Low turnout, coalition politics hover over 2026 local elections

The IEC's voter registration station at the Elias Motsoaledi informal settlement in Soweto. According to the latest Ipsos polling research, a turnout of about 42% is predictable for the upcoming local elections. This will be a huge disappointment for the IEC, says the writer.

Determining the levels of support for the parties in an election in the period prior to the election has become a science on its own. Using a representative sample of the total eligible voter population and conducting interviews with them is, at the moment, the only mechanism available to predict the possible election results.

They are done by private companies who are often also involved in market research. In South Africa, two of the most active ones are Ipsos and the Social Research Foundation (SRF), which uses a sub-contractor for its polling exercises.

All the polling results must be read with the understanding that they are subject to a margin of error, as an indication of the level of uncertainty or how much space should be left for the polling results to deviate from the final election results.

Though this is generally called a “local government election” (LGE) in November 2026, it will in fact be 257 individual elections, for each of the municipal authorities in South Africa, including the eight metro and 44 district councils.

It should be taken into account that local elections consist of two components: voting for the ward candidates and voting for the parties’ list of proportional representation candidates. The polling results do not include such a distinction but present the combined total number of votes a party is expected to receive. 

When all of these 257 elections are compiled, the national and provincial totals can be determined, but they will have no impact on the formation of the local authorities. Interestingly, the predicted party tallies in the two polling exercises are very similar.

They can also be compared with the previous LGE in 2021, and although it is not an “apples-to-apples” comparison, also with the 2024 national and provincial election (NPE) results.

What is apparent from these statistics is that the EFF and MKP will most probably continue to be in the range that they have established in the previous elections – no major changes, from a national perspective, are foreseen for them. The same does not apply to the ANC and DA.

In the previous two elections (2021 and 2024), the ANC experienced a dramatic decline in its support. Both polls predict that this trend will continue to a level in the low 30’s%. In the case of the DA, its support levels in the two previous elections remained stable, but the polling suggests that it will now increase to the mid-20s%.

An important unknown element of this election will be the voter turnout, or the percentage of registered voters who have voted. Traditionally, the voter turnout in LGEs has been significantly lower than in the NPE – as much as 20% less. It is normally ascribed to the low voter turnout of young voters.

South Africa is known for a direct correlation between voter age and voter turnout: the older the voters become, the more they vote. In the 2021 LGE (during the COVID-19 pandemic), the voter turnout was 45.8%.

What does the polling say about 2026? The Ipsos poll indicates that 33% of the respondents said that they will “very likely” vote and 43% “likely”. On the basis of these responses and the previous trends, a turnout of about 42% is at the moment predictable. 

This will be a huge disappointment for the IEC, which popularised its online voter registration facility. Moreover, their national launch on 27 May 2026 of the election campaign was a major exhibition of targeting the youth with all the media means popular amongst them.

The provinces predicted to have the highest voter turnout levels are the Western Cape and Limpopo, and the metros would be Mangaung, Nelson Mandela Bay and Cape Town. (Johannesburg, despite the Helen Zille campaign, is not amongst them.)

The most problematic provinces are predicted to be the North West, Northern Cape and Free State, and the metros would be Buffalo City, Tshwane and eThekwini. 

Regarding the ANC, the provinces where the polling indicates its best performances are: Limpopo (58%), Eastern Cape (56%), Mpumalanga (42%), Gauteng (26%) and KZN (11%). The best metros for the ANC would be: Nelson Mandela Bay (37%), Mangaung (32%), Johannesburg (31%), Buffalo City (29%). 

In the case of the DA, their best provinces appear to be: Western Cape (47%), Free State (41%), Northern Cape (30%) and Gauteng (27%). The leading metros would be: Cape Town (46%), Nelson Mandela Bay (37%), Mangaung (36%) and Tshwane (31%).

The MKP, in its first LGE, is expected to do very well in KZN (50%), followed by Gauteng (10%) and Mpumalanga (8%) – but note that it is very similar to the 2024 results, except that Mpumalanga then stood at 18%. In terms of the metros, eThekwini stands out for the party at 55%, followed by Ekurhuleni (18%) and Johannesburg (6%). 

For the EFF, this will be a critically important election. In 2024, they lost significant support in KZN and Gauteng. Their best provinces are expected to be Mpumalanga (17%), Limpopo (17%), Eastern Cape (16%), Free State (15%) and Gauteng (10%). Their best metros would be: Buffalo City (27%), Mangaung (17%), Cape Town (11%) and Ekurhuleni (10%).   

It is noticeable that the expected results of this election, and the local coalitions that will be inevitable, do not necessarily have a direct bearing on the GNU and provincial coalitions. They might indeed differ substantially from the national situation.

The predicted results make coalition government at the local level inevitable. For that reason, the current amendment to the Local Government: Municipal Structures Act (“the Coalition Bill”) becomes even more important to be finalised before the election.

** The views expressed do not necessarily reflect the views of the National Media Group.