Election Day 2012 is going to look a lot like Election Day 2008

Consider these results from a recent Pew Center poll. In this poll, Obama is 8 points ahead of Romney, close to his victory margin in 2008 (7 points). But what is especially fascinating in this poll is its internals—how Obama is faring with key subgroups of voters. Start with minorities. Obama gets 93 percent of the black vote (he got 95 percent in 2008) and 79 percent of minorities overall (he got 80 percent in 2008). (The poll does not provide data on Hispanics, but the two most recent national polls of Hispanics give him 67 percent of these voters, identical to his 2008 performance.)

He also gets 44 percent of the white vote, compared to 43 percent in 2008. Moreover, if you break the white vote down by working class and college-educated, his performance is even more impressive. Among white college-educated voters he ties Romney 49-49, compared to the 4 point deficit he ran against John McCain, and loses white working voters by only 41-55, compared to his 18 point deficit against McCain.

These demographics are excellent news for Obama. But what of the longer term? Here we must depart the realm of demographics and current polls and look to election forecasting models. The best way to sort them out is not to focus on one particular model—there are so many to choose from!—but rather to look at the factors driving these models and what they have to tell us. The one factor that stands above the rest is economic performance in the election year. Moreover, there is general agreement that what matters most is change in economic performance, not the absolute level of performance. Thus, change in the unemployment rate in the election year is more important than the level of unemployment, GDP growth is more important than the size of the economy, and so on.