The Anxious and the Climbers: Ambivalent Attitudes towards Democracy among South Africa’s Middle Class

Abstract

Albeit middle-class citizens have generally been assumed to support democracy, studies investigating this class’s political attitudes using survey data have produced mixed results. This paper argues that one reason for these ambiguities is that the middle class may not be a homogenous group. Specifically, I explore how diverging perceptions of social mobility tend to condition political attitudes towards democracy within different social strata, particularly within the middle class. Drawing upon data from the South African Social Attitudes Survey, the article finds no strong evidence of any middle-class particularism in people’s opinions about democracy. Rather, the main difference in democratic attitudes and civic values is detected between those persistently locked in poverty and the rest of the population. Interestingly, the upwardly mobile ‘escapers’ and the ‘climbers’ show significantly higher levels of trust in public institutions and tend to display higher beliefs in voting as an effective tool to influence politics compared to both the persistently poor and the ‘anxious’ middle class, between whom no significant difference is found. These empirical findings demonstrate the limits to understanding people’s political attitudes in relation to their current standard of living alone.

Acknowledgements

This work was supported by the Evangelisches Studienwerk e.V. through a doctoral studies grant. The author is grateful to Jann Lay, Stephan Klasen, Jan Hofmeyr, Sebastian Prediger, Lena Giesbert, and Rocco Zizzamia for their many helpful suggestions, and to Crain Soudien, Derek Davids, and Benjamin Roberts from the Human Sciences Research Council (HSRC) for valuable feedback and data support. Useful comments and discussions on earlier drafts of this paper were also provided by participants at the 21st Spring Meeting of Young Economists 2016, the 2016 GIGA Workshop on Inequality and Middle Class Development in Africa, and the 2016 Development Economics Conference ‘Göttinger Schule.’ The views expressed here, and any remaining errors, are the authors’.

Disclosure statement

I confirm that there are no known conflicts of interest associated with this publication.

I confirm that I am the single author of this manuscript and that there are no other persons who satisfied the criteria for authorship but are not listed.

I confirm that I have given due consideration to the protection of intellectual property associated with this work and that there are no impediments to publication, including the timing of publication, with respect to intellectual property.

Supplementary materials

Supplementary Materials are available for this article which can be accessed via the online version of this journal available at https://doi.org/10.1080/00220388.2020.1802007.

Data availability statement

The main source of data for this study is the South African Social Attitudes Survey (SASAS) conducted by the Human Sciences Research Council (HSRC). The SASAS data is public access and can be downloaded at:

http://www.hsrc.ac.za/en/departments/sasas/data.

As a supplementary data source, the South African National Income Dynamics Study (NIDS) implemented by the Southern Africa Labour and Development Research Unit (SALDRU) at the University of Cape Town has been used. The NIDS data is public access and can be downloaded at:

http://www.datafirst.uct.ac.za/dataportal/index.php/catalog/NIDS/about.

The empirical analysis presented in this study was performed using Stata 14.2. Codes will be made available by the author upon request.

Notes

1. Historically, Barrington Moore’s (1966) well-known formula ‘no bourgeoisie, no democracy’ rests primarily on the ability of the middle class to act independently of the state, specifically to form political coalitions that keep the land elite’s hegemonic ambitions in check. Accordingly, he suggests that only societies with a sufficiently strong bourgeoisie would become democratic, whereas societies where the land-owning elite is so strong that the emerging bourgeoisie has no other option but to enter into an alliance with them would turn into dictatorships (Acemoglu & Robinson, 2005).

2. To avoid confusion, I on purpose did not use the labelling of the class-sublayers suggested by Schotte et al. (2018). The reason lies in an important difference between the two approaches. Schotte et al.’s (2018) classification schema is based on people’s predicted chances of upward and downward social mobility. These are estimated using a dynamic model of observed poverty transitions. By contrast, my focus here is on people’s perception of their chances of upward and downward social mobility.

3. As the SASAS collects information on the South African adult population, all target shares have been calculated reducing the NIDS sample to all individuals aged 16 years and above. Following Schotte et al. (2018), I calculate the poverty headcount as the share of individuals with monthly per capita household expenditures below the upper-bound poverty line set at R834 in 2012 prices (StatsSA 2017). [Schotte et al. (2018) choose expenditure as the relevant welfare measure because it is generally assumed to provide a better approximation of permanent household income than the reported income.] Following this definition, 58 per cent of the NIDS adult population can be classified as poor. The index cut-off, which accordingly renders the bottom 58 per cent of the individuals in the asset-index distribution as poor, is set at a value of 0.53. [Note that this cut-off value is almost identical to the one that would have been obtained if the threshold value was calculated as the average asset index score of households falling in a R10 band around the poverty line]. Analogously, following Schotte et al. (2018), the elite threshold is set arbitrarily at two standard deviations above the mean per capita household expenditure. Following this definition, 4 per cent of the NIDS adult population can be classified as elite. The index cut-off, which accordingly renders the top 4 per cent of the individuals in the asset-index distribution as elite, is set at a value of 0.93.

4. The premise common to these approaches is that mass tendencies in individual-level attitudes have some relevance at the system level.

5. Lower levels of education had no significant effects.

6. Considering the occupational status, I furthermore find a significant difference between those in full-time employment and the unemployed, with the latter being significantly less likely to favour democratic rule.

7. In fact, 62 per cent of the escapers and 54 per cent of the climbers report feeling ‘very close’ or ‘quite close’ to the ANC, while the same is true for 49 per cent of the persistently poor, 31 per cent of the anxious and 13 per cent of the elite.

8. This implies that the existing political order may have an important effect on both people’s political views as well as the social class that they belong to. I am thankful to an anonymous reviewer for pointing this out.

9. Freedom House (2016) documents the tenth consecutive year of decline in global freedom.

Additional information

Funding

This work was supported by the Evangelisches Studienwerk Villigst e.V. through a doctoral studies grant. There has been no significant financial support for this work that could have influenced its outcome.