Vito Laterza and Stephan van Wyk | 23 August 2016
Vito Laterza is an anthropologist and political analyst. He is a postdoctoral research fellow at the Department of Social Anthropology, University of Oslo, and a research associate of the Centre for African Studies, University of Cape Town.
Stephan van Wyk is an anthropologist focusing on social and economic changes in Tshwane. He is a lecturer in the Department of Anthropology and Archaeology, University of South Africa.
The results of the municipal elections held on 3rd August have sent shockwaves throughout South Africa. The symbolic impact of the victories of the Democratic Alliance (DA) in Tshwane and Nelson Mandela Bay, as well as the African National Congress (ANC) dipping below 50% of poll support in Johannesburg, has been significant. Thanks to coalition deals, the ANC will lose control of up to four metropolitan councils.
But how bad was the hit for the ANC? And what can we say about the DA's growth, and the performance of the Economic Freedom Fighters (EFF)?
A look at the hard data across the provinces, metropolitan areas, and the key battleground of Rustenburg - will give us a better idea of some of the trends.
Nationwide, the ANC remains the party in poll position, with 176 councils in its control. This is a reduction of 22 Councils, from the 198 that it controlled after the previous municipal elections in 2011 (for a comparison of councils and seats of the main parties, see Spreadsheet 1).
The ANC has an absolute majority in 161 Councils. The DA leads in 24 councils, up from 18 in 2011. They control 19 of them. The EFF did not win any council outright, and the Inkatha Freedom Party (IFP), moved from 5 councils in 2011 to 7 in this round, all in KwaZulu-Natal. The Independent Civic Organisation of South Africa from the Western Cape, was the only other contender to win a council.
Overall, the ANC won 5,163 council seats, a net loss of 468 from 2011. The DA has 1,776 councillors (+221 from 2011). In their first appearance in municipal elections, the EFF came third nationally with 761 seats. The IFP won 432 seats (+80 from 2011). Other important parties are the Freedom Front Plus (67 seats), the United Democratic Movement (59), the African Independent Congress (55), and the now miniscule and almost marginal, Congress of the People, or COPE (45 seats, down 191 from 2011). Each of these minor parties gained less than 1% of the national vote share.
An analysis of the shares of valid votes and a comparison with the 2011 figures will give us an idea of how the three main parties performed across the provinces and in the metros (see Spreadsheet 2).
This has been the worst performance of the ANC in both municipal and general elections. Under Zuma's watch, the party finished close to the 50% mark: 53.9% of the national vote, a loss of 8 percentage points since 2011. The DA gained 3 percentage points since the previous local election, and received 26.9% nationally. The EFF closed at 8.2%.
There are significant variations across the country. The ANC losses are substantial in Gauteng: nearly 14 points across the province, and between 13 and 14 in Johannesburg, Tshwane and Ekurhuleni - in all these areas the ANC fell below 50%. They lost more in the North West (-14.6 points since 2011), with a massive drop in Rustenburg (-23.4 points). In the Eastern Cape the loss was higher in the two metros, Nelson Mandela Bay and Buffalo City, where the ruling party lost around 11 points. At the provincial level the decline was smaller (-6.1 points). While the ANC share of votes declined in the urban areas, there are signs that their rural support dropped as well. In Limpopo, home to EFF leader Julius Malema, the ANC lost nearly 12 points. In the Free State there was a decline of almost 10 points, with the Mangaung metro following a similar pattern. There were more contained losses in Mpumalanga (-7.3) and Northern Cape (-4.8). The ruling party continued to lose ground in the DA strongholds of the Western Cape (-7.4) and Cape Town (-8.4). The only place where the ANC gained was KwaZulu-Natal, with a rise of less than one point, but the eThekwini metro saw a drop of around 5 percentage points.
The DA gains in shares of valid votes were lower than the ANC losses. The DA stronghold of the Western Cape saw the most significant rise, with the DA breaking through the 60% ceiling. Over 5 points were gained since 2011 across the province and in Cape Town. Gauteng saw a significant increase of nearly 4 points, with the highest rise in Tshwane (+4.5), where the party rose above the 40% mark. The Eastern Cape also saw a significant rise (+3.5), led by an increase of 6.6 points in Nelson Mandela Bay. The DA made inroads in Jacob Zuma's home province, gaining around 6 points in the eThekwini metro since 2011, and more than 3 points in KwaZulu-Natal as a whole. Mangaung was the only metro where the DA experienced a decline (-1.1), with a minimal gain of less than half point in the Free State. In the largely rural Northern Cape the rise was close to the national average (+2.8), but in Limpopo the DA increased their share of votes by only 1.6 points. In Mpumalanga they lost a point, as in the North West, with a major loss of around 4 points in Rustenburg - in this case the rise of the EFF accounts for losses by both the ANC and the DA. Overall, it seems the DA is advancing more rapidly in the metros than in rural areas.
The EFF ran for the first time in the 2014 general election. We can only compare the shares of valid votes with the 2014 results (see Spreadsheet 3). With 8.2% of the national vote, their share increased by nearly 2 points since 2014. This is a good result for a new party, but betrays widespread expectations of breaking through the 10% ceiling. The strongest areas for the EFF are in Gauteng and its metros (around 11%), the North West (15.5%) and Limpopo (16.7%) - in the latter two, the EFF beat the DA to second place. Malema's home province saw the biggest rise since 2014 (+6), followed by Rustenburg (+5.4), where the EFF received a massive 26.8% of the votes, finishing second after the ANC, and more than 10 points above the DA. The EFF is strong in the Free State (9.7%), but also in Mpumalanga (9.4%) and the Northern Cape (8.6%) - in the last two provinces, the party experienced significant growth, over 3 points since 2014. The EFF struggles to take off in the Eastern Cape, where they received 5.2%, but had a significant rise in Buffalo City from 5.5% in 2014 to 8% in 2016. They are also weak in KwaZulu-Natal - they got 3.5% this year, but nearly doubled their share since 2014 - and in Western Cape, where they performed slightly below the 3% mark.
The EFF's performance as the third party has given them the role of kingmakers in four metropolitan councils: Johannesburg, Tshwane, Ekurhuleni and Nelson Mandela Bay. The EFF announced their vote for the DA mayors in these metros, which would allow the DA to govern, and push the ANC out - even in Johannesburg and Ekurhuleni, where the governing ANC finished first, but short of an absolute majority. The EFF has however refused to participate in formal coalitions with the DA. The DA entered into formal coalitions with smaller parties like the United Democratic Movement, COPE, the Freedom Front Plus and the African Christian Democratic Party and what seems like a loose working arrangement with the EFF in order to get DA Representatives elected into key Municipal portfolios and positions. The EFF insists that this is not a firm coalition with the DA - but a temporary situational relationship. The product of this relationship in Nelson Mandela Bay, Tshwane and Johannesburg, is that the DA with the help of the EFF, in sometimes quite dramatic and troubling circumstances, all elected DA Mayors. The Ekurhuleni Metro is likely to follow suit if the ANC is unable to cobble together a coalition or working arrangement, with smaller parties to keep a DA - EFF onslaught at bay.
A look at the percentage changes in the share of votes among the three parties gives us an idea of the potential movements between them (columns F, G and H in Spreadsheet 2). We compared the ANC decline with the sum of the DA growth since 2011 plus the EFF vote. In most cases, the two figures are close, with a difference of no more than 2 points in most provinces and metros. This is not the case in KwaZulu-Natal, where there is a bigger gap (7.5). Here the National Freedom Party (NFP), a splinter of the IFP, could not contest this election. They got 10.4% in 2011 in the province. The DA and the EFF are likely to have received some of their votes, and the IFP benefited too - as it increased its share from 15.8% in 2011 to 18.4% this year. In eThekwini, the DA growth in some areas is a pretty close match of the collapse of the Minority Front (MF), from 5.3% in 2011 to 0.5% on 3rd August. In Limpopo and Mpumalanga, COPE's collapse accounts at least in part, for the gap between the ANC decline and the sum of the DA growth plus the EFF.
There remains the question however, whether there has been a significant shift of voters from the ANC to the DA, or put it differently, whether the DA is expanding their voter base beyond the traditional demographics of a strong core of white voters, and growing support among coloured voters, especially in the Western Cape. Is it possible, as it has been suggested, that many ANC voters stayed at home while DA supporters came out in numbers?
Looking at percentage shares of valid votes does not help, because a low turnout in traditional ANC areas and high turnout in DA areas could still explain the DA gains, and part of the decline in ANC support.
At the national level, voter turnout calculated as a share of registered voters has not changed much, it was 57.6% in 2011 and rose to 58% on 3rd August (see Spreadsheet 4). There are some variations across the country, but not significant ones. Gauteng had a higher rise (+2.2), with the biggest increase in Tshwane (+3.9). But Nelson Mandela Bay, where the DA won, had a lower turnout than in 2011 (-0.7). A substantial increase in Rustenburg (+5) might explain part of the phenomenal result of the EFF there. Overall, the voter turnout percentages do not suggest significant changes since 2011. However, changes in voter turnout vary considerably across different wards in the contested metros. A more detailed statistical analysis would be required to have a precise idea of where the DA rise came from.
When we look at the actual increase in number of voters, there were 11.9% more voters casting their ballots this year than in 2011. It mirrors the increase of 11.3% in voter registration over the same period. There is not a clear pattern linking the DA growth with higher numbers of actual voters. Gauteng is 4 points above the national average, with a peak in Tshwane that saw a 21.9% increase in actual voters since 2011. But KwaZulu-Natal rose at the same levels of Gauteng, and was the only place where the ANC increased their support. In Nelson Mandela Bay, the increase in actual voters was well below the national average (5.7%). Rustenburg stands out, with a massive increase of 36.1% since 2011, linked once again to the strong EFF result there.
In order to better grasp the significance of the DA rise and the ANC decline, we need to look at the changes in voters for the two parties between 2011 and 2016 (see Spreadsheet 5).
The ANC lost around 223,000 individual voters. The DA got 820,000 voters more than the previous local election. But these figures are not adjusted for the increase in the actual number of voters, and hence tells us little about the real extent of the ANC decline and the DA growth. We adjusted the voters' change of the two main parties for the difference in turnout. The revised figures tell us that the ANC lost over a million voters (-13.1% since 2011), while the DA gained nearly 400,000 (+12.2%).
The national rate of change hides some dramatic losses for the ruling party. In Gauteng around 1/5 of voters abandoned the party, with the biggest drop in Johannesburg (-21.2%). The Western Cape (-22.4%) and the North West (-22.1%) did worse. Among the metros, Cape Town and Nelson Mandela Bay experienced the biggest decline: around a quarter of voters left the ANC since 2011. There was a steep decrease in the Eastern Cape (-14.8%), where the party lost more than half of the votes outside the metros. In the Free State (-16.6%) and Limpopo (-16.1%), the ANC declines by more than the national average. The Northern Cape was in line with the national average (-13.5%). In Mpumalanga the drop was -9%. The only province where the party experienced an increase was KwaZulu-Natal (+4.6%). They lost some points in eThekwini (-6.1%), but much less than in other metros.
The DA gains in terms of actual voters are concentrated in Gauteng, Western Cape and Nelson Mandela Bay (about 2/3 of all new voters), but these numbers confirm that the party is making significant inroads beyond the big metros and the Western Cape. The overall growth in the Eastern Cape (+13.3) is well above the rise in Buffalo City (+9.5) and Nelson Mandela Bay (+8.9%). In KwaZulu-Natal, the DA brings home an impressive 31.9% increase since 2011. EThekwini (+31.1%) mirrors the provincial average, indicating strong growth across the province. Another significant increase above the national average was in Limpopo (+21.6). In Gauteng the growth was more contained (+15.3%), unsurprisingly so given that the DA was already getting more than 30% of the share of valid votes in 2011. It is lower in the Western Cape (+9.3%) and Cape Town (+9.5%), where there is only so much room for further expansion. The Northern Cape grows below the national trend (+5.9%). The picture is not all rosy though. The DA takes significant losses in some areas, which is something of concern for a party that aims to become a viable national alternative to the ANC. In the Free State, the DA experiences a -2% decline. The party also lost support in Mpumalanga (-5.7%). The biggest drop was in the North West (-8.8%).
It is harder with this kind of data to say exactly where the DA gains are coming from in terms of demographics. The idea that the whole of the DA increase is among their traditional support base, white South Africans, does not however, seem to hold. Short of a 'militarisation' of the white vote, it is highly unlikely that the growth can be explained wholly by reference to white support. Analysts generally agree that the DA is making advances with the coloured vote, which is not surprising given the long-standing alliances with coloured communities in the Western Cape. As we said before, in eThekwini the DA experienced a surge of votes in Indian areas.
What about black voters? Is the DA eating into a base that has been seen up to now as largely beyond its reach? We have some preliminary findings that shed some light on these questions.
We selected a few wards in townships with predominantly black residents in various metropolitan areas (see Spreadsheet 6). We focused on wards with boundaries that were an exact match with 2011. There has been substantial redrawing of boundaries in highly populated areas, and we ended up with a small and not representative sample of 16 wards (5 in Tshwane, 6 in Johannesburg, 2 in Nelson Mandela Bay and 3 in eThekwini). The results are striking. While DA support in black townships remains well below provincial and national averages, the growth in most of the wards is phenomenal, both in terms of share of valid votes, and increase in actual votes.
In Tshwane, 4 out of 5 wards have seen doubling, and in one case tripling, of the share of votes from 2011 (reaching 18.9% in Ward 21), and increases in actual votes (adjusted for the difference in turnout) between 70.6% and 159.6%. In Johannesburg, increases are less marked but still strong: 5 out of 6 wards have seen actual votes rising between 33.6% and 161.4%, with ward 47 in Soweto giving 11.8% of valid votes to the DA.
Similar trends hold in the selected wards in Nelson Mandela Bay, where the starting point in 2011 is much lower. In eThekwini we also noticed very high growth rates, although the starting point in 2011 is minuscule. It is probably too early to make conjectures about how far this trajectory could be sustained.
Equally striking is the ANC decline in areas that were traditionally loyal to the ruling party. In most wards sampled, the party experienced a bigger drop than the national average, losing around 30% of actual votes since 2011 in some areas in Johannesburg, Tshwane and eThekwini.
We do not know from this data the breakdown in terms of voters' age, income and education, as there are important differentiations within the wards themselves. It would be even more difficult without a more elaborate statistical analysis to see how the DA did in higher income areas with mixed racial demographics.
The data we compiled clearly show that the ANC decline cannot be explained away as a temporary hiccup, or a strategic withdrawal of their voters. The trends are too damning to entertain this hypothesis. The DA growth seems to have more legs than a narrow media focus on the victories in Tshwane and Nelson Mandela Bay suggests.
The EFF consolidated the good performance in 2014, but we will have to wait until the next general election to have a clearer picture of their long-term trajectory.
For provincial and metro comparisons of 2011, 2014 and 2016 trends, see Spreadsheet 7.
The ANC support still looks imposing, but South Africa does not have a first-past-the-post system. If the ANC falls below the 50% mark in the 2019 general election, it could be the start of a long period of political instability without a solid coalition emerging to govern the country. The coalition deals in four metros are already giving us a taste of it. For now, the proportional representation (PR) system has favoured the DA, which is set to elect their mayors with EFF backing. But things might backfire in the future, as the EFF keep their hands free, and use their council votes to sink DA proposals that are not aligned with their (EFF) interests.
With PR, the main opposition parties have already won, when the leading party loses the absolute majority. The DA is probably not going to beat the ANC to first place any time soon, and the EFF has got some way to go before they become a well-established mass party, but both parties are rapidly gaining more influence than their percentages suggest.