Left Center Left

sco at .08 Acres gives another thorough and thoughtful analysis of Massachusetts voters, partly in response to my questions/critiques from before. I’m starting to feel we’re reenacting some distant version of the Thomas Frank-Larry Bartels debate, a humanities-social science split in perspective and assumptions. As the humanist trying to make sense of the polity, and thus necessarily out of my element, I’ll try to avoid total disciplinary hubris, but for all my points of agreement with sco, I’m still not fully convinced of his conclusions.

To sum up his argument (you should read the whole thing), there is an asymmetry between the loyalty and turnout of Democratic and Republican voters in MA gubernatorial elections. Republicans have the lowest portion of registered voters, but the highest loyalty and proportion of turnout. Democrats have over 2 1/2 times the registered voters of Republicans, but more of them cross aisles in the voting booth (24 percent versus 13 percent for the GOP voters) and a lower percentage actually come out to vote (58 percent versus 78 percent of GOP voters). Bottom of all are the Independents, with voter turnout in the mid-40s and (obviously) no party loyalty. 

From this, sco draws two major conclusions. First, he sees a "danger of crafting a strategy geared towards appeals to Unenrolled voters. Fewer than one out of two Independents even bother to vote in these gubernatorial elections, so there is a lot of effort devoted to people the least likely to turn out."  Second, the trick to winning is getting Democratic turnout up. "If we assume that one out of four Democrats vote Republican, as do 61% of Independents, Democrats still win if they match Republican turnout."

I agree that more Democratic voter turnout would be great. The thing is, we don’t know why the Democratic registered voters who stay at home for gubernatorial elections do. Until we have a better picture about that, I’m not sure that a turnout strategy is beyond the wish stage. I’m not trying to be pedantic in insisting on ever-increasing complexity to our political science models; it’s simply that once you take the step from mathematical tabulation to a political strategy, you might run into trouble. What will get voters out? A centrist strategy? A liberal-left strategy? More people running phone banks and holding signs? More money on ads? Smarter spending on ads? Divergent responses seem equally possible from the turnout-is-key thesis. And my guess is that we have a better idea, at least anecdotally, of why some Democrats and Unenrolleds voted Republican than why Dems don’t turn out for gubernatorial races.

Meanwhile, I think sco is mischaracterizing the aim-for-independents strategy. It’s not that Democratic candidates need to cater strictly to the unenrolled, it’s that Democratic candidates need to craft their appeal to included what we broadly might call "independents": Democratic voters who voted for Romney + Unenrolled voters who voted for Romney.

So we have two models, not mutually exclusive. One puts more emphasis on party identification as a fulcrum for gaining nonvoter participation at more or less constant rates of loyalty (.sco is working under the assumption, possibly yet not necessarily true, that getting GOP-levels of Democratic turnout won’t cause loyalty to decrease, that the nonvoters won’t be particularly inclined to center-right messages in gubernatorial elections.) The other emphasizes political coalitions amid more-or-less static turnout. To some extent politics are always about coalitions, and one of the overriding reasons for GOP success in gubernatorial races has been their formula for straddling center-left social positions (even Romney made gestures during the campaign) and center-right fiscal positions. Now, I can’t imagine that any good candidate isn’t going to try for strong party turnout and a broad appeal simultaneously, but the devil can be in the details.

At the very least maybe it’s helpful to think in terms of a Venn Diagram, in which we have various blocks who have different political dispositions and voting behavior. Least controversially, one could picture the 2002 electorate as follows:

This is based on 2002 election returns, and the exit polls sco cites about party loyalty in that election. Not exactly to scale of course and not all numbers add up exactly to 100%, but it gives a reasonable picture of who voted how and who’s up for grabs. I’d break down Democrat, Non-voter and Republican voters further into subgroups themselves, but without the data, that would be adding merely armchair sociology to the mix. Sadly, though, I’m not sure how to get around questions of political sociology in trying to figure out voter behavior.