Electoral Systems and the Contest for Relative Group Status in Post-Conflict Societies: The Case of the 2009 Election in South Africa

Abstract

Empirical support appears to be growing for the claimed effectiveness of proportional electoral systems in contributing to lasting peace in post-conflict societies. We raise two objections against this claim. Firstly, we argue that the theoretical rationale for this causal effect, as presented in consociational theory, is inadequate, in lacking a psychological mechanism for moderation. We argue that the capacity of constitutional arrangements to deliver parity of esteem amongst competing political groups should be added as a criterion for evaluating the effectiveness of constitutional rules in maintaining political stability. Secondly, we challenge the empirical claim that South Africa should be considered as an example which supports the argument for Party-list Proportional Representation (PR) as a conflict-regulating rule. Using data from the 2009 general election in South Africa to illustrate the nature of the engagement between ruling and opposition parties, we find that despite the entrenchment of the proportional electoral system, there is a disparity of esteem in political relationships. This necessitates a reconsideration of the purported value of PR as a conflict-regulating constitutional device.

Notes

This growing endorsement still falls short of a so-called ‘scholarly consensus’, as has been claimed by Lijphart (1991).

See Yin (2009, pp. 15–20) for a discussion on the role of case studies in social analysis and particularly Yin's description of the suitability of case studies in the expansion of theory and in the explanation of ‘presumed causal links in real-life interventions that are too complex for survey or experimental strategies’.

Post-war elections do appear to have been successful in Namibia in 1989, in Mozambique in 1994, in El Salvador in 1994 and in Burundi—which is considered to be one of the African states which most closely followed Lijphart's ‘consociational formula’ (Reilly, 2008, p. 158; Lemarchand, 2006, p. 7). However, the following examples support the argument that elections based on PR do not guarantee democratic stability: the renewed instability following the 1997 elections which brought Charles Taylor to power in Liberia (Reilly, 2008, p. 158); the lack of political tolerance and stability in Rwanda—despite regular elections and PR in government (Belloni, 2008, p. 201; Turner, 2007, pp. 1050–1051; Reilly, 2008, p. 173; Lemarchand, 2006, p. 7; Guelke, 2009, p. 108); protracted conflict between the Greek and Turkish population in Cyprus and the withdrawal of Turkish Cypriot members of the House of Representatives under a constitution providing for elections based on PR (Taylor, 2007, pp. 375–376); continued threats of insurgent violence and violent clashes during election periods in the Democratic Republic of the Congo (Taylor, 2007, pp. 347–348; Lemarchand, 2006, pp. 12–14; Tull and Mehler, 2005, p. 375).

The case for the application of social-psychological theories in the analysis of relationships and conflict between identity groups are made by numerous authors, including Horowitz (1985, 2001), Huddy (2004), and Monroe, Hankin and Van Hechten (2000). The general features of one such sub-field, that of social identity theory, are summarized in Brown (2000), Duckitt (2003) and Turner and Reynolds (2001). Definitive contributions in the field are made by Tajfel (1981) and Tafjel and Turner (1979).

The concept of ‘parity of esteem’ was first introduced into the literature on peacemaking in divided societies when it was used in the 1998 Belfast Agreement. It was not clearly defined within the text and was promptly politicized by politicians from both communities in their ongoing intergroup rivalry. See MacGinty and Du Toit (2007) for an overview.

For a detailed analysis of South Africa's election results in the first three democratic elections and the importance of race as a perceived basis for exclusion from specific parties, see IDASA (2003) and Ferree (2006).

For discussions on the application of the consociational approach in the South African context, see Lijphart (1998), Maré (2001), Norris (2008), Pottie (2001), Reynolds (1997) and Sisk (1993).

Whilst the motivations behind the racial basis of political support are questioned by some commentators, there is general agreement that South African electoral outcomes do display some important characteristics of a ‘racial census’ (see Ferree, [2006], McLaughlin [2007] and Sylvester 2009).

This absence of protection for minority representation has to be interpreted in light of the country's history of minority government and the fact that the political institutions designed during the country's democratic transition were aimed at introducing political participation to the previously disenfranchised black community.

Also see the discussion on religion, evil and liberalization in du Toit and Kotze (2011, pp. 133–58).

Additional information

Notes on contributors

Frouwien Bosman

Senior Lecturer, Department of Political Science and Development Studies, University of Stellenbosch and Huguenot College, South Africa.

Pierre Du Toit

Professor, Department of Political Science, University of Stellenbosch, South Africa. Email: [email protected]