GARETH VAN ONSELEN: The state of the ANC

Recently, the Social Research Foundation (SRF) commissioned Victory Research to undertake an extensive market survey into the political landscape (disclaimer: I am the CEO of Victory Research). With the SRF’s kind permission, what follows is the first of three essays on some of those findings, regarding the ANC, DA, and the EFF. We start with the ANC.

First, the methodological parameters of the survey. It was conducted from June 27 to July 31 2022 via mobile phone, using a single-frame, random digit-dialling sampling design, from which a probability sample was drawn. The sample comprised 3,204 registered voters and was fully representative of the voting population (that is, it mirrors the voting population and in terms of all demographic indicators: age, race, education, rural/urban, gender, etc). The margin of error for the national sample was 1.7%, at a confidence level of 95% (that is, we are confident that 95% of the time the findings won’t vary more than 1.7 percentage points up or down from reality). Finally, we upsized three sub-samples: the provinces of Gauteng, KwaZulu-Natal and the Western Cape, to make sure we could get statistically meaningful reads off those provinces. They each have a sub-sample size of about 500, and a margin of error of about 4.4%.

ANC support levels

Below are the ANC’s current support levels. The raw numbers aren’t modelled for turnout. In other words, this would be the ANC’s support if 100% of all voters voted. In reality, that never happens. So we need a way of taking out those people least likely to vote. We do that by asking respondents a series of questions about how likely they are to vote, taking those least likely to vote out of the sample and seeing what it does to the numbers:

ANC national support (raw): 46.9%
ANC national support (using 2019 turnout levels, about 66%): 49%
ANC national support (56% turnout): 52%

For the record: this is not a prediction, these are current ANC support levels, 18 months out from an election.

There are a few things to note. First, these are significantly down from the actual 2019 election, when the ANC got 57.5%. Even accounting for the margin of error (1.7%), ANC support has significantly declined and the battle in 2024 will be for the ANC to secure 50%. Second, while the ANC vote share does improve marginally on lower turnout, one must be careful in reading too much into this. For one, our modelling does not account for differential turnout: in other words, if opposition voters go to the polls in greater numbers than ANC voters. This is a strong possibility. Opposition voters have a far stronger reason to vote (the possibility of bringing the ANC below 50%). ANC voters are generally depressed. If this happens, it could well drive the ANC below 50%, if the current pattern holds.

It is also worth noting that there is nothing predictive about the turnout levels we have chosen. They are just models, of a possibility. Actual turnout on election day is the great unknown. Here we have provided just one model at 2019 levels, about 66%, and one lower, at 56%. Generally, voter apathy is increasing in SA; where it finally lands on the day is impossible to say, but critical.

Trends are also important. There is no sign the ANC’s vote share is stabilising, certainly none that it is increasing. Given there are about 18 months to the national election, if the current trend continues, the ANC vote share is likely to decline further.

Big picture takeaway: the ANC is in serious decline and it is in a fight for its life to obtain 50% nationally. How turnout breaks in 2024 is going to be critical to whether it can achieve 50%. If the ANC cannot enthuse its voters, things look bleak for the party’s national support.

The big three provinces

Here are the ANC’s current support levels in Gauteng, KwaZulu-Natal and the Western Cape (margin of error: 4.4%):

Gauteng (raw): 29%
Gauteng (2019): 27%
Gauteng (56%): 23%
KwaZulu-Natal (raw): 45%
KwaZulu-Natal (2019): 47%
KwaZulu-Natal (56%): 44%
Western Cape (raw): 15%
Western Cape (2019): 16%
Western Cape (56%): 21%

A few observations. First, the survey did not differentiate between the national and provincial ballot. These are all indicative of national support by province. The option of a provincial ballot (and vote splitting) could affect these numbers slightly, but history suggests not by much.

Second, these numbers seems to confirm another general trend: the decline of ANC support hits the party hardest in the big urban provinces. In Gauteng in particular, even accounting for the margin of error, the ANC is way below 50%. Disturbingly for the party, it is on the 50% threshold (accounting for the margin of error) in KwaZulu-Natal. In the Western Cape the ANC is finished. The party is dead in that province. It cannot operate as an opposition party, and that province would seem to confirm that, when the ANC loses power categorically, it cannot recover. It is a one dimensional party that can only function with power; without it, the complex demands of a plural democracy are too much for it.

Big picture takeaway: The ANC’s collapse in urban centres and provinces is driving its decline. Its inability to govern effectively in key metros and its failure to operate as an opposition both act to erode its support. In the Western Cape it is dead, in Gauteng it is in serious trouble and in KwaZulu-Natal it is in another fight for its life, for 50%.

Composition of ANC support

99% of the ANC’s support base comprises black South Africans. For all intents and purposes, it remains a racially exclusive and homogeneous party, driven by an implicit and explicit racial agenda inside and outside government and thus, unable or unwilling to meaningfully appeal to any racial population group outside of black South Africans.

Despite the racial composition of opposition party support being a pathological obsession for many in the fourth estate, the racial homogeneity of the ANC’s support has for decades now evaded even the most marginal of criticism, public pressure or interrogation. Outside of race, the ANC’s support splits 43%/57% between urban and rural voters, respectively. 60% of ANC voters are unemployed. 14% are aged between 18 and 29, 66% between 30 and 59 and 20% are 60 or older. 80% of ANC voters have an average monthly household income of R4,999 or less, 86% have a grade 12 education or lower and, finally, 44% are female and 56% male.

Leader favourability:

Cyril Ramaphosa’s favourability is an absolutely critical variable for the ANC. For the ANC there are always two brands at odds with one another: the party brand (severally damaged and now wholly associated with corruption and incompetence), and Ramaphosa’s brand (built around promise, competence and the idea of renewal). Whichever brand is strongest on the day, has a huge impact on the ANC’s support levels. A leader’s favourability is typically bigger than that of the party. That makes sense: they are the flag-bearer for any party, ostensibly the best it has to offer and the embodiment of its offer in the best sense. Ramaphosa’s favourability carried the ANC in 2019. Many people voted for what he represented, as opposed to the ANC itself. But his favourability, like the ANC’s support, is in decline. Here are his current levels:

Somewhat and strongly favourable: 49%
Somewhat and strongly unfavourable: 30%

But compare that to his levels just before the 2019 election. In another poll by Victory Research, using the same methodology (of 1,800 registered voters), in February 2019, Ramaphosa faired as follows:

Somewhat and strongly favourable: 60%
Somewhat and strongly unfavourable: 18%

There has been a marked decline in voters who feel favourable towards Ramaphosa (down 11 points from 60% to 49%) and a marked increase in voters who feel unfavourable towards him (up 12 points from 18% to 30%). This represents a serious problem for the ANC: Ramaphosa’s personal brand is starting to align with the ANC’s support levels, and that means he is no longer able to win people over, outside the ANC’s traditional voting base.

Some of that can be explained in part by natural attrition, all leaders lose capital over time; in part by Ramaphosa’s failures (economic decline, broken promises, a failure to act decisively on corruption and endless service delivery failures; thus the idea of renewal has become increasingly unbelievable). But whatever the reason, it matters not. The ANC’s biggest drawcard is becoming ineffectual. If that continues, it becomes difficult to imagine what, if anything, the ANC will base its 2024 election offer on, and that will make 50% all the harder to achieve.

Its promise of renewal, as epitomised by Ramaphosa is becoming unbelievable. Two other questions in the survey reveal this. First, the SRF asked all voters whether they felt whether conditions in SA had generally improved since Ramaphosa was elected president in 2017, or if they had got worse? They responded as follows:

Better: 33%
Worse: 58%

That explains why his personal brand is declining, voters think he is failing. Second, there is, however, some hope for the ANC. While voters feel Ramaphosa is failing, many think he could still succeed. We asked whether voters believe, ultimately, he will succeed or fail:

Succeed: 49%
Fail: 37%

49% isn’t great (no doubt he would have scored much higher in February 2019) but it still represents a net positive score. So, there is still some hope he can endure. If he is able to deliver anything meaningful on this front, it will rekindle enthusiasm in him. That, however, looks unlikely.

A final note on Ramaphosa’s favourability. His reputation inside the ANC is also declining. Here is how his favourability among only ANC voters has fared: February 2019

Somewhat and strongly favourable: 75%
Somewhat and strongly unfavourable: 9%

July 2022

Somewhat and strongly favourable: 66%
Somewhat and strongly unfavourable: 18%

He is still incredibly well regarded inside the party, but again, the graph is headed downwards, and if it keeps dropping, that will bring problems of its own. That is particularly true in KwaZulu-Natal:

Somewhat and strongly favourable: 35%
Somewhat and strongly unfavourable: 55%

And it is further illustrated by KwaZulu-Natal respondents attitude to the question whether Ramaphosa will succeed or fail:

Succeed: 30%
Fail: 63%

KwaZulu-Natal is a real problem for the president. It seems to have lost all faith in the Ramaphosa offer, and other forces (ethnic nationalism?) are keeping the ANC together in that province.

Big picture takeaway: Ramaphosa’s failure to deliver any meaningful reforms, means his favourability is in decline. This is a profound problem for the ANC; its primary brand offer, of reform and renewal as embodied by Ramaphosa, is crumbling. If this continues, the ANC’s own brand, of corruption and incompetence, will become dominate again, and its ability to hold supporters or convince voters invested in hope, will decline as well.

The ANC’s brand

There are two components to the ANC’s brand: the party and the ANC as a government. Both are in catastrophically bad condition. The SRF asked respondents, regardless of which party they support, which of the big three parties — the ANC, DA, and the EFF — were best at the following core governance issues. Here are the percentage of voters who said the ANC was best:

Service delivery: 24.9% (DA best at 42.3%)
Clean governance: 17.5% (DA best at 33.6%)
Good policies: 20.9% (DA best at 35.4%)
Good leadership: 35.4% (DA second at 26.8%)
Accountability: 22.5% (DA best at 34.6%)

Those are a remarkable set of (dismal) numbers. The governing party is deemed by voters, on almost any issue involving honesty, competence or transparency, to be a distant second behind the official opposition. A small minority of people — way below the ANC’s actual support levels — think the ANC is best at governing, in terms of the basics. It illustrates the gulf between the actual ANC and the idea of ANC (as capable of renewal) and suggests many people support an imagined ideal, as opposed to the real-life performance of the organisation. This is further illustrated by the assessment of the national government’s performance. The SRF asked respondents how well the national government is doing on the following issues:

Accountability:

Very well or well: 30%
Not well or not well at all: 64%
Net score (if you subtract favourable from unfavourable): -34

Fighting corruption:

Very well or well: 16%
Not well or not well at all: 83% (This is a key indicator for Ramaphosa and the perception he is delivering on this promise.)
Net score: -67

Service delivery:

Very well or well: 33%
Not well or not well at all: 62%
Net score: -29

Keeping its promises:

Very well or well: 22%
Not well or not well at all: 74% (Another key Ramaphosa indicator.)
Net score: -52

Taking decisive action to fix problems:

Very well or well: 30%
Not well or not well at all: 65%
Net score: -36

The economy:

Very well or well: 27%
Not well or not well at all: 68%
Net score: -41

Fighting crime:

Very well or well: 18%
Not well or not well at all: 82% (The national government fairs worst on this indicator and puts the positive bluster of police minister Bheki Cele into its proper perspective.)
Net score: -64

Employing excellent people:

Very well or well: 19%
Not well or not well at all: 76%
Net score: -58

That is not a comprehensive list of issues, obviously, but enough to paint a picture, and about as damning an assessment of a national government performance as you will ever read. No-one is under any illusions that the national government is fundamentally failing on every key indicator. The goodwill factor among ANC voters is a modern miracle. This belief, in Ramaphosa and the ideal he represents, is all that is holding the ANC together. As that belief is eroded, the numbers above represent the stark reality that belies it. Without the dream, there is little more than despair all the way down.

National power is also a glue for the ANC. As the Western Cape reveals, without power or the realistic prospect of it, the ANC vote share collapses. Another question illustrates this. The SRF asked ANC voters, if the ANC drops below 50% nationally, if they will continue to support it?

Yes: 74%
No: 24%

74% is a healthy majority. But if the ANC does drop below 50%, the poll suggests one in four ANC voters would reconsider their support. If 24% of ANC voters did move away from the ANC should it lose national power, that is enough to break South African politics wide open. More research is required on this, but the finding does speak to the suggestion that a number of ANC voters are holding out for a miracle and, if it is not realised, they will abandon ship.

Big picture takeaway: The ANC’s brand as a party and a government is profoundly compromised. It is regarded as incapable and its outcomes as dire. That brand is now so well set, and so comprehensive it is negated only by wishful thinking, and even the dream is now cracking and fracturing. The indicator on broken promises (above) speaks to this.

Conclusion

ANC support levels are in marked decline. The ability of Ramaphosa and his promise of renewal to carry ANC support is weakening. This is most evident in provinces with big urban centres. Without the promise of renewal, the ANC’s underlying brand is in tatters. People regard its performance as staggeringly bad. The ANC’s continued decline in terms of hard support, suggests the gulf between hope and reality is shrinking every day. If the two ever meet, the ANC’s collapse is likely only to be exacerbated.

Next week: The state of the DA

Correction: September 8 2022

Under the sub-heading “The big three provinces”, the author has amended the ANC's support in Gauteng to: