The Impact of Voter Affective Polarization on Elite Ideological Polarization

Elite ideological polarization is a growing trend in Western democracies. Yet, despite being scrutinized by students of comparative politics, accounts for divergent elite positions remain debatable. This paper identifies mass-level affective polarization as a potential explanatory factor for elite-level ideological polarization. The underlying mechanism by which voter-level affective polarization fosters party-level ideological polarization, I claim, is by weakening one of the basic voter behavior assumptions. That is, that voters give considerable weight to ideological proximity in their voting considerations. According to the median voter theorem, to maximize their share of the vote, office-seeking parties should promote policies that aim for the median voter preferences. While parties that choose to put forward more extreme platforms will gain a smaller share of the vote. However, if the electorate is affectively polarized, the relationship between voters and parties changes. Much like sports fans, affectively polarized voters develop an expressive partisan identity that is based on an emotional connection to their preferred party. Following this, I argue that affectively polarized voters attribute less weight to ideology, penalizing parties less for ideological distance. This is because disliking other parties makes one more willing to tolerate ideological distance from one’s "own" party. Consequently, at the country level, when the electorate is affectively polarized, parties are not as electorally penalized when they position themselves away from the median, and we would see more polarized party systems. To empirically examine this contention, I conduct a two-step analysis, drawing on CSES data (Module 5, 2016-2021). In the first stage, I demonstrate that highly affective polarized voters tend to penalize parties for an ideological distance less severely compared to voters with lower affective polarization levels. To show this, I run a multinomial vote model for each country separately for 29 country-election cases, and analyze the marginal effect of ideological distance for low and high levels of affective polarization. In the second stage, I run a country-level secondary analysis that uses outputs from the multinomial regressions to show that high levels of voter affective polarization are associated with high levels of elite ideological polarization. Preliminary findings support my argument.