A "rusk bar chart" comparing our voter-eligible-population (VEP) turnout to those of other countries in their most recent major election.
(This article is a follow-up to my 2019 article. )
On 29 May 2024 South Africa held a general election, the seventh since the democratic era began in 1994. All in all, 16.3 million people voted, and there were 27.8 million registered voters. So the voter turnout, as a percentage of registered voters, was 58.6%. This is the voter turnout figure that the Independent Electoral Commission (IEC) quotes on their website, and what generally appears in the media.
People use this figure (voter turnout as a percentage of registered voters) because it is easy to measure. You just need to count the number of people who voted, and divide by the number of people on the voter's roll.
But of course, what we really need to measure is the number of people who voted as a percentage of the number of people who are eligible to vote (not merely the number of people who registered).
In South Africa, every citizen aged 18 years or older is eligible to vote. We call this population the voting-eligible population (VEP).
Based on Stats SA's data and various things they have said, and a bit of spit and polish (see below), one might estimate this VEP population as somewhere between 37.3 million and 40 million people. So, voter-eligible turnout (the true turnout) is perhaps somewhere between 16.3/37.3 = 43.7% and 16.3/40.8 = 40%.
Let's just call it about 42%.
It's a terrible turnout. (It was bad enough in 2019, when it was around 49%). See figure above. (The numbers are plotted in the form of rusks to stress that in each country (not just our own) there is a fair bit of uncertainty to the VEP turnout, although this uncertainty is arguably greater in our case.)
Brazil and India really set an example for us here. Consider the case of India. They have just finished a major election a few days ago. 968 million people voted! The entire population aged 18 and over is only 992 million people (!), so 97% of voters were registered! Therefore, their voter-age-population turnout was 65%, which we will take as a (rough) proxy for their voter-eligible turnout. What can we learn from them?
We need to start by estimating the voter-age population, i.e. the number of people living in South Africa aged 18 and over, on 29 May 2024. We can use the midyear population estimates (aggregated by age) that Stats-SA usually publish each year for this. Here is a chart:
There are a couple of problems here. I was not able to obtain the 2023 midyear estimates from their website, and 2024's is not available yet.
Another problem is as follows. Stats-SA's midyear population estimate for 2022 was 60.6 million. But in 2022 we also had a census. And the result of the census were that on 2 February 2022, the population of South Africa was 62 million. So Stats-SA's 2022 midyear estimate is about 3% lower than its census estimate. A similar phenomenon happened in 2011 with the 2011 census. Perhaps we should "bump up" all the midyear estimates by 3%, to be consistent with the 2011 and 2022 census estimates which bookend them.
Then, we need to estimate the voting-age population on 29 May 2024. Let's see. It was about 40.5 million in June 2022 (maybe more like 41.8 million, if we bump up via the census discrepancy). Staring at the graph, we see it's going up around 0.4-0.5 million each year. So in mid year 2024, the VAP is somewhere between 41.3 million and 42.8 million people.
To get the VEP, we need to subtract from the VAP those who are not South African citizens. That's not easy, but we can get a ballpark figure as follows. The 2022 census said there were 2.4 million persons born out of South Africa in the population. On the other hand Statistician-General Risenga Maluleke has also estimated this number at around 4 million. Of course, this number includes those aged 17 and under, and we shouldn't be subtracting those (as they were not counted in the VAP).
So, the VEP is somewhere between 41.3 - 4 = 37.3 million, and 42.8 - 2 = 40.8 million. These are the figures I quoted above, and leads to the voter-eligible turnout of somewhere between 40 and 43.7%; let's call it 42%.
Another way to arrive at this VEP turnout figure is to use the information that the IEC gave in 2019, when they estimated that 74.5% of the VEP were registered. Suppose we assume that this percentage stayed roughly constant to 2024 (an assumption seems to be consistent with voter registration figures). Then, the VEP in 2024 would be 74.5% of the 58.6% of registered voters who voted, which comes to 43.7%.
How does our voter-eligible-population turnout compare to our neighbouring countries? Well, if we use voter-age-population (VAP) turnout as a proxy for voter-eligible-population (VEP) turnout, because it is easier to estimate and close to what we want, then we can use the website of the International Institute for Democracy and Electoral Assistance (IDEA), which has a handy world map.
According to the IDEA, our 42% eligible-voter turnout compares to 55% for Namibia (2019 elections), 53% for Botswana (2019 elections), 60% for Zimbabwe (2023 elections), 50% for Mozambique (2019 elections), and 39% for Lesotho (2022 elections).