By Nco Dube | 16 October 2026

The first warning came in August 2016. The ANC lost its outright majority in Johannesburg, Tshwane, Ekurhuleni and Nelson Mandela Bay. In three of those metros, opposition parties cobbled together arrangements that removed it from office. It proved that liberation history could no longer guarantee political power. There was excitement. But few asked the harder question: once the ANC had been removed, did its opponents know how to govern together?

Ten years later, we have a sobering answer. Herman Mashaba became Johannesburg’s mayor with support from parties that disagreed with the DA on fundamental issues. Solly Msimanga led a similarly dependent administration in Tshwane. Nelson Mandela Bay lurched through its own upheavals, while in Ekurhuleni the ANC retained power despite losing its majority. The lesson was clear. Losing a majority did not necessarily mean losing government, and securing a mayoral chain was no guarantee of being able to govern.

The coalition experiment loses its innocence

The 2021 elections pushed this uncertainty further into municipal life. The ANC dropped below half of the national municipal vote, while 66 councils were left without an outright majority, compared with 26 in 2016. eThekwini joined the growing list of metros where the ANC needed partners. Coalition government was no longer some temporary interruption of South Africa’s established political order. It had become part of that order.

A decade of coalitions also taught us that not all political instability looks the same. Some administrations collapsed over genuine policy disagreements. Others were undone by shifting support, battles over appointments or motions of no confidence. What residents experienced, however, was often remarkably similar: another new mayor promising a fresh start while municipal departments carried on with the same inherited problems. Political change at the top rarely came with the patient rebuilding of institutions beneath it.

Yet Johannesburg’s experience since 2021 suggests that political parties learnt remarkably little from the first five years. Mpho Phalatse was removed and reinstated. Dada Morero briefly took office. Thapelo Amad and Kabelo Gwamanda of Al Jama-ah followed, before Morero returned. A party with a tiny electoral footprint could occupy the mayoralty of the country’s economic centre because larger parties found the arrangement convenient. This was constitutionally permissible. Whether it served Johannesburg was another question.

And while politicians exchanged offices, residents endured water interruptions, damaged roads, billing disputes and failing maintenance. These problems did not begin with coalitions. ANC majorities had already presided over failing infrastructure and entrenched patronage. We should not rewrite that history to make coalition politics the villain of every municipal failure. But constant changes at the political top have hardly encouraged the sustained attention that repairing these institutions requires.

A weaker ANC does not mean a stronger DA

Now comes 4 November, with an electoral landscape more fragmented than either 2016 or 2021. Recent polls suggest that several metros could again produce councils without an outright majority. The News24/SABI survey placed the DA at 29% in Johannesburg, the ANC at 23%, MKP at 12%, the EFF at 11% and ActionSA at 7%. In Tshwane it measured the DA at 32% and the ANC at 24%. These are polls, not results, but the contest is no longer simply ANC versus DA.

The eNCA/MarkData survey released on 11 October recorded the DA at 37.7% and ANC at 37.4% in Nelson Mandela Bay, effectively level within the margin of error. In Ekurhuleni, it found the DA at 31.2%, ANC at 29.4%, MKP at 11.5% and EFF at 9.2%. Even if the DA improves its position in several metros, it does not follow that the ANC’s losses will deliver DA majorities. Voters are moving in different directions. Some shift to smaller parties. Others stay home. The EFF’s support is uneven across municipalities, but even a party losing ground can hold the few seats another needs to govern.

That is the peculiar arithmetic of coalitions: a party’s bargaining power can be much greater than its share of the vote.

MKP changes the political arithmetic

KwaZulu-Natal adds a complication absent from the previous two local elections. MKP secured 45.35% in the 2024 provincial election and emerged as the province’s largest party, yet an IFP-led coalition with the ANC, DA and NFP formed the government. Finishing first, we have already learnt, is not the same as governing. MKP now enters municipal contests previously shaped largely by the ANC, IFP, DA and smaller local formations. It does not simply take votes from established parties. It changes the possible combinations through which power may be assembled.

eThekwini is especially revealing. News24/SABI measured MKP at 43%, the DA at 30% and the ANC at 11%, although other polls have reported substantially different MKP figures. None of these numbers is November’s verdict. What they do indicate is an upheaval in the city the ANC once dominated. And whether MKP leads an administration, enters a coalition or sits in opposition, the real questions will not be about Jacob Zuma’s political comeback. They will be about broken water networks, sewage, municipal finances and the daily management of a major port city.

Smaller parties will have their own say in this new landscape, sometimes with only a handful of councillors. Some genuinely represent neglected local interests. Others benefit from the bargaining leverage created by divided councils. Neither their small size nor their strategic usefulness automatically establishes that their preferred candidate should run a metro. But larger parties cannot complain about opportunistic demands while freely trading positions and concessions to secure office. Every such bargain has willing participants.

The political parties also carry different burdens into November. The ANC must reckon with losing the dominance that once allowed it to govern without compromise. The DA must show that electoral growth can be translated into workable partnerships rather than merely opposition to the ANC. The EFF must decide how it uses the leverage of its council seats, while MKP faces the less glamorous task of proving that disruption can become administration. The smaller parties will need to demonstrate that their influence produces more than appointments. None can escape scrutiny by pointing to another party’s failures.

The politicians bargain. Residents pay.

There is a material dimension to this that our coalition debates too often sidestep. Municipalities control budgets, appointments, procurement and development decisions. When access to political office carries access to resources, the negotiations are not merely about competing programmes for service delivery. They are also about influence and opportunity. That does not mean every negotiator is corrupt. It does mean we would be foolish to ignore the incentives that make control of municipal government so fiercely contested.

The consequences fall unevenly. Wealthier households install water tanks and backup power. Poorer residents carry buckets, lose working hours and endure sewage outside their homes. A small business can lose a day’s trade while parties argue over who should chair which committee. Those bearing the heaviest costs are seldom the politicians responsible for the instability.

The proposed legislation regulating municipal coalitions will not be in force before the election. Written agreements and limits on repeated motions of no confidence may help, but rules cannot manufacture judgement or integrity. Stability must not protect a failing executive. A mayor who survives five years while a city deteriorates has achieved political survival, not good government.

Nor can coalition agreements substitute for functioning municipal institutions. Engineers, financial managers and administrators cannot plan properly if each political rearrangement brings pressure to reverse decisions or reshuffle leadership. We speak endlessly about which parties should govern together, but far less about how professional municipal administration can be protected from whoever wins the next council vote.

Parties entering coalition talks should be expected to publish their agreements, explain their compromises and identify measurable commitments on water, infrastructure, revenue collection and municipal administration. They need not pretend that compromise is dishonourable. In a fragmented electorate it is unavoidable. What deserves criticism is the habit of presenting a scramble for offices as a programme for governing.

Since 2016, South Africans have demonstrated that they can remove dominant parties from municipal power. On 4 November, they may again distribute that power among several parties. The greater test comes afterwards: whether the people entrusted with those votes can build administrations that last long enough, and work seriously enough, to repair what is broken.

A majority assembled in a council chamber is not, by itself, a plan to fix a city.