“Second order” Effects and the determinants of voting decisions in German federal state elections: an analysis of voting behavior in the German multilevel system

Studies on mid-term elections in multi-level systems show that it is used by the voters to get in on the crucial - punish ruling party level - mostly national. This applies also tend to state elections in the Federal Republic of Germany. With few exceptions, empirical studies to test this relationship using aggregate data. In this paper we argue that on an aggregate level often observable "second-order" effect should be the extent attributable to the individual choice decision, as should have a smaller effect on the choice of participating in the federal government parties "classical" determinants of voting behavior in state elections the greater the distance to the next federal election time. Moreover, it is argued that interest should select oriented gain importance when the result of a state election has a significant impact on the distribution of seats in the Bundesrat. We test these arguments using a data set that includes all state election studies 1985 and 2009. The analyzes show - even if one controls for a "individualization" of voting behavior - no empirical evidence for the expectations at the individual level is not a "second-order" effect nor a federal political penetration of voting behavior in state elections observable.