The ActionSA bullshit monster

Gareth van Onselen writes on the party's bizarre (and hypocritical) hostility to opinion polling

The Action South Africa bullshit monster

11 March 2024

In May 2020, sometime before Action South Africa (ASA) was officially launched, Herman Mashaba announced his new political party would contest three Gauteng metropolitan municipalities: the cities of Johannesburg, Tshwane and Ekurhuleni, in the upcoming 2021 local government elections. 

He intimated that, in the future, he would look at supplementing that list. However, he said his party would only identify, “municipalities that are strategic in nature and where we will win”. 

ASA’s official launch was in August 2020; some six months later, in April 2021, Mashaba announced that ASA would also be contesting eThekwini as well, bringing its total to four. In late July, ASA added both Newcastle and KwaDukuza to its list.

Mashaba reiterated at the time that, “we will only contest municipalities where we can prove we have sufficient support to make a significant impact” – a downgrade from “where we will win”, but potential support still remained a key ingredient in deciding where the party would expend its resources.

On Election Day, 1 November 2021, ASA ended up contesting eight municipalities, the six above and two additional district councils: Amajuba and iLembe. But those last two are of no real values for this analysis. (ASA got 2.6% and 3.4% in them respectively, for the record.) 

More to the point, here is how ASA performed on the Proportional Representation (PR) ballot in those six flagship municipalities it had selected, on the basis it had a prospect of winning/making a significant impact:

JHB – City of Johannesburg: 18.12 % (167,359 votes)

TSH – City of Tshwane: 9.28 % (62,502 votes)

EKU – Ekurhuleni: 7.36 % (49,395 votes)

ETH – eThekwini: 2.35 % (18,201 votes)

KZN292 – KwaDukuza: 8.05 % (5,812 votes)

KZN252 – Newcastle: 3.59 % (2,898 votes)

Nationally, ASA ended up with 306,167 PR votes, or 2.61%, enough to make it the 6th biggest party, with a consolidated percentage (Ward plus PR) of 2,33%.

(In fact, on the PR ballot alone, ASA was the 5th biggest party, it missed a trick there.)

“Six municipalities, 6th biggest party” would become the cornerstone of ASA’s public communication from that point on. The media – generally poor at political analysis and often disproportionately enamoured with new comers – lapped it up. Every second headline post-election screamed triumph.

Buoyed along by the post-election drama of coalitions, no one ever really stopped to properly analyse ASA’s performance (true of all parties). So effectively had the party captured its outcome in a single unthinking slogan, it negated any serious cross examination.

The idea was so successful, it has become almost part of ASA’s brand; certainly ever-present, never questioned, analysed or interrogated. As a result, in almost every interview Mashaba has given since, defending ASA or talking up its prospects, this line is front and centre. 

ASA’s “six municipalities; 6th biggest party” line is one of the greatest PR stunts in recent South African politics. But really, it is a trick of the light, an illusion. Thus, simultaneously, misleading. Here is how ASA suckered everyone, and created a monster.

1. The 2021 results

We did exceptionally well, in the history of the democratic dispensation no political party – one-year-old – has ever performed the way that we did”, Mashaba said in the response to the 2021 results. 

It wasn’t true of course, as ever with Mashaba. By way of interest, here is a list of prominent new parties, their age as of their first election, and their first result:

ASA’s was not a bad performance as new parties go – more or less on par with the ID and UDM, and a bit behind the NFP and the FF, although nothing compared to COPE or the EFF. The party did comprehensibly beat Agang’s 2014 performance though, which was something.

But ASA’s spin allowed it to bypass even comparisons with COPE and the EFF. By arguing it had competed in just six municipalities, and emphasising its JHB performance (which was legitimately impressive), it could create the illusion there was a great silent up-swelling of ASA support out there – a silent majority, that had never been formally tested.

Johannesburg alone, where ASA did best, accounted for 54.7% of ASA’s total 2021 vote share. Johannesburg and Tshwane together, 75.1%, and all three Gauteng metros – Johannesburg, Tshwane and Ekurhuleni – 91.2%.

The three KwaZulu-Natal municipalities, including a metro, accounted for just 8.2%.

The further one got away from Johannesburg, where Mashaba’s political reputation was built (as mayor), the worse ASA did. But the “6th biggest party” tag allowed it to mask that fact, and create the perception that ASA true support was powerful, broad and hidden. 

No one in the media or the political commentariat ever seriously examined the disjuncture between ASA’s performance in Gauteng and KwaZulu-Natal. But it is telling.

In truth, it was probably an error on ASA’s part to run in those KZN municipalities (Mashaba says it was at the insistence of Makhosi Khoza), because it revealed that Mashaba himself was primarily a Gauteng phenomenon. If ASA had just run in Gauteng, it would have secured 279,256 votes, enough to still comfortably make it the 6th biggest party and bolster the illusion it sought to create to even greater degree.

(6th place was actually relatively easy to achieve. 7th place in 2021 was the PA, with just 0,98% or 114,736 votes. ASA could have secured 6th place if it had only stood in JHB.)

The misleading power of perception can be illustrated another way. Let us assume the Democratic Alliance only stood in its best six municipalities in 2021 (including its four best metros). How would it have done? It is not a strictly fair comparison, but fair enough to make the point. Those six would have been:

WC032 – Overstrand: 63.66 % (21,887 votes)

WC043 – Mossel Bay: 66.15 % (23,340 votes)

CPT – City of Cape Town 58.85 % (537,284 votes)

JHB – City of Johannesburg: 25.45 % (235,120 votes)

TSH – City of Tshwane: 31.77 % (213,852 votes)

EKU – Ekurhuleni: 29.07 % (190,305 votes)

Combined, the DA would have received 1,221,788 votes (still enough for 2nd place if you include its ward performance). Its combined average for those six municipalities would have been 38%. But one’s best results are never your total results. The DA ended on 21.6%, because it did worse outside its six best, obviously. It is unlikely the DA would ever have got away with such a trick, one has to be a newcomer, which tends comes with a general exemption from critcism or media scrutiny.

The “6th biggest party” narrative also allowed ASA to have its cake and eat it. Here is the ASA party chairman at the IEC Results Centre, straight after the 2021 election results:

“Action SA has emerged as the only political party, that takes seats and takes support, in townships, in informal settlements, in suburbs and in the CBDs of different municipalities. South Africa has become used to political parties that are polarised, that have support in one area, and nothing in another. Action SA has broken that mould. We are taking support for all communities and all political parties.”

For a party that contested 3% (6 of 278) of all municipalities, in just two of nine provinces, that is quite a claim. Also nonsense. But ASA could use its spin seamlessly in whatever direction it wished: to reveal a grand national truth, or to conceal a national truth. ASA did get a broad sweep of support in Johannesburg but what the KwaZulu-Natal municipalities showed was that ASA hadn’t so much broken the mould, as replicated it. No polls shows any significant ASA breakthrough outside of Gauteng in general or Johannesburg in particular.

Thus, almost always accompanying the “6th biggest party” narrative was an insistence by Mashaba that what little public market research there was prior to the 2021 election had been grossly misleading and manipulated. To this day, Mashaba regards market research, which rarely shows ASA above 4%, as a “joke”. He insists that polls are biased and designed deliberately to under-represent ASA, and artificially boost the support of ASA’s opponents.

And so it is that market research surveys – the one thing able to cut through ASA’s spin – have become enemy number one for the party.

2. Rubbishing market research

It is hard to understand what provoked this hostility. There was no defining moment. It just seem to have emerged at some point. And grown from there.

ASA’s insistence that the 2021 polls were wrong and unfair, is perhaps best captured in this 2023 article by Mashaba, titled “Opinion polling has underestimated new political parties”. In it, he argued the following:

“In the days leading up to the 2021 municipal government elections, opinion polls after opinion polls suggested that ActionSA would struggle electorally. An Ipsos Poll on 7 September 2021 said we were likely going to attract 1% of the popular support. At the end of September 2021, the Centre for Risk Analysis proposed that our support lay slightly below 1%. But, on election day on 1 November 2021, we proved the opinion polls wrong – achieving 2.36% of national popular support despite only contesting in six out of 278 municipalities. In Gauteng alone, ActionSA achieved over 10% of the entire Province after only contesting 3 out of the 9 Municipalities.”

He said of this Machiavellian plot, “Numbers have been adjusted or displayed in such a way that it amounts to political engineering in a bid to achieve a certain outcome.” No evidence for this claim was ever provided (or ever has been).

On face value it is a bizarre argument. At worst the polls were out by about 1% when it came to ASA – an incredibly accurate result, especially as the polls he cites were in field two months before the election.

Just to put the facts on the table, there were only five polls, spanning the two and half months before the 2021 election: An Ipsos poll, one by the IRR, one by News24 and two by Ipsos/ENCA. They found the following:

All five had ASA between 1% and 2.5%. And they tended towards 2.5% as the election got closer. Certainly in the right area, and taking into account the margins of error, a very, very good base read for a small party (which are notoriously hard to pin down because they fall within the margin of error). On its 45% turnout model, News24 had ASA on 3%, 17 days out from the election. So actually over-estimated ASA’s support (by a tiny margin).

And if anyone had the right to take issue with some of those numbers, it was the ANC, which Ipsos/ENCA had 16pts too low five days out from the election; the DA, which Ipsos/ENCA had 5pts too low in the same poll; and the IFP, which Ipsos/ENCA had 4.6pts too low.

On ASA though, Ipsos was out by no more than a remarkable 0.1%, just five days from the election. 

Facts mean very little to Mashaba though. Especially when he has a grievance or conspiracy to drive. ASA also doesn’t know the first thing about market research. And that is perfectly evidenced by two of its own “polls”, released before the 2021 election.

The first, on JHB, was released on 30 September. In it, ASA said its polling put it on 31.7% – ahead of both the ANC and the DA in the metro (with a 4% margin of error). In the end, ASA ended up with 16% in JHB, in third place. Its own poll was 15pts out on its own support.

The second one, on Tshwane, was released on 26 October. In it ASA said its polling put it on 25.1%, basically neck-and-neck with the ANC and DA (also with a claimed 4% margin of error). In the end, it ended up on 8.6% in the Tshwane, in a distant fourth place. Again, its own poll was 17pts out on its own support.

In summary: 2021 independent professional polls on ASA: out by 1pt at worst; ASA’s own amateur polls: out by 15pts and 17pts at best.

Only Mashaba, could transform these facts into a narrative that the independent polls were out to get ASA. He makes Donald Trump look like Fred Rogers. One is tempted to revert to hyperbole, in trying to capture the level of ignorance and deviousness involved in this kind of sophistry. Let’s just go with stupid and dishonest. 

But effective, and that’s all Mashaba cares about. And, again, not a single political commentator has ever challenged Mashaba on this; rather, they provide platform after platform for him to spread his disinformation. BizNews is notorious for this. Again and again it has Mashaba on, and gives him free space to trash polling, without an ounce of pushback. In its defence, the BizNews style is not to interrogate, rather to let those people on talk on their own terms. But the price you pay for that is people like Mashaba can take ridiculous advantage of you, and make fools out of the public, when they aren’t being misled. So, it is hard to describe that as a public service.

3. Why the 2021 polls matter

There is another fundamental distinction that one needs to make and understand about the 2021 polls: while ASA was standing in just six municipalities, the 2021 polls were measuring its national support. Two very different things.

The truth is, the fact that ASA’s actual result (2,6%) and the independent polls (1% – 2,5%) came close, is as much a co-incidence as it is understandable. The most likely explanation is that ASA’s sample of six municipalities was big enough to accurately reflect ASA’s general support – similar, ironically, to how a poll works.

If you really want to be precise, we will never actually know how accurate the 2021 polls were, as ASA never stood nationally. And although there has been much evidence in support of their accuracy since, ASA’s own fantasy narrative has proven far more appealing.

There is a case to be made that, so successful was ASA’s spin on the 2021 elections, they have now fooled even themselves. It seems to believe that the untested ASA “silent majority” really does exist. Certainly, in the run-up to the 2024 election, there have been some truly wild predictions from ASA.

Mashaba has predicted ASA will emerge the second biggest party in SA, even the biggest (“ActionSA is not only going to emerge as the second-biggest party [in] 2024, it is going to emerge as the biggest party in South Africa.”). He has predicted 10% in KwaZulu-Natal. And ASA’s national party chairman says there is every reason to believe the party is on track for double-digits.

To date, there is no poll that has found this likely, which only seem to make ASA ever more paranoid and hysterical. True, the election is still eight weeks away, and perhaps ASA can make a signifciant late surge. That would be unprecedented, however. What the polls do confirm, is those 2021 numbers were indeed reflective of ASA national baseline support, and it seems stuck in a 1%-4% bracket.

Here is what all the public polling, from 2021 to 2024, has found for ASA nationally, in terms of its base support (that is, without turnout modelling).

Ipsos (August 2021): 1.5%

South African Institute of Race Relations (September 2021): 1.1%

Ipsos/ENCA (October 2021): 1.8%

News24 (October 2021): 1.8%

Ipsos/ENCA (October 2021): 2.5%

Local Government Elections: 2.6%

Social Research Foundation (June 2022): 4.7%

Ipsos (July 2022): 3.0%

South African Institute of Race Relations (October 2022): 3.7%

Social Research Foundation (March 2023): 1.0%

Ipsos (July 2023): 4.0%

Brenthurst (September 2023): 2.7%

Social Research Foundation (October 2023): 1.0%

South African Institute of Race Relations (October 2023): 0.3%

Ipsos (November 2023): 4.3%

Change Starts Now (December 2023): Unknown nationally but 3% in Gauteng

ENCA (January 2024): 1.4%

Brenthurst (February 2024): 2.4%

Social Research Foundation (March 2024): 2.1%

There is a margin of error on all of these, which ranges from around 2% to 6%, and you can map them all onto a graph, with the relevant trend-line, as below.

Now, it might be the case that all these international polling companies, media companies and NGOs have secretly and consistently met over a three year period, and conspired to risk their local and global reputations by manipulating their polling and committing public fraud, specifically to damage a small regional party in South Africa. Or, alternatively, it might be the case that ASA’s national base support lies between 1% and 4% and has been in that range for three years now. I will let you, the reader, decide.

But that is effectively the choice Mashaba presents you with. He assumes the public are fools, and relies on the media to facilitate his deception. In the end, he wants to look anywhere but at the data itself.

Why ASA is in such a state about market research is difficult to understand. There is a credible case to be argued that ASA is closer to 4% than 2%, especially on lower turnout, which a number of polls show it is capable of benefitting from. There is even a prospect, if everything goes right on the day, it could get to 5% or 6% as an outlier scenario. You can make that case, using actual evidence. No need to indulge fantasies or invent enemies. And 4%, or 5% would be a very good result for a small party. More or less double its 2021 performance. 

But the party seems intent on believing it is set to overtake all comers, and is primed for cataclysmic growth. And it wants everyone else to believe this too. To this end, the PR con it developed in 2021 has grown into a monster, and seems to have devoured even ASA’s in-house intellectual capacity to distinguish fact from fiction. That is what tends to happen when you are led by a demagogue: ego drives analysis, not facts.

The ASA monster now feeds exclusively on its own bullshit, and it produces as much of it as it consumes.

But what does any of that matter, when you are given an endless number of stages to perform on by the media, in an environment where most in the press don’t understand electoral numbers or market research in any substantive way, certainly not enough to critically interrogate the party’s claims or hold them accountable for their deceit. 

So, in the final analysis, one must doff their cap to ASA. Truly, a great rhetorical con. One of the all time greats. As for the monster, well, its diet of deceit is going to run out on 29 May, after which some new creature will inevitably be born and introduced to its first meal. The press will undoubtably welcome and coddle it with open arms.

(Disclaimer: I am the CEO of Victory Research, which conducted some of the polls in this analysis)

This essay is the 14th in an on-going series on Election 2024, for all other editions of this series, please click here: