Some of these charts don't show up well, so you might look here for the whole view.
The Obama campaign is embarking on a crash voter registration effort to determine whether Obama can win Arizona, a state that McCain won by 9% in 2008. While many have dismissed Obama’s chances of winning a traditionally Republican state, the demographics point toward a genuinely competitive race. Whether Obama can prod Arizona into behaving like its demographics remains to be seen, but it’s not hard to see why it’s quite plausible.
This morning, Nate Silver added his thoughts and concluded that Arizona is (probably) not a swing state. Silver argues that while Arizona may potentially be competitive, history and polls suggest that it is not a tipping point state – one that could decide a close national election. In his view, an Obama win in Arizona would be more akin to a Romney win in New Jersey – states that can be won, but probably long after the election is decided. Nate Silver is a formidable analyst and disagreeing with him gives me pause, but his analysis improperly assumes that Arizona’s partisan leanings from 2004 and current polls are indicative of its competitiveness.
Historically, Silver is right; between 1992 and 2004, Arizona leaned Republican by appropriately 7%. If Arizona could be expected to lean Republican by approximately 7%, then Arizona wouldn’t be anything like a tipping point state. But, it’s not 2004 anymore, and it hasn’t been for eight years. Profound coalitional and demographic shifts have since rejiggered the electoral map.
An examination of Arizona’s southwestern neighbors demonstrates the profound weakness of relying on historical partisanship to predict whether a state would be a “tipping point” state in the future. While New Mexico, Colorado, and Nevada hued near the national center in 2004, each lurched toward the Democrats with respect to the country as a whole.
There are a number of ways to represent the shift of these areas, but I’ll use the same one that Nate Silver does – the gap between state-level results and national results. This is a reasonable metric, since by showing the gap between the state and the country, it should roughly approximate how the state would behave in the event of a tied national election.
Back in 2004, when Arizona leaned red, Colorado, Nevada, and New Mexico were all plausible “tipping points.” In 2008, Obama’s strength among Latinos and college educated white suburbanites in burgeoning sunbelt metropolitan areas moved all of these states toward the left. The shift was substantial: 5%.
Gap Between State and National Results
2004 2008 Shift
Colorado R+2.21 D+1.69 D+3.9
Nevada R+0.13 D+5.23 D+5.36
New Mexico D+1.67 D+7.87 D+6.2
D+5.15
If such a shift had occurred in Arizona, it would have leaned GOP by 3% instead of 8%. While one could argue that, for some reason, Arizona would have been insulated from these trends, even a quick analysis concludes otherwise. As we said last week, Arizona exhibits the characteristics of a state that would have voted for Obama by 4-5%.
This wasn’t an isolated phenomenon. We can look at counties similar to Arizona’s Maricopa County, home to Phoenix and 60% of the Arizona vote. The other sunbelt metropolitan counties all lurched decisively with national demographic trends.
Gap Between County and National Results, 2004, 2008
Gap 2004 Gap 2008 Shift
Clark, NV D+7.31 D+11.73 D+4.42
San Bernadino, CA R+9.31 R+0.94 D+8.37
Riverside, CA R+14.34 R+4.95 D+9.39
San Diego, CA R+3.67 D+2.94 D+6.61
Dallas, TX D+1.06 D+8.01 D+6.95
Tarrant, TX R+22.92 R+18.96 D+3.96
Bexar, TX R+8 R+1.71 D+6.29
Harris, TX R+7.73 R+5.63 D+2.1
Maricopa, AZ R+12.18 R+17.78 R+5.6
There should be some caution in interpreting these comparisons, since many weren’t in swing states and several were in Texas, where the swing might have been exaggerated by the end of Bush’s home state advantage. But the consistency of these numbers points toward constant and underlying factors: across every sunbelt metropolitan area with a large Hispanic population, the gap between county results and the nation shifted decidedly toward the Democrats (6%), at a pace largely in line with the southwest – with the exception of Maricopa County and Arizona.
The bottom line is simple: just because Arizona leaned Republican in 2004 does not mean it will in 2012. So, will Arizona be a tipping point in 2012? It’s far from certain, but demographics say “yes,” depending on one’s definition of a tipping point state. If the gap between Arizona and the country shifts 5% toward Obama, as demographics suggest, then Arizona would now lean Republican by 3%. In my view, that’s close enough to categorize as a tipping point state, since Ohio leaned Republican by 2.6% and Florida did so by 4.45%. If those states are toss-ups, then Arizona could be as well.
While I am unconvinced by arguments about Arizona’s electoral history, there are certainly credible reasons why Arizona might not be a swing state. Unfortunately, many of these arguments have been missing from the debate, but I think they fall into two categories: 1) Arizona is unlikely to behave in according to its demographics, or 2) that the demographics point toward a state that isn’t a tipping point state.
In the first category, the crucial question is whether Obama can make up for lost time. In 2008, the Obama campaign might have registered hundreds of thousands of new voters. To get Arizona to act like its demographics, Obama must not only register the voters that would have been registered in 2008, but four more years worth of new young voters, out of state transplants, and emerging Latino voters. There are also reasons to question whether the failure to compete in Arizona might have done lasting relative damage to Obama’s standing, or, put differently, whether competing in the other swing states has given him more resilient support that he can’t count on in Arizona. These factors likely underlie Obama’s tentative approach: first determine whether the campaign can register the requisite number of swing voters, then invest money in an expensive ad campaign to hold the white vote. For the same reasons, I am reluctant to rely on current polling to gauge Obama’s eventual competitiveness.
In the second category, there are three potential arguments: 1) Obama won’t do as well as assumed with Latinos 2) Obama won’t boost Latino turnout enough, 3) Obama won’t win enough white votes. The argument that Obama won’t do well with Latinos is preposterous, in my view, and that critically frames the final two questions. Most do not understand the extent to which McCain’s victory was dependent on strong support among Latinos. Even if Obama doesn’t win a single new white voter, or doesn’t register and turnout a single new non-white voter, simply replicating his national share of the Latino vote propels Obama to nearly 48% of the vote, and there are reasons to think Obama could do better than that against Romney in a state like Arizona, the epicenter of recent immigration controversies.
If one assumes that Obama can produce increased Latino turnout, then he can potentially secure the state without any additional white voters. This is the heart of why it’s not useful to consider Arizona’s past partisan leanings: the Obama campaign just needs to activate a latent base. Whether Obama can do so is a fair question, but that’s largely dependent on whether Obama can make up for lost time.
In addition to whether Obama can register enough Latino voters, an Obama victory would still require him to win enough white votes, and that’s far from a given. Although Obama’s share of the white vote was certainly suppressed by McCain’s presence, Obama has lost considerable support among white voters nationally. Intelligent analysts can differ on whether Obama can be expected to lose, hold, or gain white votes, given the events since 2008 and the absence of McCain. Depending on the answer to this question, one can be optimistic or pessimistic about Obama’s chances in Arizona.
The current polling does not undermine these claims, as Silver suggests. First, while Silver identifies Ohio as a tipping point state based on current polling, a number of other true swing states, like Virginia, currently diverge from the national average as much as Arizona. Second, and as explained earlier, there are reasons to think that the polls understate Obama’s eventual standing, assuming he ultimately embarks on a serious effort to secure the state. Third, it’s worth noting that the most recent PPP poll shows Obama at 42% among whites in Arizona – up 2% from 2008. This poll is outdated, but if true, Obama would poised to win without any increase in Latino turnout.
Even if Silver’s characterization of Arizona’s current position with respect to the country is accurate, it doesn’t mean Arizona can’t be a tipping point state. The relationship between a state’s proximity to the national average when the race isn’t close does not preclude a state from being a tipping point state once the race does get close. This might seem counter-intuitive, but just consider this question: how would Romney gain the extra four points to produce a national tie?
There are obviously a number of routes to a tied election. Romney could gain among Latinos, college educated whites, or working class whites. Non-white turnout could fade. There are other possibilities. Depending on the answer, some states drift closer to the national average and some states drift away. If Romney closes the gap with Obama through gains among working class whites, those losses would not substantially reduce Obama’s standing in Arizona. Given Obama’s low approval ratings with working class whites, this is probably the likeliest scenario, and that introduces the real possibility that Arizona could hue close to the national average in a tight election. Conversely, a white working class state like Ohio, which currently mirrors the national average, would zip right into the Romney column.
Although demographics suggest Arizona could be a swing state, such an outcome is by no means assured. There are legitimate reasons to argue that Arizona might not be a swing state, but simply reasserting the state’s past partisan leaning just isn’t among them. Obama’s strength among suburbanites and Latinos has rejiggered the electoral map, and a state like Arizona should no longer be expected to lean toward the Republicans.