, and chairing, ward meetings on matters currently on the council agenda, even if the ward councillor was then duly invited to attend (Interview, 08–2003).

Amongst the many roles ward councillors are called upon to play is the recruitment of local labour for contractors working on development and services in their wards. One DA councillor reported finding herself having to acquiesce in a process where the PR councillor assigned to her ward had a far more hands-on role in this potentially clientelistic process than she felt was legitimate (Interview, 07–2003). Another councillor, for a predominantly Indian area, whilst at pains not to say anything critical about the ANC councillor assigned to his ward, did note with some regret that his contact with the black community had declined markedly since the ANC deployment, which perhaps reflected a certain feeling of being supplanted in his formal role (Interview, 08–2003). This symbolic role-reversal hints of course at a potential real-role reversal in future elections, an unsettling prospect for some ward councillors. In the two DA wards, there were concerns about the involvement of ANC PR councillors in voter-registration-related activities, with complaints about black voters being “imported” to register their address in large numbers in one ward (Ward 64) and of the assignment of no fewer than three ANC PR councillors to another (Ward 21) in an attempt to ensure the leader of the Metro DA’s defeat in the next election.9 In the latter case, there may be some misunderstanding involved, as two of these ANC councillors are assigned by the party to other wards, but work some of the time out of the same offices in Claremont (where they reside) as the ANC’s officially deployed representative. Because of this, however, they perhaps inevitably become involved in the life of Ward 21, where the assigned councillor made it clear to us that the round-the-clock attention to a plethora of emergencies, problems and demands involved was, at times, a quite overwhelming workload (Interview, 06–08–2003).

There are also more specifically political rationales for the local deployment of PR councillors. As described above, the demarcation of ward boundaries before the reforms of the 1990s was, in part, aimed at creating socially (i.e., racially) heterogeneous wards, an attempt to force the political system to be responsive along multi-racial lines and to move away from politics revolving around racially divided geographical spaces. However, this social and racial heterogeneity has been stretching the capacities of elected ward councillors to deal simultaneously with the different communities in their wards. These problems are partly linguistic, with, for example, local communications having to be translated into Zulu, and even where councillors from other backgrounds are fluent Zulu speakers, there are said to be cultural and political barriers making African constituents reluctant to turn to non-black or non-ANC councillors for assistance (Nxumalo, 09–09–2003; Interview, 06–08–2003). Ward 21, for example, consists of a middle class white suburban community separated from part of a former black township by a buffer zone constituted by a highway. The two communities it contains are obviously very different in outlook, political preferences, political styles, preoccupations and demands. Ward 71, in Shallcross, whilst its population is mainly Indian, has concentrations of black constituents, mainly in a large informal settlement called Bottlebrush. In these cases, as in others, the argument goes, the assignment of PR councillors is essential if minority communities are to have a representative they can easily and comfortably access and who is more likely to understand their particular situation and problems.

There are, however, suspicions voiced by parties other than the ANC that this deployment strategy is also party political in the sense of building up the strength of the ANC in those areas of the Unicity where it is weaker in electoral terms. This function of assigned councillors is not unique to the ANC: in Indian areas, both the DA and MF deploy councillors to those wards where they do not have a ward councillor. Here, whilst informants did give us some sense of how social or ethnic differentiation in Indian communities was related to party political attachments, giving voice to socially differing constituencies does not have quite the same purchase as an argument in these areas. The ANC is clearly involved in an effort to develop branches in every ward, as part of the post-2000 spatial reorganisation of its party structures, and the assignment of PR councillors is an important part of this effort at party-building and electoral mobilisation beyond its existing geographical ‘base’. ANC-assigned councillors we interviewed stressed the importance of developing ANC structures (Party branch, Youth League, Womens’ League) throughout the city and of their activities in this regard in the wards to which they have been assigned (Mtembu, 06–08–2003; Nxumalo, 09–09–2003; Interview, 06–08–2003).

However, we evaluate it in partisan terms; there are several conclusions we can provisionally come to about the functioning of the electoral system in eThekwini. A system has been constructed where by and large voters can choose which party to work with locally in all but the most formal matters of council, except perhaps in the most homogeneously ANC areas of the Metro. The “pros” of this have been rehearsed in terms of relieving the workload of (some) ward councillors and providing accessible representatives to historically disadvantaged communities embedded within wards where electoral realities have returned councillors from parties that they find it more difficult or uncomfortable to turn to for assistance.

The main difficulties from a broadly democratic point of view are as follows. Firstly, the system benefits most the party that has the most PR councillors to deploy, namely the ANC, potentially building in a further set of advantages in future elections over and above its present “formal” resources10. Secondly, and perhaps unfortunately, the blurring of the roles of ward and PR councillors in eThekwini makes it less, not more, likely that politicians from historically white and Indian parties deal with, adjust to and learn about the needs and interests of other constituents. Before reaching for metaphors such as “representational apartheid” to characterise this evolving system, one should note firstly that it is, arguably, too soon to evaluate several of the dynamics discussed in this paper and, secondly, that a system that is ‘informalised’ in certain respects is at least flexible and has the virtue of keeping the formal system of ward representation by ward councillors “afloat” in politically complex and socially and economically demanding local environments.


We have provided a sketch of some salient features of the ‘formal’ democratic system as it evolved in Durban after 2000, at both the metropolitan and ward scales. At both levels and, more importantly, between them, dilemmas arise in translating the (relatively) clear-cut results of elections into effective representational work by the metro council and the individual councillors themselves. A key theme in Durban debates is the mediation of the perceived needs and demands of ward level and metro-wide scales and of short and long term time horizons. In spatial terms, the Unicity arrangements mean that the integrated political entity of the city has grown substantially but remains internally subdivided into wards, with obvious conflicts arising between expressed demands from below and the formulation of priorities at the metro level by the local state. These tensions are endemic to urban politics everywhere, and the South African solution of structures of ward-based representative democracy mediated through broader party politics is one political format that seems to cope with the balancing acts which this necessitates in an effective and relatively legitimate, if necessarily conflictual, fashion. The mediation of time horizons is a related conundrum for city politicians and administrators. This arises most clearly in the identification of long-term matters with economic growth and short-term matters with local service delivery (which Ballard et al., this volume, discuss), although in truth, conflicts between short- and long-term priorities arise in both policy areas, at both ward and metro levels.

In this context, the wider political imperative to sustain proportionality to ensure minority representation is only one reason for the lack of locally based political representatives in South Africa, apart from in the “mixed” electoral system established at municipal level. Another is a deep concern about the potential for the development of patronage-driven local clientelistic politics if too much representation within the system is constituency-based (something which our research does indicate is of concern in existing arrangements). Yet, the value of some party linkage to constituents, as well as strong demands from citizens for immediate solutions to a host of local problems, means that some measure of constituency representation is seen as necessary and important.

In Durban, the tensions outlined in this paper are visible in contestations surrounding the roles of PR councillors. These are supposedly elected “at large” as a counterbalance to localism and clientalism, yet pressures have led to their being deployed locally by parties as supplements to the locally elected representatives. This was initiated by the ANC, the largest, and inevitably somewhat factional, ruling party with an abundance of councillors. What some in other parties see as an extension of an already dominant organizational structure and influence into “their” areas can also be viewed as an improvised response to managing significant demands for service delivery that the “formal” political system cannot channel effectively. It is also a mechanism for maintaining party discipline and lines of control by giving PR councillors some specific representational work and using this to build political support. Service delivery remains extremely important to ward councillors – in terms of their accountability to their constituents and their political (and often personal) survival. They, in particular, are caught between priorities set at the metro scale and vigorous local demands.

Because the post-apartheid urban system is socio-politically, spatially and temporally complex, so is its democratisation. City political life consists of much more than elections, councillors and party politics. It covers the machinery of local government (at municipal, metropolitan, provincial and other scales of the system), its constitution, powers, services, ongoing processes of public consultation and its growing relations with variously configured “partners” in urban development projects (the last two are explored in Ballard et al., this issue) as well as the proliferation of popular local social movements (considered in Barnett and Scott, this issue). As this paper has shown, though, city politics also includes a lively set of “informal” political practices aimed at institutionalising popular representation between elections. This has proved to be contested, inventive and unpredictable, suggesting that the city’s ‘formal’ politics is much more than a relatively unproblematic background against which the value of other emergent forms of democratic engagement can be easily assessed. It is in itself an area of local innovation and creativity, contributing to the distinctive nature of South Africa’s new urban democracy.


Although Durban is the officially recognised name of the city, the council and its jurisdiction are formally known by its Zulu name, eThekwini.


Interviews are cited by name of interviewee and date. Occasionally a more anonymous format has been used where this was requested and where more personal or sensitive local issues may be at stake.


These were delayed for KwaZulu-Natal and the Western Cape from the original timetabling of 1 November 1995. In KwaZulu-Natal, they were postponed because of the high level of political intolerance and violence, voter registration problems, disputes relating to the registration of candidates and problems and difficulties in campaigning and mobilising voters in certain areas. There had also been delays in Durban itself related to concerns over the traditional leaders and boundary disputes (On boundary disputes see Ramutsindela and Simon 1999; Narsiah and Maharaj 1999; Giraut and Maharaj 2003; Ramutsindela 2001).


Despite the consolidation achieved in the creation of transitional councils, the type of local government that materialised in terms of the LGTA was seen to fall someway short of ideal of one city, one tax base (Wooldridge, 2002:132). The separation of areas of abject poverty from areas of affluence was seen to persist in some areas (Narsiah and Maharaj 1999).


Durban Metropolitan Area Boundary Submission to the Municipal Demarcation Board, 25 August 1999, p.15.


Key individuals emerged from the House of Delegates of the Tricameral Parliament, which was a reform introduced in the final decade of apartheid to allow those designated as ‘Asian’ to elect representatives. Like the House of Representatives, for coloured people, this body was to deal with Asian affairs but was subordinate to the white elected parliament. There was no equivalent for black people.


To get a sense of the political units that made up the council, we focused attention on six wards, using these to generate some insights about political variability across what is still a very socially and racially divided city. Our selection of wards was deliberately fairly subjective and as such we can in no sense claim that they form the basis for either generalisation across the city or comparative analysis. The strategy had the more modest aim of allowing us to attain, in a city of 100 electoral wards, an appropriate level of resolution and variability to evaluate not only electoral outcomes, but also some characteristic processes of democratic representation during the 2000–2004 period. In each ward we interviewed the ward councillor, and, where possible, figures from other parties and selected groups in civil society, the latter in an attempt to validate the general picture we received from the politicians.


In the case of Ward 73, there were other complications. The resignation of the Ward Councillor was occasioned by his defection to the DA, and then to the ANC, in the course of a few weeks. As this councillor was a member of the Exco, this received a large amount of media attention, and it may have seemed prudent to the ANC not to inflame matters by running against the Minority front in the by-election.


The DA councillor was indeed replaced in this way in the 2006 election.


This is reinforced by the system of community mobilisers discussed in Ballard et al. this volume.



The research reported in this paper and in the two other papers collected here (Ballard et al., and Barnett and Scott) was undertaken as part of a joint project, New Spaces of Democracy in Post-Apartheid Durban, funded by a Leverhulme Trust Research Interchange Grant, between the Open University, the London School of Economics, and the University of KwaZulu-Natal. The authors of all three papers also thank the numerous people of Durban (political representatives, officials, activists, residents and fellow scholars) who gave very generously of their time and experience in making the research possible. The authors of this paper would like to thank others on the team for a stimulating collaboration. In particular, we would like to thank Jenny Robinson for her support and for editing this paper. Finally, we would like to thank Brice Gijsbertsen at the School of Geography, University of KwaZulu Natal, Pietermaritzburg, for compiling the map (Fig. 1).

Copyright information

© Springer Science+Business Media B.V. 2007

Authors and Affiliations

  1. 1.Department of Geography and EnvironmentThe London School of Economics and Political ScienceLondonUK
  2. 2.School of Development StudiesUniversity of KwaZulu-NatalDurbanSouth Africa
  3. 3.Department of GeographyGeography, School of Environmental SciencesScottsvilleSouth Africa

Personalised recommendations