Towards a silent revolution?

South African voters during the first years of democracy 1994 - 2006

PhD Thesis by Collette Schulz-Herzenberg
University of Cape Town (UCT)
March 2009

Professor Robert Mattes
Professor Jeremy Seekings


Click here to download the PhD thesis in PDF format.



Although South Africa has established a functioning democracy scholars continue to express serious concerns about the quality and stability of its new democratic dispensation and the potential for its consolidation. In particular, they question the extent of genuine multi-party competition due to the electoral imbalances in the political party system and the static nature of voting outcomes that have characterised South African politics since 1994.

South Africa has undergone rapid socio-economic and political transformation between 1994 and 2006, factors that many electoral analysts argue may change 'frozen' cleavage structures and electoral behaviour. These processes may therefore have important consequences for the partisan responses of social groups, the subsequent composition of partisan coalitions, and the behavioural motivations of individual voters.

This study consists of a longitudinal analysis of change in first, the level and direction of partisanship among the entire electorate, second, the social groups within the ANC's partisan coalition, and third, the motivations of individual voters by comparing different electorates over twelve years (1994-2006), using a series of cross-sectional surveys. I explore the relative influence of competing theoretical models on partisanship such as demographic and sociological factors, evaluations of government performance, party images, social networks, and the voter's cognitive skills. I also explore the relative mix and intersection of these models on partisanship. Findings are based on a range of statistical analyses such as cross-tabulations of structural and demographic variables, bivariate analyses and multivariate data analyses (logistic regression).

South African voters, much like voters elsewhere, respond to a multitude of short and long term factors. During the first decade of democracy the various theoretical models all contribute to our knowledge of partisanship in South Africa. However, they are not of equal importance. While multiple factors drive partisanship, sociological factors and party images dominate voter choice. Yet, subtle but potentially profound shifts in the motivations of voters appear to be gaining ground. Performance evaluations play an increasingly more important role over time. Cognitive awareness indirectly affects voter choice by emphasizing different types of information, but still plays a minimal role in voter behaviour.

There are several reasons to be optimistic that electoral fluidity might increase among the South African electorate. The first reason is that the overall impact of sociological factors (including race) on partisanship has declined during the first decade of democracy. Fixed social cleavages therefore hold less sway than ever before. Furthermore, South African voters do not appear to base their partisan stances on the basis of primordial racial or ethnic loyalties regardless of political performance. Instead, I find that the indirect influence of race is complex and feeds into voter decision-making via multiple channels, namely through party images and evaluations of government performance. Voters also appear to look at short term issues such as retrospective economic evaluations and future economic prospects, as well as party and candidate characteristics such as trustworthiness. I find that the influence of government performance evaluations on partisanship has increased over time, indicating that partisan choices are increasingly motivated by short-term issues. These factors are capable of producing partisan change or fluidity within the electorate, which should improve the quality of democracy in the long run as incumbents are forced to consider popular opinion. Finally, the fluctuations in partisan alignment, and the clear dealignment trend among opposition supporters, should free more voters to move their partisan support to new parties at elections.

However, several of my findings hold negative implications for electoral politics. The first concerns party images. Voters' images of the racialised nature of political parties remain quite rigid even after twelve years of democratic elections. Voters are unlikely to change their partisan allegiances because they see 'other' parties to be exclusive of their interests, or are uncertain of who they represent. In addition, the social environments of many voters generate homogenous partisan signals that reinforce existing partisan proclivities and perceptions about the racial exclusivity of political parties. In addition, opposition parties have failed to broaden their appeal among the electorate. On the whole, they remain weak and ineffective. Together, these factors are bound to minimise partisan fluidity. The chance for significant realignment is diminished and electoral predictability increases.

Finally, the study notes several concerns for the quality and stability of democracy. Any increases in dealignment is likely to have negative effects on the quality of South Africa's democracy as people become less inclined to identify with or support any party at elections. A pre-dominant party system will continue into the future but its electoral margins will probably be based on decreasing percentages of the overall voter population. The second concern relates to the continuity of a predominant party system. Elections are unlikely to act as a meaningful vehicle for popular control of government if voters are unwilling to change their partisan allegiances. In the face of continued one-party dominance, I conclude that the greatest challenge for democracy during the next decade will be maintaining high levels of incumbent responsiveness and accountability towards citizens.


Table of Contents

Glossary & Acronyms
List Of Tables
List Of Figures
List Of Graphs

Chapter 1: A Decade Of Multi-Party Elections And Social Change

Chapter 2 Theoretical Perspectives On Voting Behaviour

Chapter 3 The Central Argument

Chapter 4 Research Design And Methodology

Chapter 5 The Partisan Responses Of Social Groups

Chapter 6 The Influence Of Sociological Factors

Chapter 7 The Influence Of Performance Evaluations

Chapter 8 The Influence Of Party Images And Social Networks

Chapter 9 Cognitive Mobilisation

Chapter 10 Towards An Integrated Model Of South African Voter Behaviour

Chapter 11 Conclusions: Voter Behaviour And Prospects For Multi-Party Democracy

Appendix 1 National Public Opinion Surveys
Appendix 2 The South African Population (Census 2001)
Appendix 3 Additional Data For Chapter 5
Appendix 4 Additional Data For Chapter 8
Appendix 5 Additional Data For Chapter 10




Click here to download the PhD thesis in PDF format.

Proudly hosted by Skipjack IT Solutions.