Since former president Jacob Zuma’s announcement last October of his involvement in the uMkhonto Wesizwe Party (MK party), there has been a lot of noise about the party’s impact on KwaZulu-Natal and national politics.
This seemed to reach fever pitch recently, when a poll was released by the Social Research Foundation (SRF), indicating that the MK party could get as much as 24% in the KZN provincial election on 29 May. According to the same poll, ANC support was down to 25%, with the DA% at 15%, the EFF at 5% and the IFP at 24%
“The ANC’s house is on fire; it is finished. They are in shock,” one analyst warned. “This indicates a total game change in South African politics,” another said, pointing to this poll as proof that the MK party had “halved ANC support in the province in the last four months”.
Although there can be little doubt that Zuma’s involvement with the MK party will draw a significant number of votes in KZN, some of these predictions and analyses are problematic.
It is important to note that the SRF poll sampled 820 respondents telephonically. In a province with more than 5.7 million voters that is an extremely small poll. Telephone polls are generally less reliable than in-person polling, thus the self-declared margin of error was 5%. (This means that you could add or subtract 5% to any result.)
However, let’s assume that this poll is a true reflection of current voter support in the province. Is it then correct that the MK party halved the ANC support in just four months?
Well, yes and no.
Last year Ipsos released a huge, in-person poll. The fieldwork was done in June and July 2023 – months before the MK party registered at the IEC and Zuma’s public involvement with the party. The Ipsos poll indicated that the ANC was already then sitting at 22%, the DA at 13%, the EFF also at 13% and the IFP at 17%.
So, the ANC support was in fact the same – or even a few percentage points lower in June last year than the SRF poll two weeks ago. Significantly, the Ipsos poll indicated that almost 20% of respondents said that either they didn’t know who to vote for or didn’t want to answer. A further 8% said that they would not vote.
Disillusioned ANC voters
What we can assume is that many of the “unsure or won’t answer” from the middle of last year were disillusioned ANC voters who have now found a home in the MK party. The point being that the ANC has known for more than a year that it is in big trouble in KZN, and Zuma’s “Lazarus” performance has only added to their trouble – it is not the cause of it.
Perhaps because of the ANC’s bad performance, the EFF has also shown steady growth since the 2019 elections when it received 4% support. According to Ipsos, it was sitting at 13% in the middle of last year. If the recent SRF poll is correct, it would indicate significant losses for the EFF, almost certainly to the MK party. This would make sense since many disillusioned ANC voters would have gravitated to the EFF and – given their similar ideologies – are now happy to support the MK party.
So, what are the implications of all of this for the province? There is no doubt that neither the ANC nor the IFP would be able to get an outright majority; therefore, there will have to be a coalition arrangement. At this stage, an ANC-IFP coalition could just about reach the 50% mark. An IFP/DA or IFP/Multiparty Charter coalition will fall short of 50% as would an EFF/MK party or ANC/EFF coalition.
Personally, I believe that an ANC/IFP/DA coalition would be ideal, not only for unity in that province, but also for the desperately needed improved service delivery.
Although most of the focus has been on KZN, it is worthwhile remembering that there are also other provinces with significant numbers of Zulu-speaking voters who could be attracted to Zuma’s MK party. Just over 19% of the people of Gauteng and 24.5% of Mpumalangans are isiZulu speakers.
If the MK party indeed succeeds in getting 24% of the vote in KZN, that would result in about 5% or 20 seats in the National Assembly – about half of what the EFF has currently. However, if the MK party can also garner some support in Gauteng and Mpumalanga, it could take another few percentage points off the ANC support nationally, which might just see the need for an ANC/IFP/DA coalition at the national level. DM
Methinks you’ve been in Ireland too long,
If MK makes an impact on the IFP vote does that make a DA-IFP coalition impossible? Maybe that’s one of the MK desires…or unintended outcomes
A very small poll of barely 800
As Ramaphosa is likely to use the ANC’s dip below 50% nationally as an excuse to resign and return to his farm, there is no doubt that Mashatile will welcome MK with open arms as a coalition partner – birds of a feather, etc etc. Which will leave Julius and the Charter fuming with impotence as RSA slides even further into …
Different points of view on politics always welcome, but be respectful and not personal. “Dearie” is not appropriate nor is targeting someone’s name.
Methinks that the DA is becoming ever increasingly irrelevant and will no longer be the official opposition after the May elections! Very sad but the reality, where are the current leadership??
There is a leadership. But the wrong one. They just don’t seem to understand the electorate and how to appeal to it. They act as though the whole electorate is non African middle class.
“What we can assume is that many of the “unsure or won’t answer” from the middle of last year were disillusioned ANC voters who have now found a home in the MK party.”
No, no, just no. You can’t assume that at all.
Election polls/voter opinion surveys typically have different sets of results that they present. Firstly there is the raw data — the unadjusted responses in the survey, which typically includes refusals to answer, “don’t know” answers, and declarations of plans not to vote/not having registered to vote. In some polls, the latter group might be screened out before even participating. Then there is the reworked/modeled outcome, which typically involves eliminating likely non-voters, and allocating undecided voters to their most likely preference. If party support totals up to 100%, then the dataset has been reworked. At that point, it is a different set of results entirely, and you can’t simply compare the one set of data to the other — it is apples and oranges.
100% of registered voters in KZN didn’t suddenly stop being coy about their vote preference and/or lose all indecision between Jul & Feb. But here we have yet another pundit talking through their hat while comparing one poll’s modeled outcome to the raw data of another poll.
Since IPSOS did not rework their poll results on a provincial basis (only nationally), we are only left with the option of comparing raw data to raw data. But even that is problematic without more info on SRF’s methodology.
in 2017 Melanie Werwoerd was one of those commentators who said Cyril had virtually no chance of winning the election, and NDZ would win – so be careful to take everything she says seriously.