On race and crime: A reply to a reply

Gavin Silber and Nathan Geffen respond to James Myburgh's critique

This is a reply to James Myburgh's response to our article Race, class & violent crime in South Africa: Dispelling the ‘Huntley thesis'.

A summary of the debate up till now:

1. Brandon Huntley was granted asylum in Canada in 2009 based on the argument that whites are disproportionately targeted as the victims of crime in South Africa. A Canadian panel found that the ANC government was failing to protect the white minority from criminal violence perpetuated by black South Africans.

2. 142 academics wrote an open letter criticising the outcome of the Huntley case in which they claimed that blacks bear the brunt of crime in South Africa (see here).

3. Myburgh disputed the claim by the academics, using national victimisation surveys to argue that whites bore the brunt of crime (see here).

4. Our article defended the academics' claim. We argued that the balance of the --admittedly incomplete-- evidence shows that blacks bear the brunt of violent crime in South Africa, which we argued was the reasonable interpretation of the academics letter (see here).

5. Myburgh then wrote a response to our article (see here).

The crux of the debate

Myburgh argues that "the authors (Silber and Geffen) read into the academics letter a claim they did not make", i.e. that they were talking about violent crime. But Huntley's asylum application was about violent crime, not about all categories of crime and the academics letter dealt with the Huntley case, so as we argued this was a reasonable reading of the academics letter. We have held that it is disingenuous and even dangerous to fail to distinguish between crime in general and violent crime, as it is the latter that has a greater adverse effect (albeit impossible to quantify) on both the individual and society at large.

As we also made clear, we have primarily (but not exclusively) discussed murder as a proxy for violent crime because data on other forms of crime is so unreliable, largely due to the poor quality of police data and under-reporting (particularly for other violent crimes such as rape and assault).  Despite being novices in this field, we are not the first to contend this.  Ted Leggett - a research expert at the United Nations Office on Drugs and Crime (UNODC) and a former ISS senior researcher - has worked extensively with South African crime statistics, and has noted that "murder is the surest indicator of the real level of violence in society, because it is almost 100% recorded by police. The public feels compelled to account for dead bodies."[i]

The essence of this dispute to date can be best illustrated as follows: Huntley's asylum application was dependent on the claim that whites are disproportionately affected by violent crime. If (a) the victims of violent crime are distributed proportionately across races or (b) blacks are disproportionately affected, Huntley's claim is false. We argued and continue to argue that the balance of evidence shows (b), i.e. that blacks are indeed disproportionately affected by violent crime (using mainly homicide as a proxy). Myburgh's arguments are sympathetic to Huntley's position.

MRC intimate partner violence data

Myburgh points out that the MRC data we presented on women victims of murder was for intimate partner violence only. He is correct and this was one of two errors on our part for which we apologise. Nevertheless, he then provides data showing that black women compared to white women are disproportionately the victims of murder, whether the data is broken down by intimate partner violence, by other perpetrators or both. While Myburgh admits this for coloured women who are the victims of non-intimate perpetrators, he incorrectly states that the data suggests that black women victims of non-intimate partners are not disproportionately murdered. We explain his error in an endnote.[ii]

CSVR homicide victim survey

Myburgh points out that CSVR homicide victim data we quoted was for specific high-risk areas in KZN and Gauteng and not proportional to the population. This was our second error, for which we apologise. However, again Myburgh provides the correct data which still supports our position and not his, albeit less convincingly.

Cape Town homicide data

Myburgh states that the "authors cite murder and rape statistics from Cape Town as a 'case study.' Given the political and demographic exceptionalism of that city is difficult to see why they do so other than that this very uniqueness supports the pre-ordained conclusion they are straining to reach."

Beyond Cape Town, we cited national unnatural deaths including assault data, as well as data from Kwazulu-Natal and Gauteng. But critically, Cape Town is the city where Huntley lived and was allegedly disproportionately targeted because of his race. So our use of Cape Town data is appropriate.

Myburgh criticises us for stating: "five police districts account for over 44 per cent of murders - Nyanga (13,18 per cent), Harare Khayelitsha (8,67 per cent), Khayelitsha (8,47 per cent), Gugulethu (7,58 per cent), and Delft/Belhar (6,1 per cent)." He argues, "but they fail to take us into their confidence by telling us what proportion of the population of Cape Town actually lives in these particular areas". We did not present this data in isolation; we included it with other data showing unequivocally that blacks are disproportionately the victims of murders in Cape Town. Myburgh appears to acknowledge this.

It is difficult to calculate accurately the populations for these areas as they all contain vast informal settlements and existing census data is dated and unreliable.  The best illustration we can find, which lists causes of death as a percentage of the local population, can be viewed at the City of Cape Town's Health Information website (unfortunately the latest available data is from 2004), which clearly illustrates where the burden of violent crime resides.[iii]


Myburgh concludes his response with a number of less relevant and occasionally perplexing claims.  He correctly states that many victims of violence are themselves involved in violent crime. Indeed, the high prevalence of gangs in black and coloured communities and the incentives that pull youths into gangs is indicative of how fractured black townships are because of the socio-economic factors we touched upon in our article.

He further warns that "so-called ‘social fabric crimes' are not really relevant in determining whether individuals from vulnerable minorities are seen as legitimate or soft targets by predatory criminals".  These assertions show a bias that fails to recognise how poverty, inequality and social turmoil characterises violent (and quite possibly property related) crime in South Africa.

Surely whether one is a attacked in one's own home, or because of involvement in criminal activity or on the street by a stranger, one is still a victim of crime. At any rate this has no bearing on the key point of debate, i.e. whether whites are the disproportionate victims of violent crime.

Targeted crime against whites?

Myburgh quoted a statement by Dullah Omar to support his statement that it is "an open question whether [the ANC's] racialist propaganda would have given the green light to criminals to cross over the colour line en masse." We responded that it was "far-fetched to imply that an obscure statement by a minister would have been sufficient to encourage cross-racial crime."

Myburgh responded, "But I fail to see why it is inherently implausible. Would Silber and Geffen also claim that Thabo Mbeki's equally 'obscure' statements on AIDS and anti-retrovirals had no effect on people's behaviour on the ground or on the conduct of ANC apparatchiks in state and society?"

Myburgh and one of us, Geffen, have independently and extensively documented Mbeki's statements and, as importantly, his actions in support of AIDS denialism. As Myburgh is undoubtedly aware from his own important work, they were definitely not obscure and they were part of an unequivocal and calculated ideological position. It is testimony to the obscurity of Dullah Omar's statement that Myburgh originally attributed it to the wrong minister.

We acknowledge that it is possible that Myburgh believes this was not an isolated statement, but we are unaware of evidence supporting an increase in crime specifically targeting whites since Omar's statement was made. Surely such a serious insinuation or suggestion from Myburgh must be backed up with substantial evidence.

Perhaps such evidence exists (we are sceptical), but Myburgh should present it or else not make the allegation. This would require collecting a compelling corpus of statements or actions by senior political leaders as well as data showing that on a large scale whites are being specifically targeted because of their ethnicity. The NVSs are not enough, as they can only represent perception of crime.


Myburgh's response to us has presented data that predominantly supports our argument not his. The balance of the admittedly limited evidence shows that blacks are disproportionately affected by violent crime, which is certainly at the crux of the Huntley case and surely the open letter by the academics, even if they were not explicit about it. 

As we have said before, everyone in South Africa is exposed to crime. Fighting over resources on the basis of race - an implicit consequence of Huntley's asylum case and its aftermath - is unproductive and divisive.  This is best exemplified by the racist comments in response to both our and Myburgh's articles on Politicsweb, appropriately described as moronic by Myburgh.

In any case, with the exception of xenophobic crime against Africans from outside South Africa, it is arguably class and income inequality that are the key factors differentiating the types of crime people experience. Their high correlation with race should not be confused with racially based crime.[iv]

Click here to sign up to receive our free daily headline email newsletter


[i] Leggett, T. (2003) Murder is real measure of violence. Sunday Times. 28 September.

1.[ii]  Myburgh cites data showing that the rate of murder of females by non-intimates is 8.49/100k for Africans and 5.87/100k for whites. The difference is statistically significant. Myburgh's error appears to stem from the fact that the proportion of African women murdered compared to other race groups is in proportion to their population, but this has only worked out this way because of the particularly large proportion of coloured women murdered. Note that in our articles we have used the standard meaning of black, i.e. Indians, Coloureds and Africans.


[iv] Two further less important points we wish to respond to:

(i) Myburgh claims that a quote by Michael O'Donovan was taken out of context, but this was not the case.  We noted that O'Donovan points out that NVSs are useful insofar as they reflect respondent's beliefs, but should not be used in isolation to objectively gauge the level and distribution of crime. We further pointed out that O'Donovan noted certain anomalies that cautioned against applying NVS data to the public as a whole.  Myburgh's intial article provided no such warnings, when he provided selected quotes by the same author stating that "contrary to popular perception, the likelihood of being a victim of crime rises with income", and concludes by asking whether minority groups might be more vulnerable to crime.

(ii) Myburgh presents 2007 Stats SA mortality data on unnatural deaths which essentially shows the same thing about unnatural deaths from the earlier survey we used. There is no disagreement between us here. We used the earlier data because it was part of a manually tabulated analysis (not since repeated to our knowledge) and included a breakdown by race of death by assault which we have not been able to find elsewhere.