How to analyze second-order election effects? A refined second-order election model

Abstract

The second-order election (SOE) model assumes that voters in subordinate elections tend to turn out in lower numbers and support opposition, small and new parties to the detriment of parties in national government. This model has been successfully applied to European and subnational elections taking place in Western Europe but it fares far less well in explaining electoral outcomes in Central and Eastern Europe (CEE). I refine the model by differentiating between six party types (large and small government and opposition parties and new and no-seat parties) and by introducing the state of the economy (economic growth, inflation and unemployment) as explanatory variable instead of time elapsed between first- and second-order elections. An analysis on 488 regional elections taking place in 6 CEE countries shows that second-order effects relate to the state of the economy that has a different impact depending on party type. These results strongly suggest that regional elections in CEE are second-order but in order to be able to trace SOE effects the SOE model needs to be refined.