The distribution of carbon emissions across South African households

South Africa is the world’s most carbon-intensive non-oil-producing developing country[1]. In 2011, we emitted an average of 9.1 tons of CO2 (tCO2) per person – almost double the world average of 4.9 tCO2[2]. As we transition to a low-carbon economy, it will be necessary for South Africans to reduce their carbon footprint. However, if our transition is to be just and fair, not all South Africans should be held equally responsible for doing so.

There are many ways to measure carbon emissions. Before we look at how they are distributed across South African households, it is helpful to consider first the different ways in which emissions can be measured.

We typically measure and report production-based emissions. These emissions include all carbon emitted due to the production of goods and services within our borders. When we adjust these emissions to account for trade (i.e. adding in and removing emissions ’embodied’ in imports and exports), we get what we refer to as consumption-based emissions.

Consumption-based emissions are more suitable if we want to understand how the lifestyle and consumption choices of households contribute to our total carbon emissions. When we make this adjustment, our per capita emissions almost halve from 9.1 tCO2 to 5.6 tCO2. This suggests that a significant proportion of our total carbon emissions are exported to the rest of the world (i.e. the emissions are generated to produce goods and services consumed outside of South Africa).

This estimate of consumption-based per capita emissions is an average, which masks the underlying distribution of emissions across households. In the figure below, I plot the distribution of households’ average consumption-based emissions (per capita) by their consumption decile they fall into. A decile corresponds to each of ten equal groups into which South African households can be divided according to their consumption levels. For example, decile 1 corresponds to the poorest 10 per cent of household while 10 corresponds to the richest 10 per cent of households.

This figure is helpful since it shows how emissions are distributed across households. We can also determine which South Africans over- or under-contribute towards carbon emissions relative to the national average[3]. We can do this by comparing the average per capita emissions in each decile relative to the national per capita amount, which we can think of as each household’s “fair [4]. If a household’s emissions lie below the “fair share” amount, they are an under-emitter. Similarly, if a household’s emissions lie above the line, they are an over-emitter.

It is immediately apparent that there is a great deal of inequality in our carbon emissions, which are heavily skewed towards richer households[5]. Strikingly, the average per capita emissions for the top decile of South African households is almost 4 times higher than the national average. Moreover, if every South African adopted the same consumption patterns as the richest 10 per cent of South Africans, we would be the 5th highest per capita emitter globally – ranking only after oil-producing and refining countries such as Qatar, the UAE, Kuwait, and Singapore[6].


We can also observe that the bottom 70 per cent of South Africans are under-contributors, while the top 30 per cent of South Africans are over-contributors. This is most extreme in deciles 9 and 10, which consume between 2 to 4 times more than their “fair share” of carbon emissions. In fact, this group, which makes up 20 per cent of our population, are responsible for almost 60 per cent of South Africa’s consumption-based carbon emissions.

This has several important implications.

Firstly, a crucial part of meeting our climate goals involves richer South Africans taking active steps to reduce their consumption of carbon-intensive goods and services. For example, households that fall into deciles 9 and 10 will need to reduce their emissions by roughly 50 and 75 per cent, respectively, should they want to reach their “fair share” amount of emissions. This will involve fundamentally changing consumption patterns and making deliberate choices to reduce wasteful and unnecessary consumption.

Secondly, given the strong correlation between poverty alleviation and economic growth – which has historically been carbon-intensive – meeting our development goals will necessitate allowing the poorest South Africans to increase their carbon emissions in the short- to medium-term if we stay on our current carbon-intensive development path. For example, the poorest 10 per cent of households could increase their emissions by over 4.5 times their current amount and still only be emitting as much as the average South African household.

Thirdly, we must be careful of placing the sole responsibility of reducing our emissions on South African consumers. Given that production-based emissions are 30 per cent higher than consumption-based emissions, this responsibility must be equitably shared amongst producers and rich consumers alike.

Finally, in a carbon-constrained world, the carbon-intensive growth path that we have historically relied on to achieve the levels and patterns of consumption that rich South Africans enjoy is no longer tenable. Should we wish to achieve our climate and development objectives, it is important that we rapidly transition to a low-carbon economy while at the same time finding a more equitable distribution of carbon emissions between richer and poorer households.

We should, therefore, not misconstrue increased per capita emissions for poorer households as a failure to meet our climate goals per se. Doing so would be unfair. Rather, any increase in the per capita emissions of the poorest South Africans should be considered within the context of what is happening to our overall emissions and whether we, as a nation, are on course to meet our international commitments with respect to preventing catastrophic climate change.



Summary (approximately 150 words)

In this post, I investigate the distribution of carbon emission across South African households. I show that carbon emissions are unequally distributed, with rich South Africans emitting between 2 to 4 times the national average. Strikingly, if every South African adopted the same consumption patterns as the richest 10 per cent of South Africans, South Africa would be the 5thhighest emitter of carbon emissions globally. This highlights several important points: Rich South Africans need to actively reduce their carbon emissions; an increase in emissions by poorer households should not be misconstrued as a failure to meet our climate goals; responsibility to reduce emissions should be shared between producers and rich consumers alike; and, if all South Africans are to enjoy the consumption levels of rich South Africans, it is necessary to transition to a low-carbon economy.



[1] Arndt, C., Davies, R., Makrelov, K. and Thurlow, J., 2013. Measuring the carbon intensity of the South African economy.South African Journal of Economics,81(3), pp.393-415.


[2] Ritchie, H. & Roser, M. (2017) – “CO₂ and Greenhouse Gas Emissions”. Published online at Retrieved from: ‘’ [Online Resource]

[3] The total carbon emissions attributable to households comprise of both direct and indirect carbon emissions. Direct emissions include, for example, the burning of coal or wood for energy use. Indirect emissions include all the emissions that are emitted to produce a good, as well as all the emissions that were emitted to produce the intermediate goods that went into the production of the final good that a household consumes.

[4] The term “fair share” in climate change literature refers to the equitable distribution of emissions that countries can safely emit while still meeting their nationally determined contributions in line with the Paris Agreement, as well as their development goals. It is likely that our national average is greater than our true fair share. It is therefore technically incorrect to refer to our national average as our true fair share of emissions. In this blog I make no attempt to argue what our true fair share should be. This is a simplification I make in the interest of clarity and to demonstrate that when thinking about emissions reductions we need to consider factors such as inequality and what is fair.

[5] Following Arndt et al. (2013), I use consumption expenditure as a proxy for income. This is a reasonable assumption if we assume that consumption expenditure is a useful proxy for lifetime earnings (Poterba, 1989). A household is therefore considered “rich” if they have a high level of consumption expenditure.

[6] Note: This ranking is based on consumption-based per capita CO₂ emissions in 2011. Although it’s has no oil reserves, Singapore is Asia’s major oil trading and refining hub (EIA, 2016), hence it’s high ranking.