Contamination in Reverse: The Pandora’s Box of Partisanship and Economic Voting during the Global Financial Crisis

“You live or die as a government in the economic field” opined former Australian Prime Minister Bob Hawke. And while many studies of voter behaviour validate the idea of economic voting (e.g.: Lewis-Beck & Stegmaier 2013), the main critique remains that economic perceptions are highly contaminated by partisanship bias (e.g.: van der Brug, van der Eijk, & Franklin 2007) and thus economic assessments do not have any independent effect of their own. In this paper, we challenge the view that economic perceptions are endogenous and therefore subject to partisanship bias. We argue that partisanship in the 21st century is much less stable and is more liable to shift from election to election, thus raising the possibility that endogeneity concerns might be overblown. Further, we posit that, in times of severe economic shocks, such as the recent Global Financial Crises, partisanship is likely to shift substantially, and thus we expect to see a drop of party identifiers for dominant governing parties, a variable that is very robust otherwise. If this pattern is observed, a legitimate argument can be made that economic perceptions come first in the causal chain. To test our assumptions, we rely on the Comparative Study of Electoral Systems (CSES) data. We focus on countries before and after the GFC and using multilevel models, we find that the economic crisis resulted in a reduction in partisanship for governing parties but that its effect varied dependent on the economic context. Our results have implications for our understanding of economic voting and the GFC on voters.