The 2024 elections will celebrate 30 years of democracy in South Africa.
Three surveys released over the course of the past seven days all point to a loss of its parliamentary majority by the ANC in the coming general elections, and to an increasing likelihood of a coalition government at national level.
The polls — by the Social Research Foundation (SRF), Ipsos and the Brenthurst Foundation/SABI Strategy Group — also all indicate that the governing party is on the brink of losing control of both Gauteng and KwaZulu-Natal to the opposition Multiparty Charter for South Africa.
All three polls were conducted among samples of registered voters across South Africa and are aimed at providing a snapshot of the political mood in the country, but critics have claimed an alignment with the opposition parties on the part of some of the agencies.
The research groups say their polling simply reflects the attitudes and responses of respondents and that the findings are aimed at informing the choices of voters, rather than influencing them.
The SRF recently published surveys that suggested support for the ANC would drop ahead of the 2024 national and general elections, pegging its potential share of the vote next year at 45%.
The SRF’s latest poll on KwaZulu-Natal predicted that the ANC is unlikely to retain its majority in the province, with or without a high turnout of voters. It suggested the ANC would drop to 40% in the province with a 66% voter turnout and receive 41% with a low turnout of 49%.
SRF associate Gabriel Makin said the agency’s focus on KwaZulu-Natal was motivated in large part by the recent string of by-elections that saw an Inkatha Freedom Party resurgence.
“We wanted to see; does that trend in by-elections reflect overall in the trends that we see in the province because obviously, when you look at by elections, you can only glean so much because they’re very localised,” Makin said.
He said the reason for the survey was not only election polling but also to test the overall climate of opinion in matters such as the Ingonyama Trust.
“Similarly, we see that there’s kind of a lack of public opinion polling on stuff like the importance of traditional leaders and the Zulu royal household in politics. So those more personal opinions are what really motivates us to conduct these polls,” he said.
The process of conducting polling includes phoning a demographically and geographically representative sample of voters.
“It has to be random so that we don’t have any bias slipping into the surveys,” Makin said, adding that the SRF team tried to reach a sample of respondents that would match the reality on the ground in terms of registered voters in the province.
“The surveys themselves can last anywhere between 25 to 40 minutes, depending on the speed at which people answer but we ensure that we get as many people to go through all of the polls as possible,” he said.
In February, Andrew Kerr, a professor at the University of Cape Town’s School of Economics, criticised the SRF in an opinion piece published by Business Day.
He said SRF chair Frans Cronje’s relationship with the Institute for Race Relations (IRR) as a former director called into question the polling group’s independence. The IRR has in the past been rumoured to have links to the Democratic Alliance.
“Based on its director and who reports the results in Business Day, the SRF does appear to be linked to the IRR. That is important information because some of the SRF surveys ask people about who they might vote for or their views on political coalitions, and the IRR might not be impartial about these matters,” Kerr wrote.
Makin said the SRF works to ensure that it doesn’t introduce bias into its reports.
“We only report on the data,” Makin said, also denying the SRF had any affiliation to the DA. “We’ve been called many things in my time at the SRF. DA supporters is just one. We’ve also been told that we are trying to protect the ANC because we wrote a piece that was in favour of the general Energy Action Plan and Minister [Gwede] Mantashe attempts to focus on the coal power plants.
“I think it’s probably a good indication of our political neutrality that it isn’t clear exactly, that after every poll, there is a different group that is angry at us and a different group that is happy with us. And to be completely forthright we are open to conducting meetings with any political party that wants to talk to us.”
Rural provinces prove difficult in conducting polls, Makin said, adding that the methodology used by SRF, and in part the language barrier, can be a problem.
“There is a certain degree of suspicion that people might have towards a random call coming from a person you don’t know, saying that they’re affiliated with an organisation that you’ve never heard of, and then asking you very intense questions about your political opinion,” he said.
Market-research company Ipsos’s knowledge director, Mari Harris, said polling was a mixture of art and science.
Its latest poll was released on 25 October, based on 3 600 interviews in June and July.
The Ipsos research puts support for the ANC at 39%, DA (16%) and EFF (15%) among the registered voters.
Ipsos’s Pulse of the People series conducts face-to-face interviews with respondents who are asked to mark their choices on an imitation ballot paper on the computer.
“This is done to simulate a real ballot paper and ensure a secret ballot,” Harris said.
Looking at a different analysis of these results, the Multiparty Charter for South Africa led by the DA looks to draw about 31% to 33% of the vote at this stage, the poll found.
Ipsos is the most widely regarded polling research organisation, operating in 96 countries and often used by political parties as a yardstick for elections.
Harris said although a lot of polling organisations in the United States were allied to political parties, South Africa’s polls were mostly independent.
“One has to be very, very careful when you look at these things. First of all, I would say, who is this organisation? Where are they affiliated? Who paid for this? It does happen that an organisation can pay and they sort of determine what is asked.”
She said it was important for polls to be designed to be unambiguous. Harris said people need to keep in mind the date of the poll because opinion can change rapidly.
She added that it was important to specify the sample size interviewed.
“Your sample size determines your sample error. Every sample has an error so you need to specify your sample error in your technical detail as well. Sample error is influenced by your sample size, the methodology used and your response from it.”
The Brenthurst Foundation survey made similar findings, polling the potential collective performance of the Multiparty Charter for South Africa parties in addition to the individual performance of them and the ANC.
The foundation was set up in 2004 by the Oppenheimer family, and has been active in the country — and on the continent — in the political space, and has been accused by its critics of pushing a neoliberal, pro-United States agenda in both. Its director, Greg Mills, is a former adviser to the Nato command in Afghanistan.
Brenthurst director Ray Hartley said the poll was a follow-up on one they had done a year ago and had incorporated reactions to the multiparty charter because it was a new development on the political scene whose impact needed to be assessed.
“I don’t think it was particularly driven by the moonshot coalition as such. We are trying to have some kind of consistent, external polls conducted from outside the country,” Hartley said.
The research had been conducted for the Brenthurst Foundation’s own research purposes, and to “have some kind of confidence about what is happening in our political sphere,” Hartley said.
Hartley said the multiparty charter initiative was “interesting” and increased political competition in South Africa. “We believe that greater political competition is needed in this country. It makes all the parties perform better if there is increased political competition, and the moonshot certainly does that. As to whether we endorse the moonshot, I don’t think that is really relevant. That’s really up to the voter.”
The Brenthurst Foundation had decided to work with SABI, which had conducted a 15-minute telephone interview with each respondent, because it was not working with any South African party during this election cycle.
Hartley said there was no involvement of former DA officials in the Brenthurst Foundation and SABI.