Overview
There are grounds for tracing service delivery protests back to the apartheid era, and a strong case can be made for linking them to discontent that was noted in surveys conducted in the late 1990s and to the social movements that emerged in the years after 2000 (Seekings 2000, Nthambeleni 2009). However, analysts agree on dating the contemporary phenomenon back to 2004 (Atkinson 2007, p. 54, Booysen 2007, p. 24, Pithouse 2007).
In defining the object of investigation, Booysen (2007, p. 21) writes of ‘grass-roots protests against both the quality of service delivery and public representation of grass-roots’ service delivery needs'. Pithouse (2007), who draws on detailed knowledge of shack-dwellers' protests, rejects this ‘economistic’ approach, arguing that the protests are about ‘citizenship’, understood as ‘the material benefits of full social inclusion … as well as the right to be taken seriously when thinking and speaking through community organisations.’ Perhaps the distinction between the two approaches is more one of focus and level of analysis than a substantive difference about the collection of events that requires explanation. Atkinson's interest is in ‘social protests – many of them violent – that wracked black and coloured townships …’ (Atkinson 2007, p. 54). This neatly sidesteps the debate about whether ‘service delivery’ is a defining characteristic, but it opens the scope too broadly. The xenophobic violence of May 2008, for instance, had very different dynamics, and strikes and other occupation-related protests (such as those by police, soldiers, students and street traders) are also distinct. This analysis will not, however, ignore the xenophobia and worker solidarity present in some of the protests that concern us here, or reject the possibility that there may be underlying causes linking the various actions.
It appears that what we are attempting to grapple with is locally-organised protests that place demands on people who hold or benefit from political power (which includes, but is not limited to, local politicians). These have emanated from poorer neighbourhoods (shack settlements and townships rather than suburbs). Perhaps this is best captured by defining the phenomenon as one of local political protests or local protests for short. The form of these actions relates to the kind of people involved and the issues they have raised. They have included mass meetings, drafting of memoranda, petitions, toyi-toying, processions, stay-aways, election boycotts, blockading of roads, construction of barricades, burning of tyres, looting, destruction of buildings, chasing unpopular individuals out of townships, confrontations with the police, and forced resignations of elected officials.
The varied nature of such protests makes them difficult to quantify. One potential source is data collated by the Incident Registration Information System (IRIS), which is maintained by the South African Police Service (SAPS) Crime Combating Operations' Visible Policing Unit (VPU). This includes a subset on ‘public gatherings’ (Vally 2009, p. 10). The definition here of the term ‘public gathering’ derives from the Regulation of Gatherings Act 1993, which recognised freedom of assembly and protest as democratic rights, and sought to ensure that these were practised in a peaceful manner (State President 1994, Duncan 2009, p. 4). ‘Gatherings’ were not defined in the Act, although the term included ‘processions’ (also undefined), and according to Duncan (p. 6) events involving 15 people or fewer were excluded, as these were regarded as ‘demonstrations’ (again undefined). From a list of ‘prominent reasons’ for gatherings that the VPU provided to Centre for Sociological Research (CSR) researcher Natasha Vally, it is clear that a large majority of such events were protest-related (Vally 2009, p. 11). The reasons included ‘demand wage increase’, ‘solidarity’, ‘dissatisfied with high crime rate’, ‘resistance to government policy’, ‘mobilising of the masses’, ‘in sympathy with oppressed’, ‘service charges’, and, finally, ‘sporting event’. While many gatherings were probably local political protests, the quantity of these as a proportion of the total is unknown. Contrariwise, some of the actions defined above would not have been included in the IRIS data. Notwithstanding these qualifications, the data presented in the tables below provide some indication of the scale of the protest movement. Data for 2008/9 are not yet available.
| 2004/05 | 2005/06 | 2006/07 | 2007/08 | Total | |
|---|---|---|---|---|---|
| Gauteng | 1932 | 2205 | 1888 | 1451 | 7476 |
| Western Cape | 557 | 511 | 577 | 642 | 2287 |
| KwaZulu-Natal | 1891 | 2529 | 1774 | 1146 | 7340 |
| Limpopo | 660 | 915 | 665 | 642 | 2882 |
| Eastern Cape | 754 | 1383 | 1626 | 733 | 4496 |
| North West | 1108 | 1341 | 1159 | 1502 | 5110 |
| Free State | 506 | 728 | 713 | 483 | 2430 |
| Mpumulanga | 295 | 336 | 337 | 4 | 972 |
| Northern Cape | 301 | 489 | 427 | 400 | 1617 |
| Total | 8004 | 10,437 | 9166 | 7003 | 34,610 |
| Source: Vally (2009), based on IRIS data. |
Table 1 provides the IRIS data for all gatherings broken down by province. The VPU recognised that some of these figures were unreliable (attributing this to the impact of institutional restructuring conducted after 2006), and the Mpumulanga figure for 2007/8 is clearly inaccurate (Vally 2009, p. 8). Similar data had been provided to the National Assembly for the middle two years, but the table presented on that occasion had five numbers that were different (Internal Question 43/2007, see Vally 2009, p. 18). These were for the Western Cape – 672 in 2005/6 and 687 in 2006/7; for Free State – 673 in 2006/7; and for Mpumulanga – 501 in 2005/6 and 547 in 2006/7. The reason for difference between the two tables is unexplained, and it is not known which of the two is more reliable, but the disparities are not great. The total gatherings in the National Assembly statistics are 10,763 for 2005/6 and 9446 for 2006/7. In proportion to both total and urban populations, the two wealthiest provinces, Gauteng and the Western Cape, had relatively fewer gatherings than the others, and North West had the highest total.
The VPU distinguishes between gatherings that are ‘peaceful’ and those that are ‘unrest-related’, but it only provides national-level statistics under these two categories. The numbers are given in Table 2. Again there is a discrepancy between the figures that were provided to Vally and those presented to the National Assembly. The VPU did not provide definitions of ‘peaceful’ and ‘unrest-related’ gatherings and it is not known how the distinction relates to that between ‘legal’ and ‘illegal protests’. In 2004/5, the last year for which the data are available, the government said that there were 5085 legal and 881 illegal protests (Booysen 2007, p. 23, citing Cape Argus 13 October 2009; see also Atkinson 2007, p. 58). It is probable that ‘unrest’ was more likely to occur in those instances where a gathering occurred without police permission, and thus was generally regarded by them as ‘illegal’, because in these cases the police would be entitled to use force to disperse the event (Vally 2009, p. 12).
| 2004/05 | 2005/06 | 2006/07 | 2007/08 | Total | |
|---|---|---|---|---|---|
| Peaceful | 7382 | 9577 9809 | 8486 8703 | 6304 | 31,749 |
| Unrest-related | 622 | 860 954 | 680 743 | 699 | 2861 |
| Total | 8004 | 10,437 10,763 | 9166 9446 | 7003 | 34,610 |
| Source: Vally (2009) based on IRIS data. Figures presented to National Assembly for 2005/6 and 2006/7 in italics. |
Given the imprecise nature of the concepts used by IRIS and a lack of confidence in the quality of data collection, and given that no comparable figures are available for other countries, one cannot make any strong claim based on the statistics presented here. For this author, however, the numbers seem very high, and many times greater than the kind of figures presented for protests around the world (Walton and Seddon 1994, Dwyer and Seddon 2002, Seddon and Zeilig 2005).1 If it is assumed that nearly all the unrest-related gatherings were protests of some kind, which is a reasonable assumption, then in an average week over the period 2004 to 2008 there were more than ten protests involving ‘unrest’. Many of these would have been labour-related, but it seems likely that a high proportion were local political protests, and one can certainly agree with Duncan (2009, p. 4) when she draws on the same statistics to propose that ‘protest action became a significant feature of political life in South Africa during Mbeki's term of office.’
But what of the post-Mbeki era?
Another source of data is the ‘Hotspots Monitor’ maintained by Municipal IQ, a private research company (see Figure 1 below). This covers ‘major protests’ around service delivery, and includes most such events covered by the media. Comparing the numbers from this source with those from IRIS, there is clearly a considerable amount of activity that is not being picked up by reporters. The data are interesting, however, because they capture newsworthy protests – generally sizeable events involving a measure of unrest.
The graph highlights an initial peak in 2005, followed by a lull in 2006, the year of local government elections. The elections took place in March, and the nadir may not have been picked up in IRIS statistics because these are presented by financial year (i.e. from 1 March). The implication is that, for a period, activists put their energies into the elections, and/or that broader populations placed some hope in the possibility of electoral politics addressing their concerns. In 2007, the number of protests shoots up again. There is then a slight tailing-off during the period that the Zuma campaign gathers pace and leads into the national election of April 2009. The figures for 2009 are staggering, and only five of the protests took place before the election. There were a further 15 protests in the next four months (approximately), with the remainder, 63, occurring between the Municipal IQ reports dated 21 July (2009a) and 17 November (2009b). Thus, according to this source, there were more service delivery protests in the first seven months of the Zuma administration than in the last three years of the Mbeki administration. Unlike the lull that followed the 2006 elections, the 2009 election was followed by a storm.2
Existing analyses and Mbeki-era protests
There is a flourishing literature on South Africa's social movements, a developing debate about participatory planning and a growing body of work on xenophobic violence; for recent examples see Dawson (2008), Duponchel (2009) and Sinwell (2009). By contrast, academic writing on local political protests is scanty.
Booysen (2007) draws on national survey data and case studies conducted in five municipalities where service delivery protests had occurred in the two years preceding the March 2006 election, and proposes that:
The South African local electorate thus appears to believe that ‘voting helps and protest works’ when it comes to deciding on a repertoire of action to optimise service delivery in communities. … Results [of her research] indicate that communities continue believing that the ANC remains the party that is best equipped to take care of service delivery. (Booysen 2007, p. 31)
Pithouse (2007) takes this last point further. He argues that ‘there is a pervasive sense that the state disrespects people by lying to people at election times and by failing to listen to them at other times.’ Where councillors are present, he says, they ‘most often function as a means of top-down social control aiming to subordinate popular politics to the party’. There are doubtless some councillors who live alongside their constituents and do their best to articulate popular concerns, but where they are absent, or operate in a top-down or corrupt manner, space is opened up for new leaders to fill the void.
Drawing mainly on media reports, Atkinson's (2007) account largely confirms that of Booysen, but provides some new emphases. The problem is not merely with a lack of services, she observes, but with something not picked up in roll-out statistics: poor maintenance and management of those services (lack of repairs, problems with billing and other issues). In her view, more attention should be paid to technical issues and skills development. She also notes that even where councillors are not corrupt in the criminal sense, they engage in various forms of self-enrichment and lead ‘fat-cat’ lifestyles. She reminds us that some harbour significant political ambitions (which might, one assumes, impact on their concern to serve higher party interests rather than their communities). Pressure to use politics to advance personal interest is exacerbated by high unemployment, she asserts. This final point leads Atkinson to propose, moralistically, that ‘councillors need to have a built-in sense of self-control’, and also, surely anti-democratically, that councillors ‘should be able to derive income from other means … such as employment or private enterprise’. In general, her analysis tends towards technocratic conclusions. For instance, she sympathetically quotes the Minister of Housing, who bemoaned the problems caused by poor communication between the three spheres of government (national, provincial, and local).
Four detailed case studies have been conducted. Two of these were undertaken by the Centre for Development and Enterprise (CDE): Khutsong, located within Merafong Municipality, and Phumelela Municipality in the Free State. A general conclusion of the CDE report is that:
It is not the intensity of the disturbances that should concern us most … [but] that they have been widespread, repeated over a long period, and triggered by a variety of grievances, thus illustrating a persistent and general malaise in our system of governance. (CDE 2007, p. 53)
Phumelela includes three small towns (Memel, Vrede, and Warden), and protests, sometimes violent, took place in each of them between mid-September and early November 2004. According to the CDE's researchers (2007, p. 12), there was ‘some contact among the groups in the three towns’, and protests which had occurred earlier in September in Harrismith, about 50 kilometres away, had a ‘demonstration effect’. Significantly, leadership was provided by ‘unemployed people and youths of school-going age’, who constituted ‘Concerned Youth Groups’, as they called themselves. Unusually, one of the marches was joined by whites from the local ratepayers' association. The CDE report adds two new dimensions to the analysis of local protests. First, poor policing was blamed for exacerbating problems. Second, the three towns had experienced an influx of migrants from rural areas, who placed additional strains on resources and relationships. Otherwise the case study provides further evidence for the two themes that are now familiar: poor services (roads, dirty water supply, and insufficient sports facilities were specifically mentioned), and pitiable local government (including nepotism, lack of transparency and indifferent, incompetent, and contemptuous officials). In terms of developing a prognosis, the Phumelela case (CDE 2007, p. 54) offers an interesting insight:
The unrest led to government action to remedy the local state of affairs. Individuals were removed from office, national government departments provided assistance, and administrative and managerial chaos was cleared up. However 18 months after the unrest, residents of the former townships were still waiting for significant improvements in their living conditions. The individuals on whom they had focused their frustrations might have been removed from office, but had gone on to equal or better posts elsewhere.
The other two studies were undertaken by the Centre for Development Support (CDS) based at the University of the Free State (Botes et al. 2007). These covered Phomolong in the Matjhabeng Municipality, again in the Free State, and the Nelson Mandela Bay Metropolitan Municipality (i.e. Port Elizabeth). The Phomolong protests took place between 4 February and mid-April 2005. Botes et al. (2007, p. 17) describe the beginning of the rebellion: ‘Protesters brandishing toilet buckets and banners, and armed with pipes and sticks, sang protest songs and blocked the streets.’ In the wake of this protest people employed outside the area were prevented from going to work, and schools were closed for a month. There was also some looting, with, according to one source, pillaging of Indian-owned businesses encouraged by local businessmen. The action only abated after the removal of a particularly unpopular councillor, which had been a prominent community demand (though possibly one inspired by rivals within the African National Congress [ANC]). Other major concerns included the continuation of the bucket system and complaints about housing delivery. There was a feeling that the municipality discriminated against Phomolong in favour of Welkom, the area's main town, with this reflected, for instance, in complaints about the award of tenders. Botes et al. show that while bucket sanitation had declined between 1996 and 2001, it was more prevalent in Phomolong than in the municipality as a whole, as was informal housing. As an important underlying cause, the authors point to rising unemployment related to the decline of the mining industry, but the protesters actually complained that jobs were going to people from Lesotho rather than to locals. Three points are emphasised in the study: first, the importance of a lack of service delivery; second, the failure of the specific councillor to address concerns; third, the prominence of unemployed youth in the struggle.
Port Elizabeth was hit by protests during May 2005. These were widespread, occurring in a least six locations, but damage to property was relatively minor compared to elsewhere, and mainly involved burning of tyres. Botes et al. (2007) note, once more, the importance of service delivery (notably around housing issues) and unresponsive local government (including unfulfilled promises). Yet again, unemployed youth were to the fore in the struggles, but so too were school students (whose schooling was disrupted), and there was also participation by older women and men. Much of the action involved shack dwellers. The authors also mention police violence (which occurred in Phomolong as well), and they record suggestions that factionalism within the ANC (possibly South African Communist Party (SACP) and unions versus others) may have stimulated the protests. Finally, they propose that the high level of socio-economic inequality that exists in Port Elizabeth, and large-scale rural–urban migration, were significant structural considerations.3
The case studies tend to substantiate the main theme raised in the general literature: that the protests were principally about the lack of basic services and inadequate local administration. But the role of unemployed youth and even school students now comes into view, and high unemployment is surely a critical factor. Related to this, the importance of comparative local disadvantage, and hence inequality more broadly, need to be considered. There is also some suggestion of protests being linked to power struggles within the ANC, and of poor and/or violent policing contributing to the violence.
Zuma-era protests
For accounts and analyses of the protests that have occurred since the election of Jacob Zuma as president in May 2009, we must turn to press reports and our own research. We described the latter as ‘rapid response research’. A team of twelve researchers investigated five protests, doing so between a few days and about a month after the main events. The five included Piet Retief (where protests peaked at the end of June 2009), Diepsloot (early July), Balfour (mid-July), Thokoza (late July), and Standerton (throughout October).4 Thokoza and Diepsloot are located in Gauteng and the other three are in Mpumulanga, which has seen a high proportion of the Zuma-era protests. While the research lacked the depth and breadth of typical academic studies, and we regarded its conclusions as tentative, it had the benefit of immediacy. Also, in filling a gap between reportage and rigorous case studies, it enabled us to intervene in public debate. The primary methods were interviews (mainly with participants because we were particularly interested in motivation), observation (principally meetings), and analysis of memoranda. The research was backed up by a workshop held at the end of October and attended by protest activists. This occasion, which we filmed, enabled us to test and modify our analysis.
The form of the new protests was similar to the earlier ones. Generally (but not always), a memorandum was formulated at a community meeting and delivered to the local municipality (though, in the case of Piet Retief, it was addressed to the provincial premier). There were processions (with marchers sometimes wielding pangas, axes, and knobkerries), stay-aways, the construction of simple barricades within townships and/or informal settlements, occasional blockading of major highways, much burning of tyres (perhaps an emblematic feature of the protests), confrontations with the police, some arson of symbolically significant buildings (e.g. council offices and councillors' houses), and often the looting of shops. Police brutality has been common, often precipitating counter-violence by protesters, and in Piet Retief at least two protesters were shot dead.5 However, in some cases the police lacked the resources to quell disturbances, leaving people in control of their own streets.
The range of demands from the new protests was also broadly similar to those in the Mbeki era, though the specifics varied. Housing has been an important concern – both a lack of formal accommodation and the extremely poor quality of much subsidised (i.e. Reconstruction and Development Programme (RDP)) housing. In Balfour, the main memorandum demanded such things as a training centre, a township police station, a mini-hospital and internet connection in the library, as well as clean water, the paving of roads, and street lighting. However, it is quite possible that governance issues have become more pronounced. Indeed, in Piet Retief and Standerton, possibly the two biggest protests, services hardly figure in the formal demands, except as adjuncts to a focus on maladministration. The Piet Retief memorandum mainly covered misappropriation of funds, with much of it couched in terms of demands for information (though see Dlamini 2009a, pp. 1–15 for a broader account). In Standerton, the main issues were around corruption, although press interviews suggest that poor services were also a factor. Accountability has emerged as a major theme in the unrest.
Many of the protests have occurred in shack settlements and others have emanated from poorer parts of townships. In the case of Thokoza, hostel residents were also involved. However, it would be an oversimplification to explain the protests in terms of poverty as such. There has been a general, albeit modest, decline in absolute poverty since 1994 (Presidency 2009); many of the poorest parts of the country, notably the rural areas, have been free of protests; and, to the contrary, there has been considerable unrest in Gauteng, where services tend to be better than elsewhere in the country. Comparative poverty, that is, impoverishment relative to some locally important reference group, is probably more significant. According to Municipal IQ (2009b):
An analysis of Census 2001 data (the latest available ward data) reveals that protests take place in wards: (1) that have higher unemployment rates than the average ward in their municipalities … [and] (2) that have worse access to services than the average ward in their municipalities.
Income inequality must rank highly among the structural determinants of the protests. After 15 years of democratic rule, the country's Gini coefficient remains one of the highest in the world.6 Recent government statistics show that between the end of apartheid and 2007, South Africa's Gini coefficient dropped slightly, but also that the income of the poorest 20% of the population declined from 2.7% to 2.3% of total income (Presidency 2009). Increasing unemployment, which has followed in the wake of the global financial meltdown, will have expanded inequality and intensified problems of survival. According to the Quarterly Labour Force Survey (Statistics South Africa 2009; see also Letsoalo 2009), total employment dropped by 475,000 in the first six months of 2009, and by a further 484,000, 3.6% of the workforce, in the three months that followed. By the end of September 2009, the unemployment rate, already high, had increased to 31.1% (once discouraged work-seekers were included). But these figures do not capture the full extent of the employment problem, for it is highly likely that the ‘underemployed’ (casuals, subsistence traders etc.) tend to be poorer than the unemployed (Wale and Alexander 2009).7
The importance of unemployment is reflected in the predominance of young adults, particularly those who are unemployed (or underemployed), in the struggle. In the case of Balfour, the memorandum came from ‘the youth’, and the protest was led by young people, women as well as men, who described themselves as ‘community leaders’. These leaders were fluent in English and had a good command of strategy and tactics, and the main spokesperson was a graduate from the University of the Witwatersrand. From some of them we learned that housing is a generation issue, because many twenty-somethings would like to start their own family, yet are forced to stay in their parents' backyards. Two other factors have contributed to the development of a new generation of leaders. First, at least in the minds of some of the Balfour youth, older community leaders were often compromised by business interests linked to municipal tenders and by their membership of the ANC. Second, unemployed youth had both a shared identity and the time to organise. While the significance of generation has been more pronounced in Balfour than elsewhere, it has been an important dynamic in the protest movement generally.
The very sharpness of the recent growth in unemployment may help to explain the current spike in protest action, but we should be wary of pushing the joblessness argument too far. Unemployment peaked in 2002, when the expanded rate (including discouraged job-seekers) reached 41.8%. This was two years before the start of the local protest phenomenon, and for its first four years the rate of unemployment was in decline. Moreover, employment has not been a major issue raised in the protests, and where it has been it is generally formulated in relation to the allocation of jobs.
For a fuller understanding of the timing and scale of the movement, politics must be brought back into the equation. Here, a consideration of Zuma's leadership of the ANC is critical. There is no doubt that his defeat of Thabo Mbeki at the Polokwane congress, held in December 2007, was widely and enthusiastically welcomed by poorer people in general (see, for example, Ceruti 2008). Analysts contrasted the pro-Zuma delegates with those who supported Mbeki, noting the high proportion of unemployed and trade unionists among the former, and the more middle-class nature of the latter. Moreover, it seems highly likely that local protests and strikes fed into the anti-Mbeki campaign.8 For the next nine months the party and the government were split between the Mbeki-ites and the Zuma-ites, with Mbeki eventually resigning as President in September 2008 (replaced initially by Kgalema Motlanthe). It is plausible that from then until the April 2009 election there was a decline in the number of protests. We can now return to the ‘dual repertoires’ analysis. While a high proportion of the electorate supported the ANC in the hope that it would bring improvements to living conditions, they also did so because there was no serious alternative. Other formations were either marginal or regarded as unsympathetic, with the latter including the white-led Democratic Alliance and, in the last national election, the pro-Mbeki Congress of the People (COPE). Given massive disappointment with Mbeki, especially over service delivery, electing a new ANC leader was a perfectly rational step. Protest will have assisted Zuma's rise, but in the immediate pre-election period the main battle was to secure the victory of the ‘pro-poor’, Zuma-led ANC. Thus far, Booysen's argument holds up, albeit in a form that requires an appreciation of intra-ANC battle.
How, then, should we analyse the politics of unrest since the election of Zuma? This is only possible if we recognise the importance of the distinction between national government (led by Zuma) and local government. Only closer examination will reveal whether unpopular councillors were actually Mbeki-ites, but, in any event, opposition to local authorities is possible without it destabilising support for Zuma and his administration. Provincial governments, also elected in April 2009, are generally seen as extensions of national politics. In our interviews we found no evidence that Zuma, or the ANC in general, were held responsible for people's problems, and some interviewees argued that the timing of the protests was linked to having a government that, at long last, would listen to people and address their complaints. When national leaders have appeared at protest hot-spots they have been well received (rapturously so in the case of ANC Youth League leader Julius Molema when he joined Standerton crowds celebrating the overthrow of their mayor).
Gwede Mantashe, general secretary of the ANC, has said (in an address to a South African Municipal Workers' Union conference): ‘You find that in the majority of cases a march is led by members of the ANC.’ The ANC carried out its own study of recent protests (which it refused to make public), so it probably has good evidence for his statement. Moreover, our own investigations tend to confirm its validity. In Balfour, most of the leaders were ANC sympathisers, and in Piet Retief the memorandum of demands ended: ‘We would also like to state it clear[ly] that the … citizen's concerned group are members of the ANC.’ However, it would be wrong to see the protests as a battle between good/the people/national leaders and bad/Mbeki supporters/local councillors. The internal politics of the ANC are more complicated. At another extreme they involve battles over patronage: access to tenders, jobs and so on (see, for instance, Dlamini 2009b). Somewhere between these poles lie fights between different components of the ANC and its Alliance partners. For instance, the Diepsloot protest recorded in our report was probably led by the local branch of the South African National Civic Organisation. In practice, the politics of the unrest can simultaneously involve various strands, with different activists having different motivations. So, a particular protest might contain both a popular antipathy to corruption and a struggle for patronage. To complicate matters further, activists of other parties, notably youth members of the Pan Africanist Congress (PAC) in Balfour and COPE members (ex-PAC) in Piet Retief, have backed the protests. The ‘dual repertoires’ thesis has now broken down, for it cannot capture the messiness of the politics of local protests, let alone their broader dynamics.
Returning to the local/national distinction, there is a suggestion that the nationally focused issues that began to emerge in the Mbeki-era protests, notably around cross-border disputes such as that in Khutsong, have been tempered in the new period, with, for instance, housing, partially a national responsibility, being represented as a local problem. If so, this would fit our account, with Zuma-ite ministers given space to deflect responsibilities that were not available to their predecessors.
Such openings have been useful when it comes to calming the revolt, as seen, for instance, in Balfour, where this author attended a meeting held in the town hall on 27 July, a week after the start of the unrest. The platform consisted of Malusi Giqaba, Deputy Minister of Home Affairs, and members of the provincial executive, and the floor included about 60 representatives of various parties and interests. The platform deployed a number of tactics. First, they simply listened, showing respect but also assessing the temper of the meeting and noting differences of opinion. Second, they said, in effect, ‘We have only recently been elected and were not aware of your problems, which we will certainly address.’ Third, they requested that the community pull back from further action for five days (with the implication of state violence if it did not). Fourth, they established a task team that would consider the various issues in the memorandum and where these should be addressed (municipal, provincial, or national government). This served to marginalise the youth, who could only be a minority in the task team, giving encouragement to older leaders, particularly those members of the Alliance who had not been compromised by holding office. Fifth, attempts were made to split the youth forum (which included ANC and PAC members). The meeting was followed by an unannounced visit by Zuma. Disagreements within the ANC, which had been an important aspect of the unrest, were reworked to divide the community and stabilise the situation. This was only possible because the ANC nationally retained legitimacy, because the government was prepared to make some concessions, and because divisions within the community were multiple.
Some further aspects of the unrest are worthy of consideration. The first of these is the issue of xenophobia, which was highlighted by several reporters and some politicians as an important dynamic in the protests. It is tempting to sense an echo of the May 2008 pogrom, when more than 70 people were killed (about a third of them South Africans) and thousands of foreign residents were chased from their homes. There is no doubt that xenophobia is ubiquitous in South Africa (and spread fairly evenly across all social groupings), or that poverty and competition for resources, factors in last year's violence, were also present in this year's unrest. Moreover, one can find numerous examples of xenophobia in the protests. Its principal form has been the looting and torching of non-South African-owned shops (which have become a feature of post-apartheid South Africa, notably in the townships of Mpumulanga and the Free State).9 Sometimes these shops are owned by new Chinese immigrants and sometimes by Somalis, but most of those affected in Piet Retief were Pakistani-owned and most in Balfour were owned by Ethiopians.
In Piet Retief and Balfour we interviewed the shopkeepers who had suffered. Their accounts and opinions were very similar. They sympathised with the protesters; believed they had good relations with local communities; felt that the violence was not intentionally xenophobic, but rather the behaviour of an opportunistic minority; and planned to return to their shops. In Balfour, after the unrest had subsided, a rally was held at the local stadium where foreign shopkeepers were welcomed back to the township and invited to read their ‘holy book’ (Keepile 2009, and this author's response in Mail & Guardian, 6 November 2009).
Two of the young leaders of the Balfour revolt explained that, early in 2008, local businesspeople had called a community meeting with the aim of turning people against the foreign traders. The two countered arguments around drugs crime and competition with practical suggestions, and, so they claim, dissipated potentially dangerous xenophobia. After the May 2008 xenophobic violence, the Balfour community organised a sports day that brought together South Africans and foreign nationals, many of whom were Zimbabweans. Teachers and others with a little money provided a braai. Such was the strength of opposition to xenophobia that some refugees from other townships fled to Balfour. In Thokoza, we asked to be introduced to some Zimbabwean residents, but most of the people we then met turned out to be South Africans. From the Zimbabweans we did speak to, we learned that some of their compatriots had joined the protests alongside established residents. Our conclusion, then, is that while there was violence against the property of foreign traders during some of the protests, the character of the unrest was entirely different to that of May 2008 (see also Alexander and Pfaffe 2009).
Secondly, we should consider class dynamics within the struggle. The emphasis on favouritism in the award of tenders and evidence that shopkeepers may have encouraged some of the xenophobic violence, suggests that the township petty bourgeoisie may have been influential. While this argument might have some merit, it should not be overstated, partly because the protests have been mass events, with impoverished youth to the fore, and partly because the size of the petty bourgeoisie that resides in townships should not be exaggerated, as people with money generally move out.10
The response of workers and their unions has been mixed. The Congress of South African Trade Unions (COSATU), the main union federation, has been rather quiet in public. Its media office produces about two statements a day on average, but only one relevant item could be found for 2009. This was a short section of an address by COSATU General Secretary Zwelinzima Vavi, presented to the South African Municipal Workers' Union (SAMWU) conference on 3 November. In this, Vavi stops short of supporting the protesters and limits himself to blaming ‘false promises of the previous ANC government’, the privatising and outsourcing of basic services, and deployment of unqualified people to senior positions. SAMWU, the main union with members employed by the municipalities, has been more sympathetic, arguing for instance in a press statement that:
Our members on the ground know and experience the frustration of the poor. It is entirely consistent with the ideals of a democratic society for the poor to protest and collectively raise their demands. That is why SAMWU supports those campaigning for an increase in services, and will continue to urge the government to make good on its promises to deliver them. Scores of perfectly legal protests on the question of service delivery have been met with hostile and unnecessary police violence. (SAMWU 2009)
In Balfour, the 27 July meeting mentioned above was not attended by SAMWU representatives. While it was proceeding, the union's members toyi-toyied outside the municipal offices as part of a separate strike action, which did not appear to be backed by members of the wider community. However, this is merely one example, and some of the SAMWU members, when interviewed, identified with the township protest. A general response among the leaders interviewed was that workers were members of the community and backed the actions (indeed their tone sometimes implied that the answer should be self-evident). Also, as we have seen, the protests have often involved calls on workers to stay away from work, though we have not yet investigated the impact this had. We did, though, come across one remarkable example of solidarity, which occurred in Standerton. Here there had been a long-running strike at the local poultry factory, which employed some 4000 workers, most of whom were members of a Federation of Unions of South Africa affiliate (who were on strike), with a minority of workers belonging to a COSATU affiliate (who were not). Strikers addressed community meetings and community activists addressed those of the strikers. On one occasion community members joined a mass picket at the factory, where barricades were erected and there were clashes with the police.11
More evidence is required before one can develop a satisfactory analysis. However, it is possible that COSATU is influenced by its close relationship to Gwede Mantashe, Chair of the South African Communist Party (SACP) and Secretary General of the ANC, who has been prominent in attempting to address issues related to the unrest.12 In contrast, SAMWU, although a COSATU affiliate, has taken a non-partisan position in elections and its members see local government problems at first hand. On the ground, workers are generally (though not always) residents of townships, and easily sympathise with other members of their communities. However, as workers they tend to organise at work and in unions, which involves a spatial separation from the township. Time is taken up with work itself, and sometimes union activism, rather than township mobilisation.13 Arguably, while there are class interests that can unite workers and non-working or underemployed township residents, there are also divisions, especially in terms of use of time and organisation.
Thirdly, we noted an apparent lack of connectedness between the various protests, though our evidence is slim and may be biased by our sample. There is probably a ‘demonstration effect’, with activists in one area following events in another, mainly through the media, but this is not the same as sharing experiences, let alone coordinating action. If this assessment is correct, it may be associated with the localism of the demands being raised, but also with leadership being provided by a new generation of activists, who are not the kind of people to have a wide geographical spread of contacts. More broadly, with the exception of the Standerton example, there appears to be a disjuncture between the various struggles affecting South Africa in recent years. These have included many militant protests dispersed by the South African police (usually using rubber bullets), which have involved workers, students, street traders, metropolitan police in Johannesburg and soldiers in Pretoria.
Finally, resonances from the apartheid period are striking. The battles have been fought by residents of formal townships and informal shack settlements, which are still spatially and socially separated from the suburbs (even if this is now more a class divide than a racial one). As before, the dynamics of the protests, including spatial location and demands raised, distinguish them from the struggles of workers. Moreover, people are pressing issues around housing and service provision that democracy has failed to address, and their concerns are still ignored by uncaring local officials. This has rekindled the socio-political divide between ‘citizen and subject’ (to use Mahmood Mamdani's formulation), with people responding by attempting to exert political influence through the development of a collective, community voice, as distinct from formal local politics. In addition, similar tactics are deployed, including the use of stay-aways and barricades with flaming tyres. The police response, too, is reminiscent of the apartheid era.
Conclusions
Since 2004, South Africa has experienced a massive movement of militant local political protests. In some cases these have reached insurrectionary proportions with people momentarily taking control of their townships, and it is reasonable to describe the phenomenon as a rebellion of the poor. There are strong similarities linking the Mbeki-era and Zuma-era protests, notably in relation to issues raised, which are principally about inadequate service delivery and lack of accountability by local councillors. A significant difference is that the new government has greater legitimacy. Ironically, perhaps, this may have encouraged a heightened level of protest, with people believing that Zuma is more likely to address their demands.
The protests reflect disappointment with the fruits of democracy. While some people have gained, the majority are still poor. Levels of unemployment are greater than in 1994, and income inequality remains vast. People can vote, but all too often elected representatives are self-seeking and real improvements are few. Many problems can be traced back to post-apartheid government policies that can be described as ‘neoliberal’. Privatisation of local services opened up new opportunities for private accumulation by councillors and their cronies. Inadequate investment in public goods has produced a shortage of people with the skills necessary to administer local government and maintain municipal services (especially outside the metropolitan areas). There has also been underfunding in key areas, particularly housing. Critically, neoliberalism has sustained massive inequality, which, linked in particular to policies associated with black economic empowerment, has added to feelings of injustice. In practice, local councillors are often at the receiving end of problems that began elsewhere. Moreover the structure of governance makes matters worse. The ANC's policy of ‘deployment’ combined with its political hegemony can mean that councillors are more concerned about the support of party bosses than that of their electorate. Members of parliament (MPs) and members of provincial assemblies (MPAs) are elected through party lists, so they do not have constituencies of their own. Unsurprisingly, they appear to have been absent from the scene of the protests, thus failing to reduce the gap between national and local politics.
The key difference between the apartheid years and now is the existence of a democratic government retaining widespread legitimacy. Politics operates at several levels, including the influence that national leaders have over local party structures. As a consequence, the government still has some scope for manoeuvre, as our Balfour example showed. In particular, the ANC can remove corrupt and unpopular councillors, and the 2011 local government elections will provide a focus for such activity. Replacement of councillors can provide an opportunity to co-opt popular leaders of the protests, but it is also possible that self-interested councillors will be replaced by other self-interested councillors, and, as Mantashe observed, removing too many councillors can lead to a loss of necessary experience (Maravanyika and Davis 2009). In addition, whilst the government would like to improve service delivery, it is constrained by the conservatism of its economic policies and by the current recession. While there has been some reallocation of resources, the government's main response has merely been that of pushing for greater monitoring and evaluation of elected officials and administrators.
One scenario is that, through a combination of political manoeuvre (listening, co-opting, dividing and other strategies), minor improvements (for instance, library internet connection in Balfour would cost relatively little) and repression, the government will be able to assuage the unrest. This author's sense is that it is doing too little, too late. It does not have the resources or the will to address the housing crisis, to rapidly improve skills, to significantly expand the number of jobs and markedly reduce inequality, or even to eradicate corruption in local government. A second scenario is that, having fought hard and won very little, people will collapse into political apathy and/or engage in another round of xenophobic violence. The third possibility is that struggles will generalise, developing more interconnections between townships and between township struggles and other arenas of conflict. This is what happened under apartheid, where the local battles of the early 1980s paved the way to national coordination. It is also what occurred in Zambia and Zimbabwe, for example, where food riots on a much smaller scale than South Africa's local protests contributed to the development of new parties (Alexander 2000). There is a dialectical relationship between the scale of a mass movement and its demands, but these are also influenced by political formations. For now, in South Africa, it is more likely that action will feed into the SACP and reform of the system than into the development of a revolutionary movement. But the position can change, particularly given the relative youth of the new generation of activists.