This is a paid article which your subscription is allowing you to read.

Helen Zille, whom this newspaper has endorsed, has put up a heroic political fight to save Johannesburg. Yet her party is still polling at levels it first reached in the city a decade ago.

In the 2016 local government election, the Democratic Alliance (DA) won 38.4% of the vote in Johannesburg. In 2021, that fell sharply to 26.1%. Through much of 2026, however, the party has been polling around 40.0%, with the latest October survey putting it at 36.0%. All of that is set out in the chart below, alongside the performance of the DA’s key rivals. The average columns include data beyond the three 2026 polls shown in the chart.

Given the parlous state of Johannesburg, the DA’s numbers demand some interrogation. Why, after a decade of sharply deteriorating municipal government, has the party failed to make a more decisive breakthrough?

The first factor is competition. The DA’s sharp decline in 2021 arose to some extent from the emergence of ActionSA, a direct competitor for DA votes. That party won 16.1% of the overall vote in Johannesburg in 2021 and has polled between 9.0% and 10.0% in the city in 2026. Writer Richard Wilkinson has previously argued that The Common Sense may have overpolled ActionSA, and readers can find that excellent critique here.

A second factor is the DA’s own record in Johannesburg. It is sometimes forgotten that the party has governed the city at intervals over recent years. Add those periods together, and the DA has run Johannesburg for roughly three to four years, albeit under often chaotic coalition circumstances in which residents’ lives did not materially improve. Voters may therefore be less inclined to regard the party as an entirely untested alternative to the African National Congress (ANC).

A third factor is that the DA remains, in the main, a middle-class party. As this newspaper has previously shown, its support is now predominantly black, with 36.0% coming from narrowly defined black voters, 35.0% from white voters, and the balance from coloured and Indian voters. The DA is therefore not a white party, as sections of the legacy media continue to try to portray it. It is instead a party of the middle classes, and the growth of those classes may be an important determinant of the DA’s electoral reach. In a city characterised by mass unemployment and poverty, that may impose a natural ceiling on its support.

A fourth factor is that, at the national level, the DA has not been wildly impressive since entering the Government of National Unity (GNU). Broader support for the GNU has also slipped as the public has come to fear that its performance will not deliver materially different results from those of the ANC government before May 2024. The DA’s participation in national government has therefore not necessarily translated into greater confidence in its ability to govern Johannesburg.

A fifth factor, as Wilkinson alludes to, is that the polls themselves may be off. The Common Sense and the Social Research Foundation (SRF) poll widely and have considerable experience in doing so, but remain open to the possibility that their data may be wrong. With relatively few voters participating in some of South Africa’s municipal contests, including Johannesburg, a swing of just 50 000 or 100 000 voters turning out or staying home could make the difference between the DA winning or losing control of a major metro.

All these factors (less the data being wrong, as we are confident it is broadly indicative of what the final result will be) likely play a role in explaining the DA’s stubbornly flat support.

But beyond them, another factor may be considerably more important.

Regime change, which South Africa is currently undergoing, is a two-stage process. The first stage sees people lose faith that the old regime can defend their interests. The second sees them acquire the belief that a new regime can do so.

South Africa has, in the main, completed the first stage. The ANC’s electoral collapse demonstrates that millions of voters have lost confidence in its ability to protect their interests or improve their circumstances. But the second stage remains incomplete. Millions of households, particularly those below the middle classes, have yet to develop the conviction that the DA, or indeed any other alternative, can defend their interests more effectively.

That helps explain why South African politics has become stuck in a political no-man’s land, with no party commanding a majority and none seemingly on course to secure one. The ANC has lost the confidence of much of the country, but its principal opponents have not yet won that confidence for themselves.

Much the same is likely occurring in Johannesburg. The DA’s problem is not principally that voters remain loyal to the ANC, it is that too many of those who have abandoned the ANC remain unconvinced that the DA represents a sufficiently credible alternative.

That is the final hurdle Zille must overcome. The collapse of the ANC has created the opportunity for political change in Johannesburg. But the collapse of an old political order does not, by itself, establish a new one.