How big is South Africa's middle class - especially its black component - and who belongs to it? Here are five main ways in which it can be measured.
A decrease in poverty and growing middle class are widely seen as desirable features of an economy.
In South Africa, the growth of the black middle class in particular is used as a yardstick of sorts of the government's economic policy since 1994.
The expansion of the middle class is not necessarily an end in and of itself. Rather, it is thought to create knock-on effects, such as growing spending power, a better educated and trained workforce, more investment, more job creation and so on.
In the South African context, it is also seen as important that the country's middle class change its demographic makeup to include more of those disadvantaged in the past.
Is South Africa's middle class growing?
This is where things get a bit tricky. The middle class is often talked about in a concrete manner and yet it can be, and is measured and defined, in innumerable ways. A few examples:
The African National Congress' 2018 January 8th statement included a line that the black middle class had grown from 1.7 to 6 million in the previous five years. (Note: President Cyril Ramaphosa skipped over this part when he delivered the speech.)
In 2017, the New York Times presented two different claims - in the same article - about the black middle class in South Africa, saying it doubled between 1998-2008 and "more than doubled" between 2004-2013.
As a result, saying that the middle class has grown is equivalent to saying democracy has strengthened - it seems like a good thing, but it doesn't mean much if the specific context and content of the terms aren't explained.
As Dr Justin Visagie, a research specialist at the Human Sciences Research Council, told Africa Check: "I don't think there is a straight answer when it comes to defining the middle class - it depends on your perspective and what you are interested in."
Here are just some of the ways in which the middle class can be measured and how such measurements impact the numbers:
Because "middle class" (or any class) lacks a broadly accepted definition, other than a vague sense that it sits somewhere between the rich and the poor, we can all decide for ourselves what it means - and we do.
A straightforward approach to finding the size of the middle class is to ask people what class they identify themselves as belonging to. The World Values Survey, where a global network of social scientists interview people from almost 100 countries around the world, collects data on "beliefs, values and motivations". One of the questions asked is: "Which social class would you describe yourself as belonging to?"
Class identification Total SA population SA's black population
Upper class 1.8% 1.0%
Upper middle class 12.1% 7.4%
Lower middle class 17.6% 17.0%
Working class 25.3% 24.0%
Lower class 38.2% 45.7%
Don't know 4.9% 4.9%
Source: World Values Survey, 2013
If we collapse upper and lower middle class, a total of 29.7% of South Africans and 24.4% of black South Africans identified themselves as belonging to the middle class in 2013. Of that middle-class group, 61.7% were black.
The major drawback of this approach is that the answers are subjective, which inherently has some potential weaknesses. People don't always answer truthfully, or they might answer differently if the question were asked in another way.
To illustrate the point, using the same survey data, 24.5% of those identifying as middle class said they or their family had "often" gone without enough food to eat in the past year. Regularly going without enough food is unlikely to be part of the picture many people have of a middle-class lifestyle.
The literal middle
We might also be interested in a more objective determination of middle class. This is usually done by identifying a characteristic such as personal income, and finding out who falls approximately in the middle of the range of the whole population.
Various methods for deciding the literal middle exist, but it generally involves identifying an income range from the median income. (Note: Median income is the amount where half of people earn less than that, and half more.)
Different studies have put that range between 50-75% of median income at the low end and 125-225% of median income at the high end. Whoever falls within that range is counted as middle class.
Using data from the 2008 National Income Dynamics Study (NIDS), Visagie showed about a third of households fell within his selected band of 50%-150% of the median per capita income per month. 83.9% of this middle class were identified as black.
This may seem like a surprisingly large proportion to anyone who knows something about income distribution by race group in South Africa, and it is. This demonstrates one of the major drawbacks of using the actual middle to measure the middle class in a highly unequal country like South Africa.
When a large proportion of the population have relatively small incomes, which we would expect when income inequality is high, the median income will also be relatively small. Indeed, the actual middle as calculated here had a monthly income of between R380-R1,140 per person in 2008.
Even at the top end of this income bracket, we are not talking about an affluent group of people. And again, it is difficult to imagine this is the group most people think of as "middle class".
A universal range
Identifying the middle class by income or wealth using a universal range is a popular approach.
This method makes it easier to compare countries. And the middle class produced won't be so unrealistically poor as when the literal middle is used in countries where income is very unevenly distributed.
Credit Suisse. The financial services company Credit Suisse employs this approach in their Global Wealth Report. It determines the global middle-class to be that group of people whose net worth is between US$10,000-US$100,000 (approximately R119,500-R1,195,000 currently).
While Credit Suisse does not reveal the data sources used to estimate people's net worth (only saying it comes "from sources believed to be reliable"), it estimated the size of South Africa's middle-class at 28.5% of the population in 2017. (Note: Credit Suisse did not provide a breakdown by race.)
African Development Bank. The African Development Bank also suggests a global middle-class, though in their study of the middle class in Africa the calculation is based on daily consumption rather than net worth.
The bank counted people as middle class when they consumed between US$2-US$20 (approximately R24-R240) per day in 2011. (Note: Like Credit Suisse, the bank also did not make its data sources public.)
By this measure, a rather impressive 43.2% of South Africans were middle class.
The bank did distinguish between a floating class (consumption between US$2-US$4), lower middle class (US$4-US$10) and upper middle class (US$10-US$20), but all were grouped as middle-class in this study. Without the floating class, South Africa's middle class comprised 19.8% of the population. (Note: The bank also did not provide a breakdown by race.)
The major drawback of these methods is that universal income or wealth ranges are crude at best. They do not take into account the cost of living or other country- or region-specific differences, so what buys a middle-class lifestyle in one place may not do so in another.
We could also point to the relatively wide bands both Credit Suisse and African Development Bank put forward and ask whether these truly encapsulate people living in similar circumstances.
A net worth of R119,500 is arguably categorically different from a net worth of R1,195,000. A daily consumption rate of R24 per day equates to R720 per month; R240 per day equates to R7,200.
The lives of people at the two extremes of both these ranges are likely to be fundamentally different, so grouping them under the same class seems misguided.
Visagie demonstrated yet another way to identify who belongs to the middle-class by looking at occupation. This method gets closer to identifying the affluent middle class than some of the more straightforward measurements already discussed.
This approach identifies as middle class those households whose highest income earner has, what Visagie termed, a "typically middle-class occupation", like managers and senior officials. He defined these households as the "relatively affluent middle."
By this method, the "relatively affluent" middle class comprised 29.9% of South African households in 2008, but in comparison to the literal middle (also about a third of the population) has a far higher monthly income bracket of R1,400-R10,000 per capita. In terms of racial composition, the proportion of black South Africans in the relatively affluent middle was at 52.3%.
The income band of the relatively affluent middle is certainly narrower than those we have seen generated by the other approaches. Arguably this provides a better fit for that part of the South African population who are not the poor, but who also are not part of the very rich.
On the other hand, we are still left with a relatively disparate group in income terms, which means that the experience of daily life of those at either end of the spectrum is unlikely to be similar.
What sets the middle class apart from the working class or poor when it comes to their daily life, then? The approach taken by researchers from the Southern African Labour and Development Research Unit (SALDRU) in a working paper was to define the middle-class by the (im)probability of falling into poverty.
In other words, the middle-class is understood to have security unavailable to the working-class or poor. This sets them apart from those who are no longer poor, but who are still vulnerable to falling back into poverty should they lose their job or suffer a similar setback.
This method appears to provide a more meaningful understanding of the middle class than looking just at income, for example. As the authors point out: "Being able to afford a certain basket of goods at a given point in time does not yet give any indication of whether the same will be true in the next period."
Using National Income Dynamics Study (NIDS) data from 2014/2015, the authors looked at various features "associated with the probability of being poor in South Africa".
With the maximum risk of falling into poverty set at 10% over the next six years, the share of the South African population identified as middle class was 13.5%. To put it slightly differently, 13.5% of South Africans had less than a 10% chance of falling into poverty.
By this method, 48% of the South African middle class was identified as black, up from 39% in 2010/2011.
How big is South Africa's (black) middle class?
The examples discussed, which are by no means an exhaustive list, show that the middle class can vary quite widely depending on how one chooses to define and calculate it.
In this factsheet alone, the South African middle class has varied from 13.5% to 43.2%, and the share of black South Africans forming part of it between 48% and 83.9%.
Method Middle class as % of SA population Black South Africans as % of middle-class population
1. Self-identification 29.7% 61.7%
2. Literal middle 31% 83.9%
3. Universal range 28.5% / 43.2% n/a
4. Occupation 29.9% 52.3%
5. Vulnerability 13.5% 48%
It is clear that there's no consensus, either on the definition of the middle class or its size in South Africa. As a result, we should be careful when making or assessing claims about how big it is and who belongs to the middle class.
Researched by Markus Korhonen