Return to: Search Page or to: Table of Contents Vol. 20, Issue 2

Matthias Fatke, "Allure or alternative? Direct democracy and party identification," Party Politics, 20 (March 2014), 248-260. [Available at ]

First paragraph:
This article investigates whether and how party identification is influenced by direct democratic institutions. The concept of party identification is often praised as the most important discovery in explaining electoral behaviour (Green et al., 2002; Weisberg and Greene, 2003). Representing 'the holy grail of electoral research' (Dalton 2009: 628), it is shown to be one of the most consistent and influential factors of electoral behaviour ever since the early findings by Campbell et al. (1960). But emerging evidence that party identification has been suffering a steady decline in recent decades obviously begs the question: what is responsible for increased levels of dealignment among voters? (Dalton, 2007; Dalton and Wattenberg, 2000). Despite the great importance of the concept, there is in fact very little knowledge as to what influences and shapes party identification. Be it coincidence or not, at about the same time the dissemination of direct democratic institutions and procedures, too, notably grows throughout the world (Butler and Ranney, 1994; Matsusaka, 2005; Scarrow, 2001; Schmitter and Trechsel, 2004).

Figures and Tables:
Table 1. Random-intercept logit models of party identification.
Figure 1. Predicted probability of party identification given direct democracy.
Figure 2. Moderating effect of popular votes.
Figure 3. Robustness test through manual jackknife procedure.
Figure 4. Party identification in bilingual cantons. Note: Numerals in parentheses indicate number of respondents.
Table 2. IV probit models of party identification with index of direct democracy instrumented by share of peasants.

Last Paragraph:
(first paragraph of conclusion) The concept of party identification plays a major role in the study of electoral research (Green et al., 2002; Weisberg and Greene, 2003). It is of central interest to political science to know why and to what extent individuals affiliate with political parties (Dalton, 2007). But despite the phenomenon of declining party identification and increasing dealignment among voters (Dalton and Wattenberg, 2000), little systematic evidence exists as to which factors influence individual party identification. Our article contributes to improving on this lacuna by considering the educative effects of direct democracy (Smith and Tolbert, 2004). Thereby, it represents the very first analysis of the influence that direct democratic institutions exert on party identification. Theoretically, two competing hypotheses sound plausible. On the one hand, direct democracy might strengthen political parties and promote the need for cues so that voters succumb to the allure of partisan attachment. On the other hand, direct democracy might provide an alternative to the representational function of political parties thus rendering party identification less essential. Our empirical analysis of individual data from the Swiss cantons gives clear support to the alternative hypothesis. In Switzerland, individuals are significantly less likely to feel close to a political party if they have permissive direct democratic institutions at their disposal. Surprisingly, however, this effect diminishes if individuals are constantly exposed to popular votes.

Last updated March 2014