Ross Douthat has a smart post up on his blog today about the Obama campaign strategy. He points out that Obama’s ad campaign has been overwhelmingly negative, and that the main focus of the attack on Romney is to attack him as a Wall Street fat cat job killer. This would seem to be a strategy aimed at working-class whites. But, on the other hand, Obama has also been moving leftward, and loudly, on a host of social issues, which would seem to be a strategy aimed at educated liberals, and to provide Romney with an opening to win those same working-class whites that Obama can’t afford to lose. What gives? Ross says:

Viewed from a certain angle, these two approaches would seem to be at cross-purposes. (Why pander to blue-collar whites in your ad buys if you’re just going to alienate them on immigration policy or welfare?) But if the goal isn’t to win disaffected working class whites so much as to render Romney sufficiently radioactive that they mostly just sit things out in disgust, then the two-track approach makes considerably more sense. In that scenario, Obama doesn’t need these voters to like him, so he can afford to direct his policy pandering elsewhere; he just needs them to dislike his opponent enough to declare a plague on both houses and stay home.

On the one hand, this strikes me as a basically sensible way of understanding the race. President Obama is trying to excite his base and destroy his opponent’s reputation among that segment of the population that is the key to his opponent’s possible path to victory. Sounds like politics.

But there’s another way to understand the dichotomy. Take a look at this graph:

That’s the national unemployment rate versus the average unemployment rate across six Great Lakes states: Illinois, Indiana, Michigan, Minnesota, Ohio and Wisconsin, in each case as of May of the relevant year – the most recent state data I had for 2012 was for May. (For simplicity, I just took an arithmetic average of the unemployment rates, not an average weighted by population, but eyeballing it suggests that the shape would be pretty similar if I calculated the average correctly.) The list includes Democratic-leaning and Republican-leaning states, and I modestly suggest that, in electoral terms, “working class whites” really means “working class whites in states like Michigan and Ohio” since Obama was never going to win working class whites in Tennessee or Georgia.

It’s a noteworthy chart, because what it shows is that, before the financial crisis, the region had a somewhat higher unemployment rate than the United States as a whole, and that the gap grew as the financial crisis and the subsequent recession deepened. But in the three years since May of 2009, the region’s performance has significantly outpaced the national average, dropping nearly 3% while national unemployment dropped less than 0.7%.

Why has the Midwest outperformed the nation in the Obama years? I’m sure there’s no consensus among economists about this question. But I expect the Obama Administration to argue that the auto bailout – which Romney opposed before he retroactively supported it – was a crucial component to the degree of recovery that has been achieved. Correct or not, it’s going to be a powerful argument, and the “Romney is a Wall Street fat cat who only cares about skimming profits and not about saving jobs” is a crucial piece of the narrative. “Obama has written off working-class whites” is a tough argument to square with his obvious determination to win Ohio, a state which is 86% white where Obama has been leading consistently the entire campaign.

So if the old Rust Belt has outperformed during the recovery, what part of the country has underperformed? Well . . .

California’s unemployment rate started out, pre-crisis, pretty close to the national average. Like the Great Lakes region, California underperformed, employment-wise, during the crisis – in fact, it did performed even worse, with unemployment going up 6.4% from May 2006 to May 2009 (versus 5.2% across my Great Lakes 6 and 4.2% across the nation). But California also underperformed during the recovery phase. From May 2009 to May 2012, California’s unemployment rate dropped only half a percent, versus 0.7% nationally (and 2.9% across my Great Lakes 6 group).

What’s the story Obama can tell to the citizens of California about what he’s done for them lately? He hasn’t done anything much about the enormous number of mortgages that are still underwater. But he has made birth control a civil right.

Obviously, California doesn’t account for the entirety of the disparity between Midwestern and national unemployment trajectories – but it’s a big state; it’s a big chunk of the reason. And California is also a safe Democratic state. But that’s because it’s ground zero for the Emerging Democratic Majority strategy of assembling a winning coalition of educated whites and minorities – the Democratic side of the culture war-related sorting of the population into political tribes. And some of the other states that are underperforming the national recovery that are more important to Obama’s reelection effort – Colorado, FloridaNevada, New JerseyNorth Carolina – have some demographic characteristics in common with California. The winning Democratic coalitions in these states – when and if they materialize – are the McGovern coalition, just like in California.

It’s possible that Obama has written off blue-collar whites, and that his negative campaign against Romney is just about making Romney toxic. But it strikes me as more probable that the Obama campaign has different messages for different segments of the population and different parts of the country.