by Karl Kurtz
What can state policymakers do to increase voter turnout? The answer is, "Not much," because voting is more a question of motivation on the part of citizens than it is of rules and structures provided by the legislature. But there are some things that states have done or could do that might increase voter turnout, at least on the margins.
I was asked last week to lead a discussion with NCSL's Redistricting and Elections Committee on this subject. To prepare for that session, I reviewed a lot of recent political science literature on voter turnout and thought that I would share the results here in The Thicket.
The first point is that registering to vote and casting a ballot are much easier acts today than they were a generation ago. Widespread reforms like the federally enacted "motor voter" law and state initiatives like early voting, "no excuse" absentee balloting, election day registration and Oregon's balloting by mail system have made it easier than ever to cast a vote.
Yet the rate of voter turnout has remained almost constant since 1972, averaging 56% of eligible voters in presidential elections and 40% (39.9% in 2006) in midterm elections. This is also contrary to the popular view that turnout has been declining.
Read more about turnout statistics and state policies that affect turnout after the jump.
George Mason University Professor Michael McDonald points out that most calculations of turnout rely on voting age population (VAP) as the base (denominator) on which turnout is popularly calculated. He argues effectively that voting age population, which includes non-citizens, felons and the mentally incapacitated (all of whom are not eligible to vote) and excludes those living abroad (who are eligible to vote), needs to be corrected and that turnout calculations should be based on voter eligible population (VEP) instead.
Using painstaking statistical methods (pioneered by my primary graduate school adviser, emeritus professor Walter Dean Burnham--sorry, I had to get that one in), McDonald shows that all of the supposed decline in voter turnout since 1972 is caused by increases in the proportion of non-citizens (about 9% of voting age population today) and felons (about 1%) in the population. You can find information on McDonald's statistical calculations, graphs of voter turnout over time, and his article (with coauthor Samuel Popkin) on "The Myth of the Vanishing Voter" here.
The Center for the Study of the American Electorate reaches similar conclusions about voter turnout statistics.
Nonetheless, voter turnout in the United States is lower than that of every democracy in the world other than Switzerland. There are a number of structural characteristics of American government that help to explain this fact. First, both our separation of powers system and our federal form of government mean that American voters have to cast votes for far more offices than do those in parliamentary systems with unified executive and legislative elections. Second, we are the rare (only?) country in which registration is the responsibility of the voter rather than the government. In most other countries that have national identification systems, your national ID card allows you to vote. In short, in academic lingo, the "cost" of casting a vote in the United States is greater than in other countries.
There's not much to be done about these basic characteristics of American government, but there are other elements of our electoral system that can be controlled by Congress and state legislatures. One of the most notable reforms of the electoral process was the "motor voter" law, more formally the National Voter Registration Act of 1993, This law, enacted on the basis of a Michigan model, requires states to allow voters to register at government agencies like drivers' license bureaus. Scholars like Georgetown University's Michael J. Hanmer (link to bio and email but no electronic version of his study) conclude that the motor voter law has had a slightly positive effect on turnout but not nearly as much as the advocates of the law anticipated.
Seven states have enacted election day registration (EDR) (sometimes called same day registration), which allows voters to register right up until the time they vote. Three of the states--Maine, Minnesota, and Wisconsin--enacted these laws in the 1970s as part of the general movement for civil rights and greater enfranchisement. Idaho, New Hampshire and Wyoming, though, passed their EDR laws in response to motor voter: a provision of that law allows states to opt out of providing registration at drivers' license and motor vehicle registration bureaus if they have EDR.
Georgetown's Hanmer has also studied EDR and concludes that, like motor voter, it does not have nearly as much impact on turnout as some claim. At best he says that this provision can increase voter turnout by about 3-4%. However, Mary Fitzgerald of James Madison University finds that EDR is particularly popular with young people and increases youth voting by 14 percentage points in presidential years and 4 points in midterm elections. Neither she nor Hanmer reconciles the differences in findings between young voters and all voters.
The seventh state to enact EDR was Montana in 2005, so this provision was in effect for the first time in last month's election. In an email message that I received soon after the election (but not posted on the sender Demos' web site), two former secretaries of state from Montana and Connecticut say that 3,700 voters took advantage of this opportunity in the 2006 election and that this made the difference in the narrow election of U.S. Sen. Jon Tester (not to mention control of the Montana legislature). The difficulty with this claim--and with assessing the impact of election reforms like these in general--is that there is no way of knowing how many of these 3,700 voters might have registered under the normal schedule if EDR had not been available to them.
The last category of election reforms that states have enacted to encourage voter turnout is early and absentee balloting. All 50 states provide for absentee balloting, but only 29 permit unrestricted (or "no excuse") absentees to vote. Early voting is distinct from absentee balloting in that voting is allowed at actual polling places using the same machines as on election day. Political scientists' assessments of these provisions is that their effects on turnout are at best ambiguous and probably have little or no positive effect.
Because this is already an overly long post, I will save some additional suggestions for what states can do to improve voter turnout for a second post on this subject.
[Photo courtesy of Flickr.]