By Elnari Potgieter, Paul Berkowitz and Ebrahim Fakir | 14 July 2016
Declining voter participation (turnout), voter apathy (ie a low voter turnout and/or a low percentage voting-age population turnout), or a voter stay-away - has been said to have a possible negative effect on percentage support for the ANC - in particular in key metros (such as Nelson Mandela Bay and Tshwane; Municipal IQ 2016; McKinley 2014).
Support for political parties and voter participation in the Nelson Mandela Bay Metro can be investigated by looking at patterns from past elections. This may offer some clues in assessing possible support patterns prior to the 2016 local government election.
Political Party Support Patterns: Percentage Support and Seats allocated
- The dominant parties in Nelson Mandela Bay during the period under review was the ANC and the DA. The DA's support increased during the 2009 election, despite reasonably good support for COPE during that election. However, the ANC's support dropped significantly during the 2009 national election. The ANC's support continued to decrease in 2011 (local) and 2014 (national) elections, whilst support for the DA increased significantly from 2011 and has been on the up in both national and local elections.
- It would appear that the ANC lost support to COPE during the 2009 elections, and that COPE subsequently lost support to the DA in the following (2011 - local and 2014 national) elections - potentially with previous ANC supporters not returning from COPE to the ANC.
- During the 2014 national elections, the EFF managed to win 4.3% of the votes cast in the Nelson Mandela Bay Metro. This is lower than the national support the EFF received 635% (IEC 2016).
- Support to other smaller parties, such as the ACDP, UDM, ID (which collapsed in to the DA) and PAC, has declined over the period under review - although they each manage to retain one seat (a positive result of the compensatory PR electoral system, which distributes the ward votes cast for these parties).
In Nelson Mandela Bay, there is a clear signal of movement towards consolidating a political party system around two/ three major parties. The 2016 local government election will therefore either consolidate this trend, or introduce some fragmentation with the introduction of new parties and some independent candidates. This may have some implications for coalition formation - depending on how the new entrants perform.
- The total amount of seats for Nelson Mandela Bay increased from 108 in the year 2000, to 120 in the year 2006 and 2011.
- In the year 2000, the ANC had 41 more seats than the DA in total. In the following year, the ANC had 51 more seats than the DA in total (following the increase in the amount of seats in the metro). This decreased to the ANC having 15 more seats than the DA in total in the year 2011. Should the decline in ANC support continue, this gap in amount of seats between the ANC and DA can become even smaller - depending on support to the EFF and/or new parties and independent candidates.
- COPE won six seats in the 2011 local elections, with 4.95% of votes (please see Table 1) . Given the precipitous decline of COPE in 2014, coupled with the entry of the EFF and the United Front of the Eastern Cape - a strong showing from COPE is unlikely.
- The UDM's total amount of seats decreased from two in the year 2000, to one in the year 2006 and 2011. The ACDP and PAC won one seat each despite percentage losses (please see Table 1 in this article) and in spite of the fact that more seats have become available since 2006. Other smaller parties, like the VF+, A-Team, and the Congress Movement of the Coloured People in South Africa have each had one seat in different local elections. But since 2011 none of these parties feature and are unlikely to do so in 2016.
Balance of Power in the Metro
Since the 2006 local government elections there have been 60 wards and 60 PR seats in the Nelson Mandela Bay Metropolitan Municipality comprising 120 Council seats for the Metro in total. This figure will be maintained for the 2016 municipal elections.
This means that 61 seats are needed for a clear majority in the metro, either for a particular party or - a coalition.
The ANC has held a majority in the metro since 2000, but this has been reduced significantly since 2006. In the 2006 elections the ANC won 81 out of 120 seats, or just over a two-thirds. This fell to just 63 seats in 2011.
There have been a few by-elections in the metro since 2011 where the ANC had its majority whittled down further - the party currently holds 62 of the 120 seats, or a majority of just two seats in the council.
These results do not take into account the possible influence or effect of the EFF on party support patterns in the Metro. The party didn't contest the 2011 local government elections or any of the subsequent by-elections. Based on the 2014 results, the EFF might expect to win at least five seats in the council provided that the voter support of all those who voted EFF in 2014 in the Metro remain consistent and faithful in turning out and voting EFF.
Before examining how voter turnout patterns have influenced party support patterns, w now examine voter turnout.
Voter Turnout, Registration and Eligible Voting-Age Population Voting (2000-2014)
Table 3 below summarises the voting-age population (i.e. those residents aged 18 and over) and the number of registered voters for each of the elections since 2000.
1. By Voting-Age Population we mean every South African citizen over the age of 18. Any person 18 years and older may register to vote in South Africa. They can, however, only vote if they have registered (with a South African ID book), after which their names will appear on the voters' roll.
2. By Registered Voters we mean every South African citizen over the age of 18 who registered to vote and appears on the Voters Roll.
3. Percentage of those who voted in a particular election as a proportion of the total number of those registered to vote.
4. Percentage of those who voted in a particular election as a proportion of every South African citizen over the age of 18.
Approximately 85 percent of people aged 18 and over have been registered to vote in each election in the Metro.
Voter turnout is generally higher in national / provincial elections (2004, 2009 and 2014) than in municipal elections (2000, 2006 and 2011). There has been a trend of decreasing participation in national / provincial elections, with turnout falling from 82 percent in 2004 to 81 percent in 2009 and 75 percent in 2014.
Voter turnout for the 2011 municipal elections was relatively high, at 65 percent.
Voter Participation and Support to the ANC / DA (2000-2014) - a relationship?
Comparisons have been made between voter turnout and party support in South Africa's Metros - stating that a decline in Voter Turnout correlates with a decline in support for the ANC. Given the data above, it can be stated that there has been a decline in Voter Turnout in national elections in Nelson Mandela Bay. Similarly, there has been a decline in the percentage voting-age population from the 2009 elections in the 2014 national elections. However, Voter Turnout and percentage of the Voting-Age population in local elections has increased from 2000-2011. From 2000-2014, in both national and local elections, support for the ANC has decreased and support for the DA has increased.
If we compare voter turnout and support for the ANC in consecutive national and local elections separately, the following is evident. Between the national elections of 2004 and 2009, voter turnout decreased marginally by 1%, yet ANC support decreased from 69.28% to 50.14% (a decrease of 19.14%). From the 2009 to 2014 National Elections turnout decreased by 6%, while ANC support declined with 0.97% from 50.14% to 49.17%. The ANC's most significant decline was thus during the national elections of 2004 - 2009, when COPE won 17.02% of the votes cast during the national election of 2009 in the Nelson Mandela Bay Metro.
With respect to local elections, voter turnout in 2000 was 62%, decreasing by 6% to 56% in 2006, and then increasing by 9% in 2011 to 65%. ANC support in 2000 was 66.16%, which increased slightly to 66.53% in 2006, and then decreased by 14.62% to 51.91% in 2011.
Therefore, in National Elections - there might be a relationship between a decline in voter turnout and decline in voter support (although the decline in ANC support and Voter Turnout is not correspondingly proportionate). In local elections however, this is not the case.
Despite a decrease in voter turnout between 2000 and 2006, support for the ANC increased slightly. From 2006 to 2011 - when voter turnout increased, support for the ANC declined significantly.
The assertion that a lower voter turnout will translate into a lower proportion of support for the ANC thus does not always hold true for elections in Nelson Mandela Bay, and will need to be tested on a ward specific basis.
It is difficult to compare election results across two different types of elections (Local and National). This is so because the electoral systems used in each are different, and because voter turnout is significantly higher for national/provincial elections than for municipal elections. Because of this difficulty, we have opted to compare voter turnout in a specific ward and party support levels in specific wards for the 2011 and 2014 elections, separately. In other words, the voter turnout in a particular ward for a specific election has been compared to party support in that ward for a specific election. Before we provide that comparison below, we provide a figure which demonstrates the voter turnout figure per ward relative to the metro average, which allows us to identify wards which have significant deviations from the Metro average (above or below 10%).
The two figures below demonstrate that Voter turnout across Nelson Mandela Bay wards has been fairly uniform in past elections. Average voter turnout in the metro in the 2011 municipal elections was 65 percent, with the ward with the lowest turnout having a 48 percent turnout, and the ward with the highest turnout at 81 percent voter turnout. The vast majority of wards (45 out of 60) had a turnout that was within five percentage points of the average, and almost all wards (56 out of 60) had a turnout within ten percentage points of the average:
This pattern is repeated in the 2014 elections:
Only four out of 60 wards had voter turnouts that deviated from the metro average by more than ten percentage points. If these four wards (Wards 4, 9, 43, and 52) are excluded then the level of voter activity is very similar across the remaining wards.
The scatterplots below show that there is a weak relationship between voter turnout in a ward relative to the Metro Average and support in that ward for the ANC and DA - and the nature of that relationship can change if outlying values are excluded from the analysis These scatter plots demonstrate the relationship between one variable (voter turnout) and another variable (party support). All the data used to create the scatter plots are tabulated from the results available on the IEC website.
The two scatterplots show that there is a negative relationship between ANC support and voter turnout, and there is a positive relationship between DA support and voter turnout. In other words, a ward with voter turnout that is higher than the metro average is more likely to support the DA, and a ward with voter turnout lower than the average is more likely to support the ANC.
As mentioned above, the relationship is not particularly strong - and it doesn't hold true for every ward. Almost a third of the wards in the metro show the opposite relationship: they're either ANC-leaning wards with turnout above the average or DA-leaning wards with turnout below the average.
In conclusion, voter turnout is not always a good predictor of how a particular ward will vote in Nelson Mandela Bay.
IEC 2016 National and provincial election results [Database], http://www.elections.org.za/content/Elections/National-and-provincial-elections-results/ (accessed 14 July 2016).
MCKINLEY, DT 2014 "Low Voter Turnout: The Real Story of South Africa's National Elections". SABC Breaking News, 12 May, http://www.sabreakingnews.co.za/2014/05/12/low-voter-turnout-the-real-story-of-south-africas-national-elections/ (accessed 14 July 2016).
MUNICIPAL IQ 2016 "Election Apathy a Danger for the ANC", http://www.municipaliq.co.za/index.php?site_page=article.php&id=84 (accessed 14 July 2016).