May 12th, 2014
I usually stick to public policy
wonkery, leaving political analysis to those more interested in “political analysis” than I. The election results in Johannesburg are however very surprising, and I have been grappling to understand the results. The analysis on Johannesburg election results have however focussed on understanding the impact of issues such as Nkandla, e-Tolls, service delivery protests and other issues. No doubt, each of these issues matter. However, we may better explain the Johannesburg results by focussing on demographics and data. The purpose of this post is not to provide a definitive answer on why Johannesburg votes the way it does, but rather to indicate the importance of understanding the relationship between demographics and election results. The example taken is the correlation between unemployment rates and voting outcomes for the 2014 elections.
Take Away: Correlation between voting patterns and unemployment
A striking correlation is how closely unemployment rates in each ward, match election results. The Democratic Alliance wins wards with higher levels of employment, and the African National Congress wins wards with higher levels of unemployment. Importantly, correlation is not causation.
Elections Map: Dusty “Green” Streets vs. Leafy “Blue” Suburbs
The first map uses 2014 elections results and is drawn from News 24 elections app.
The map shows a picture of Johannesburg as a divided city. The picture of a divide between the north and south of Johannesburg is something any Johannesburg resident experiences. Well, at least those of us that travel around the city. It is still amazing how many people (and political parties) from the “South” remain in the “South”, and how many in the “North” remain in the “North” The election results as shown in the map is however very surprising, as I would not have expected the results to align so closely to the spatial shorthand of “dusty streets in the South” and “leafy suburbs in the North”.
Unemployment Map: Unemployed South vs Employed North
The election results in Johannesburg piqued my interest. Could it really be that the election results were following demographic patterns? The wonderful innovation that are data applications allows for some quick and dirty (but plausible) reflections. Statistics South Africa provides map data based on the Census 2011. The only readily available map (related to income) was on unemployment. The map is reproduced below.
The map is a little difficult to read due to the colour scheme. The darker the area on the map the higher the rate of unemployment. The important thing to notice is that there is a significant overlap
- between higher levels of employment and wards won by the Democratic Alliance.
- between higher levels of unemployment and wards won by the African National Congress.
Does this explain the results?
The correlation between unemployment and voting preferences is just one tiny part in understanding the results. The intent here is not to explain the results, but rather to nudge other analysts (who do more election analysis than I will do) to begin looking at the relationships between election results and available data more closely. This would get us closer to understanding why people voted the way they did in Johannesburg, and in Gauteng.