Today, a poll found Obama leading by two points in the generally uncompetitive state of Arizona, where neither campaign has aired advertisements. Given that Obama's trailing nationally and locked in a tight race in states like Nevada and Colorado, Obama probably doesn't have a lead in Arizona, so let's just cut that argument out of the discussion.
While we're dealing with a surprising poll from a state where both sides can agree that Obama's probably not in the lead, let's take this moment to recall that occasionally we'll get unusual poll results and that there's nothing wrong with it. Keep this poll in mind next time Obama's up 14 in Wisconsin or down 7 in Florida.
But although Obama is unlikely to win Arizona, it's not hard to see how the state could be closer than many imagine, even in a tight national race.
1) If McCain wasn't from Arizona, Obama probably would have won Arizona in 2008.
Here’s the simple way to look at it: Obama improved by about 10 points over Kerry’s performance nationally and 14 to 16 points over Kerry’s performance in Nevada, New Mexico, and Colorado. Obama also improved by double digits in fast growing, Sun Belt metropolitan areas with large Latino populations similar to Phoenix, like Dallas, Houston, Las Vegas, San Diego, Riverside-San Bernadino, or San Antonio. If McCain hailed from another state and Obama made commensurate gains in Arizona, Obama could have lost by one percentage point or won by as much as six points.
Here’s the complicated way to look at it: McCain’s home state edge and moderate reputation earned him an outstanding performance among Latino voters. Obama won just 56 percent of Latino voters in Arizona, much like Kerry's in 2004. If Obama had won 67 percent of Arizona Latinos, as he did nationally, Obama would have won 48 percent of the vote without higher Latino turnout or any gains among other demographic groups. McCain performed unusually well among other demographic groups, as well. Obama performed worse than Kerry among Arizona whites, even though Obama gained 2 points among whites nationally and even more than that in Nevada and Colorado. If Obama won Latinos by the same extent that he did nationally and then simply did as well as Kerry did among other demographic groups, the state would have been a deadheat. Realistically, Obama would have made the gains among other demographic groups necessary to put him over the top, especially if a full campaign effort produced higher minority turnout, which was lower in Arizona than the other southwestern battleground states.
2) Demographics raise the possibility of a close race in 2012.
Caveat: this is a scenario for illustrative purposes, not a description of *will* occur.
A. Obama performs as well among non-white Arizonans as he did nationally in 2008, or perhaps even better among Latinos. Given that Arizona is ground-zero in recent immigration debates, it’s not hard to envision how Obama could do better among Arizona Latinos than he did nationally. Indeed, a recent Latino Decisions poll showed Obama with 80 percent of Latino voters and today's survey showed Obama at 77 percent. Realistically, Obama won’t do as well in the exit polls, which tend to show Democrats doing worse with Latino voters than Latino Decisions surveys. But if Obama does as well among non-white voters as he did nationally without any additional gains among Latinos, he would still hold 47.8 percent of the vote. If Obama did a net-10 points better among Arizona Latinos than he did nationally (72 percent in the exit polls), Obama would hold 48.6 percent of the vote in Arizona. And if the exit polls actually showed Obama winning nearly 80 percent of the vote, then Obama could fight to a true tie (about 49.5 percent of the vote) with gains among Latino voters alone.
B. Obama maintains ’08 support among white voters or increases non-white turnout by enough to compensate for losses among whites. Both halves of this condition are possible. The possibility that Obama maintains his support among Arizona whites, despite losing white voters nationally, is more plausible than it is elsewhere since Obama actually did worse among Arizona whites than Kerry did in 2004, presumably because of McCain’s home state advantage.
But if Obama lost additional support among Arizona whites (which also seems quite possible) Obama could make up ground with a more diverse electorate. While Arizona does not track voter registration by race or ethnicity, the U.S. Census suggests that Latinos now represent 18 percent of Arizona’s registered voters, up from 14 percent four years ago. Using U.S. Census data has limits, but the Latino share of registered voters has probably increased over the last four years and it could help make up for losses among non-Latino voters. But a two point increase in the Latino share of the electorate is necessary for Obama to compensate for a one-point loss in support among white voters, so he can’t afford for his support among white voters to fall far from forty percent.
3) Obama’s decision not to contest the state suggests that Arizona probably won’t go his way
The Obama campaign toyed with competing in Arizona and even reconsidered the state in September, suggesting that they were intrigued by a case probably not too dissimilar from the one advanced above. But whether it was insufficient funds, insufficient Latino enthusiasm, or a big drop in support among Arizona whites, the Obama campaign appears to have concluded that Arizona is not likely to go their way in the fall. For that reason alone, it seems unlikely that Obama is ahead by a couple of points in Arizona—even before considering Romney’s narrow national advantage or a tight race elsewhere in the southwest. Without McCain on the ballot, increased Latino support can and turnout can easily get Obama to 47 or 48 percent of the vote, but the final two percent get quite difficult, requiring either a massive swing among Latino voters (40 points), a big increase in Latino turnout, and resilient support among Arizona's white voters. But if Obama did get into the state, it wouldn't be exceedingly hard to see why. The conditions necessary for a close race in Arizona are all quite plausible, it's just a question of whether they exist in reality, not just on paper.