Sincere or second-order? Exploring sincere and strategic voting behaviour in regional elections

For a long time, the dominant framework to differentiate between the national and regional electoral arena has been the second-order election (SOE) approach. Following the SOE model voters treat non-national elections differently because there is less at stake since the outcomes of these elections do not have a direct effect on national level policy making. Given that there is less at stake during these regional elections, voters who might cast a strategic vote during national elections, are more likely to cast a sincere vote during regional elections. Especially small and new parties should benefit from sincere voting during regional elections. Even though this is one of the central propositions in the SOE model, the hypothesis that voters are more likely to cast a sincere vote in regional elections and cast a strategic vote during national elections has received little academic attention. Most attention has been paid to ‘protest’ voters who use a second-order election to voice a signal of discontent to the parties in national government by voting for a party that is in national opposition. Furthermore, voting for new parties in SOEs is also explained by ‘experimental voting’ which is less ‘risky’ in non-important elections. In other words, we do not know if voters express party preferences during regional elections, if they cast a protest vote based on national politics, or if they would like to give new parties a chance. This study examines the conditions under which voters are more likely to vote less sincerely (and more strategically) in national than in regional elections. This study applies a quasi-experimental approach to study strategic voting behaviour based on a set of strategically selected regional and national election results in 15 Western European countries between 1950-2020. The results imply that previous studies have overestimated the extent to which regional elections are conceived by voters to be second-order.