Introduction

Political and social transformation have affected South Africa profoundly. New and non-racial forms of democratic government are being established, and reconstruction and development have begun. But the process has been far from painless: while political violence is generally declining with parts of KwaZulu-Natal the major exception the transition to democracy has been characterised by rising levels of crime.

Maintaining order during this period of reconstruction and democratisation is vital to the health of South African society. The failure of the police to counter higher levels of crime has led to disillusionment with the new government on law and order issues, and a potentially greater willingness among citizens to take the law into their own hands. As disorder has increased, state policing agencies have undergone fundamental changes. Also, new private and local authority policing institutions are beginning to solidify.

This article is an attempt to analyse these processes. It traces the decline in political violence that marked the end of South Africa's transition to democracy, and tries to explain growing levels of crime and their relationship, if any exists, with political and social transformation. It focuses in particular on attempts to control crime since the Government of National Unity (GNU) took office in May 1994. It includes an overview of changes in the South African Police Service (SAPS), and new initiatives in policing. Finally, changing forms of private and self-policing are examined, and the implications for the future of policing in South Africa are considered.

From violence to crime?

The transition to democracy was accompanied by intense conflict: about 16 000 people died between 1990 and 1994 in internecine violence,2 most of them in the urban complexes.3 Since the 1994 election political violence has diminished, but crime has continued to increase, as it did during the first four years of the 1990s.

Political violence began before 1990, but increased especially from June 1990 in KwaZulu-Natal and on the Witwatersrand. Deaths in KwaZulu-Natal rose from an estimated 800 in 1989 to more than 1 500 in 1990, dropping slightly in 1991 and 1992 but increasing to a high of 2 009 in 1993.4 Violence began to increase in the year negotiations began, and peaked in the year they ended.

Better policing and monitoring seems to have curbed some of the violence. However, it changed its pattern, not its extent. By 1993 pitched battles between antagonists had waned, only to be replaced by drive-by shootings and hit-and-run attacks.5 The visible parties to the conflict were replaced by invisible ones. The result was an even more lethal conflict: by 1993 more people were being killed than wounded.6 The resumption of all-party talks in March, after a dispute about state involvement in the conflict, coincided with a sharp rise in violence, after it had decreased for the first part of 1993: monthly deaths for the first two months of the year were little more than half of 1992's monthly average.7

Explanations for the violence varied. Some claimed a conspiracy on the part of the apartheid regime. Violence, they tried to show, peaked when there was a breakthrough in negotiations; either its architects were those who did not want a settlement, or it resulted from a strategy of the National Party (NP) to weaken its bargaining partners.8 There was extensive evidence of state involvement: senior police officers had been implicated in weapons smuggling and providing support to the Inkatha Freedom Party (IFP), while the police had been shown to be biased against the African National Congress (ANC) and its allies.9 But there is no evidence (yet) of a link to the highest levels of the NP, nor could this theory explain why the conspiracy did not try later to disrupt an election that was likely to and did end the NP's reign. As it was, the complexity of the conflict confounded simple conspiracy theories that sketched all perpetrators of conflict as mere puppets of Pretoria. Where there were conspirators, they encouraged and worsened existing conflict; they did not create it out of nothing.

Others claimed the violence was primarily caused by political rivalry. Indeed, by the beginning of 1993 the political dimensions of the conflict were almost beyond dispute, as participants clearly labelled themselves as being either ANC or IFP. Violence in KwaZulu-Natal had assumed war-like proportions, with opposing political groups battling for territorial control on the urban periphery, while around Johannesburg internecine conflict simmered as streets, blocks, houses and hostels were claimed as the territory of one or other political party.10

However, to view the violence purely as a political clash would obscure important underlying causes. There was evidence, for instance, that violence often began for a variety of other reasons many related to local disputes or grievances that acquired political labels over time. Monitoring evidence also showed that the political nature of the conflict was not always clear, with one survey showing that only about 7 percent of violent acts were reported as ANC/IFP clashes in 1991.11 Furthermore, only about one quarter of those involved in conflict in certain areas of KwaZulu-Natal in 1990 could identify political leaders, and explain the ideologies of the parties whose interests they were meant to be defending.12 In turn, some of the violence took place not only between opposing political groups, but also within them.13

Conflict was also most common in areas on the Witwatersrand and in KwaZulu-Natal that were characterised by extreme poverty. It often seemed to take the form of a battle for access to meagre resources on the urban periphery, a fight between the `haves' and the `have-nots', even though they were indicated as belonging to different political parties. This suggested at least that structural and material factors were involved that would not necessarily have changed after the transition to democracy had been completed.14

The election did lead to a decrease in conflict, at least in the short term. The participation of the IFP in the poll meant that all parties participated, with the exception of the white far right. The IFP's late entry ensured that there was little time for the party and its rivals to campaign in each other's territories.15 Fatalities dropped from more than 500 in the month before the poll to less than 100 on average in the post-election months.16 The election did not end violence altogether. By mid-1995 KwaZulu-Natal was simmering once again. In August that year nearly 100 people were killed and more than 200 houses burnt down, with monitors describing the province as being in a `situation of near anarchy'.17

The election also gave rise to new but interconnected fears that the country was sliding into criminal anarchy. South Africa's crime problem is not recent: past commentators already labelled South African society as `crime-generic'.18 The decade 19801990, in which the apartheid state was most strongly challenged, showed significant increases in crime. According to police figures, serious offences rose by 22 percent, and less serious ones by 17 percent; murders increased by 32 percent, rape by 24 percent, and burglary by 31 percent.19

The increase in levels of crime peaked in 1990, the year in which the political transition began. Recorded levels of almost all crime showed absolute increases for the period 1990 to 1994. While the murder rate declined by 7 percent, in line with declining levels of political violence (from 16 042 fatalities in 1990 to 14 920 in 1994), other crimes increased significantly during this period: assault increased by 18 percent, rape by 42 percent, robbery by 40 percent, vehicle theft by 34 percent, and housebreaking by 20 percent.20 There was also an increase in the crimes of the affluent: although no accurate figures were available, white-collar crime increased significantly over this period.21 Trends throughout the country were not uniform, with the greatest increases occurring in urban complexes around Johannesburg, Durban, Port Elizabeth and Cape Town.22

Such a rise in criminality is not exclusive to South Africa. Comparative evidence suggests that crime increases significantly during periods of political transition that are coupled with instability and violence. This was the case in Eastern Europe during the transition to democracy, and in the final days of the Soviet Union and the first days of the Soviet republics. The Russian case contains added complexities: economic and political collapse has created a vacuum into which organised crime has expanded and in which current and former members of the security forces are active.23 In Northern Ireland it is feared that the promise of political compromise, the concomitant loosening of tight and opposing community bonds, the greater number of (armed) ex-fighters on the streets and less vigilant policing may herald an increase in crime.24

African evidence on the subject is scarce: apparently, crime increased markedly in Namibia before and just after its first inclusive election.25 The bonds keeping societies intact are loosened in periods of instability, seemingly making crime more likely.

While political violence has declined in some areas of South Africa, it may have left a legacy with important potential consequences for crime trends. Campaigns to retrieve illegal arms have been largely unsuccessful: slightly more than 1 000 firearms were surrendered in two amnesties during 1994.26 The relatively easy availability of licences to legally own weapons a consequence of permissive gun ownership laws increases the stock of weapons available to criminals, who may acquire them legally or steal them from legal gun owners. Violence has also weakened social control, producing marginalised groups that rely on conflict for a livelihood.27 This results in increased levels of crime, as disaffected individuals (often township youth) become engaged in it. This points towards likely trends in crime rates in the near future.

Perhaps most importantly, the rising crime rates are almost certainly related to political, social and economic trends that existed before the formal transition began, and have been accentuated by it. Evidence suggests that crime rates in townships have been high for many years,28 but that racial segregation largely insulated more affluent (white) areas from its effects. It seems likely, given the gulf between township residents and the police, that most crimes in these areas were unrecorded.

The erosion and subsequent collapse of apartheid boundaries allowed crime to move out of the townships and into the suburbs, where it was more likely to be recorded. Greater affluence in these areas, and in parts of black townships where barriers to black economic empowerment disappeared, increased the rewards available to criminals and the incentive to engage in more organised and sophisticated, forms of crime.

Policing the transformation

The apartheid order generated crime rather than controlling it. The police were agents of a state that created crimes through its efforts to erect moral, economic and political boundaries between the races. The official historian of the South African Police (SAP) has acknowledged that during the apartheid era, only one in 10 members of the force were engaged in detecting and investigating crime.29 Combating crime was secondary to policing apartheid and maintaining internal security.

As the police force was so heavily politicised during the apartheid era, the recent political transition also demanded a transition in the force. But, unlike the South African National Defence Force, the SAPS cannot turn inward for a period of reform, and then return to the public arena to perform its task. Transition impedes the capacity of the police to combat crime, while crime in turn impedes their capacity to transform the force. Current analyses of changes in the police force warn against a situation where "restructuring of the police is treated in virtual isolation from the societal conditions within which the institution is embedded".30

Public concern at the growth in crime rates coincides with a period in which the SAPS has been severely stressed as a result of the transition. Inevitably, the police transition has not been smooth. Visible conflict within the force, revelations of past police brutality, continuing corruption and ebbing police morale evident in the 142 police suicides in 199431 are not conducive to the efficient maintenance of law and order. It is hardly surprising, then, that the latest figures show that the police solved on average only half of all recorded violent crimes in 1994.33 Crimes committed by police officers have also undermined public confidence in the force: one in four officers in the Johannesburg area were investigated for suspected criminal activity in 1994.34 Much will still have to be done before a national policy on crime prevention and control becomes a reality.

The transformation of the police force has been under way for some time. The old SAP sought to control change by restructuring the force before the new political authorities took office. Changes since then have concentrated on civilianising the SAPS, and making it more service-oriented. Key among these has been a focus on `community policing', aimed at encouraging greater co-operation between police and public in the fight against crime.

The process has not been uncontroversial, and many commentators argue that it has only been partly successful. Many believe community police forums are controlled by the police, and the public and the police themselves frequently see community policing as a `softly-softly' approach to the problem of crime. However, there is not necessarily a contradiction between greater police accountability to the citizenry and effective crime control. In the light of the brutal history of policing in South Africa, a police force more responsive to community needs is not only an appropriate strategy to fight crime, but also essential for ensuring a police agency with which the public can co-operate.

There are obstacles to this transformation. Continued violence on the urban periphery in KwaZulu-Natal suggests that community policing strategies will be hard to implement when paramilitary methods are needed in the short term to police the conflict.35 In general, if police respond to demands to tailor their operations to democratic norms, immediate gains in crime control are not guaranteed, and any benefits are only likely to become evident in the longer term. This imperils attempts at police reform, since it will only yield limited if any reductions in crime for the time being, with a potential growth in pressures for a return to the authoritarian style of the past. Or it may divert resources badly needed elsewhere to policing matters, while current comparative evidence suggests that increased expenditure on policing does not automatically translate into reductions in crime rates.36

In comparison to other societies, there is a long-standing assumption that South Africa is inadequately policed. Contrary to popular perceptions, the SAPS is proportionately not much smaller than police agencies in other countries. The combined police strength is 3,1 active force members for every 1 000 people, just slightly less than the European average of 3,5. However, policing resources have been concentrated almost exclusively in white areas. Until recently, estimates suggest, 80 percent of policing resources were concentrated in suburbs and city centres, while formerly black, `coloured' and Indian areas were policed by the 20 percent, with African townships receiving an 8 percent share.37 South Africa is not only inadequately policed, but police resources are also spread unevenly.

Some senior police officers have conceded that the SAPS will have to reconsider its distribution, providing a more comprehensive service in townships and leaving the suburbs to private security companies. `Normal' models of policing have largely been confined to white areas; the suburbs have been well stocked with police services, while the townships have been policed more for purposes of control than for crime prevention. With the focus beginning to shift, it is clear that the SAPS is not only ill-prepared to police crime in the townships as it has to move away from the past instruments of political control, such as informer networks but it also does not have the required resources.

Adequate reform of the SAPS will entail a fundamental shift of resources towards the townships, implying that the suburbs will receive less policing. It is also possible that some SAPS resources will have to be concentrated on areas of focus policing, such as drug trafficking, allowing `visible policing' to be undertaken by other agencies. The implication is a further growth in local authority and private policing. As an SAPS document candidly admits: "It is general knowledge that the different socio-economic classes of a national population will have different security needs. It is also assumed that the safest measures are available to those who can afford it, because it is expensive. This can lead to public policing for the needy, and private policing for the affluent. If this is the case, what is the future of a national police service?"38

The future of safety and security?

Sections of the citizenry are undoubtedly unwilling to wait for the uncertain benefits of police transition, and have made their own arrangements to protect themselves and their possessions. The result is a substantial growth in private security services, demands for the establishment of local authority policing agencies, as well as possible new forms of self-policing all of which could have an important impact on how South Africa is policed in future.

Surprisingly, the political transition seems to have weakened citizens' anti-crime initiatives in some areas, such as neighbourhood organisations, and self-defence and self-protection units.39 As these responses are not necessarily subject to democratic rules, this seems to be a gain for democratic order, further reinforced by more recent initiatives in which civil society groups have united to fight crime without resorting to armed force.40 However, community policing initiatives, together with declining confidence in the ability of public institutions to maintain order, could prompt a resurgence in self-policing. While this will require careful management, these initiatives could strengthen police activities, and in a context that enhances democracy. This will be subject to the provision that where they emerge, they will be regulated and encouraged to co-operate with the police rather than ignored or simply suppressed.

Besides this, a debate has begun on the desirability or otherwise of local authority policing. The interim constitution provides for local police, although they are confined to `crime prevention'.41 Given the proposed relationship between local police and local government, accountability may be easier to attain. Comparative and domestic examples suggest that local policing could help to counter crime in the city centres, while performing a useful public service role. But opportunities for corruption and abuse of power still exist. Municipal policing could strengthen democratic order, provided diverse local forces will obey common democratic norms. However, the danger remains that if national government is too prescriptive of the standards local authority policing has to achieve, local forms of policing could be privatised, and lines of democratic accountability may become blurred.

Indeed, private security has clearly benefited from an reduced reliance on public policing. The industry has grown hugely; it initially expanded by about 30 percent a year, and now outnumbers the public police.42 It is seeking to assert itself as a source of effective protection against crime. But it is not a homogeneous sector; different components provide various services, and due to internal competition it has been difficult to consolidate and regulate. The relationship between private and public policing may be more complex than a mere withdrawal of the state from some areas of control, to be replaced by private means. The instruments of public policing have not contracted during the period when private policing institutions have grown.43 Rather, private security enterprises have expanded as a result of increases in private property ownership44 and heightened perceptions that society is unsafe.45 In turn, societies such as South Africa emerging from authoritarian rule often have the personnel to staff growing private security sectors as a result of the numbers of demobilised combatants, often with few other marketable skills, available for employment.

The South African private security industry itself was partially spawned by apartheid policies. The security apparatus encouraged its growth, so that state resources could be concentrated on policing political dissent. The industry has a continuing close relationship with the state: private security puts mechanisms in place guards, alarms and detection devices to gather information that can be fed to the police: rather than decreasing demands on the police, private security may overburden it in some areas.46 Besides this, the market for the industry is expanding into areas that have previously been the exclusive domain of the public police. In many areas, `panic buttons' are no longer connected to police stations already inundated with calls it is estimated that there are 400 000 panic alarms in Johannesburg alone but are becoming the almost exclusive realm of private reaction units.47

In some instances, private policing has directly replaced the public police. A small number of municipalities charge private security to local authority rates and taxes, and award contracts to security firms to `police' individual suburbs. City centres have not been excluded from this trend: for example, a private consortium polices a number of blocks in the city centre of Johannesburg. The growth in private security enterprises is also not exclusively confined to the wealthy parts of the urban environment. Former ANC combatants have established security firms in township areas to protect delivery vehicles, and are selling their services to blocks of township residents.48

This expanding industry may harbour dangers to South Africa's developing democracy. These are inherent in the fact that security companies, mainly staffed by officers from the former SAP or similar apartheid agencies, will potentially protect the rights of their clients at the expense of everyone else, and further entrench the divide between the (privately policed) minority and the majority. The private security industry in itself does not constitute a threat to democratic order, as it is too fragmented and competitive and needs to ensure a healthy relationship with the state. Indeed, the growth of the industry may contain some advantages. Private security could absorb redundant personnel left over the rationalisation of the military and the disbandment of the liberation armies, and it could reduce attempts at self-policing that have a greater potential to threaten civil liberties.

The blessing, however, may be distinctly mixed. There is evidence that residents of the suburbs may react to crime by seeking to insulate themselves physically from the poor, who are seen as its perpetrators. This is reflected not only in the use of private security, but in the changing architecture of cities, that are increasingly characterised by walled suburban complexes. This could entrench a form of social distance that would impede attempts to create a common South African loyalty. Private security could threaten citizens, as is evident in reports of private security officials being responsible for cases of abuse,49 and could entrench a divide between those who can hire personal protection and those who cannot. In the suburbs, the likelihood that the former will be mainly white and affluent and the latter mainly poor and black may enhance racial and social barriers and increase the possibility that security will become, or at least be seen to become, a weapon used by the former against the latter. This would erode a fundamental norm of democratic societies that policing is uniformly available to all, and its powers exercised through universally applicable laws.

Notes and references

Edited version of a paper prepared for a UNESCO conference on urban transformation in Africa, Nairobi, Kenya, 1995.



M Shaw, `The bloody backdrop: negotiating violence', in S Friedman and D Atkinson (eds), South African Review 7 The small miracle: South Africa's negotiated settlement, Ravan, Johannesburg, 1994.



S Bekker (ed), `Capturing the event: conflict trends in the Natal region, 1986-1992', Indicator SA Issue Focus, 1992.



A Minnaar, `The impact of political violence since 1990 on the transition to democracy in South Africa, unpublished paper', June 1994.



J Rauch, Drive-by shootings, Centre for the Study of Violence and Reconciliation, University of the Witwatersrand, 1993.



Ibid.



Human Rights Commission Monthly Repression Report, February 1993. See also Report of the Independent Board of Inquiry, February 1993.



D Everatt and S Sadek, The Reef violence: tribal war or total strategy, CASE/HSRC, Johannesburg, 1992.



Africa Watch, The killing in South Africa: the role of the security forces and the response of the state, New York, 1991.



For an overview of the conflict in KwaZulu-Natal, see J Aitchison, `The civil war in Natal', in G Moss and I Obery (eds), South African Review 5, Ravan, Johannesburg, 1989. See also R Taylor and M Shaw, `Interpreting the conflict in Natal', Africa Perspective 2(1), 1993. For violence on the Witwatersrand, see R Taylor, `The myth of ethnic division: township conflict on the Reef', Race & Class 33(2), 1991.



A Louw and S Bekker, `Conflict in the Natal region a database approach', in Bekker, Capturing the event, p 44.



S Stavrou and A Crouch, `Molweni: violence on the periphery', Indicator SA 6(3).



Opposing groups of youth aligned to the ANC, for example, have clashed in Bambayi near Durban and Sebokeng in the Vaal Triangle.



M Morris and D Hindson, `South Africa: political violence, reform and reconstruction', Review of African Political Economy 53, 1992.



When the ANC did canvass in IFP territory the weekend before the poll, two of its canvassers were killed.



Human Rights Commission, Monthly Repression Report, October, 1994.



Weekly Mail & Guardian, 2228 September 1995.



Brewer, `Crime and control', in J Brewer (ed), Restructuring South Africa, MacMillan, London, 1994.



M Shaw, Partners in crime? Crime, political transition and changing forms of policing control, Research Report no 39, Centre for Policy Studies, Johannesburg, June 1995.



Ibid, pp 1718.



Ibid, p 12. The exact extent of white-collar crime is difficult to determine, given that most companies deal with the problem internally.



Ibid, pp 2327.



M Walker, The Soviet Union, Collins, London, 1989.



The Economist, 15 October 1994.



C Tapscott, `Crime in independent Namibia: social equity and the crisis of expectations', paper presented to a conference on managing crime in the new South Africa, Pretoria, 46 August 1992.



A comparative perspective on the incidence of crime in South Africa, SAPS Centre for the Analysis and Interpretation of Crime Data, Pretoria, February 1995, p 44.



M Shaw, Crying peace where there is none? The functioning and future of Local Peace Committees of the National Peace Accord, Research Report no 31, Centre for Policy Studies, Johannesburg, August 1993.



Brewer, `Crime and control'.



M Dippenaar, The History of the SAP: 1913-1987, Promedia, Silverton, 1988. See also J Brewer, Black and blue: policing in South Africa, Clarendon Press, Oxford, 1994.



E van der Spuy, `Transforming the police, policing the transformation', South African Sociological Review 6(2), 1993, p 68.



Weekly Mail & Guardian, 1117 November 1994.



Rapport, 26 February 1995.



The Citizen, 4 April 1995.



Sunday Times, 12 March 1995.



Community policing has been successful in few places in KwaZulu-Natal and then only in areas where substantial police resources have been devoted to the project.



P Engstad and A Evans, `Responsibility, competence and police effectiveness in crime control', in R Clark and J Hough (eds), The effectiveness of policing, Gower, Farnborough, 1980.



Financial Mail, 10 February 1995.



South African Police Service, Report by Regional Commission `H2', Witwatersrand, p 18.



The Leader, 27 May 1995 and Sunday Times, 30 April 1995. See also Shaw, Partners in crime, 1995, pp 4143.



The Western Cape has seen a number of such initiatives, of which the Western Cape Anti-Crime Forum is the most prominent example.



See M Tarbitt, `The functioning of metropolitan and municipal police services in a new constitutional dispensation', South African Police Service, Organisation and Work Study Efficiency Services, undated.



Ratios vary from 3:1 to 5:1 depending on whether in-house security personnel are counted. See G Cawthra, Policing South Africa: the SAP and the transition from apartheid, David Phillip, Cape Town, 1993, p 71.



M Shaw, `Policing for profit?', Crime and Conflict 1(1), Autumn 1995.



C Shearing and P Stenning, `Private security: implications for social control', Social Problems 30(5), June 1983.



See D McDowall and C Loftin, `Collective Security and demand for handguns', American Journal of Sociology 88, 1983.



See N South, Policing for profit: the private security sector, Sage, London, 1988.



Security Focus 12(7), 1994, pp 1819.



P O'Leary, `Hijacking: MK enters the fray', Fleetwatch, November 1993.



The most notorious case is that of Sybrand `Louis' van Schoor, a security guard and ex-policeman who shot dead 41 alleged burglars over a few years. After each incident, magistrates found that he had acted within the law; he was not once cautioned by the police or the courts.