Introduction

In the absence of regular polling data, analysts look at by-election results to determine voters’ attitudes to political parties and to democracy in general. Since the November local government election (LGE) there have been 43 ward by-elections, with 12 wards changing hands.

Ward by-elections are not a representative sample of a general election: they usually take place in competitive wards and municipalities and they are usually triggered by councillor deaths, resignations and expulsions. A large proportion of councillor deaths are due to violence and assassinations, and many resignations / expulsions occur because councillors leave their original parties to campaign for new ones.

Still, there is valuable information that comes from by-elections. Over the last nine months the by-election results have highlighted some concerning trends, including increased fracturing of the vote and an increase in voter apathy, particularly in the metros.

By-election winners and losers

The Sankey diagram shows the flow of seats from winning parties in the November LGE and the by-elections. Hover over a single path to see the number of wards moving from one party to another. 

Most of the wards were retained by incumbent parties. The ANC won 20 of the 27 wards it was defending, the DA won seven of its eight wards, the IFP three of four wards and the NFP defended its solitary ward.  

The ANC was the biggest loser over the period, losing seven wards to the EFF (3), the African Unified Congress (AUM), DA, IFP, and an independent candidate. The party won a DA ward.  

The DA won three wards off the ANC, PA, and the Karoo Gemeenskap Party (KGP), losing one ward to the ANC. The IFP won two wards from the ANC and Al Jama-ah (ALJ) but lost a ward to the NFP.  

On aggregate the EFF was the most successful of the established parties, gaining three wards. The DA gained two wards and the IFP and NFP each gained a ward. The ANC ended up six wards lighter.  

Close races and competitive wards  

The number of wards changing hands does not tell the full story. In 20 of the 43 wards – almost half – the winning candidate received less than 50% of the vote in one election. In ten wards the winners in both the by-election and the LGE won less than half of the vote.

The bar chart shows the 20 most competitive ward by-elections. Hover over the bars for more information.  

The hardest-fought by-elections are usually in the same wards and municipalities that were competitive in the 2021 LGE: northern KwaZulu-Natal, the West Coast district of the Western Cape, and wards in metro municipalities. 

In the five most competitive wards the by-election winner won less than 40% of the vote. This suggests that the vote is split at least three ways and ANC-IFP-NFP tussles have been a trend in northern KwaZulu-Natal since 2011.

However, when the ANC clung onto a Soweto ward with under 32% of the vote – a ward it had won with 59% in the 2021 LGE – a new possibility emerged. The EFF (24%) and ActionSA (22%) carved away at the ANC’s safe margin. Could this herald more knock-down fights in Gauteng wards?

The vast majority of ward seats are won by either the ANC or the DA. The growth of ActionSA in Gauteng and the PA in the Western Cape is challenging this trend.

Voter turnout: driven by competition and dulled by apathy

The scatter plot compares voter turnout in the 2021 LGE with turnout in the by-election. There is a positive correlation between the two: a ward with high turnout in the LGE is likely to have high turnout in the by-election Similarly, low turnout begets low turnout. 

On average, turnout in a ward for the LGE is a few percentage points higher than in the by-election. The average, however, hides the differences between wards that are retained by incumbents and wards that are won and lost. Click on the legend at the bottom to turn the won / lost and retained wards on or off.

The wards below the trend line had much lower turnout in the by-elections than the LGE. Seven of the eight wards with the lowest by-election turnout are in metros (hover over a dot for more information). Turnout of 30% or lower is a concern and is probably caused by voter apathy.

All of the wards which changed hands are above the trendline, some significantly so. Only two of these wards had lower turnouts in the by-elections than in the LGE. It is likely that heavy campaigning by parties in these competitive wards increased voter turnout.