First-order, second-order or third-rate? A comparison of turnout in European, local and national elections in the Netherlands

•We examine whether perceptions of how much is at stake in an election affect turnout.•We compare three different elections (European, local and national) using Dutch panel data.•Aggregate results support the ‘less at stake’ argument of the second-order theory: higher stakes respond to higher turnout.•Individual level results only partially support the ‘less at stake’ argument, suggesting a context-dependent impact.•Campaign exposure has a positive effect on turnout, and the effect is greater in second-order campaigns.

Second-order elections are characterized by low turnout. According to the second-order theory this is because people feel there is less at stake. This study tests whether the less at stake argument holds at the macro and micro level using panel survey data obtained in three different Dutch elections. Furthermore, it examines whether campaigns' mobilizing potential differs between first- and second-order elections. We find that at the macro level perceived stakes and low turnout go hand in hand and differ strongly between national, local and European elections. At the micro level the impact of perceived stakes on turnout is limited and contingent on the type of election. Also, campaign exposure affects turnout, but the effect is substantially larger in second-order contests.