Analysing South Africa’s election earthquake

Source: News 24, Dawie Scholtz

There were two key drivers in the 2024 elections: Voter turnout and voting patterns across demographic groups. (Gallo Images/Sydney Seshibedi)

Dawie Schlotz reflects on two key drivers of the 2024 election results – voter turnout and demographics – and what impact each had on the final shock result which saw the ANC losing its majority for the first time in 30 years.

---

A deep assessment of election results requires thorough examination of two key drivers: The relative level of turnout across various demographic groups, which determines the overall make-up of the electorate; and the voting patterns across various demographic groups.

Before diving into the numbers, a high-level observation: In the early years of our democracy, turnout and voting patterns were strongly correlated with race. It was by far the strongest predictor of voting intention. Today, race is still strongly correlated with electoral choice, but home language has emerged as an increasingly influential additional variable to consider. The analysis below, therefore, follows this logic where relevant.

A final methodological note is that the demographic breakdowns below are created by aggregating voting districts with similar demographics where a single group is clearly in the majority. For example, in attempting to analyse the “white” electorate, we do so by aggregating all voting districts that have strong majorities of white voters. This does not, however, mean there are only white voters in those samples. So when we say “white”, we mean “majority white voting districts”, which also include many other voters. These demographic estimates should, therefore, be read as imperfect or indicative only.

• Turnout

Overall turnout was 58.5%, making it the lowest national and provincial election turnout on record. Our previous lowest, in 2019, was 66%. Taken together with our lowest turnout in a local government election (in 2021, at 46%), it’s clear that South Africa’s voter participation rate is steadily decreasing – an important and concerning topic for another day.

The overall level of turnout, however, broke down very differently between demographic groups.

This is a crucial set of variables that determines the makeup of the overall electorate. In 2019, white turnout outstripped black turnout by ~14%, creating a significant “differential turnout”, thereby inflating the relative size of the white electorate and assisting the parties that do particularly well among white voters (chiefly the DA and Freedom Front Plus).

This differential is a critical variable closely watched by election analysts. In 2024, the overall level of turnout decreased, but it disproportionately reduced in the black electorate, effectively increasing the differential to ~16%. This is the largest differential I have observed in a national election (I’ve been watching closely since 2014) and was significantly helpful to the DA; and also quite damaging to the ANC’s overall percentage.

Looking even deeper into the turnout data reveals a further problem for the ANC. When you split out the black electorate by province and assess the relative turnout levels of black voters by province, it’s clear that turnout was highest within the black electorate in the provinces where the ANC was struggling most. It’s also clear that the ANC base was not fully enthused about turning out, particularly in the Western Cape, Eastern Cape, Limpopo and North West.

The low level of ANC base turnout in the Eastern Cape and Limpopo was a big driver of the poor overall ANC result. The picture would have been fundamentally different if the ANC was able to inspire a 60%-plus turnout among black voters in Limpopo, Eastern Cape, North West and Free State.

• Voting patterns

For this aspect of the analysis, we assess each major group individually.

First, this is a reminder that these numbers are from “majority white” areas. A minority of black, coloured and Indian voters is growing in these voting districts as they steadily diversify over time.

The DA clearly recovered in this portion of the electorate by regaining Afrikaans voters from the FF Plus (slicing the FF Plus vote in half). The turnout pattern also suggests that some of the ANC and EFF voters in this election did not vote.

Finally, there was perhaps a minor “Ramaphoria” element in 2019 in some sections of the white electorate, which had clearly evaporated by 2024.

It is also useful to look at the white Afrikaans electorate in isolation since that shows the impact of the FF Plus’ reduction more starkly. Isolating just for majority white Afrikaans voting districts, the results were as follows:

Here, the DA recovery was most dramatic, effectively reclaiming ~10% of white Afrikaans voters from the FF Plus. This was an important driver of the DA’s overall increase in the election.

Here, the DA suffered significant losses compared with 2019, with those losses going to the Patriotic Alliance (PA).

The ANC also lost support to the PA.

The DA is likely relieved that the losses were limited to -12% since trends from by-elections before the election suggested they may have been higher.

There is one very interesting sub-pattern hidden in this data, which is that the coloured electorate in Cape Town (and a few other major towns in the Western Cape) is fundamentally different and stronger for the DA than the rest of the national coloured electorate.

It should be noted that there is significant fragmentation in this portion of the electorate, with many other smaller parties not shown here, but taking somewhat significant portions of the vote in certain areas. For example, the National Coloured Congress did quite well in the Cape Flats.

There was a significant erosion of ANC support in favour of the Umkhonto weSizwe (MK) Party in a large portion of voting districts in KwaZulu-Natal.

The DA also strengthened its position at the expense of the ANC.

A fascinating result! Quite a few important observations to be made here:

First, and this may be stating the obvious, the ANC’s collapse to 53% support among the national black electorate was unprecedented, perhaps even unthinkable, before 29 May. This was clearly driven by the 19% swing to MK Party.

Second, the IFP and DA figures are not a typo. They are exactly at the same level nationally as in 2019. In the case of the DA, they are down to the decimal (4.4%). There are important movements at provincial level, where the DA showed good growth in some areas, but which were evened out by losses in others.

Finally, ActionSA’s performance will be a massive disappointment to the party, securing only a quarter of the support the DA was able to muster.

It becomes especially interesting to observe the breakdown of the voting patterns by province.

Here, we see a significant variation and regionalisation in the voting patterns. The most important driver is the significant swing among Zulu speakers from the ANC to MK Party.

An additional point is that the ANC’s losses were strongest in urban areas, meaning more urbanised provinces such as Gauteng saw a drop in ANC support.

The DA’s support levels by province have been largely stable or have only seen minor movements, with the exception of the Free State, where the party achieved a historic double-digit vote among black voters – the first I have ever observed.

KwaZulu-Natal was a particular disappointment for the DA, where it dropped from 4% among black voters in 2019 to just over 2% in 2024 – most likely due to the MK Party surge.

One final note here on the MK Party is that while its surge was most concentrated among Zulu speakers, the party also made significant inroads in communities that are not Zulu-speaking (in Gauteng and Mpumalanga), as illustrated by the breakdown for Mpumalanga below.

The 14% that the MK Party achieved in the non-Zulu-speaking black electorate in Mpumalanga was unexpectedly high and devastating for the ANC’s result. A similar trend was observed in Gauteng.

Five points to take into account

The election outcome and its key drivers can be summarised into five key messages:

• Lower overall turnout (58%) and a very significant differential turnout, where 71% of registered white voters turned out to vote and only 55% of registered black voters turned out to vote, drove the DA’s percentage higher.
• A significant swing from the FF Plus to the DA among the white Afrikaans electorate, combined with a swing of former white, English-speaking “Ramaphorists” from the ANC to DA drove a recovery for the DA among white voters.
• A significant swing away from both the DA and the ANC in the coloured electorate towards the Patriotic Alliance occurred nationally, but was particularly strong outside of the urban Western Cape.
• There was an enormous swing from the ANC to the MK Party in the black electorate, concentrated in the Zulu-speaking electorate in KwaZulu-Natal, Gauteng and Mpumalanga, but also present to a lesser extent among non-Zulu-speakers in Gauteng and Mpumalanga.
• Varying swings in the black electorate from the ANC to all the other opposition parties across the other provinces, in the range of -2 to -10 in each province, were also observed.