Change in Opinion, Change in Vote? Proximity Considerations and Electoral Volatility

While scholars generally agree that levels of electoral volatility are rising in most established democracies, the implications of this trend are much more debated. On the one hand, optimists argue that rising levels of volatility are a consequence of high educated and high politically interested voters making up their mind more freely. On the other hand, pessimists fear that vote-switching is mostly driven by a lack of interest and knowledge about politics, leading to uninformed and random switching. Despite a large literature on this fundamental question, the debate is still ongoing. In this paper, I suggest that for studying this question and for settling this debate, we should investigate how stable and volatile voters respectively make up their mind when deciding what party to vote for. More specifically, I focus on the importance of proximity considerations in guiding the vote choices of stable and volatile voters. When investigating whether proximity between parties’ and voters’ issue positions explains whether or not they switch parties from one election to another, it furthermore is important to take into account the possibility that voters’ opinions change as well. Proximity voting has been indicated to be one of the determinants of voting behaviour that allows for democratic representation (Przeworski, Stokes and Manin, 1999). Indeed, democratic theorists argue that it is essential that the electorate expresses its preferences during elections by casting a vote for a party that takes positions that are proximate to their own issue positions. If volatile voters achieve a higher level of proximity than stable voters, the obvious conclusion would have to be that volatility contributes to this ideal of responsive government. Investigating the role of proximity voting on party switching and doing so in a dynamic fashion, requires high quality data that include both measures of opinions on issues as well as the party choice at different points in time. The data of the Belgian Election Panel (BEP) meet these demands, as they consist of a panel dataset spanning two consecutive election cycles, including the question which issue is the most important for the voter to make up her party choice, and which party she thinks is best in handling this issue. This combination of questions enables to investigate whether changes in opinions lead to changes in votes, and thus to provide new insights in the debate between optimists and pessimists of electoral volatility.