Most of South Africa's councils must now be assembled by bargaining. After the 4 November 2026 election, only 74 of 213 councils returned an outright single-party majority; 139 did not. The ANC, on 30.76%, leads 128 municipalities but holds an outright majority in 53, so in the rest of the councils it leads it needs partners.
How far did the ANC fall?
The Witness reported the declared national result on 8 November 2026. ANC 30.76% and 3,246 council seats, DA 24.64% and 1,695. MK came third on 9.76% and 785 seats, ahead of the EFF (6.22%) and ActionSA (5.67%). About 14.9 million people voted.
The ANC is still first. It no longer looks like a party that can expect to govern alone. Wikipedia's tally of the 2021 election, ward and proportional votes combined, gave the ANC 45.59% and the DA 21.62%. On that basis the ANC has lost about 15 points in five years and the DA has gained about three, though the bases of the two counts may differ slightly.
Some numbers are still labelled provisional. On 8 November the data site ESP showed ward-ballot shares of 30.6% for the ANC and 24.5% for the DA, with 23,694 of 23,696 voting districts counted. The differences from the declared totals are a fraction of a point.
Why are so many councils hung?
The Witness counted 74 councils with a single-party majority and 139 without. The ANC is the largest party in 128 municipalities but has an outright majority in 53. In the rest of the councils it leads, it cannot govern alone. The DA leads in 32 and controls 12 outright. MK and the IFP each lead in 19, with outright control in three and five respectively.
The system explains the spread. Councils mix ward seats with a proportional component, which The Witness puts at 40%. When the vote fragments, the councils fragment with it.
What do the rules require, and what do they leave out?
The clock is short. Daily Maverick reported in 2021 that the Municipal Structures Act allows 14 calendar days, not working days, to constitute a council. The first business is to elect a Speaker, who then takes over for the election of an executive mayor and deputy mayor.
Failure carries a penalty. The same report said a council that stays unconstituted “may ultimately be dissolved and an election rerun held”, and it cited Nquthu in 2017 as the precedent.
Daily Maverick's account sets a deadline and a penalty. It describes no rule on who may bargain with whom or what a pact must contain. Reform has been on the table. The Dullah Omar Institute summarised a 2024 bill that would bar the removal of a Speaker or mayor for two years after an election unless there was a serious breach, and would set a 1% threshold for winning a seat. That was a bill in June 2024. Its 2026 status is unconfirmed, so neither provision can be counted as law.
Which cities are in play?
Not all of them. Eyewitness News called Cape Town “all but a done deal” for the DA on 5 November, while noting that 13 of the Western Cape's 25 councils were hung. eNCA's provisional results page, fed by the IEC, showed the DA on 60.86% in Cape Town against 11.42% for the ANC.
Elsewhere the same page showed no such dominance. In Johannesburg the DA had 40.33%, the ANC 16.77% and ActionSA 15.46%. In Ekurhuleni the DA had 42.39%, the ANC 15.57% and ActionSA 15.04%. In Nelson Mandela Bay it was the DA on 41.25% and the ANC on 26.55%. eThekwini showed no data, and Buffalo City and Mangaung were not listed.
These are provisional vote shares, not seats. The DA leads in all three without a majority; who holds the mayoralty depends on bargaining.
What does Lijphart add?
Arend Lijphart's Patterns of Democracy, published by Yale University Press (second edition 2012), examines thirty-six democracies from 1945 to 2010 to ask what type of democracy works best. Our shorthand for his contrast is winner-takes-all government against government by negotiated power-sharing.
Our reading is that South Africa's councils now sit at both ends of that scale at once. Cape Town, on the early count, is a majoritarian outcome: one party, one government. The 139 hung councils point the other way, towards bargaining, because no party won a majority under a mixed ward and proportional system.
The gap is in what usually accompanies consensus. South Africa's local tier has the proportional voting without the written coalition rules, and what it has instead is a 14-day deadline and the threat of dissolution. Proportionality has been adopted. The rulebook for living with it has not.
What is the strongest objection?
Our reading of the best objection is that we are stretching Lijphart. Power-sharing, on this view, should be designed, not forced by arithmetic. A council that has to patch together a mayor before the clock runs out is not designed for anything. It is deadlocked, and the better description is instability, not consensus.
The objection has a second edge. Written rules can entrench a weak mayor: the 2024 bill's two-year bar on removal would do exactly that. A council that cannot dump its leader may be stable and badly governed.
We accept the first point: bargaining here is forced, not routine. The real disagreement is over which failure costs more. We judge a dissolved council, with no government at all, the costlier one. Those who prefer a free hand to remove mayors would weigh it the other way. Cape Town is a reminder that the majoritarian outcome is still available where one party wins by a wide margin.
Our position is that the next risk lies in the law, not the electorate. A hung council is what the voters returned. A dissolved council would be what the rules produced.
Other systems face the same question of how a fragmented vote becomes a government: see our briefing on New Zealand's level vote, and our piece on the US midterms and the case for gridlock.
Watch for the first Speaker elections in the large hung councils, and for any council that reaches the 14-day limit without a mayor. Nquthu in 2017 shows what that costs.
Where to read next
- The Witness on the declared national results and council-by-council control.
- Daily Maverick on the 14-day clock and the dissolution rule, from 2021.
- Dullah Omar Institute on the 2024 amendment bill.
- Yale University Press on Lijphart's Patterns of Democracy.