Does motor voter work? Evidence from state-level data

Using time-series cross-section data on state-level registration and turnout, "motor voter" registration programs mandated by the National Voter Registration Act are found to be effective in enhancing voter participation in the states that have already adopted them. A "duration" specification for motor voter is introduced as an alternative to a simple dummy variable specification, as the typical length of states' driver's license renewal cycles implies that not all drivers have had an opportunity to register via motor voter until several elections after implementation of the program. Registration effects are estimated to reach about 13 percentage points, with a turnout impact of roughly half of that level. As very few states to date have "mature" programs in effect, most of this potential impact of motor voter has yet to be realized. Little evidence is found that mail-in or agency registration, also mandated by the NVRA, affect participation rates.