Hello all, welcome to the home of SA election data. I am going to treat this account like a live blog almost, talking about the data I have, how things are being developed and of course the outcomes. If you are interested in that kind of stuff, I would love to talk to you about it. Insights, thoughts, requests all welcome. All of the final data will appear on my blog - https://t.co/2w4hrlZAmD - it already has a ton of stuff up on it, and I am always growing and developing it. So, if this kind of thing appeals to you, welcome!
@camerondugmore@MAlvarezRivera I see being pedantic about helpful generalisations is important to you. You are correct. The EFF has contested 85% of all by-elections since its formation. Well done. Glad I can provide 100% clarity on that entirely inconsequential phrase.
The May edition of the SA By-election track is out (If you want to subscribe, send me an email, it's free), but thought it would be worth looking at the performance of two parties in particular: The EFF and the FF+. Both those parties currently enjoy a certain narrative - the EFF, that is is in decline, on the back of defections and Malema's conviction; the FF+ - that it is in the ascendancy, on the back of a few prominent results and social media sentiment. So far as actual by-elections go however, (and that is all I am talking about here) both are incorrect. Let us start with the EFF. There are two EFF graphs attached - it's monthly average (12_MA) and its swing average (by-election result vs previous LGE result). It's monthly average it strong, a slow gradual upward tick. No sign of implosion (and the EFF has a very big sample, it contests everything). It's swing average it also positive, by almost 2pts. In fact, it has never gone negative. So, it appears much more solid than it is given credit for. Whatever its public problems, they are not manifesting in its by-election performance at all. The FF+ is the opposite - significant online positivity, but poor by-election track. The FF+ achieved an all time high at the beginning of 2023, but since then, it has been tracking consistently down. Its monthly average has dropped 10pts since 2023, down from 24pts to 14pts. The FF+'s swing graph really tells that story, a catastrophic decline, from +16pts just before the 2021 points all the way down to -0.3pts in May 2026. In layman's terms: the FF+ has systematically been performing worse that its previous LGE result for five years and now, the first time since 2011, that cumulative decline has resulted in its average going negative. What is interesting about both of these, is how they bear very little resemblance to the narrative, largely driven on social media, that defines both parties. In the same way, they are good reminder that Twitter is not a good guide to real world. in the real world of actual by-elections (as opposed to Twitter polls), the EFF is rock solid, and even climbing marginally; the FF+ is in serious decline. The FF+ result is perhaps even more significant, remember the EFF contests everything, the FF+ is very selective and only competes in its strongholds, so this result is properly doubly troublesome for it. There is a lot of time to go until the 2026 LGE and a lot can happen and with these small parties - small shifts can have big implications for them. But as far as trends go, these are pretty rock solid. A final word, perhaps these tracks constitute the biggest indictment of the media, who generally and unthinkingly replicate online narratives as the expense of hard facts and real world indicators but, I suppose, that is par for the course. I mean, ask yourself, what kind of pressure would the DA leader be under, if the FF+'s by-election track was his by-election track record? At any rate, something for everyone to chew on, as you try to distil fact from fiction and data from Twitter, in the coming months.
Emfuleni 10. So, I think, from a DA perspective, the situation is this: Emfuleni demonstrates that there is a tipping point. It seems to be closer to 20% than 10%, which is what the DA had in Evaton prior to the by-election, but not necessarily in rest of municipality. Nevertheless, the trends are clear throughout Emfuleni: profound ANC collapse among black voters, DA systematic albeit slow growth in same market; 17% in Evaton, 10% elsewhere. So, again from the DA's perspective, there is a tipping point. It seems to be around 20% as general rule. It might not happen in this upcoming election on a large scale. Perhaps there are a range of 'all-black" VDs elsewhere that sit in tipping point range, but they will be particular and not general. But, and this depends on the DA maintaining its long term trajectory, if Evaton is a guide (and I admit we are dealing with a sample of one), a tipping point does exist for a big switch. To trigger it, it seems the DA needs two things: further ANC decline, plus a big campaign to make an impact. If this theory is correct, the 2026 LGE, which will constitute such a trigger, will see the ANC, in a surprising number of "black" wards, where the DA is around 20%, lose control. If it is incorrect, it won't happen this election or there will be too many wards where the DA is not near the "trigger threshold". Either way, it is clear the tipping is close, and the DA is getting closer to finding it.
Emfuleni 1. So where to start? Well, I think a good general map is helpful. Here is Emfuleni municipality. Evaton, the ward where the recent by-election took place is pretty much at 12 0'clock. Zoom in. Maybe 12.05. Midvaal (DA run) is on the right.
Emfuleni 9. The DA up to 10% across all those 95% black VDs in Emfuleni. The ANC collapse among black voters in Emfuleni also clear. Same ingredients. Problem is, we don't know if DA would spike to 30% in those VDs if by-election held in them.
Ok. A good solid seven or eight affirmatives. I will take it. I am going to set out some thoughts. Am not quite sure how to do this, I am looking up stuff in real time, so it might be a bit messy. I will try and make sure each tweet follows the other.
Yesterday’s by-elections are the last for May, busy updating the by-election track. It is only available by email. If you want to subscribe, send your name and the email address you want it sent to, with the title SABT, to [email protected] Will send out May edition tomorrow.
Just to also say: it also speaks to the lack of agility on the ANC’s part. You need to do everything you can to drive economic growth, one path to exemption for Starlink might be exactly your concern - make connection costs for rural schools low, say, and you can come in. Something like that. So many options here to make this work, and get benefit from it. But the rigid ideological hatred of Musk means the interests and needs of people and the SA economy, is always a secondary consideration. At least it is not for the DA. Personally I would like to see a poll, that asks rural South Africans: would you like the option of access to high speed internet. And see what they say. I suspect the answer will be very different to a Twitter poll.
I suspect Redi Thlabi’s article on Starlink carries the kind of Twitter headline (“My piece on why Starlink IS a national security issue for South Africa”) that causes 90% of people to retweet or condemn it based on that alone. But on actually reading it, it is striking how weak it is. For one, it does not provide an answer to the headline. There is no evidence that Starlink is a national security risk anywhere in the article, in any shape or form. Remarkably, there is in fact not a single fact about Starlink. Staggering when you think about the response to it. Instead, most of it is about social media, the dangers of propaganda, on platforms like X and TikTok, and how various governments have or have not responded to those platforms. But on high speed internet access, not a word. That is because all that stuff is irrelevant to high speed internet – it just makes your internet go faster. What makes the article even more bizarre, is that the companies responsible for these platforms, X, Facebook, TikTok, etc, all already operate in South Africa. So, I mean, given her passion for national security, that is what she really should be focusing on. But it’s the patronising paternalism that really grates. Starlink is a product, like Discovery Health or your Apple iPhone. If it was allowed to be switched on here, every consumer, just like every other product, would be faced with a choice: read the terms and conditions, and sign up, or don’t. Whatever you want. And, like all other products, it would subject to laws and regulations, and held to account for wrongdoing. What Thlabi is really saying, in a sort of “mother of the nation” mode, is: “you people are not capable of making this choice, you need me, Redi Thlabi, to make it for you, by advocating for the choice not to be available to you at all.” Of course, she doesn’t do that for any other product. Banks, Facebook, Insurance firms, they are all interested in your personal data. They all tell you what they will or won’t do with it. It’s a decision you make 100 times a year in different ways. Starlink may do this, I don't know, not sure why it would need to, but even if it does, it would give you a choice - sign up if you like our conditions, or don't. As for politics all of those companies have owners with views. Some of them with quite mad views. But it is only on this one that Thlabi is going to fight, not for your right to choose, but to ensure you can’t choose: because if you are given that choice – being the idiot you are, you will make the wrong decision. This is an ego out of all control. Really. And arrogant too. Thlabi starts by dismissing out of hand any other consideration. She says, “We are told this is about rural children accessing the internet, about innovation, about economic growth” and then says she knows “the real question”. Oh really? I think rural children, the unemployed and the economically marginalised would disagree. I think they could tell her the “real question” very quickly: how can I get a job? Thlabi has a job, and internet access. She also thinks she knows what is best for you. Much like the ANC. And much like the ANC, she has offered up a massive smoke screen filled with rhetoric and hyperbole, that deals with nothing but her own animosity towards Elon Musk and, in doing so, shows nothing but contempt for the choices of people themselves. https://t.co/BEkKY63ONn
I don’t think it is a silver bullet, I don’t think such a thing exists. But I do think it would be positive, not negative. Cost is an issue, agreed. But institutions, schools, SMEs can afford it. Like all economic incentives it requires the right supportive environment (this is a broader ANC policy question). Whether or not Starlink is economically viable in SA we would need to see, but that is a business decision for Starlink. Either way, it would open up a closed market, increase choice and opportunity, to whatever degree and spark competition - so that internet speed becomes more competitive. These are all good things to my mind. But what I don’t think anyone is in a position to argument that nothing is better than nothing, that doesn’t help anyone.
Worth taking a moment to contemplate the opposite. Only the DA endures this kind of pressure. There is no expectation that the ANC, EFF or MK ever win a “white” ward. The fact that they are racially homogeneous organizations is hard wired into SA political analysis and accepted uncritically as normal.
For the first time ever, the DA has won a township ward populated entirely by black voters.
I remember in 1997 how the DP won a string of by-elections in Gauteng that heralded a sea-change among white voters.
There is now a shift happening among black voters. It may not be on the same scale as 1997, but there is a shift, and it’s happening.
Doing all these party support by population group maps quickly reveals that it is really only about the DA. Every other party of any significance is mono-racial. The DA really is unique. The only substantive party that can claim, on hard evidence, to have a diverse support base.