The back and forth last week between me, Curt Gans, and Barry Burden about early voting and turnout on Rick Hasen’s election law listserv was a useful discussion of how various reforms have or have not increased turnout.
Michael Hanmer of University of Maryland is coming out with an excellent book on the subject, and I’ve had another email exchange with him on the same subject. He’s agreed to let me share it here:
It just so happens that I have a forthcoming book that examines why most election reforms, even election day registration, have relatively modest effects on turnout.
I took a look at the Demos document and they take state turnout (using Michael McDonald’s data) for ID, IA, ME, MN, MT, WI, WY, and NC compared to the rest of the states, finding a 7 point difference in the favor of the EDR states. I didn’t check their calculation and assume it is correct but I have two concerns with what they have done.
First, this sort of analysis ignores the fact that these states, for the most part, already had high turnout rates before they adopted EDR. In my forthcoming book (hopefully next month), using a difference in difference approach I estimate that EDR had about a 4.5 point effect in MN and WI that was durable but find no real increase in the states that adopted EDR to avoid the NVRA (ID, NH, and WY). I use individual level CPS data so I could not examine IA and MT as the data were not available when I was writing.
Second, if one takes the approach Demos does then I think ND should be included and NC excluded as NC only has EDR during the early voting period; Demos acknowledges the details of the NC law but treats it the same as EDR in the other states.
Here is the citation to my forthcoming book: Hanmer, Michael J. 2009. Discount Voting: Voter Registration Reforms and Their Effects. New York: Cambridge University Press.