How Different Party Electorates Shape the Magnitude of the Second-Order Election Effects: Comparing the Regional and National Elections in Poland

Despite several limitations, the second-order election model has been the default starting point used to compare different electoral arenas. This was the case for the European Parliament elections and for the regional ones compared to first-order national elections. Aggregate electoral data were typically used to assess the second-order nature of elections. Although the absolute distinction between first- and second-order elections is debated, with some parties strategically prioritising second-order elections, the electorate's perspective on this issue remains largely unexplored. Therefore, our goal in this paper is to shed light on how different party electorates shape the magnitude of second-order election effects. We do it by analysing the correspondence between the recent national and regional elections in the example of Poland. Our dataset comprises official electoral statistics obtained from Poland's State Election Commission (PKW), integrated with a comparable, original post-election survey administered roughly one month following the national and regional elections. The analysis controls for several factors, such as economic voting, previously argued to moderate second-order effects on regional elections across Central and Eastern Europe. Our findings emphasise that the size of the locality plays a crucial role in explaining the magnitude of second-order effects, with smaller localities exhibiting turnout in the second-order election closer to that of the first-order election. We show how and why the importance of regional elections varies between different party electorates. In general, our contribution provides valuable conceptual and empirical insights into second-order election effects and the nature of multi-level electoral politics.