What does the ‘middle class’ mean in a polarised, developing country such as South Africa?

Reads 27,059
In a developing, highly unequal country such as South Africa, it is unlikely that a definition of the middle class that is based on an income threshold will adequately capture the political and social meanings of being middle class. We propose a multi-dimensional definition, rooted in the ideas of empowerment and capability, and find that the ‘empowered middle class’ has expanded significantly since 1993. It also is much larger than when measured in terms of income.


This article critically considers the reigning notions of what it means to be middle class in a polarised, developing country such as South Africa. Historically, class and class differences are often associated with differences in education, occupation and income. In Marxist and structuralist political-economic analyses, class and class differences are tied to power. Mainstream definitions of the middle class assume that this group of educated, high-income, skilled individuals are a cohesive middle group whose members associate with one another, have significant buying power and have shared political interests.

In line with these theories, the growth of the middle class in South Africa is viewed as symbolising the retreat of poverty, the growing availability of good jobs and the rise of a new consumer market. It symbolises economic progress and development and is seen to contribute to social stability and a more effective democracy. This middle-class perspective acts as a spur for the nation to invest in public infrastructure, roads, schools and health services.

At first glance, the concept of the middle class clearly appears to be significant for a young democracy like that in South Africa which also has a legacy of poverty and inequality. However, one may ask whether the concept of the middle class, as conventionally conceptualised, can be applied to a developing country with significant inequalities in income and assets. In other words, is the idea of a ‘middle class’ appropriate to the assessment of social development and economic progress – and can it guide the formulation of social and economic policy – in South Africa?

The concept of the middle class originated in developed countries and is tied to the structure of these societies. However, due to extreme inequality, social polarisation and the effects of political fragmentation, our country lacks a homogenous cohesive middle group. This complicates and weakens the application of the traditional concept. An alternative approach may be required.

Standard approaches using income levels

The standard approach to defining the middle class uses a lower and upper bound of household income per head in the context of the overall distribution of income. All individuals that have a monthly income above the defined lower bound (threshold) and below the upper bound are then classified as belonging to the middle class.

However, as shown in a recent Econ3x3 article by Visagie (2013a), South Africans in the actual middle of the income distribution (i.e. close to the median) tend to be low skilled, have low income and few assets. When the concept of the middle class is transferred from a developed country to the circumstances prevalent in a developing country, a tension emerges between its link to the income midpoint and its connotations of empowerment, good education and skills, stability, comfortable living standards, savings and assets. The people who are situated in the middle of South Africa’s income distribution are still quite poor and vulnerable and can hardly be described as empowered or comfortable.

It is unlikely that any income-based definition of the middle class will correspond meaningfully to the political and social meanings of the term ‘middle class’ or adequately capture them.

Towards a different approach

Our research proposes a fresh perspective. It is based on the observation that the income-based ‘middle-class’ in South Africa is a broad and layered group which is not as homogenous or cohesive as it has been portrayed in public discussions. This is supported by findings by Schlemmer (2005) and Burger et al. (2014) which indicate that those who would be classified as middle class according to standard notions of class often do not see themselves as being middle class. Moreover, apartheid has left us with a strong association between class and race that has been remarkably persistent and enduring. Qualitative work by e.g. Khunou (2012) and Krige (2012) show that this association still influences individual notions of what class means. This further undermines any thought that there could be a single, unified middle-class identity.

Our main point is this: if one wants to transfer the concept of a middle class in a meaningful way to the milieu of a developing country, i.e. whilst retaining the most central meaning and usefulness of the term, then essentially one has to capture the level or degree of individual empowerment that existsand not just how much (disposable) income they earn.

The concept of empowerment is aligned with human development and the characteristics associated with self-sufficiency and each person’s capacity to be in control of his or her own life. Such an approach would better capture the pathways for the social and political benefits associated with the ‘middle class’ as conventionally understood. For example, people with better education or better access to reliable information are thought to be more able to navigate their own destinies and be less vulnerable to shocks like economic ups-and-downs and workplace, personal, domestic or national calamities.

More importantly, this approach to a ‘middle class’ embodies Amartya Sen’s theory of capability (1989; also Robeyns 2005). In his view, the condition of poverty is not so much a material deprivation as a lack of freedom and, specifically, a lack of power or capability to pursue and realise one’s own goals. This also implies a much broader approach to development.

According to this perspective, the middle class would be defined as the people in South Africa who are empowered and capacitated and who are free to develop and realise their aspirations. Practically this means that they have attained key capacities and capabilities above a certain minimum or threshold, as discussed below.

In this context, the distinction between the middle class and the elite is of little importance. In any case, in South Africa most definitions of the elite would refer to a social layer that is so thin as to be difficult to analyse with the available household data. Therefore, contrary to the traditional model of a bottom, a middle and a top, we propose a simpler, dual structure – one that focuses on the distinction between the disempowered and the empowered (who would then comprise the ‘middle class’). We think that the most important divide is between those who are empowered and those who are not and who consequently often rely on external support – whether from family or friends or, frequently, from the state.

Income is sometimes used as a proxy for empowerment. However, to consider income only would be very narrow and, frankly, inadequate because it represents only market-based channels for the realisation of aspirations – buying goods, services and experiences. Thus, for instance, it does not reflect the effect of free public services (water, sanitation, health clinics, electricity, etc.) on human empowerment, neither does it capture the impact and value of political and social freedoms. Additionally, there are questions as to whether market prices adequately reflect the value that people attach to specific capacities. An example comes from recent research on happiness, which shows that being employed has a more direct influence on the life satisfaction of an individual than the effect of receiving income (Frey & Stutzer 2002).

Measuring the empowered middle class

We propose that for developing countries it may be sensible to abandon an approach which is based on income in favour of a richer, multi-dimensional approach – one that approximates the idea of being middle class by using several characteristics typically associated with ‘middle class’ empowerment, such as having comfortable living standards, stability and being highly skilled.

We estimated the size and growth of the ‘empowered middle class’ in South Africa between 1993 and 2012 using the 1993 Project for Statistics on Living Standards and Development and the National Income Dynamics Surveys of 2008 and 2012. Because our measure looks at empowerment and capacity – which are naturally a function of age – we have limited our sample to the adult members of a household (i.e. 18 years or older).

Specifically, these data sets allow us to define an individual as belonging to the middle class if the person meets all of the following criteria, which would mean that he or she:

  • has achieved literacy (at least 7 years of education);
  • has access to key public services: electricity, clean water and sanitation (at least a ventilated pit latrine);
  • is employed or living in a household with an employed member (i.e. having a link to the labour market);
  • has access to information (via the media); and
  • lives in a household with a stove and a refrigerator.

Only when an individual meets all these criteria can the person be categorised as empowered (or middle class).

Unlike many multidimensional poverty indices, we do not add up attainment scores across dimensions in conjunction with a threshold value. We argue that each of these dimensions is required for empowerment – conversely, that the absence of attainment in any of these dimensions would significantly constrain or disempower an individual, and thus disqualify him or her.

Employing this definition, we end up with a ‘middle class’ or ‘empowered group’ that is much broader than income-based measures of the middle class – but also, we would argue, is a more appropriate and meaningful barometer of social and political change in our country.

While it would be a stretch to assume that this is a cohesive group, the use of several dimensions instead of one means that we capture a number of overlapping layers in our definition, which may plausibly enhance cohesiveness.

Results: a growing empowered ‘middle class’

The empowered, capacitated middle class in South Africa has expanded from 28% of all adults in 1993 to 48% in 2012. These percentages indicate, firstly, an empowered middle class that is much larger than a middle class that is defined in terms of income. Secondly, the percentages indicate that the extent of the growth in the empowered middle class since 1993 is much higher than the growth in income-based measures of the middle class.[1]  

A decomposition on a racial basis reveals that the main change has occurred in the share of black South Africans that belong to this empowered group, i.e. rising substantially from 12% to 40% of black adults. Again, this change is much larger than when one employs an income-based definition of the ‘black middle class’ (see footnote 1).

The most fundamental causes of the growth in the empowered middle class are the large improvements in South African adults’ access to services. Looking at each of the empowerment dimensions for 1993 and 2012, we find the following trends. Access to clean water has increased from 86% to 95% of adults, access to adequate sanitation from 83% to 95% and access to electricity from 48% to 90%. Similarly, the prevalence of literacy has risen from 70% to 82%. The proportion of adults who have a stove in their households has risen from 47% to 85%, while ownership of a fridge has increased from 43% to 76%. Access to information has increased from 89% to 91% – a small change but one starting from a high initial level.

The discrepancy between the capabilities-based and income-based measures of the middle class suggests that these improved capabilities have often not been rewarded with higher incomes in the labour market. This may be related to the low quality of health and education services or the disappointing labour market performance over this period. The slow improvement in links to the labour market – from 73% to only 75% – over this period of almost 20 years is most disconcerting (especially since our criterion does not even require that an individual be directly employed, but only that somebody in the household be employed).


Given the weaknesses of using income as a measure of whether someone belongs to the middle class in a country characterised by great inequality in terms of people’s earnings, we propose the adoption of a multi-dimensional definition based on the ideas of empowerment and capabilities.

Our empirical results suggest that, in terms of this new definition, South Africa has made significant progress between 1993 and 2012. The indicators of empowerment provide evidence that, contrary to much public opinion, the post-apartheid period has produced a substantial and broad improvement in empowerment. We find that there has been a substantial increase in the number and proportion of South Africans that can be regarded as reasonably self-sufficient and empowered and also in the proportion of black South Africans that belong to this group.

However, the lacklustre performance of the labour market has restricted opportunities for individuals to translate empowerment and untapped capabilities into productive employment and earned income – which is likely to be an on-going constraint on development and on the expansion of the empowered middle class.   


Burger R., McAravey C. & Van der Berg S. 2014. The Capability Threshold: Re-examining the definition of the middle class in an unequal developing country. Mimeo, Department of Economics, Stellenbosch University.

Burger R., Steenekamp C., Van der Berg S. & Zoch A. 2014. The middle class in post-apartheid South Africa: Examining and comparing rival approaches. Mimeo, Department of Economics, Stellenbosch University.

Frey B.S. & Stutzer A. 2002. What can Economists Learn from Happiness Research? Journal of Economic Literature (XL:. 402- 35).

Khunou G. 2012. What middle class? The shifting and dynamic nature of social life: life histories of two women from Gauteng. Mimeo, Department of Economics, Stellenbosch University.

Krige D. 2012. Being black and middle class in contemporary Johannesburg: the narrated life story of MR. Mimeo, Department of Economics, Stellenbosch University.

Robeyns I. 2005. The Capability Approach: a theoretical survey. Journal of Human Development 6(1): 93-114.

Schlemmer L. 2005. South Africa's new middle class. In: Ann Bernstein and Sandy Johnston (Eds.). The next decade: Perspectives on South Africa's growth and development. Chapter 5 (112-140). CDE.

Sen A.K. 1989. Development as Capability Expansion. Journal of Development Planning 19: 41–58.

Visagie J. 2013a. Who are the middle class in South Africa? Does it matter for policy? Econ3x3, 29 April (http://www.econ3x3.org/article/ who-are-middle-class-south-africa-does-it-matter-policy)

Visagie J. 2013b. Is the middle class becoming better off? Two perspectives. Econ3x3, 15 July. (http://www.econ3x3.org/article/middle-class-becoming-better-two-perspectives)


[1] We also estimated an income-based middle class and find that according to this definition the middle class has risen from 28% of the adult population in 1993 to 41% in 2012, while the share of black adult South Africans who are classified as middle class according to this definition has increased from 12% to 32% (Burger et al. 2014). Visagie (2013b on Econ3x3), using  a different method and period and different thresholds, reports findings indicating that the income-based, combined middle-and-upper class comprised 20.4% of the total population (of all ages) in 1993 and 24.1% in 2008 –  likewise reflecting a much smaller change than in the empowered middle class.

Download article

PDF icon Burger 2014 Rethinking middle class FINAL.pdf
Post a commentary

This comment facility is intended for considered commentaries to stimulate substantive debate. Comments may be screened by an editor before they appear online. To comment one must be registered and logged in.

Another possibility is that the notion of 'middle class' is not particularly useful in a country such as South Africa.   The normal meaning of the word middle class is that it refers to a social stratum that exists in the middle of the social hierarchy.  But the class you have described exists at the top of the social hierarchy. It is not a middle class. It is Seekings and Nattrasss' Upper Class.  

Thank you for the comment. We explicitly acknowledge, as a point of departure, that in developing countries there is a tension between the meaning of middle class and being in the middle of the income distribution. That is why we proposed an alternative, i.e. the concept of an 'empowered class' or group. Because most of the interest in the concept of a middle class is motivated by the underlying meaning and the associated literature on social/economic/political benefits, we opted to use a 'redefined'. approximated meaning of the term. But the criteria have been specified at relatively low levels to ensure that this 'empowered class' as a whole would not be seen as an 'upper class' (although it will include a relatively small number of people who are 'upper class'). As noted, our major interest is in highlighting the threshold between the disempowered and the empowered - and its potential importance in analysing social and economic change in South Africa. 


Interestingly, your results seem to match a 2013 poll of attitudes by futurefact in which 48% of South African adults classify themselves as 'middle-class' (and another 16% as 'upper class'). Whatever that means to each respondent is another question. http://futurefact.co.za/futurefact/south-africa%E2%80%99s-first-generati...

This comment facility is intended for considered commentaries to stimulate substantive debate. Comments may be screened by an editor before they appear online. Please view "Submitting a commentary" for more information.