THE STRATEGIC TIMING OF DIRECT DEMOCRACY

This paper focuses on the strategic timing of elections by agenda-setters in direct democracy settings. Because concurrent elections affect turnout, scheduling referenda for different elections will produce different median voters. I hypothesize that agenda-setters with power over the timing of a referendum will schedule the referendum in conjunction with the other set of races that produce a policy closest to their preferred outcome. Consistent with the theory, I show that Wisconsin school boards' use of special elections for school referenda are related to differences in the revealed preferences of voters in low- and high-turnout elections.