The Capability Threshold: Re-examining the Definition of the Middle Class in an Unequal Developing Country

Abstract

We argue that a multi-dimensional approach to categorizing the middle class is more appropriate for a polarized developing country and propose an alternative measurement rooted in the ideas of empowerment and capability. We find that the “empowered middle class” has expanded significantly since 1993 and also constitutes a larger share of vulnerable subgroups such as blacks, female-headed households and rural inhabitants. Differing trends between the middle class categorized based on income and based on capabilities are attributed to improved capabilities that have not been rewarded with a proportional increases in access to the labour market. It is disconcerting that links to the labour market improved only slightly and this is attributed to sluggish labour market growth and low quality of education. It is concerning that vulnerable individuals harbour unrealistically high expectations of the social mobility of their households and appear to not understand the determinants of social mobility and labour market prospects. This is attributed to heightened expectations following the political transition, but also the continued disconnection and marginalization of vulnerable subpopulations from the mainstream economy.

Keywords:

• Middle class
• Capability approach
• South Africa
• Inequality
• Income
• Polarization

About the Authors

Ronelle Burger is an associate professor at the Economics Department of Stellenbosch University. Her interest is in poverty and exclusion in African countries, with a focus on the role of government and nonprofit service delivery. Her research is closely aligned with policy processes because it aims to inform and support initiatives to fight poverty and inequity. She has consulted to National and Provincial government and also multilateral donors such as UNICEF and the World Bank.

Camren McAravey is an Emergency Administration and Finance Manager at Médecins Sans Frontières. She holds a Master’s degree in Economics from Stellenbosch University and has participated in a student exchange with Georg-August Univeritaet Goettingen in Germany. Her areas of interest include development economics and socio-economic development.

Servaas van der Berg is professor of Economics at the University of Stellenbosch in South Africa, and holds the National Research Foundation’s South African National Research Chair in the Economics of Social Policy. His research covers economic development, poverty and income distribution, the economics of education, health and welfare, and fiscal incidence. He acts as consultant to international, regional and other institutions and often provides policy advice to government departments.

Notes

1. The term empowerment is commonly used to indicate both a process (of empowering groups or individuals) and an outcome (a person or group is empowered).

2. Access to electricity was measured by considering whether electricity was used for lighting. We find that even if electricity is available, households often prefer to use alternative energy sources for cooking and heating because they are cheaper or the complementary goods required to use electricity for these purposes (e.g. electric stove or electric heaters) may for many households be too expensive to purchase.

3. Adequate sanitation excluded only the no sanitation category and the bucket system.

4. We did not penalize households that consisted entirely of students and/or pensioners.

5. Comparing 1993 and 2008 income estimates for the same data sources Leibbrandt and Levinsohn (2011) find that while there has been a strong mean increase in per capita real household incomes, the median increase has been small because gains were concentrated amongst a small subset. They attribute this to slow job growth during a period of strong macroeconomic performance.

6. In line with the definitions used in the two surveys and conventional categorization in South Africa, the category “black” includes African individuals and the category “coloured” refers to mixed race individuals.

7. These estimates were lower in 2010 and 2012, but remained very high with a respective 73% and 80% of respondents expecting to climb at least one rung of the ladder over the next 5 years.

8. We also investigate partial correlations between social aspirations, per capita income and a set of individual characteristics and geographical variables including race, age, gender, rural location and province. We acknowledge that because aspirations is the difference between current self-assessed position (on the imagined rungs of the ladder) and the future position, a higher current self-assessed position can contribute to lower mean aspirations, leading to a negative relationship between self-assessed initial position and aspirations. Intuitively, if individuals assess themselves to already be at the top end of the ladder, this limits their ability to express ambitions for further social mobility on this six-rung ladder scale. To ensure that we are incorporating such distortions, we include indicator variables for each of the self-assessed initial starting positions. It is important to acknowledge that this could overcorrect, i.e. also capture higher ambitions amongst those who initially see themselves at the bottom of social ladder. We also experiment with eliminating individuals assessing themselves to be at the top of the distribution initially (top two rungs of the ladder). We further limit the sample to individuals between 26 and 40 to consider whether these results hold for a subset where one would expect higher social mobility due to higher annual salary gains and lower unemployment probabilities following the first few years of work experience. These regressions are shown in columns 2 and 3 respectively of Table 7.

9. Achieved social mobility is measured by creating ladder rungs from household per capita income and measuring individual movements between 2008 and 2012 for the subset of individuals observed in both these waves. Expected social mobility is estimated as the expected number of ladder rungs that the individual expect their household to climb over the next five years in 2008 (i.e. between 2008 and 2013). We limit the sample to individuals between 28 and 50 and exclude those on the top rung (perceived top rung for expectations and actual top one-sixth for per capita income).

10. While South Africa saw a lowering of poverty over this period, this is generally associated with the introduction and expansion of the child support grant (Van der Berg, Louw, and Yu 2008).