Farm murders: Fact-checking the fact-checkers

James Myburgh examines the claims that there is nothing out of the ordinary about the murder of white farmers in SA


Over the past year the efforts by AfriForum and others to once again raise awareness of the ongoing epidemic of farm attacks and farm murders in South Africa have attracted a highly dismissive response from left-leaning opinion both in South Africa and the West. One part of the push back, as Marie-Louise Antoni has noted, has been to focus excessively on the completely hyperbolic, and easily refutable, reports of “white genocide” (mass murder) in South Africa, run in the right-wing US and European media. The other, led by “fact checking” publication, has been to raise numerous questions around the real seriousness of the underlying problem.

Africa Check has over the past couple of years repeatedly sought to contest any claim that the farm murder rate is many multiples that of the murder rate for South Africa as a whole, instead arguing that “calculating a farm murder rate in SA is near impossible.” The BBC’s “Reality Check” meanwhile has also argued that we have “no clear idea about the murder rate on South African farms. And because of that, the claim being made by protesters about farmers being more likely to be murdered is not supported by reliable data.”

These claims have been presented as authoritative by those commentators and journalists trying their best to turn the media’s attention away from this hard and morally discomforting issue. The essential message conveyed is that while white farmers are indisputably being murdered in numbers (often brutally and senselessly), “there’s no evidence to support claims that white farmers are more targeted than anyone else.”

The key question is, is this really true?

Murder rate calculations

While these “fact-checkers” have raised legitimate points around the methodology for calculating the murder rate for white farmers, it is unclear why instead of trying to come up with a better estimate they went with the assertion that this was, alone among all the many great mysteries of the universe, fundamentally unknowable.

Determining a murder rate is not, in theory, difficult to do. All one needs is a numerator (the number of murders) and a denominator (a reasonable estimate of the population affected). One then divides the former by the latter and multiplies this by 100 000. So, for the overall murder rate, the South African Police Service (SAPS) recorded 19 016 reported murders in South Africa between April 2016 and March 2017. Statistics South Africa estimated the population of the country at 55 908 900 in mid-2016. This means that the overall murder rate nationally was 34 per-100 000 people.

When it comes to farm murders the SAPS defines farm attacks as “acts of violence against person/s on farms and small holdings refer to acts aimed at person/s residing on, working on or visiting farms and small holdings, whether with the intent to murder, rape, rob or inflict bodily harm.” This definition explicitly excludes “smallholdings where no agricultural activities take place” and which are predominantly residential. Both the SAPS and TAU SA keep statistics and both counted 74 farm murders in the 2016/17 reporting period, though these were not all the same cases. These are also minimum not maximum figures.

TAU SA keeps detailed records – name, location, relationship – of farm murder victims, so the numerator is readily available. It is also possible to determine the race of the victims from this data. Although white farmers are not the only victims of farm murders, they do predominate. In 2016/17, according to the TAU SA data, 66 of these 74 murders were clearly of white farmers or family members, and 51 of white farmers alone.

The problem comes with the denominator. As Africa Check and Reality Check both pointed out the most recent estimate for the number of all (not just white) commercial farmers in the country – 32 375 - comes from the 2007 Census for Commercial Agriculture. On the one hand the number of units has been declining, from 45 818 in 2002 and 57 980 in 1993, and this is a trend that was likely to have continued. On the other hand this census only covered operations that had registered for VAT, so would have excluded a number of smaller operations. Reality Check states that:

“We simply don't know what proportion of people on South Africa's farms is white. The closest we get is the 2016 Community Survey of agricultural households. Of the nearly 200,000 households whose agricultural activity takes place on farmland, about 54,000 have a white person at the head of the household. The problem is, we don't know how many people are in each household or how many of the households are racially mixed. So we don't know how many white people there are on South Africa's farms. And that means we can't work out how likely those white people are to be murdered.”

In fact, contrary to these claims, StatsSA’s 2016 Community Survey did produce estimates of the size of that section of the white population that is affected by farm attacks. It found that there were 47 218 white-headed households engaged in some kind of agricultural activities on “farm land” – including both commercial farm land and small holdings. Of these 24 959 were involved in commercial agriculture and 11 007 were small holders, with the others farming just for leisure, food or for some other reason. The white farming community in South Africa is an aging one, and the average size of these households stood at 2,64.

One is thus looking at a population of around 47 218 “white farmers”, of whom half are fully engaged in commercial agriculture, and a population of 125 000 when it comes to “white farmers and their families.” This matches the universe of white South Africans affected by farm murders, in terms of the standard SAPS definition. If TAU SA’s numbers are right you are thus looking at an overall murder rate for the white farming population of around 51,2 per 100 000 (1.6 times the national average) and of “white farmers” alone of around 108 per 100 000 (3.2 times the national average.)

Table 1: Overall murder rate vs. that of white farming community in 2016/17 reporting period


Population RSA mid-2016

55 908 900

Murders 2016/17

19 016

Rate per 100 000


Population white farmers + families 2016

125 000

Murders 2016/17


Rate per 100 000


Ratio to 2016/17 murder rate


White heads of households on farm land 2016

47 218



Rate per 100 000


Ratio to 2016/17 murder rate


Sources: SAPS, StatsSA, TAU SA.

The East-West divide

This comparison of the murder rate of “white farmers” with the overall murder rate nationally is however inadequate. Overall averages can create a distorted impression if low levels of victimisation in certain areas offset high levels in others, and vice versa. There is in fact a huge difference in the number of farm murders depending on whether you look at the situation in the west or east of the country - two halves of South Africa that have quite distinct histories, populations, and land settlement patterns.

According to recently released SAPS statistics, over the past five years sixty-percent of all farm attacks, and 56 percent of farm murders, occurred in just three provinces: Gauteng, KwaZulu Natal and the North West. This is also supported by the following map by AfriForum which plotted where each farm murder they (and TAU SA) recorded occurred in the 2016/17 reporting year. If you draw a line due north up from Port Elizabeth to Mafikeng, you will see that seventy of seventy-four recorded murders occurred east of that line, and all of four to the west of it. This despite the fact that most individually white-owned land (by extent) falls in the western part of the country.

Map: Farm murders by location in 2016 /17

Source: AfriForum. Green line added by Politicsweb.

An analysis of the provincial murder rate vs. the rate of murders of white farmers by province allows one to control for this geographical effect to some extent. In the three more “westerly” provinces - Western Cape, Northern Cape, Eastern Cape – along with the Free State, there were eight white farmers killed, according to TAU SA’s count, at a rate of 36 per 100 000, only 0,8 times the overall murder rate (47,3 per 100 000) for this area. By contrast in 2016/17 there were 43 “white farmers” murdered in the five more easterly provinces combined - Gauteng, North West, Limpopo, KZN and Mpumalanga - at a rate of 173 per 100 000. This was 6.2 times the murder rate (28 per 100 000) for this area. The high rate in the east offsets the low rate in the west to bring the average down to 108 per 100 000.

Table 2: Murder rate of white farmers by province vs overall provincial murder rates:


Overall provincial murder per 100 000

White heads of farming households 2016 StatsSA CS

White farmers murdered – TAU SA

Ratio per 100 000

Ratio of rate of murder of white farmers / rate of murders by province



6 588






5 700






4 181






3 678






4 701






6 402






7 377






3 898






4 693






47 218






24 848






22 370




Sources: TAU SA, SAPS, StatsSA

Types of murder

There is yet another consideration however that needs to be brought into account. If you are comparing rates of murder of white farmers and their family members with the overall national murder rate you are not in reality making the appropriate comparison. All murders may be lumped in the same category in the crime statistics in South Africa, but they are the product of several very different types of violence. These affect different communities in different ways, have different causes and drivers, and require different police (and societal) responses to deal with them. The profound and enduring trauma of losing a beloved family member to murder is however the same whatever the circumstances.

In South Africa, as in many other countries[1], a substantial proportion of murders are committed either by perpetrators who know the victims, often intimately, and/or as the result of arguments and brawls, often under the influence of drugs or alcohol. The SAPS definition of farm murders itself explicitly excludes “cases related to domestic violence or liquor abuse, or resulting from commonplace social interaction between people.” Such “social murders” should not be included in the comparison of rates.

Gang violence, in which gang members kill each other or innocents are caught in the crossfire, accounts for another major share, and is a major contributor to the horrifically high murder rate on the Cape flats. Yet another share is taken up by mob justice murders where communities turn on and kill suspected criminals or other perceived enemies of the people.

If you want to properly judge the relative level of victimisation of white farmers at the hands of “criminals” then, you need to compare like-with-like. In other words what is the murder rate of white farmers relative to the rate of the most relevant sub-category of murder? The argument has repeatedly been made that farm murders are just the product of “ordinary crime”, namely aggravated robbery, and are not (in any way) politically or racially driven. If one accepts this categorisation, if not the underlying premise, then the most appropriate comparison is clearly with the rate of murders committed during aggravated robberies, and particularly home and business robberies.

Calculating this raubmord rate, to borrow the German term, is not all that simple. South Africa’s standard crime statistics reporting does not provide breakdowns of the differing circumstances of reported murders, though it urgently needs to do so. However, a national docket analysis of 2 912 murder cases reported between April and September 2015 was done by the SAPS, and published along with the release of the 2015/16 crime statistics. In 1 727 (59,3%) of cases it was possible to establish the circumstances of the death. In 58,9% (1 017) of these cases these murders were “social murders”, the product of drunken arguments or lovers quarrels, and so on. Almost 10% (171) were mob or vigilante killings, 2,1% (36) were found to be gang related. 352 cases of murder (20,4%) were committed during the “commission of another crime”: 101 during street robberies (5,8%), 74 in residential robberies (4,3%), 70 in business robberies (4,1%) and 71 during carjackings (2,1%).[2]

One needs to be cautious about projecting these ratios onto the 40% of murders where it was not possible to establish the circumstances leading to death, and from there onto all murders as a whole. “Social murders” are probably over-represented in the known category, as these would have a high clearance rate. The ratios when it comes to residential, business robberies and carjackings are probably similar, given the relative ease in establishing motive even if the perpetrators are not arrested. It would be most difficult to establish a motive relating to murders stemming from street crimes, where the body may be only discovered after daylight breaks, and where there are no witnesses able or willing (in the case of gang killings) to give evidence.

One could cautiously posit the following: A plurality or a slim majority of killings likely fall into the “social murder” category. A third of murders may occur during the “commission of other crimes”, many of which were during street robberies. Just under a third of these – a tenth overall – would be murders stemming from residential or business robberies. Gang violence related murders and vigilante murders would make up much of the remainder. If these proportions applied to the 19 016 murders reported in 2016/17 then around 2 000 would have been committed in the course of residential and business robberies and a further 4 400 or so during street robberies and other crimes. The raubmord rate for the population as a whole would be in the vicinity of 12 per 100 000 for all aggravated robberies, and 4 per 100 000 for home and business robberies (combined).

This would mean that the murder rate of white farmers and their families is about four times a plausible national raubmord rate for all aggravated robbery, and thirteen times the rate for home and business robbery-murders only. For white farmers alone these multiples can be doubled.


The above all suggests that there is ample evidence that white farmers, particularly in the east of the country, are being “targeted” at an extraordinarily high rate. This is especially so if you make the most appropriate like-for-like comparison when it comes to the type of murder. Indeed, it is difficult to see what could possibly have motivated reputable “fact-checking” websites to go out and insist on the contrary.

There were 24 murder-robberies in England & Wales (population 56 000 000) in the year ending March 2017, a raubmord rate of 0,04 per 100 000. If the robbery-murder rate was ever to come close to reaching the same level there, as on white farms in South Africa, there would be well over 20 000 such homicides a year. In the United States the figure would be over 100 000.

If such a catastrophe was ever to occur in their own societies it is difficult to imagine that the first impulse of such fact-checkers, and the intellectuals citing their work, would be to try and strenuously downplay the enormity of the problem, in an effort to hobble any sustained political efforts to try and deal with it.


[1] For example in the US in 2015 there were 13 455 recorded homicides, according to the Federal Bureau of Investigation. In 7 023 (52%) of cases the relationship of the victim to the perpetrator could be established. The vast majority of these victims knew their killer. In 24,5% (1 721) of the 7023 cases the victims were family members of the perpetrators, in 9,2% (648) they were in a relationship, in 46,7% (3 279) they were friends or acquaintances, and only in 19,6% (1 375) were the killers strangers. In the 8 089 (60%) cases where the circumstances of the killings could be established, 15,9% (1 284) were gang or drug related, 42,7% (3 454) were the product of arguments, brawls, love triangles etc. and only 9,3% (754) occurred during robberies, burglaries or thefts.

In Wales and England meanwhile there were 571 homicide cases, in the year ending March 2016. In 80% of cases the circumstances were known. 48,5% (277) resulted from a quarrel, a revenge attack or a loss of temper. This proportion was higher where the principal suspect was known to the victim (57%), compared with when the suspect was unknown to the victim (37%). Only 4% of homicides (23 offences) occurred during “furtherance of theft or gain.”

[2] SAPS, Annual Crime Report 2015/2016, Addendum to the SAPS Annual Report, 31 August 2016. This docket analysis also found that 35,7% of murders were committed in streets or roads, 11,3% in open spaces, 9,5% in bars, pubs, shebeens or taverns, and only 20% at or in the residences of the victims. In the 38,1% of cases where it could be established from the dockets whether the victims had consumed alcohol, or not, 56,8% were found to have done so. In 50,2% of these cases the murders were committed over the weekend, in 73,5% of cases they occurred between the early evenings to early morning (6pm to 9am). 41,2% of murder victims were aged between 20 and 29 and 69,7% between 20 and 39 years.