It takes a while for our states to tally up the total number of votes cast in the presidential election, so only now is it possible to say something reasonably close to definitive about turnout in the 2008 general election. Overall turnout, as currently reported, seems to be 130.7 million, about 7 percent more than the 122.3 million of 2004. That's a much smaller percentage increase in turnout than between the 2000 and 2004 elections, which was 23 percent. Turnout increase in 2004-08, at 6.9 percent, was higher than population increase 2003-07 (as estimated by the Census Bureau and which I take as a reasonable proxy for the population increase from which increased numbers of voters could be drawn) of 3.9 percent, but not by much.
One of the things that have struck me as I have been crunching numbers from the 2008 election returns is how variable turnout increases/decreases were. In states that were seriously contested in 2008 but not 2004, turnout tended to be way up (example: Indiana, plus-12 percent, in a state with 2.6 percent population increase), but in states that were seriously contested both years, turnout was not up and, in one significant case, was down (example: next-door Ohio, minus-3 percent, in a state with a 0.3 percent population increase).
So, I decided to try to quantify these developments in what is probably an oversimplified way for each state by subtracting from the percentage increase (or decrease) in total voters 2004-08 the percentage increase (or decrease) in estimated population 2003-07. It might be better to use the percentage increase or decrease in estimated population 2004-08, but the Census Bureau isn't releasing the 2008 population estimates until later this month.
Here's a table that shows some of the results, with the percentage of the popular vote that identify themselves as black for each state (2007 figures for those who say their race is wholly or partially black), with appended comments.
Turnout 2004-08 minus population increase 2003-07, with black percentage
|Turnout Black percentage|
|District of Columbia||14.9||56|
The obvious story here is that a combination of enthusiasm for Barack Obama among black voters and/or massive and sophisticated organizational efforts by the Obama campaign produced substantial increases in black turnout in states with large numbers of black voters that were not targeted in 2004. North Carolina, by the way, had the largest turnout increase in the nation in 2004-08, 23 percent—which was the same as the national increase in 2004-08. My preliminary examination of county returns from Virginia, North Carolina, and Georgia shows that the Obama campaign (and/or spontaneous enthusiasm) resulted in marked increases in black turnouts in counties that had never seen serious Democratic organizational efforts before. Much credit is due to the Obama campaign, which eked out 1 percent victory margins in North Carolina and Indiana. But one must add that spontaneous increases in black turnout are evident in the numbers for jurisdictions that no one thought were target states—D.C., South Carolina, Mississippi, Alabama, Maryland. About Rhode Island, you can conclude what you will.
In these states, the numbers are less spectacular and mostly less dispositive. The Obama victory in New Mexico seems to have owed much to a shift toward the Democrats among Hispanics, who form nearly 40 percent of the state's electorate—and who for the most part are not immigrants or people with recent immigrant family backgrounds. They liked George W. Bush pretty well; John McCain, not so much. The surges in turnout in Montana and Missouri came close to, but not quite up to, delivering 14 more electoral votes for Barack Obama.
I have omitted the states where the numbers come out between 0.1 and 3.9, on the theory that turnout there increased roughly in tandem with population increase, and so these numbers don't tell us much one way or the other about increased or decreased black or other turnout. And then, there are the states that come out negative or zero when you subtract percentage population increase from percentage turnout increase—in other words, where turnout increased less than population.
Some are states where turnout has long been among the nation's highest levels—Minnesota and Idaho, for example. Many are hopeless for the Democrats (Wyoming, Alaska, Arkansas, and Kentucky in this election; Nebraska, Oklahoma, West Virginia, Utah) or, after mid-September this year anyway, for the Republicans (Minnesota, Wisconsin, Oregon). South Dakota had the rip-roaring Senate race between Tom Daschle and John Thune in 2004 and no Senate race this year; hence (in my guess) the relative decline in turnout in a state people figured would go Republican for president both times. The one puzzler is Ohio, a hugely targeted state in 2004 and 2008. My working hypothesis: Republican turnout, pretty high in 2004, was depressed in 2006 and 2008 by disillusionment not only with Bush and the national Republicans but with former Gov. Bob Taft and the Ohio state Republicans, who had controlled state government for 16 years, which turned out mostly to be years of economic and fiscal decline. In Ohio, Obama got a few more votes than John Kerry; John McCain, many fewer than George W. Bush. Ohio alone among states targeted in both 2004 and 2008 had fewer people turning out to vote this year. I see this mostly as a state where 2004 Republicans were no longer motivated to turn out—although in some Appalachian and working-class counties, you could also see (in smaller numbers total) 2004 Democrats who were no longer motivated to turn out.
A final thought: Americans seem to be getting more adept at tactical voting—or so my hypothesis goes. Twenty years ago, when there were many fewer statewide polls in the presidential race, we had many more states that were closely divided and less divergence in turnout increase/decrease among the states—at least, so my memory goes. This year, with many more state polls, and with voters thus much more able to know whether their states were seriously in play or not, we seem to have had much more differences in turnout increases/decreases—whether because of the campaigns' organizational efforts or because of potential voters' spontaneous decisions of whether to vote or not, or because (as I suspect) of some combination of both of the above.
Americans, in other words, seem to be engaging more in tactical voting/nonvoting. Better information produces, in this theory, different turnout patterns across state lines. In this situation, a key to turning defeat into victory is to make a serious effort to expand the electorate in states that were not previously seriously contested, as the Obama campaign did in Virginia, North Carolina, Indiana, and (not quite successfully) Georgia.
- Read more by Michael Barone.