Drivers of household carbon emissions in South Africa

Based on consumption survey data and carbon intensity measures of various goods and services, I estimate that in 2010/11 South African household consumption resulted in a total of 205 million tonnes of CO2 (tCO2). In figure 1 below, I break down total emissions into seven different types of goods and services that households spend their money on, as well as the carbon intensity of these goods (measured as the tonnes of CO2 emitted per R1000 spent on that good).

It is worth mentioning here, that when I talk about carbon emissions I am talking about the sum of both direct and indirect carbon emissions. The former includes emissions that are directly released through consuming a product or service. For example, when you drive a vehicle, you directly emit carbon through the combustion of petrol – these are direct emissions. Alternatively, when you buy and wear a shirt, you are not directly emitting carbon emissions. In this case, we measure indirect emissions: the emissions emitted during the production and transportation of that shirt to its point of use.

Figure 1: The carbon intensity of different consumption categories (tCO2) (left) and relative contributions to total household consumption-based emissions (%) (right)

Source: 2010 Income and Expenditure Survey (Stats SA, 2013) and Arndt et al.( 2013). See my previous blog post for more information on how household carbon emissions are estimated.

The bars on the left of figure 1 show that the most carbon-intensive consumption categories are energy, followed by private transport. Energy is by far the most carbon-intensive consumption group – more than 10 times more carbon-intensive than private transport. This is unsurprising given that most of our energy comes from coal-fired power plants.

Interestingly, but not unexpected, private transport is almost twice as carbon-intensive as public transport. We can also see that 36% of total emissions come from the consumption of energy, followed by housing[i] (22%) and private transport (16%). Collectively these three items account for three-quarters of total household consumption-based emissions in South Africa.

Figure 2. Relative contribution to total carbon emissions of different consumption categories by expenditure decile (%)

Source: 2010 Income and Expenditure Survey, (Stats SA, 2013) and Arndt et al.( 2013).

In figure 2, I break down the relative contributions of each consumption item to total emissions by expenditure decile. This gives us an idea of the emissions patterns for poorer and richer households.

We can see that for poorer households, energy, food, and clothing contribute to most of their emissions, contributing to roughly 80% of the total emissions for the poorest 10% of households. However, as households become richer, relative emissions shift towards private transport and housing, while emissions from spending on essential goods and services like energy, food, and clothing decline.

Overall, spending on energy, food, and clothing drives the carbon emissions for poorer households while spending on energy, private transport and housing drives the carbon emissions for richer households. The fact that housing contributes significantly to total emissions for richer households explains why housing is the second largest contributor to total emissions, even though it is relatively less carbon-intensive than energy and transport.

Understanding what drives emissions for richer and poorer households is important from a climate policy perspective. For example, in 2019 South Africa implemented a carbon tax on producers which places a tax of R120 per tCO2 that they emit. If we assume that this tax is fully passed through to consumers, energy will likely experience the largest price increase, given its carbon intensity. Since spending on energy, as well as spending on other essential goods like food and clothing, contributes to the largest share of poorer households’ carbon emissions, this tax is likely to place a disproportionate burden on poorer households relative to richer households.

To ensure that a carbon tax is not regressive, one solution could be to limit the price increase of energy for poorer households by recycling carbon tax revenues appropriately and perhaps considering a higher carbon tax on private transport. This could help ensure that the carbon tax is progressive and does not place a disproportionate burden on poorer households, while still incentivising a shift towards lower carbon-intensive products.


ADDIN Mendeley Bibliography CSL_BIBLIOGRAPHY Arndt, C., Davies, R., Makrelov, K. & Thurlow, J. 2013. Measuring the carbon intensity of the South African economy. South African Journal of Economics. 81(3):393–415. DOI: 10.1111/j.1813-6982.2012.01324.x.

2013. 2010 Income and Expenditure Survey [Dataset]. Pretoria: Statistics South Africa Available: [2021, June 01].


[i]Emissions from housing includes all direct and indirect emissions emitted during the construction and maintenance of housing, as well as indirect emissions associated with the manufacturing and distribution of household furniture and appliances.