Separated powers, divided government, and turnout in US presidential elections

Theory: Turnout is determined in part by the motivations of voters to affect the course of public policy. Institutional arrangements that condition the link between the vote and its policy consequences impact these motivations. One such institutional arrangement is the separation of powers. Countries with separated powers should see lower turnout, and anything that causes the extent of separation to vary should give rise to corresponding variations in turnout. Hypotheses: Divided government, by temporarily increasing the extent to which powers are separated, will reduce the motivations of voters to participate in subsequent elections. Methods: A time-serial model of turnout is developed and tested on data for U.S. presidential elections from 1840 to 1992. Findings: Divided government does influence turnout, controlling for changes in electoral laws and the closeness of elections. Moreover, the effects are cumulative, with turnout declining by about 2% for each consecutive presidential election conducted under conditions of divided government. This finding may help to explain declining U.S. turnout in recent years. More importantly, it validates the conjecture that the separation of powers itself has a depressing effect on turnout, which helps to explain why turnout is lower in the U.S. than elsewhere.