Voting and violence in KwaZulu-Natal’s no-go areas

Coercive mobilisation and territorial control in post-conflict elections

Maria Schuld studied political science with a focus on conflict studies at the Free University of Berlin, Germany. She graduated in 2013 on violence in post-conflict transitions to democracy. This article was written when she was a visiting scholar at the Centre for Civil Society of the University of KwaZulu-Natal. She wishes to express special thanks to Patrick Bond and Shauna Mottiar.


Post-conflict elections have become an important tool of international conflict resolution over the last decades. Theoretical studies usually point out that in war-to-democracy transitions, military logics of territorial control are transformed into electoral logics of peaceful political contestation. Empirical reality, however, shows that the election process is often accompanied by various forms of violence. This paper analyses post-conflict elections in war-to-democracy transitions by comparing support structures for conflict parties as well as their coercive mobilisation strategies in times of violent conflict and post-conflict elections. It does so through a single case study of KwaZulu-Natal. This South African province faced a civil war-scale political conflict in the 80s and early 90s in which the two fighting parties – the African National Congress (ANC) and the Inkatha Freedom Party (IFP) – used large-scale violence to establish and protect no-go areas of territorial control. This study finds that in the first decade after South Africa’s miraculous transition, these spatial structures of violence and control persisted at local levels. Violent forms of mobilisation and territorial control thus seem to be able to survive even a successful transition to democracy by many years. Measures to open up the political landscape, deescalate heated-up party antagonisms and overcome geopolitical borders of support structures seem to be crucial elements for post-conflict elections that introduce a pluralist democracy beyond the voting process.


Over two decades ago, a new instrument appeared in the toolbox of international conflict resolution: elections. As a central element of democratisation, elections were supposed to bring peace and a legitimate democratic regime at the same time. Expectations were especially high in the context of Huntington’s ‘third wave of democratisation’ and theories of the ‘liberal democratic peace’. However, scepticism about this practice became more and more pronounced as at second glance, the ‘arranged marriage between peace and democracy’ (Ohlson and Söderberg Kovacs 2009:165) seems to be full of contradictions.

On the one hand, empirical reality showed that elections in post-conflict environments are ‘an exceptionally risky venture’ (Reynolds and Sisk 1999:14, see also Mansfield and Snyder 2005). According to the Uppsala Conflict Data Program Peace Agreement Dataset, half of the peace agreements followed by elections between 1989 and 2004 resulted in violent conflict (Jarstad 2009:161). Elections thus do ‘not appear to be an instrument for enhancing the durability of peace’ (Collier, Hoeffler and Söderbom 2007:470). In addition to post-conflict risks, electoral violence accompanied around 80% of all elections in Africa (Lindberg 2006). On the other hand, theoretical considerations put the cure-all power of post-conflict elections into question. Elections are a highly competitive mechanism to distribute political power, and especially war-torn societies dealing with legacies of violence are vulnerable to the effects of such inclusive conflict resolution mechanisms (Höglund 2009; Reynolds and Sisk 1999). There are various incentives for armed groups turned into political parties to continue violent strategies in addition to parliamentary competition (Mehler 2007; Höglund 2008:83; Höglund, Jarstad and Söderberg Kovacs 2009:353; Laakso 2007; Dunning 2011). Furthermore, peace and democracy seem in some regards to be mutually exclusive goals (Jarstad and Sisk 2008). In short: post-conflict elections and violence stand in a complex relationship, and there is a ‘policy need for improved knowledge on the role of elections in building peace’ (Jarstad 2009:153).

While determinants for success and failure of post-conflict elections have been studied in detail through quantitative analysis,1 a more distinguished perspective on the different qualities and the specific logics of post-conflict periods is only recently emerging (Höglund and Söderberg Kovacs 2010; Richards 2005; Mac Ginty 2008; Suhrke 2012). This study wants to add to scholarly knowledge on post-conflict elections in war-to-democracy transitions by comparing support structures for parties in conflict as well as their coercive mobilisation strategies in times of violent conflict and post-conflict elections. It thereby follows up on the often posed argument around post-conflict elections. Electoral strategies – so it goes – are the opposite of violent strategies. While in war configurations, violence seems to be the central mechanism to accumulate territorial control (and thus political power), democratic systems are said to support peaceful voter mobilisation for electoral competition. This study aims to give evidence for a revision of this view. It argues that support structures as well as strategies to maintain support from the population can be very similar in both settings. The findings presented here add to the recent scepticism towards post-conflict elections as an all-in-one cure against violent conflict.

The article employs a single case study of KwaZulu-Natal. This South African province faced a civil war-scale political conflict between the African National Congress (ANC) and the Inkatha Freedom Party (IFP) in the 80s and early 90s, in which a geographical jig-saw-puzzle of party strongholds was established by violence. Political support for one of the opposing parties became obligatory and strictly related to territory. Four rounds of mutually accepted post-conflict elections as well as four rounds of local elections forced the ANC and IFP into a regional power sharing configuration. After a theoretical introduction, this case study will follow three steps: First, the establishment of these ‘no-go areas’ of coherent support will be illustrated, with an emphasis on the role violence played. Second, violence in relation to post-conflict elections shall be analysed to show how patterns remained very similar to the ‘years of political conflict’. Third, the development of voting patterns in a number of former no-go areas in South Africa’s post-conflict elections will be compared. Here too one can see that voting patterns continued to match with the power configurations maintained by the former ‘patron’ of an area. The last part discusses the consequences of these findings and draws conclusions for post-conflict peacebuilding.

Territorial control and voter mobilisation: Theoretical background

This paper argues that territorial control in times of civil war might show strategies of support mobilisation as well as geographic patterns of support very similar to those in the first decade of post-conflict elections. The relationship of violence, support and military control on the one hand, and of violence, voter mobilisation and post-conflict elections on the other, have been subject to previous theoretical work that shall be briefly presented here.

Territorial control is a main feature not only of conventional war, but also of civil war. It refers to military control in terms of successful defence against the intrusion of enemies, but also to a certain degree of political control. Territorial control means that ‘an insurgent organization which controls a given area […] operates as a counter-sovereign authority, a “counter-state”. It provides protection, administers justice, collects taxes, and applies its social program’ (Kalyvas 1999:295). A minimum of support from the civilian population is therefore a crucial condition for local warfare. Compliance can thereby be induced by fear through violence. But as Stathis N. Kalyvas’ in-depth study, The logics of violence in civil war, of 2006 shows, (indiscriminate) violence against the population is a risky strategy because of its delegitimising effects. That is even more true in the shadow of elections: ‘A rebel group that relies almost exclusively upon coercion to control territory is unlikely to be successful in the postwar electoral period’ (Allison 2010:107). Hence, to get and maintain control of a certain area, armed groups usually apply a dual strategy of contractual and coercive behaviour towards the residents (Metelis 2010:4–5). Territorial control in times of civil war is thus not merely a military business, but relies on support from the population.

In opposition to military logics, electoral logics appear to be characterised as a peaceful contestation in which ‘ballots take the place of bullets’ (Allison 2010:107). The electoral process is often seen as an alternative to violence and violence around elections is consequently regarded as incompatible with democracy (Chaturvedi 2005:189; Fischer 2002). But democratisation is by no means a genuinely peaceful process. Elections are a highly conflict-inducing political instrument:

Elections, as competitions among individuals, parties, and their ideas, are inherently just that: competitive. Elections are, and are meant to be polarizing (Reynolds and Sisk 1999:18).

This stands contrary to the consensus-orientated aims of conflict resolution:

The very nature of elections has the potential to instigate conflict… To win support, differences are emphasized rather than the common elements which bring people together (Höglund 2009:421).

The competitive nature and polarising effects of elections play a crucial role in post-conflict societies. Here the given starting conditions are a weak economy, weak state institutions, insecurity, a patrimonial system of loyalties along lines of polarised identities and a culture of violence entrenched in social behaviour (Höglund, Jarstad and Söderberg Kovacs 2009; Höglund 2009; Steenkamp 2005). These conditions, as well as the high stakes in post-conflict elections, make the use of violence as a competitive means much more likely than expected by many conflict resolution practitioners. Violence does not only directly prevent elections from being carried out in a free and fair manner, it can also provide a background setting of fear and intimidation that influences voting patterns.

This study focuses on electoral violence defined as acts or threats of physical harm or intimidation perpetrated to affect the results of elections. This definition is more limited than the ones commonly used (e.g. Fischer 2002; Höglund 2009; Sisk 2008) to exclude violence employed to boycott the democratisation process (also known as ‘spoiler violence’). Electoral violence as used here only refers to contexts in which the elections themselves are an undisputed mechanism of political contestation.

We can conclude that the ‘military’ logic of civil war and the ‘peaceful’ logic of post-conflict elections are surprisingly similar in that political power requires a minimum of support by the population. In times of war and after, armed groups have incentives to use violence as well as peaceful means of mobilisation to win this support. While in times of war the population’s collaboration needs to be won to gain territorial control, in post-war elections their vote is the crucial factor. This assumption serves as the theoretical frame under which the following analysis will be conducted.

The theoretical model must be treated with care because it reduces highly complex actors to coherent rational entities with a general strategy. The empirical reality of civil war however tends to reveal an interwoven network of local-level conflicts underneath the ‘master cleavage’, and a larger variety of violent actors (see for example Kalyvas 2003; Autesserre 2010; King 2004). As we will see in the following, this is especially true for the case of KwaZulu-Natal (KZN) where the local level plays a central role in how violence is produced (Schuld 2012a; Benini et al. 1998; Krämer 2007). Although the party frame plays a crucial role in most of KZN’s conflict narratives, it seems to make little sense to talk about the ‘ANC’s or IFP’s strategy to gain political power’ once we have a closer look into the complex network of local power struggles. An actor’s ‘strategy’ as mentioned above does therefore not necessarily describe a group’s coherent plan to reach a certain goal. It rather refers to the mechanism resulting from a whole variety of actors and their respective strategies that sum up to a seemingly concerted agenda. The ‘rational-actor’ approach thus remains a theoretical construction.

The violent establishment of no-go areas during political conflict in KwaZulu and Natal

South Africa’s transition to democracy was marked by high levels of violence, peaking in the years after Nelson Mandela’s release from prison in 1990.

The ANC as well as the apartheid government, between whom the roadmap towards a new political system was exclusively negotiated, both used a dual strategy of violence and political means to ensure that the new system would meet their respective interests. Especially black townships throughout the country faced violence in terms of sabotage acts, police repression and hit-squad activities on a civil war scale.

The area of what is today KwaZulu-Natal (KZN) was one of the hotspots of violence: of the estimated 25 000 people who died in the violent conflict between 1985 and 1994, as many as 15 000 were killed there. Violence in this region had its own dimension and dynamics. The ANC’s liberation struggle against the apartheid system translated into ‘black-on-black’ violence between vigilantes of the conservative Inkatha party and ANC comrades. Inkatha (who turned into the Inkatha Freedom Party/IFP in 1990) governed the KwaZulu homeland and was initially aligned with the ANC apartheid resistance. In the late 70s the ideological distance between the two movements grew. Inkatha stood for independent traditional rule of a ‘Zulu kingdom’ as opposed to the ANC’s struggle for a national democracy. In the wake of their growing antagonism, Inkatha increasingly collaborated with groups from the right side of the spectrum, like the various police units active in KwaZulu and Natal as well as the National Party, taking the role as a ‘surrogate force’ of the apartheid government to enforce order in the region (Aitchison 1993:234). The violent conflict for political control between the ANC and Inkatha became the dominant frame under which the plenitude of local groups (vigilantes, the United Democratic Movement, the Congress of South African Trade Unions and many others) as well as violent entrepreneurs (warlords, taxi companies, ethnic groups, gangs, self-defence and self-protection units) fought their local wars. Fatality figures from ‘political conflict’ did only drop as late as the end of the 1990s, but violence continued in many forms and areas (Schuld 2012a; 2012b).

Violence played a major role in the IFP’s and ANC’s efforts to gain territorial control, which was closely related to ideological support. In the late 80s, when large-scale violence broke out in the Durban townships and later in the Midlands, no-go zones were formed through strategies of eviction and forced recruitment. ‘The general pattern of this time was that vigilantes attacked and areas defended themselves’ (Leeb 1988). Meer (1989) describes in detail how groups of Inkatha supporters armed with traditional Zulu weapons swept through residential areas, burning and stoning houses and beating up residents. In many violent incidents recorded in the Natal Violence Monitor, people were asked for their political affiliation, forced to sing songs of a certain party or simply threatened into leaving the area (see also Minnaar 1992a; Aitchison 1990). Recruitment campaigns were especially carried out at schools, from where male youth carried political alignments into the households (Bonnin 1997).

By the beginning of the 90s, neutrality had become impossible as being neutral meant being on the side of the enemy. Politically homogeneous areas emerged as people supporting other parties were driven out of the respective strongholds (Minnaar 1992a:5). In this way ‘a geographic area would become associated with a particular political group’ (Bonnin 1997:28). Coercive means to gain territorial control thus led to geographically coherent support bases in which structures of governance like taxation (on the ANC’s side ‘to support the armed resistance’) and institutionalised people’s courts were established. Violence was especially directed against the respective enemy. Bonnin (1997) describes how empty houses at the boundaries were patrolled by young men watching out for attacks. Township sections became inaccessible for people from the ‘other side’. The geographic divide deeply affected people’s daily routines. Taxi drivers crossing borders risked their lives (Kentridge 1990:36). Working across the road, loving a person from a different section, coming home from a birthday party after dark could mean death (Bonnin 2006). Although forced political affiliation turned out to be stronger in some areas than in others, by 1993 the region of what is today KZN had become a ‘jigsaw puzzle of party political strongholds and no-go-areas’ (Truth and Reconciliation Commission 1998:248).

Coming back to the question of actor coherence, it is yet an open question as to how far the party leadership, namely Mangosuthu Buthelezi and Nelson Mandela, were actively instigating the violence. Both parties show a mix of antagonistic incentives. On the one hand military power might have worked as a useful threat in the national negotiations (in which the IFP struggled to participate), and territorial control promised a pole position in the shadow of upcoming elections. On the other hand, the demonstration of stability was crucial for the ANC to calm white South Africans’ fear of a potential black government, and for Inkatha to advance their agenda of a strong independent Zulu authority. This view is supported by the various peace process efforts undergone by both parties and their leaders. Consequently, many scholars argue that on the local level violence ‘developed a momentum of its own’ (Simpson 1993, see also Johnston 1996:179; Woods 1992). As discussed above, both parties’ strategy of gaining and maintaining territorial control through the establishment of geographical areas of coherent support appears to have been the product of complex networks of actors and motives.

After the 1994 elections: Violence, territory and voter mobilisation

With the regime change through South Africa’s first democratic elections in 1994, local governance posts were formally distributed through an electoral process as opposed to military power and territorial control. Elections as a political institution were initially highly contested from the side of the IFP. Inkatha used what is often termed ‘spoiler violence’ to boycott South Africa’s first democratic elections on 27–29th April 1994. The IFP had the bitter experience of being excluded from the ANC-NP negotiations about a new South African interim constitution. While the ANC advertised a nation-building process for a united and non-racial South Africa, the IFP strongly supported a federalist system with an independent homeland of KwaZulu for a traditional rule of the ‘Zulu kingdom’. This position was not only consistent with the official party objectives and its mobilisation along ethnic lines, but also reflected the IFP’s relative weakness in other parts of the country. When elections were announced for 26 April 1994, the IFP refused to join and high levels of violence followed (Truth and Reconciliation Commission 1998:319). A week before the elections, the IFP leadership suddenly decided to take part. The speculation on motives ranges from Inkatha’s increasing isolation in the fragmentation of the right (Johnston 1996) to successful mediation (Griggs 1996), and a ‘deal’ at the highest level of leadership to ensure undisrupted elections (Aitchison 2003). The last version is clearly supported by the great number of irregularities during the elections (Muthien and Khosa 1998; Reynolds 1994) as well as the striking result of 50,3% or exactly 41 out of 81 provincial seats for Inkatha. Whatever motivated the IFP to join the elections, political violence decreased significantly in that week. From April 1994 on and up to today, elections or election results have not been disputed on a large scale.

We can therefore say that electoral logics replaced war time logics of informal control in KZN. As argued here, violent means to ensure the population’s support for and compliance with one of the conflicting parties remained very similar to the patterns before 1994. In 1994 the Independent Electoral Commission (IEC) reported on 165 complaints of no-go areas in South Africa (Du Toit 2001:66). For the parties participating in South Africa’s first ‘free and fair’ election, ‘[f ]ree canvassing in the province was impossible’ (Lodge 1999:122). Taylor (2002) conducted a detailed study of three massacres in Richmond, Shobashobane and Nongoma, showing how in these areas violence as a means for territorial control and power peaked in the years after 1994. Shobashobane, for example, is an ANC controlled ward surrounded by seven IFP dominated wards. On Christmas Day 1995, a concerted IFP mob attack on the isolated area, with a passive police and escape routes systematically cut off by roadblocks, left 19 people dead. Nongoma and Richmond are examples of intra-party power struggles caused by defections, again determining control over territory and people. Krämer (2007) analyses post-1994 violence in Inchanga township and finds that fighting between strictly divided ANC and IFP supporting sections continued until 2000. But territorial control was not necessarily primarily in the hands of one of the two parties. Some geographical areas were controlled by strongmen like warlords, traditional chiefs or gangs that were only superficially related to a certain party (Shaw 2002:72). For example, in KwaMashu (Durban), township sections were controlled by different gangs, formed partly for reasons of self-defence and gradually taken over by criminal elements. Although ‘by 1997 political violence between the IFP and ANC had all but completely abated’, the antagonistic gangs were mostly affiliated with the ANC (Nebandla 2005). Examples of hotspots of violence in the late 90s and early 2000s dominated by warlords are the above mentioned Richmond area under the (later expelled) ANC mayor Sifiso Nkabinde (Taylor 2002; Dunn 1998) and the shack settlement Lindelani under IFP warlord Thomas Shabalala. Reports on local strongmen’s regime of terror and intimidation in rural Macambini can be found from as late as 2004 (De Haas 2009b).

These cases represent extreme examples of violent structures after 1994 where fighting still continues between geographically separated areas of support for either the ANC or the IFP. In other areas, post-conflict elections were more peaceful and people supporting different parties lived next door to each other. Like the cases selected in the third part, these examples merely serve to illustrate how support structures for armed groups or political parties are able to continue.

The landscape of territorial control shifted in the last decade. The ANC was finally able to win a majority of votes in more and more wards, so that in 2004 they won the provincial elections by a small margin and limited IFP supporting areas to the KZN northern heartland in 2009. According to Booysen (2011) a reason for this final victory is the success of the ANC’s liberation narrative and incorporation tactics as opposed to Inkatha’s ‘overcome’ ethnic approach. What is important here is that apparently more and more IFP no-go areas became penetrable for ANC campaigning and political activity, a fact also confirmed by township residents from Inchanga and KwaMashu in informal interviews. Although ‘pockets of territory in which non-democratic and violent modes of political behaviour persist’ (Bonnin 2006), the overall picture is that of an increasing intrusion through former borders, accompanied by a shrinking IFP power base. The ANC is on its way to becoming the unchallenged ruling party in KZN.

The decline of military forms of territorial control and high levels of violence, however, did not imply that violent mobilisation strategies disappeared. Electoral violence is a common feature of elections in KZN (Höglund and Jarstad 2011). Unlike in other African cases of electoral violence, the election days remained relatively calm (Du Toit 2001:66; Zulu et al. 2009) and results were mainly unchallenged. Electoral violence in KZN rather took the form of the disruption of rallies, prevention of electioneering in politically sensitive areas, attacks on party supporters and intimidation (Piper 2004a; Dugard 2003). Accordingly, violence mostly ‘consisted of public confrontations between ANC and IFP supporters over election related events’ (Piper 2004a). Both parties frequently complained about the high levels of intolerance. The IFP for example referred to an ANC members’ demonstration at which coffins representing the IFP were burnt, and an IFP campaign in certain areas was prevented by roadblocks and car stoning as well as the attacking of an IFP councillor in Wembezi Township (Piper 2004b; De Haas 2009a). The ANC in turn alleged that Inkatha supporters were responsible for blocking access to areas, turning down posters as well as assaulting and sjambokking ANC supporters (De Haas 2009b). Attacks on political rallies thereby have a long tradition. Back in the late 80s and early 90s, political gatherings as well as funerals of victims from political violence were often disrupted by armed groups (Schuld 2012a).

A more severe form of electoral violence have been the notorious political assassinations that frequently happen up to today. Patrick Bond (2011:15–16) names a number of recent cases, especially in which leaders of ANC and IFP break-away parties were killed; obviously for political reasons. COSATU secretary-general Zwelinzima Vavi stated that 13 members of the COSATU/ANC alliance were murdered ‘owing to political infighting’ (South African Institute of Race Relations 2011:707). Especially in the run-up to elections in 2004 and 2006, ‘numerous killings of high-profile politicians belonging both to the African National Congress (ANC) and the Inkatha Freedom Party (IFP)’ were reported (South African Institute for Race Relations 2005; see also Mottiar 2006; De Haas 2008). Also, assassinations appear to be a heritage of the pre-1994 period, in which the killing of political leaders – through petrol bombing of houses, drive-by shootings or hit-squad attacks – was one of the characteristic forms of violence (Schuld 2012b).

As in the years of political conflict, the question of systematic coordination of violence remains largely open. Some clashes between supporters of different parties might well result from personal hostilities developed through decades of violence, or as Piper (2004a) puts it, were ‘cast as spontaneous conflicts driven by local people’. Reports on assassinations,2 denied access to areas and disruptions of campaigns, however, point to an at least minimally coordinated shape of electoral violence. Whatever the degree of coordination may be, elections obviously remain subject to violent contestation, or, in other words, the ‘democratic way’ to gain political control of an area is still contested by violent means. Electoral violence thereby shows continuities from former violent patterns. It seems as if in KwaZulu-Natal, certain violent methods of support mobilisation slipped through the overall pacification to live on in the new electoral environment.

The above analysis leaves us with a number of open questions. How is military territorial control before 1994 related to voting patterns after South Africa’s transition to democracy? Did pre-1994 territorial control directly translate into voting patterns or did support distributions change due to elections? How long were the parties able to maintain the support of the population in their respective strongholds? An attempt to answer these questions will be derived from a brief analysis of voting behaviour in some reputed former no-go areas. Information is taken from the South African Independent Electoral Committee (IEC). In order to limit the analysis and since South Africa’s system of municipal elections is complex, only national election results will be considered. The 1994 elections are for two reasons not included in this analysis: First, the interim electoral commission responsible for organising the elections faced many difficulties and results are not available online for each voting district. Second, irregularities on the voting day like prepared ballot boxes, distribution of voter cards to children and the forced removal of observers seriously question the results for KZN (Reynolds 1994; Hamilton and Maré 1994).

A first look indicates that the general pattern of IFP/ANC support survived the last 18 years, with the ANC gaining more and more support in former IFP strongholds. In the early 90s, the townships close to urban white residential areas of Durban and Pietermaritzburg were largely ANC controlled, while the IFP dominated the rural areas of KwaZulu. The South Coast region was highly contested as affiliations changed rapidly (Minnaar 1992b; Griggs 1996). After 1994, the South Coast showed ‘relatively homogeneous but sharply divided IFP and ANC zones’ (Taylor 2002). In the townships, the IFP created outposts of control through intense recruitment in male labour hostels.

Until today hostels are said to be IFP affiliated. This rough picture is reproduced in voting patterns of the 1999 national election (see figure 1). Rural areas, especially in Northern KZN were homogeneous IFP strongholds, while the urban Durban-Pietermaritzburg corridor was ANC controlled. The South Coast shows a patchwork of political affiliations. In the following elections, IFP support was reduced to the northern inlands as the ANC widely won support.

When comparing areas of territorial control with voting patterns after the end of apartheid in more detail, four obstacles appear. First, information about the actual location of these areas is crucial. Like other geographies of (in) security, the map of political alignment in KZN was only known to insiders and difficult to reconstruct without extensive fieldwork. Information on geographic control therefore has to be distilled from case studies and monitoring reports. Second, areas often have informal names that are difficult to trace on official maps.

Voting and violence in KwaZulu-Natal fig1
Fig. 1: Election results of 1999, 2004, 2009 national elections in KwaZulu-Natal (Source: IEC 2012)

Third, voting districts, the smallest unit for which election results are available, are not congruent with former no-go areas. On the contrary, boundaries often go through such areas or combine several antagonist territories. Fourth, voting districts’ demarcations were exposed to frequent change by the IEC and thus difficult to compare over time. Attempts were made to avoid these pitfalls by the selection of cases and by checking against other indicators of voting behaviour.

The first case under investigation is Inchanga township, located in a semi-rural area between Pietermaritzburg and Durban. Krämer (2007) gives detailed information about the political landscape (see figure 2).

Voting and violence in KwaZulu-Natal fig2
Fig. 2: Left: Sketch of informal sections in Inchanga (Krämer 2007). Right: Inchanga voting districts (Source: IEC 2012)

In the Fredville area, all sections except Tin Town were ANC dominated, while, as a result of severe fighting with neighbouring sections, Tin Town residents chose to support Inkatha. Voting patterns clearly reflect this division. In the 1999, 2004 and 2009 national elections, the ANC won more than 90% in the voting districts (VDs) 43400124/641/731/821/573 surrounding the area of Tin Town, with tendencies rising. In 2009, the Congress even won a share of almost 98% in VD 43400573 and 43400821, leaving the IFP with around 1,5%. Tin Town is a part of voting district 43400584 in which the IFP won 76% of the votes in 1999, leaving the ANC with only 20%. The registered population in the area increased enormously in the last decade. In 2004, the ANC could catch up with 31,64% (against an IFP dominance of 64,96%) and finally won in the VD with 87,71% in 2009 (IFP: 9,35%). What is interesting is the fact that the share of IFP votes continuously amounts to around 200 people. It is therefore likely that with the influx of more and more residents, the stable Tin Town IFP support was overruled rather than weakened.

The Shobashobane area in southern KZN presents another interesting case. Shobashobane is an ANC stronghold of about 240 people in an area amounting to not more than 7 square kilometres and surrounded by IFP strongholds. It basically matches VD 43992535 (see figure 3).

Voting and violence in KwaZulu-Natal fig3
Fig. 3: Shobashobane voting districts (Source: IEC 2012)

For lack of sufficient information on the 1999 elections and in order to compare ward results, municipal elections will be consulted here as an exception. The election results report of 2000 clearly shows that all voting districts but VD 43992535 had a significant IFP majority, while in this particular area 91,88% voted for the ANC. In the 2006 local elections, all wards of the Izingolweni municipality showed equal shares of almost 50% for both parties, and in 2011, the ANC clearly led the wards. National election results produce the same picture. In the overall growing voter share for the ANC, Shobashobane still remains the most striking example.

A third case is Nongoma in northern KZN, a traditional no-go area for the ANC according to Taylor (2002). Two out of three local strongmen decided in March 1999 to defect to the ANC to enhance their leverage against the third, and opened an ANC branch in the area. This strategy led to a cycle of assassinations and counter-assassinations, and Nongoma mainly remained in the hands of the IFP. Unfortunately Taylor does not give further account on the more detailed aspects of political geography, which allows only for a rough analysis. The central part of the municipal area (see figure 4) indeed shows relatively homogeneous high levels of IFP support in 2004 (municipal average 93,04%). The ANC average lies at 3,88% and the party only managed results close to 10% in VD 43503432 and 43530016/38. 2009 shows a linear increase for the ANC in all VDs, with the municipal share rising to 15,61%, while IFP votes drop to 83,07%.

Voting and violence in KwaZulu-Natal fig3 fig4
Fig. 4: Nongoma voting districts (Source: IEC 2012)

The IFP stronghold Nongoma accordingly lasted a decade after 1994, but became penetrable for the ANC during the last five years. A very similar pattern is true for Ulundi, 50km South of Nongoma, where according to Nebandla (2005) the IFP headquarters were located. Also here persistently high levels of IFP support (around 93% in 1999 and 2004) were broken by 14,92% votes for the ANC in 2009, leaving the IFP with 83,62%. In the prominent ANC stronghold Ndaleni in Richmond in turn, proportions of ANC vote remain constantly high at about 96%, while the IFP has only been able to secure little more than 1% since 1999.

The analysis of post-conflict voting patterns in representative examples of KZN’s former no-go areas shows that support in the pre-1994 period directly translated into voting behaviour. Up to the 2004 elections, both parties won a striking 90% and more in their strongholds. In recent years however, the ANC was able to win voter support in IFP dominated areas, while at the same time maintaining their high performance in their own districts. An additional important observation drawn from this analysis is that parties apart from the ANC and IFP could hardly secure more than a 1% share of the votes.

The only exception is the Democratic Alliance, a party mostly voted for in white residential areas. On the whole, however, 18 years after the democratic transition the struggle for power in KZN still follows the ANC-IFP lines exclusively, and the ANC seems to win the fight.


This study has investigated patterns of support and support mobilisation around the post-conflict elections in KwaZulu-Natal, South Africa. During the country’s violent transition phase, support of a certain party was strictly associated with the geographical area a person lived in. Violence played an important role in the establishment and maintenance of these so-called ‘no-go areas’. After the 1994 elections, a number of paramilitary structures of territorial control remained intact for many years and accounted for constantly high casualties from political violence in the region. Territorial control after 1994 also directly translated into voting patterns as the ANC and IFP were able to win between 90 and 100% of the votes in their respective strongholds, even as late as 2004. Elections were (and on a lower scale still are) accompanied by electoral violence in the form of rally disruptions, intimidation and political assassinations, showing a striking continuity from violent forms before 1994. Violence thus remains a means of political mobilisation and polarisation.

A conclusion can be drawn for the ‘risky business’ of post-conflict elections. As remarked by others, electoral logics do not necessarily lead to peace, but are prone to various forms of violence. Violence does thereby not only influence election results through its manifestation of direct electoral violence or even resumption of violent conflict. It also creates and maintains a political geography of support structures that can be reproduced by post-conflict elections, in turn reproducing the use of violence. In some places of KwaZulu-Natal, this led to the survival of sharp boundaries between areas in which one of the two antagonist parties was supported, and conflicts continue to evolve around party divisions. The establishment of a diversity of political alternatives – since it is a vital factor for a post-conflict democracy to include non-armed political movements – seems to be severely suppressed. These observations feed the scepticism against the ‘cure-all’ power of post-conflict elections and point towards the importance of the question of how armed groups’ support structures turn into electoral support. Measures to open up the political landscape, deescalate heated up party-antagonisms and overcome geopolitical borders of support structures seem to be crucial elements for post-conflict elections that introduce a pluralist democracy beyond the voting process.

The case study, however, also underlines the important role non-violent mobilisation could play for armed groups turned into political parties. Especially in the long run, ideological mobilisation seems to be more than a mere supplement to coercion. The ANC was able to develop a nationwide narrative that links the party to the armed struggle and revolutionary victory. As Booysen (2011) points out, this is a main pillar of the ANC’s ongoing success story in times of increasing numbers of service delivery protests. The ANC furthermore succeeded in undermining the IFP’s claim of exclusively representing the ‘Zulu culture’, and South Africa’s third post-apartheid president Jacob Zuma often refers to his Zulu background by wearing the traditional dress of Zulu chiefs. The ANC’s increasing power in the last decade seems to finally put an end to the IFP-ANC stalemate in South Africa’s eastern province. Consequently, violence as a political strategy can never be analysed without taking self-presentations, narratives and other factors of ideological mobilisation into account – as a view of post-conflict elections without the frame of violence only produces a limited image.

Altogether, an academic effort to better understand the relationship between violence, territorial control and political mobilisation is necessary in order to advise conflict resolution actors on the risks and complexities of post-conflict elections.


  1. Aitchison, John ed. 1990. The Seven Days War: 25–31 March 1990. The victim’s narrative. Centre for Adult Education, University of Natal. Durban, University of Natal.
  2. Aitchison, John 1993. Numbering the dead. The course and pattern of political violence in the Natal Midlands 1987–1989. M.A. thesis, University of Natal, Durban.
  3. Aitchison, John 2003. KwaZulu-Natal: The pre-election wars of the 1990s. In: Greenstein 2003, pp. 73ff.
  4. Allison, Michael 2010. The legacy of violence on post-civil war elections: The case of El Salvador. Studies in Comparative International Development, 45, pp. 104–124.
  5. Autesserre, Severine 2010. Peacetime violence. Post-conflict violence and peacebuilding strategies. Program on States and Security. Available from <> [Accessed 11 June 2012].
  6. Basedau, Matthias, Gero Erdmann and Andreas Mehler eds. 2007. Votes, money and violence: Political parties and elections in Sub-Saharan Africa. Uppsala, Nordiska Afrikainstitutet.
  7. Benini, Aldo, Anthony Minnaar and Sam Pretorius 1998. Persistent collective violence and early warning systems: The case of KwaZulu-Natal, South Africa, Armed Forces & Society, 24 (4), pp. 501–518.
  8. Bond, Patrick 2011. Welcome to Durban. In: Bond, Patrick ed. Durban’s climate gamble. Trading carbon, betting the earth. Braamfontein, UNISA Press.
  9. Bonnin, Debby 1997. Spatiality in the construction of identity: African women and political violence in KwaZulu-Natal. Society in Transition, 1 (1–4), pp. 27–42.
  10. Bonnin, Debby 2006. Legacies of political violence: An examination of political conflict in Mpumalanga Township, KwaZulu-Natal, South Africa. Transformation: Critical Perspectives on Southern Africa, 62, pp. 59–83.
  11. Booysen, Susan 2011. The African National Congress and the regeneration of political power, 1994–2011. Paper presented at a conference on One hundred years of the ANC, Johannesburg, 20–24 September.
  12. Chaturvedi, Ashish 2005. Rigging elections with violence. Public Choice, 125, pp. 189–202.
  13. Collier, Paul, Anke Hoeffler and Mí¥ns Söderbom 2007. Post-conflict risks. Journal of Peace Research, 45, pp. 461–478
  14. De Haas, Mary 2008. Yet another year of the gun in KZN. 23 December 2008. Violence Monitor, December. Available from: <> [Accessed 22 June 2012].
  15. De Haas, Mary 2009a. Pre-election violence. April 14, 2009. Violence Monitor, April. Available from: <> [Accessed 22 June 2012].
  16. De Haas, Mary 2009b. Pre-election update. April 20, 2009. Violence Monitor, April. Available from: <> [Accessed 22 June 2012].
  17. Dugard, Jackie 2003. A future settlement. In: Greenstein 2003, pp. 22ff.
  18. Dunn, Kate 1998. Why the killing (and making up) in KwaZulu-Natal. Christian Science Monitor, 90 (175), April.
  19. Dunning, Thad 2011. Fighting and voting: Violent conflict and electoral politics. Journal of Conflict Resolution, 55 (3), pp. 327–339.
  20. Du Toit, Pierre 2001. South Africa’s brittle peace: The problem of post-settlement violence. New York, Palgrave.
  21. Fischer, Jeff 2002. Electoral conflict and violence: A strategy for study and prevention. IFES White Paper, February 5. Washington, D.C., International Foundation for Electoral Services (IFES)
  22. Greenstein, Ran ed. 2003. The role of political violence in South Africa’s democratisation. Johannesburg, CASE publication.
  23. Griggs, Richard 1996. Geostrategies for both power and peace in KwaZulu-Natal: A neorealist analysis. GeoJournal, 39 (4), pp. 387–396.
  24. Hamilton, Georgina and Gerhard Maré 1994. The Inkatha Freedom Party. In: Reynolds 1994, pp. 73–88.
  25. Höglund, Kristine 2008. Violence in war-to-democracy transitions. In: Jarstad and Sisk 2008, pp. 80–101.
  26. Höglund, Kristine 2009. Electoral violence in conflict-ridden societies: Concepts, causes, and consequences. Terrorism and Political Violence, 21 (3), pp. 412–427.
  27. Höglund, Kristine and Anna Jarstad 2011. Toward electoral security: Experiences from KwaZulu-Natal. Africa Spectrum, 46 (1), pp. 33–59.
  28. Höglund, Kristine, Anna Jarstad and Mimmi Söderberg Kovacs 2009. The predicament of elections in war-torn societies. Democratization, 16 (3), pp. 530–557.
  29. Höglund, Kristine and Mimmi Söderberg Kovacs 2010. Beyond the absence of war: the diversity of peace in post-settlement societies. Review of International Studies, 36, pp. 367–390.
  30. Hug, Simon 2011. Elections, political parties, and civil war. Paper prepared for presentation at the Computer Supported Cooperative Work (CSCW) Workshop on ‘Agents of Violence and Institutional Constraints’, University of Mannheim, April 28–29.
  31. Independent Electoral Commission (IEC) 2012. Electoral Commission of South Africa: Elections. Available from: <> [Accessed 22 June 2012].
  32. Jarstad, Anna 2009. Post-accord elections and armed conflict. In: Swain et al. 2009, pp. 153–163.
  33. Jarstad, Anna and Timothy Sisk 2008. From war to democracy: Dilemmas of peacebuilding. Cambridge, NY, Cambridge University Press.
  34. Johnston, Alexander 1996. The political world of KwaZulu-Natal. In: Johnson, R.W. and Lawrence Schlemmer eds. Launching democracy in South Africa: The first open election, April 1994. New Haven/London, Yale University Press.
  35. Kalyvas, Stathis N. 1999. Wanton and senseless? The logic of massacres in Algeria, Rationality and Society, 11 (3), pp. 243–285.
  36. Kalyvas, Stathis N. 2003. The ontology of ‘political violence’: Action and identity in civil wars. Perspectives on Politics, 1 (3), pp. 475–494.
  37. Kalyvas, Stathis N. 2006. The logic of violence in civil war. Cambridge, Cambridge University Press.
  38. Kentridge, Matthew ed. 1990. An unofficial war: Inside the conflict in Pietermaritzburg. Claremont, David Philip.
  39. King, Charles E. 2004. The micropolitics of social violence. World Politics, 56 (3), pp. 431–455.
  40. Krämer, Mario 2007. Violence as a routine: Transformations of local-level politics in the disjunction between centre and periphery in KwaZulu-Natal (South Africa). Köln, Rüdiger Köppe Verlag.
  41. Laakso, Liisa 2007. Insights into electoral violence in Africa. In: Basedau et al. 2007, pp. 224–252.
  42. Leeb, Wendy 1988. Daily life, daily death in Pietermaritzburg. Unpublished paper available from the Gerhard Maré Collection at the Alan Paton Archive in Pietermaritzburg.
  43. Lindberg, Staffan 2006. Democracy and elections in Africa. Baltimore, Johns Hopkins University Press.
  44. Lodge, Tom 1999. Consolidating democracy: South Africa’s second popular election. Johannesburg, Witwatersrand University Press.
  45. Mac Ginty, Roger 2008. No war, no peace: The rejuvenation of stalled peace processes and peace accords. Basingstoke, Hampshire, Palgrave Macmillan.
  46. Mansfield, Edward and Jack Snyder 2005. Electing to fight: Why emerging democracies go to war. Cambridge/London, MIT Press.
  47. Meer, Fatima ed. 1989. Resistance in the townships. Durban, Madiba Publishers.
  48. Mehler, Andreas 2007. Political parties and violence in Africa. Systematic reflections against empirical background. In: Basedau et al. 2007, pp. 194–223.
  49. Metelis, Claire 2010. Inside insurgency: violence, civilians, and revolutionary group behaviour. NY/London, New York University Press.
  50. Minnaar, Anthony ed. 1992. Patterns of violence. Case study of conflict in Natal. Pretoria, Human Sciences Research Council.
  51. Minnaar, Anthony 1992a. ‘Patterns of violence’: An overview of conflict in Natal during the 1980s and 1990s. In: Minnaar 1992, pp. 1–25.
  52. Minnaar, Anthony 1992b. ‘Undisputed kings’: Warlordism in Natal. In: Minnaar 1992, pp. 61–93.
  53. Mottiar, Shauna 2006. KwaZulu-Natal. EISA Election Update, (1), pp. 37–45.
  54. Muthien, Yvonne and Meshack Khosa 1998. Constructing a regional political identity in South Africa. Social Identities, 4 (3), pp. 457–472.
  55. Nebandla, Injobo 2005. Freedom from Strife? An assessment of efforts to build peace in KwaZulu-Natal. Braamfontein, Centre for the Study of Violence and Reconciliation.
  56. Ohlson, Thomas and Mimmi Söderberg Kovacs 2009. Peace through democratisation: The challenges of postwar democratisation in weak and war-torn states. In: Swain et al. 2009, pp. 165–182.
  57. Piper, Laurence 2004a. Politics by other means: The practice and discourse of violence in KZN. EISA Election Update, (4), pp. 22–24.
  58. Piper, Laurence 2004b. The end of the KZN as we know it: The rise of the ANC in 2004. EISA Election Update, (8), pp. 20–22.
  59. Reynolds, Andrew ed. 1994. Election ’94 South Africa. The campaigns, results and future prospects. Claremont, David Philip Publishers.
  60. Reynolds, Andrew and Timothy Sisk 1999. Elections and electoral systems. Implications for conflict management. In: Reynolds, Andrew and Timothy Sisk eds. Elections and Conflict Management in Africa, pp. 11–36. Washington, D.C., US Institute for Peace Press.
  61. Reynolds, Andrew 1994. The results. In: Reynolds 1994, pp. 173–210.
  62. Richards, Paul ed. 2005. No peace, no war: An anthropology of contemporary armed conflicts. Athens, Ohio University Press.
  63. Schuld, Maria 2012a. Violence in post-conflict transitions to democracy: A phenomenological analysis of violent event types in KwaZulu-Natal, South Africa. Diplom Dissertation, Freie Universität Berlin.
  64. Schuld, Maria 2012b. Old wine in new skins? Post-conflict violence in KwaZulu-Natal, South Africa. Unpublished manuscript, June 2012.
  65. Shaw, Mark 2002. Crime and policing in post-apartheid South Africa, Transforming under fire. Cape Town, David Philip Publishers.
  66. Simpson, Graeme 1993. Gewalt in Südafrika. German translation of ‘Explaining endemic violence in South Africa’. Weltfriedensdienst Quersbrief, 3, pp. 10–15.
  67. Sisk, Timothy 2008. Elections in fragile states: Between voice and violence. Paper prepared for the International Studies Association Annual Meeting San Francisco, California, March 24–28.
  68. South African Institute of Race Relations 2005. South Africa survey 2004/05. Johannesburg, SAIRR.
  69. South African Institute of Race Relations 2011. South Africa survey 2010/11. Johannesburg, SAIRR.
  70. Steenkamp, Christina 2005. The legacy of war: Conceptualizing a ‘culture of violence’ to explain violence after peace accords. The Round Table: The Commonwealth Journal of International Affairs, 94 (379), pp. 253–267.
  71. Suhrke, Astri 2012. The peace in between. In: Suhrke, Astri and Mats Berdal eds. The peace in between: Post-war violence and peacebuilding. London/New York, Routledge.
  72. Swain, Aahok, Ramses Amer and Joakim í–jendal eds. 2009. The democratisation project. London, Anthem Press.
  73. Taylor, Rupert 2002. Justice denied: Political violence in KwaZulu-Natal after 1994, African Affairs (101), pp. 473–508.
  74. Truth and Reconciliation Commission 1998. Truth and Reconciliation Commission of South Africa Report, Volume 5. Cape Town, CTP Book Printers.
  75. Woods, Gavin 1992. Natal Violence: A contemporary analysis of underlying dynamics. In: Minnaar 1992, pp. 37–48.
  76. Zulu, Paulus, Zolani Dlamini, Lungisile Khuluse and Nation Mabaso 2009. The 2009 KwaZulu-Natal Election: A drop in levels of intimidation and violence? Report of the African Alliance for Peace (AfAP) 2009 Observer Mission. Durban, ACCORD.


  1. For a comprehensive literature review, see Hug 2011.
  2. That have surprisingly never been systematically investigated by academics or civil society organisations (Schuld 2012b).