Why history has repeated itself The security risks of structural xenophobia

• September 2009 25 At a closed presentation earlier this year (2009), a provincial Disaster Management executive responded to criticism of South Africa's nonadherence to certain humanitarian assistance standards by stating that Disaster Management had followed a high-level decision that assistance to foreign nationals displaced by the May 2008 xenophobic attacks should not exceed the living conditions of South Africans living in poverty. At first glance, this seems a reasonable position to take. Its common-sense quality flows from the following logic: 1) Citizens are the true political objects of the state and therefore most deserving of its assistance. 2) It would not be fair for noncitizens to benefit from forms of state assistance that are not available to citizens. 3) It would be natural for citizens to react with anger if noncitizens were seen to be favoured (indeed, it was argued that perceptions of South Africans' relative deprivation when compared to non-nationals had caused the attacks in the first place). The results: 1) citizen anger could worsen an already explosive situation of anti-foreigner violence, and 2) citizen anger could turn the electorate against elected leaders at various levels.

At a closed presentation earlier this year (2009), a provincial Disaster Management executive 2 responded to criticism of South Africa's nonadherence to certain humanitarian assistance standards by stating that Disaster Management had followed a high-level decision that assistance to foreign nationals displaced by the May 2008 xenophobic attacks should not exceed the living conditions of South Africans living in poverty.At first glance, this seems a reasonable position to take.Its common-sense quality flows from the following logic: 1) Citizens are the true political objects of the state and therefore most deserving of its assistance.2) It would not be fair for noncitizens to benefit from forms of state assistance that are not available to citizens.3) It would be natural for citizens to react with anger if noncitizens were seen to be favoured (indeed, it was argued that perceptions of South Africans' relative deprivation when compared to non-nationals had caused the attacks in the first place).The results: 1) citizen anger could worsen an already explosive situation of anti-foreigner violence, and 2) citizen anger could turn the electorate against elected leaders at various levels.
The rationale is seductive.But consider, for a moment, that the very same logic was used to justify the displacement of foreign nationals in the first place.Non-nationals were perceived, rightly or wrongly, to occupy a more favourable social and economic position to poor citizens, or, through their activities, to frustrate South Africans' opportunities to rise in the social and economic hierarchy. 3In fieldwork conducted by the Forced Migration Study Project (FMSP) during 2008, 4 many of the South African participants who shunned the violent means used by those perpetrating the xenophobic violence supported the end result of displacement, on the assumption that their life chances would be improved in the absence of 'competition' from

Why history has repeated itself
The security risks of structural xenophobia non-nationals.When the response to popular violence against a stigmatised group takes its rationale from the perpetrators, the outlook for justice and security is poor.And indeed, although 62 people died in the 2008 crisis, over a year later only one murder case has resulted in conviction. 5

THE STRUCTURAL PREJUDICE AGAINST NON-CITIZENS
The South African Constitution and the Promotion of Equality and Prevention of Unfair Discrimination Act, 2000, specifically prohibit forms of prejudice such as racism or sexism.But a prejudice for citizens and against non-citizens is intrinsic to the very structure of the contemporary, territorial state.In the state as a political form, a relationship is naturalised between a relatively centralised political authority, a territory, and the population of that territory, defined through citizenship laws.Intrinsic to this system are those who are not the natural residents and beneficiaries of the state.Permanent residents, and those with a variety of permits or statuses, including refugee status, reside in the territory in a relationship that is by definition 'unnatural' -foreigners become citizens only through the process that is literally called naturalisation.And even then they may once again be at risk of being rendered unnatural if they do not conform to standards set by the state.
In other words, prejudice against foreigners is naturalised through the state as a political form.The moral panic that presents xenophobia as a psychological pathology often obscures the structural nature of this prejudice.Furthermore, the inherent favour the state shows its 'natural' citizens necessarily forms a fault line that threatens the realisation of human and constitutional rights that should apply without reservation.
In this article we examine the South African state's responses to the internal displacement of non-citizens by anti-foreigner attacks.We show how the naturalised prejudice of the state and its bureaucracy towards citizens -its structural xenophobia, as it were -has informed the response to violent xenophobia in the country.Finally, we illustrate the obstacle this prejudice poses to effective prevention of future attacks, and the threats to security and justice that it perpetuates for all residents of South Africa.

STATE RESPONSES TO XENOPHOBIC VIOLENCE AND DISPLACEMENT
The South African state's responses to the xenophobic violence of May 2008 were characterised by a number of failures thatdeliberately or otherwise -supported the intentions of perpetrators, further criminalised the victims, and promoted future security risks for non-nationals (and by extension, as we saw in the May 2008 attacks, the South African communities in which they live).We discuss these failures below.They include failure to prevent violence, failure to protect victims during the attacks (both from victimisation and from deportation), failure to prioritise prosecution of the offenders throughout the judicial process, and a reluctant approach to humanitarian assistance and reintegration.

Failure to prevent violence and protect victims
May 2008 was not the first time xenophobic violence occurred in many of the affected areas.For instance, prior attacks on non-nationals had occurred in Alexandra (1994-95); 6 Diepsloot (20046 Diepsloot ( , 2006, March and April 2008); 7 the East Rand (1996, 1999); 8 the West Rand (2001, 2007); Johannesburg CBD (1997); 9 and areas of the Western Cape including Hout Bay (1994), 10 Milnerton (2001), 11 and Masiphumelele and Knysna (2006). 12That many attacks recurred in places that had experienced similar breakdowns in order in the recent past is strongly suggestive of a failure to manage risk in these areas.This prompts the question, if citizens had been the victims of the attacks, would risk management have been equally poor?As South Africa has since 1994 seldom experienced ethnic-based violence of the scale and nature perpetrated against foreigners, it is difficult to give a definitive answer to this question.But the seemingly natural character of anxiety about foreigners, and the tendency to see risk primarily in terms of the presence of non-nationals rather than the prospect of attacks against them, may encourage a greater tolerance of anti-foreigner initiatives at a local level.For instance, at a meeting in Alexandra prior to the May 11 attacks, a policeman told community members that 'people must decide on how they deal with someone who has entered his kraal and took his cattle.' 13 The failure to manage risk in the face of threats to foreigners' security is seen once again in areas where, prior to the May 2008 attacks, police and/or local authorities were aware of the meetings in which attacks were organised, or were even present when communities threatened to take the law into their own hands, but launched no preventative response.This was the case in Alexandra, for instance, 14 where police appeared to be well aware of the meetings in which attacks were organised.Said one senior police official: Prior to the attacks, there was a meeting on the 10th of May, 2008 and it was decided that they will attack around the hostel and the shack area.This was not the first meeting; it was a follow up meeting. 15sewhere, respondents observed the role played by participatory governance structures -which exist to enhance democracy and security -in encouraging attacks.This role was sometimes one of omission rather than commission, but as such demonstrates the risks of failing to muster the will to prevent anti-foreigner violence.As a respondent in Madelakufa II noted: 16 This would not have happened were the street committees doing their job.They were in meetings with the street committees, and they did this thing, and the street committees just stood back […] remember that in Madelakufa I, this thing did not happen because the street committees and members of the community stood together and said no one was going to come in and kill another human being.But here since the street communities did not stand up to the violence, people came and did what they liked with us.
Community members interviewed in affected areas generally reported that the local police, as representatives of the state, were reluctant to intervene on behalf of victims during the attacks.On the one hand, local police were underequipped to respond to large-scale violence and could not be everywhere at once, but on the other, some were seen to be intimidated by the prospect of opposing what appeared to be the general will of the people (for good reason -in Itireleng, a police officer who fired rubber bullets was later assaulted by community members).This created public perceptions that police supported the antiimmigrant agenda carried out through the attacks, as evident in the following exchange among participants in the male focus group conducted in Sector 2, Alexandra: RESPONDENT A1 They were there.Nothing could happen if police did not want it to happen.Police were watching.If police were tight like the police that came after, nothing would have happen.RESPONDENT A2 The campaign to remove foreigners went on till the early hours of the morning.The police were always present.RESPONDENT A2 It is because police were already fed up with foreigners and so, they somehow supported us.
Although some respondents attributed protection failures to the fact that police were outnumbered and overwhelmed, others had witnessed particular officers joining the mobs, or actively assisting in the looting of goods. 17This may have fuelled the impression that state actors were complicit in the anti-foreigner campaign.Moreover, many evacuations were done at the expense of protecting non-nationals' property and livelihoods, leading to rampant house robbery, arson, malicious damage to property and appropriation of homes. 18This suggested that local police and elected authorities lacked the will to invest resources in the return of non-nationals at a later date.In Sector 5, Alexandra, where the ward and block committees actively prevented the looting of non-nationals' shops and homes, individuals sheltering at the police station were able to return to their shacks and trading posts in the immediate aftermath of the attacks.In areas where property was not protected, foreign residents were displaced permanently or for a longer period, 19 and their long absence facilitated opportunistic crime even after the initial wave of violence. 20 the time then-President Thabo Mbeki approved army deployment to end the disorder on 21 May 2008, 24 people had been killed and 24 000 were reported displaced in the preceding ten days of seemingly uncontrolled violence, which was made highly visible by constant media coverage. 21By the time a provincial disaster was declared in Gauteng on May 30, at least 342 shops belonging to foreigners across the country had been looted, and 213 burnt down, decimating the livelihoods of hundreds of households.A total of 1 384 suspects had been arrested in connection with the attacks (the resulting cases will be discussed later in this article). 22It is important to note here the research suggesting that the state's late and indecisive response to prevent and stop the violence further heightened security risks to non-nationals and their communities of residence, by encouraging outbreaks in communities that were not initially affected. 23Arguably, it also encouraged the unmonitored flight of victims, who became invisible and thus untraceable once cases went to court.

Deportation of victims
That victims of attacks are seen first and foremost as outsiders to the national community, rather than as members of the local communities they were displaced from, is evident in police recourse to 'voluntary' and involuntary deportation as a response to victimisation.In March 2008 the Department of Home Affairs (DHA) sent immigration officers to arrest undocumented victims who had taken refugee at Laudium police station after attacks in Itireleng. 24milar arrests and deportations were made at least once during the May 2008 attacks: on 15 May 2008, 32 non-nationals who were attacked in Olifantsfontein on the East Rand were charged with illegal immigration and scheduled for deportation.On the same day, Minister of Home Affairs Nosiviwe Mapisa-Nqakula said foreigners made vulnerable by the attacks would not be deported, regardless of their immigration status. 25his suggests an incoherent attitude within the Home Affairs bureaucracy, or at least a failure to effectively communicate with police as the implementing arm for border control activities.Also, halting deportations appeared to be a philanthropic move, and as such its security implications went unrecognised: deporting victims and witnesses is not only an affront to the vulnerability of a population in humanitarian need, but also a stumbling block to the judicial response and hence to future security.
If the security implications of deportation had been recognised, perhaps we would have seen fewer instances of government representatives organising 'voluntary' deportations and buses to return the willing to their countries of origin.This was reported in the press as early as the second day of attacks in Gauteng. 26In Ramaphosa the local councillor was involved in organising buses for 'voluntary' deportation, 27 and in KwaZulu-Natal at least 300 people had been 'voluntarily' deported from Cato Manor and Greenwood police stations in Durban by 25 May 2008. 28is is not to suggest sinister motives on the part of officials who organised 'voluntary' deportations, often at the request of victims, but simply to question what may have appeared to be a 'logical' solution to the dilemma facing non-nationals.Aside from the impossibility of a truly 'voluntary' return in a context of escalating violence; little prospect of return to communities of residence; and a lack of protection and welfare guarantees; it is important to recognise the disservice to the cause of justice that was done through deportations, whether 'voluntary' or otherwise.A report by the Department of Justice and Constitutional Development specifically links the NPA's withdrawal of the majority of xenophobiarelated cases to the fact that 'witnesses became missing or left the country.' 29 The irony is that the 'assistance' rendered to foreign nationals willing to return home in the heat of the violence may in fact have helped to eliminate the prospect of justice for the 40 per cent of cases withdrawn.In this way, by its evacuation and deportation efforts, the state supported the intention of perpetrators to remove foreigners from communities, contributed to stalling judicial measures against perpetrators of violence, and, where deportation was involuntary, criminalised undocumented victims rather than their assailants.

Failure to prioritise prosecution
Since 1994 foreign nationals have repeatedly been attacked in South Africa, but few perpetrators have been charged, let alone convicted.This has two effects.First, where attacks are due to mobilisation by particular figures, 30 these individuals and their agendas remain a risk for the stability of communities in which they operate.Second, a message of impunity is communicated, which at best removes a disincentive to perpetrators and would-be perpetrators, and at worst legitimises attacks by suggesting that the state supports an anti-immigrant agenda.
Thus, flawed state responses to initial attacks may promote the risk of future violence.In some instances, state agents have actively protected those accused of anti-foreigner violence.In Masiphumelele outside Cape Town, for instance, the former Western Cape Provincial Premier, Ebrahim Rasool, the MEC for Community Safety, Leonard Ramatlakane, and the Ocean View police reportedly intervened to secure the release of business owners who had been arrested after xenophobic violence in 2006. 31Little wonder, then, that violence recurred in 2008, motivated once again by business interests. 32 some cases, suspects in the May 2008 violence were released without charges, due to community protests and mobilisation.In the case of Itireleng for example, charges against 11 suspects, including members of the informal leadership structure accused of instigating the May attacks in that area, were withdrawn after a protest march to the court on the day of the hearing.The local councillor had advised members of the community that if they 'tell the police that they did this as a group not as individuals,' police would comply with a demand for the suspects' release. 33e NPA committed itself to prioritise and fasttrack cases resulting from the May 2008 xenophobic attacks, and its progress indicates that it followed through.When compared with somewhat dated South African Law Reform Commission (SALRC) research conducted in 1999, the case finalisation rate for the prioritised xenophobia cases after May 2008 is more than double that for randomly sampled violent crimes (28 per cent versus 11 per cent).Only ten per cent of cases that are just over a year old are yet to be tried, as opposed to 75 per cent of two-year-old cases found in the SALRC sample.However, the case withdrawal rate is four times higher among the xenophobia cases than among the SALRC sample (40 per cent versus ten per cent), and for this reason, the higher levels of finalisation do not necessarily indicate a successful judicial response -as evidenced by the single murder case that has resulted in conviction thus far. 34is demonstrates the integrated nature of judicial processes, which cannot succeed in the wake of a disaster unless specific measures are taken in advance to track the whereabouts of mobile victims and witnesses.It is our argument that such measures were not taken while victims of attacks were sheltering at police stations, or during the steady attrition from camp-based shelters, because of the tendency to see non-nationals primarily in terms of their non-citizenship and thus as outside the mandate of government agencies (other than the DHA).Investigating after the fact, under perpetual resource constraints and in the absence of mechanisms for tracking the displaced population, is bound to be futile.As noted further on, the failure, especially in Gauteng, to launch a concerted reintegration campaign probably also impacted upon levels of case withdrawal and successful conviction.
Another reason that has been given for the 208 case withdrawals reported by the NPA is failure to obtain an interpreter. 35This seems a very poor reason for withdrawal of a case, and evidence that these cases were insufficiently prioritised in provinces such as Gauteng, where the special courts envisioned by the NPA 36 never materialised.failed to recognise, or was indifferent to, the effects on judicial outcomes of forced mobility caused by crises such as this.Consequently, there has been a high degree of impunity, despite the efforts of the NPA.As other analysts have observed, 40 this can only encourage the ill-intentioned to attack foreigners and outsiders.

Reluctant approach to humanitarian assistance
Humanitarianism is based on principles of human rights and non-discrimination, as is the South African Constitution.It follows that foreign citizenship should not be used as a legitimate basis for discriminatory treatment of any kind.But in fact, issues of immigration status impacted on government responses to the May 2008 attacks to quite a substantial degree.
First, despite the obvious fact that the DHA has limited capacity to coordinate and carry out its existing functions of immigration managementlet alone the capacity to mount any kind of social assistance response -other government organs were keen to see the DHA as the lead department for the humanitarian response. 41Almost a year later, Disaster Management staff continued to express this sentiment to FMSP researchers in one province. 42On the most basic level, this represents the tendency to see victims of xenophobic attacks primarily as outsiders to the state rather than as residents of the country in need of protection.
Responding to the xenophobic attacks became a hot potato between local and provincial governments in more than one province. 43The worst affected was Durban, where by June 2008 no provincial disaster had yet been declared and the City refused to assist displaced people, saying it was not their responsibility.Eventually, after being assaulted by security guards, homeless displaced people living on the steps of Durban City Hall were moved to Albert Park, where they slept without shelter until a humanitarian agency set up a tent for them. 44The City's failure to see foreign nationals as their responsibility meant that there was no attempt to risk-manage their reintegration, or address their perceived impacts on local communities.The It is essential to prioritise convictions for crimes resulting from forms of mobilisation that result in large-scale attacks, arson and destruction, and a general breakdown in the rule of law.These pose risks beyond the targeted individuals, especially in informal settlements where the dangers of fire are aggravated and the dense, poorly lit environment hinders policing.The May 2008 attacks created an opportunity for a number of criminals to carry out looting or assassinations under cover of the general chaos.As conviction rates in South Africa are generally low, the higher rates obtained in the prioritised xenophobia cases -especially for rape 39 -are worthy of celebration, and it is surprising that more has not been made of these convictions by the media.But the disproportionate withdrawal rate suggests that priority status was given to these cases too late to ensure that justice was served, and that the system security risks of this failure became all the more clear in January 2009: two foreigners died when a mob attempted to evict them from a building just metres from the police station in Albert Park. 45report commissioned by Oxfam to evaluate the humanitarian response to the attacks records anxieties within government about public perceptions of aid to non-citizens. 46Concerns did not only revolve around the possible risks of these perceptions, but also around political legitimacy: municipal and provincial officials feared voter reactions if they were seen to be sympathising with or allocating resources to foreign nationals.
Arguably, this was a contributor to the generally inadequate response to the protection needs of displaced non-nationals, specifically in terms of their reintegration or resettlement into South African communities, as discussed below.But inadequate and inconsistent provision within temporary shelters established by government caused attrition in the camp population, including 'voluntary repatriation', and does not seem to have been monitored. 47Unmanaged early departure from the camps -and later departures as government deliberately reduced service provision to create a push factor out of the Gauteng camps 48 -thus posed a security threat to non-nationals returning to still-unstable communities.It also impeded the justice process when witnesses could not be found, or complainants became reluctant to pursue cases. 49Nor were the movements of the displaced population monitored, even in the less chaotic encamped stages of the displacement.
Inadequate monitoring and evaluation of the reintegration process, 50 and little, if any, attempt to prepare host communities for the return of displaced foreigners, meant that victims returned to situations of unresolved tension, where popular leaders who played a role in the attacks continued to dominate some communities. 51In this intimidating context it is unsurprising that some complainants who had 'reintegrated' into communities from which they were displaced, dropped their cases.A number of displaced foreigners were killed or injured on attempting to reintegrate, 52 while others were forced to pay protection fees. 53Repeated attacks that destabilise areas and put whole communities at risk have been the long term consequence of continuing tensions and impunity for rogue leaders -for instance in Masiphumele and Du Noon in late 2008 and early 2009.
The state's obvious lack of will to assist or meaningfully engage with the displaced population in the camps led to tensions between nationalities, and between displaced persons and government officials.These tensions, and the increasing politicisation of the camp population, plagued the social world of the camps and led to a number of security concerns, particularly at the Soetwater, Glenanda (Rifle Range Road) and Acasia settlements.Hunger strikes, suicide threats, fighting and hostage-taking became commonplace.
The priority given to ending humanitarian assistance to non-nationals, particularly in Gauteng, provides further evidence of a disinclination to assist those who are not seen to be the legitimate object of government welfare attention.The Province flouted a court ruling by closing the camps at the end of September 2008. 54

Lack of commitment to reintegration
There was much talk of reintegration after the May 2008 attacks, but the shape this took was a far cry from the way it is defined by the UNHCR.Broadly defined, reintegration refers to return and acceptance of a person as a participant member of a community.As such, the term is used to describe the re-entry of formerly displaced people into the social, economic, cultural and political fabric of their original community.According to the UNHCR, reintegration, as it refers to returning refugees, requires access to reasonable resources, opportunities and basic services to establish a selfsustained livelihood in conditions of equal rights with other residents and citizens. 55Reintegration should therefore be distinguished from return (the process of going back to one's place of original residence) or resettlement, which refers to the process of starting a new life in another part of the country. 56uccessful reintegration after conflict-induced displacement requires a number of sine qua non voluntary and participatory processes.These need to be informed by an adequate understanding and resolution of root causes of the conflict/displacement; evidence of the rule of law to quell fears about Internally Displaced Persons' (IDP) security and future; 57  In addition, post-response evaluations have noted the greater priority given to closing camp-based shelters rather than identifying safe communities for reintegration or resettlement, as well as the failure to adequately track the reintegrating population. 59It is not surprising, then, that violence has recurred in several communities of 'reintegration,' and where it has not recurred as yet, there is no certainty about how long the apparent peace will last.

CONCLUSION
This article demonstrates the importance of a strong will at every level of government to assist and protect victims of xenophobic attacks, thereby securing long term security and justice for all.We argue that though it may appear 'natural' to view victims of xenophobic attacks primarily as 'foreigners' rather than as fully fledged members of the communities in which they are attacked, viewing them in this way leads to a series of flawed responses at a variety of levels of government that, seen together, confound the judicial process.The resultant impunity allows individuals who are threats to community safety to remain at large, sometimes in positions of informal leadership, and thus perpetuates the risk of future instability.A few basic recommendations can be drawn for future responses to xenophobic attacks.
1) A risk management framework and early response mechanisms must be developed to ensure that threats to evict foreigners are registered and that related anxieties are addressed.Tensions must be lowered by finding peaceful solutions rather than heightened by denying the validity of locals' concerns, however misled.2) Awareness must be raised among police of their key role in ensuring successful prosecutions from the onset of attacks.3) Conflict resolution negotiations should steer clear of satisfying demands for impunity.
Where participatory governance structures are responsible for attacks, whether by omission or commission, they should be investigated and guilty parties prosecuted if appropriate.4) Local police, who are embedded in the politics of their areas and may be intimidated by the informal forces at play, are not the ideal force to protect victims of attacks.When attacks occur, police from other localities should immediately be deployed for protection activities.5) To the greatest extent possible, victims' homes should be protected from criminal onslaught during their displacement.This preserves victims' livelihoods and communicates the message to attackers that their agenda is unsupported and that the state intends to return displaced persons to the community.6) Documented and undocumented displaced persons should be registered in the immediate wake of attacks with a view to future tracing of witnesses and complainants (police or DHA officials are not best placed for this task).Undocumented witnesses or victims must receive guarantees that they will not be deported as a result of their involvement in reported cases.Details should include next of kin whose places of residence may be more permanent.7) Priority should be given to xenophobia-related cases, and investments made in obtaining interpreters to lower the case withdrawal rate.8) Deportations should be avoided and where victims are voluntarily returned to their countries, registration on the tracing list should be a precondition.9) Reintegration or managed resettlement planning should commence immediately, with the involvement of displaced persons and the local community.In partnership with UNHCR, government must systematically assess areas for amenability to reintegration or resettlement.Areas of resettlement or reintegration and new addresses should be recorded on the tracing list.10) Successful convictions of perpetrators should be communicated to the media in a manner that promotes visible coverage, as a strategy to counter impressions of impunity at a wider scale.
To comment on this article visit http://www.issafrica.org/sacq.phpNOTES 1 Xenophobia: 'Reintegration conditions good', SAPA/IOL, 25 July 2008, http://www.iol.co.za/ index.php?set_id=1&click_id=13&art_id=nw20080725131647926C795199.(last accessed 29 August 2009).2 Because this statement was made at a closed meeting, we feel it would be a breach of good faith to specifically identify this individual.3 These perceptions were commonly held among 400 participants in 11 sites where FMSP studied the causes of xenophobic attacks between 2007 and 2008 (see also footnote 4).