On 22 August 2018 US President Donald Trump tweeted: “I have asked Secretary of State @SecPompeo to closely study the South Africa land and farm seizures and expropriations and the large scale killing of farmers. ‘South African Government is now seizing land from white farmers.’ @TuckerCarlson @FoxNews”.

These claims were then subjected to a “fact check” by the Washington Post. This asserted in turn that “Fringe groups in South Africa and the United States say these white farmers are targeted and killed at disproportionately high rates. There’s no evidence for this claim, experts say, but it feeds the white-supremacist agenda by stoking racial resentment and division.”

It awarded Trump Four Pinnochios for making this claim.

The following is the subsequent correspondence between Editor of Politicsweb, James Myburgh, and the Washington Post The Fact Checker.

Letter from Dr James Myburgh to Glenn Kessler and Salvador Rizzo, 24 August 2018

Subject: WP fact check on farm murders in SA: A complaint

Dear Glenn,

I read with interest your "fact check" on Trump's tweet on farm murders in South Africa. It cited research I did in the following way:

" James Myburgh, a South African researcher whose work was referred to us through Tupy, wrote that considering the total number of white farm owners, the more than 60 murders recorded in 2016-2017 would be far above the norm by his calculations. However, he acknowledged that the official farm-murder statistic was flawed. This is the best-case scenario for Trump: that some educated guesswork based on flawed statistics sort of backs him up, maybe."

Try as I have to make sense of this paragraph I have absolutely no idea what Salvador Rizzo is trying to say here. It is also a misleading and garbled summary of my research. Our official statistics agency estimated that the white farming population in South Africa was around 125 000 people in 2016. This is a reasonable figure and consistent with other data.

In 2016/17 there were 66 murders affecting that population counted by [AfriForum]/TAU SA*. Again, this number is consistent with the SA police service figures. (They don't provide a racial breakdown however so I could not use them for the analysis). You are thus looking at a farm murder rate of white farmers and/or their family members of about 50 per 100 000 people. These are, by definition, murders of people on their property usually by gangs of criminals. 

Africa Check contacted me at the time and asked for all my underlying data, which I provided to them. They have not subsequently disputed this estimate. Nor has anyone else. Until more definitive research comes along this is the best estimate there is. 

66 murders per year may not seem like a great number, but it is a very small population. If you had a similar rate of people being attacked and killed on their properties in the US that would be over 150 000 people a year. These farm murders have been occurring, at similarly high levels, for just under three decades in South Africa.

It is entirely your call as to whether these figures support or contradict Trump's tweet. But there is no reason not to lay out the numbers fairly and accurately.  Further, the claim that there is "no evidence" that white farmers are being targeted and killed at disproportionately high rates is simply not true, as is the claim that only "fringe groups" in South Africa are saying this. 

I have long been a great admirer of your and the Washington Post's work in this field, so it is a disappointment (to say the least) to read such a "fact check" with such glaring falsehoods. 

Best regards, 

James

Reply by Salvador Rizzo of the Washington Post, 24 August 2018:

Thanks for your email. I wrote the fact-check you're referring to and I think the paragraph you take issue with is clearly and fairly written. We report on statistics often and believe that there are different tiers of quality and vetting for statistical data and analysis and that your estimates fall in a lower tier.

You calculated more than 60 murders of white farmers in the 2016-2017 period, you found that this was disproportionately high considering the total number you estimated, you acknowledged the official police statistics were flawed, and that's what we reported. We have a responsibility to ensure that spotty information like this is correctly described and appropriately contextualized.

Whether you take issue with our conclusion at the end of that paragraph that the best-case scenario for Trump is that some educated guesswork may or may not support his claim, that is an accurate description of what your calculations indicate, considering your acknowledgment of the deficiencies in the underlying data and the fact that farm murders as reported by the police and others have been declining for years.

I reviewed the back-and-forth over the quality and completeness of each dataset and am satisfied that none of them gives an accurate measure. As you note, a racial breakdown is not included. The TAU data seem incomplete and subject to arbitrary influence, relying as they do on social media posts and other forms of word of mouth and anecdotal evidence. I'll end by noting that we linked to your work and acknowledged your argument, not what I'd call unfair.

Regards,

Sal Rizzo
Reporter, The Fact Checker
The Washington Post

Reply by James Myburgh to Salvador Rizzo, 24 August 2018

Dear Salvador,

Thank you for your prompt reply, and taking the time out to respond to my concerns. I appreciate it. 

As you know the figure for farm murders comes from TAU SA. They literally counted 66 cases of murders of white farmers and their family members in 2016/17. They recorded their names and the circumstances of their deaths, and even have the police case numbers. Almost all of these cases were also reported in the press. There is some margin for error in categorisation, but that is it. If one could cross check with the cases in the police database then the numbers would just be more complete and possibly higher as well. 

The whole point of the Africa Check piece that you cite is that there is great uncertainty not with the number of farm murders, but the population affected. Here I was able to get data from the StatsSA Community Survey of 2016. This is one down from our census. The calculation of the rate - that I did - is derived from the best available data that exists currently in South Africa (however imperfect). This is not "educated guesswork."  

The further two points I made were that, firstly, farmers in the eastern half of the country are far more likely to be attacked and killed than in the western half. And, secondly, that the national murder rate is not the most appropriate point of comparison, but rather a sub-category of the murder rate. Both are completely valid.

It was here that the SAPS data is spotty, but I tried to use the most conservative assumptions possible to accommodate for this. The principle is also fairly universal. Being killed on your property by strangers is a highly feared - but very rare - form of murder. Except for white farmers in South Africa. You can reasonably argue whether farmers and their family members are eight, or ten, or twelve times more likely to be killed by criminals on their properties than the average South African. But to claim that there is "no evidence" that they are being disproportionately targeted for attack is completely ludicrous.

If you could have found better data or estimates you could have published it in your article, but you didn't. Instead you quote the former US Ambassador Patrick Gaspard and Mark Pitcavage at great length. I have great respect for them individually, but they have clearly prejudged the matter, and have no knowledge or understanding of the actual issues at play here. 

You also say that "farm murders as reported by the police and others have been declining for years." Your source for this is a Guardian report citing an Agri SA report citing SAPS figures, which I actually was the one to first publish. They are not reliable, for the last few years, as explained here.

I would humbly posit that the reason you did not accurately summarise my argument in the text of your analysis is simply because it got in the way of your predetermined conclusion.

There are actually two fringe viewpoints in this debate. On the one extreme are the racial lunatics - in your country - who claim that there is mass murder going on of white farmers in South Africa. On the other, are those who claim - like yourself - that there is "no evidence" that white farmers in South Africa are being targeted and killed in disproportionately high numbers. Both are ideologically driven.

You'll understand the bizarre phobias of the white supremacists in your society better than I can, and I won't speak to that. On the other side, there is a very strong underlying sentiment among many white and Western intellectuals, that when a white African farmer gets killed he is just getting paid what he deserves. Essentially, all the sins and crimes of the era of European world conquest are placed upon such individuals, and their "punishment" (through murder or dispossession) then triggers a semi-religious feeling of absolution from white guilt. Whenever there is any suggestion that white farmers should or could be considered as "victims" this elevating feeling of virtue is threatened, and there is an emotional, and not always very rational, push back against that. Again, you'll understand the impulse at work here far better than I do. 

As you know - but didn't mention in your article - since the unbanning of the liberation movements in South Africa in 1990 there have been around 2 400 people killed and many others seriously injured in well over 15 000 farm attacks. These numbers dispose of the claim that this qualifies as "mass murder”. But the brutal nature, and sheer accumulated number, of these attacks and murders are on any other measure horrific.

By writing with partiality about the issue in the way that you did you only undermine your own credibility. And by making light of the situation you are giving acres of space to your white racist fringe to exploit it. Given the current populist threats to liberal democracy in the West this seems a particularly ill-advised approach to take.

Best regards,

James

Follow up letter to Salvador Rizzo and Glenn Kessler, 27 August 2018:

Dear Salvador and Glenn,

From your non-responses I presume that my objection to your "fact check" on South African farm murders is not going anywhere further.

If that is indeed the case I may go ahead and make my disagreement with your article public, and publish this correspondence. Please advise however if you are still planning to conduct a further review of the piece, or have any objection to that. 

Even if I feel you have fallen short here - I respect the work you do generally. It is critically needed in this age.

Best regards,

James

Reply by Salvador Rizzo, 27 August 2018:

Dear James:

 We respect your right to voice your disagreement in public. We didn’t have predetermined conclusions for this fact-check. During our reporting, we were convinced by the rationales laid out by Africa Check and the BBC, which both found it was not possible to assess accurately the number of farm murders in South Africa using currently available data. To the extent that it might be possible, we note that both the SAPS and the Agri SA numbers have been declining in recent years.

 We read your analysis with interest and thought it merited discussion. We continue to stand by our assessment. The acknowledged gaps in quality in the underlying data you used, the fact that the official farm-murder statistic has been in decline, and the lack of consensus around your conclusion gave us pause. We didn’t foreclose the possibility you might be right, although we expressed our skepticism appropriately. As I wrote to you previously, we report often on statistics and data analysis and we could not responsibly tell our readers to rely only on your analysis, considering the issues we identified.

Sincerely,

Sal Rizzo
Reporter, The Fact Checker
The Washington Post

ENDS