Does executive autonomy reduce second-order election effects?

The second-order election (SOE) model expects voters to punish parties in national government and reward opposition, small and new parties because there is 'less at stake' in an SOE. One key assumption that is rarely studied is whether SOE-effects depend on the extent to which voters can impact the selection of the executive in an SOE. This article argues that executive -autonomy - i.e. the extent to which executives are independent from the parliament regarding their formation, termination and execution of their competences - increases the impact of authority. Executive autonomy reduces SOE-effects when authority is high but increases SOE-effects when authority is low. An empirical analysis of 41,603 vote share swings for 4733 parties competing in 2665 elections held in 282 regions in 14 European countries between 1945 and 2019 confirms the hypotheses. These results have important implications for electoral democracy and party competition at the regional, national and European level.