It has since been drawn to our attention by James Myburgh of www.politcsweb.co.za that the TAU data is compromised by a significant undercount of as much as 7:1 in terms of the actual number of farm attacks. Our initial analysis has therefore been revised accordingly. The revised analysis follows below. The table below compares the farm attack rate for farmers and their families to the house robbery rate, aggravated robbery rate, and combined aggravated robbery and murder and attempted murder rates for the broader population. It is based on four assumptions.
The first is that there are 47,000 commercial farmers in South Africa. This is based on StatsSA’s Census of Commercial Agriculture 2007. The second is that the average household size for farmers is 3. This is the average household size for white South Africans. The third is that, following from James Myburgh’s analysis, TAU undercounts farm attacks by an average of 7 to 1. The fourth is that attacks on smallholdings should be included as attacks on farms.
The table shows that based on those four assumptions the number of attacks on farmers and their dependants is 422/100,000. The number of house robberies in the broader population is 33/100,000. In this comparison farming families are 13 times more likely to be attacked than other families. However, farm attacks also include all forms of robbery and not just house robbery. For that reason the table also compares the farm attack rate to the broader armed robbery rate in the country. Here the farm attack rate, which remains at 422/100,000, must be compared to the national armed robbery rate of 200/100,000. In terms of this comparison farmers and their families are twice as likely to be attacked as other citizens. Farm attacks also involve murder and attempted murder.
For this reason, the table also compares the farm attack rate to the joint aggravated robbery/murder/attempted murder rate for the broader population. Here the farm attack rate, which remains at 422/100,000, can be compared to the national murder and aggravated robbery rate of 260/100,000. Even on this analysis, farmers remain more vulnerable to attack by a ratio of 1.6 to 1. However, the argument can be made that smallholdings are basically large urban plots and not really farms. For that reason, we have published the table below that compares farm attacks, excluding attacks on smallholdings, to criminal attacks in the broader population. The other three assumptions made about the data (above) are unchanged.
This table shows that when smallholdings are taken out of the equation the farm attack rate declines to 262/100,000 farmers and their families. Compared to the national house robbery rate of 33/100,000 farmers are on this measure eight times more likely to be attacked than other citizens are. However, the table shows that when this new farm attack rate is compared to the broader aggravated robbery rate of 200/100,000 this ratio comes down to 1.3 to 1. Furthermore, when the farm attack rate is compared to the broader aggravated robbery/murder/attempted murder rate of 260/100,000 that ratio reaches a level of parity of 1 to 1. In other words, on this measure farming families are no more likely to be attacked than other citizens are. We have also conducted an analysis of the murder of farmers, based on TAU’s more complete murder figures. In order to produce a figure for the highest possible murder rate the table below assumes that only the 47,000 farmers, and not their families, are targeted in farm murders.
On this measure the table shows that in 2011 some 68/100,000 farmers in the country were murdered. This compares to a national murder rate of 31/100,000. On this analysis, farmers are twice as likely to be murdered as ordinary citizens. However, while farmers are particularly likely to be murdered, it is true that family members may also be murdered in attacks. It is for this reason that we have prepared the final table below, which compares the murder rate for farmers and their families to that of other citizens.
This table shows that in 2011 the murder rate for farmers and their families was 33/100,000. The murder rate for the broader population was 31/100,000. On this comparison farmers and their families are not more likely to be murdered than other citizens. Keep in mind that our analysis has excluded the 220,000 emerging commercial farmers, some of whom are large producers, who may also be attacked. We have also excluded from this analysis farm managers and extended families that may live on farms and become victims of attack. Our figures therefore suggest worst-case scenarios for farming families.
What conclusions can be drawn?
The first is that the past week has seen a sharp increase in the quality and level of analysis of farm attacks in South Africa. It has also focused considerable media attention on the problem. Both of these are good things.
The second is that it is possible to argue that farmers are uniquely vulnerable to attack contrary to our initial conclusions drawn from the incomplete TAU data. This is especially so where a straight comparison is drawn between the house robbery rate and the farm attack rate. The same is true for the murder rate of farmers (excluding family members).
The third is that by changing some assumptions it is possible to reach somewhat different conclusions. For example, it is reasonable to compare farm attacks, which is a term inclusive of robbery and murder, to the broader robbery and murder rates of the society. When this is done farmers remain more at risk, although by a smaller measure. However, it is when the assumptions shift to exclude smallholdings that the picture changes significantly. Then the comparisons suggest that farmers are not uniquely vulnerable to violent attack. However, even this result should not be taken to suggest that farmers are safe. Analysts of our data must consider that South Africa’s crime rates are uniquely high. Our murder rate, for example, is 500% higher than that of the United States and 3000% higher than parts of Western Europe such as the United Kingdom and Germany.
This brings us to our fourth conclusion that all South Africans face an extraordinary criminal onslaught. People who have previously regarded farm attacks as a somewhat distant problem, nothing to do with them, must realise that they arguably live in as much peril as farmers do. It is clear therefore, regardless of how they rank compared to other citizens, that as a best-case scenario, farmers live a perilous existence in a largely lawless society.
The fifth conclusion is that, unlike urban middle class residents, farmers do not have the benefit of armed response companies or nearby police stations. Rather they depend on their own defences to secure themselves and their families. In addition, no analysis of the security position of farmers would be complete without noting the role of the State in exacerbating their vulnerability. The closure of the commandos and the Government and the African National Congress’s ambivalence at best toward the incitement by some of their own members to kill farmers creates an environment in which South Africa’s farmers are likely to be killed.
The final conclusion is that there can ultimately be no solution to farm attacks without a broader solution to the general problem of criminal violence in our society. In other words, farmers will not be safe until other citizens are safe and vice versa. This is a point that advocacy groups can use to great effect in their important campaigns to alert policy makers and the international community to the murder of farmers in South Africa.