On Tents

I usually have tremendous respect for the analytical skills of Charles Blow, but his latest column strikes me as being Just Plain Bad. It has two basic premises: the Democratic Party is becoming more and more the party of liberals, which in the short term alienates it from independents, and the liberals in the Democratic Party show a tendency toward "eating their own." Each argument has, as I see it, one fundamental problem: drawing lines through the electorate differently doesn't actually change the electorate, and the second argument is simply and utterly incorrect.

First off, consider the following metaphor. In golf there are basically three kinds of grass-covered ground that run between the tee and the green, namely fairway, rough, and intermediate rough, which is usually a narrow strip between the other two. Those hitting the fairway tend to score best, and those in the rough tend to score worse. So suppose we want to divide things into only two categories, "hit the fairway" and "missed the fairway." Our first inclination is to put intermediate rough with the "missed the fairway" category. Suppose that, of 100 golfers, 60 hit the fairway and averaged -0.2 strokes to par; 10 hit the intermediate rough and averaged +0.1 strokes to par; and 30 found the outright rough and averaged +0.4 strokes to par. With this categorization scheme, our "hit the fairway" category averaged -0.2 strokes to par, and our "missed the fairway" category averaged +0.325 strokes to par. But what if we decide that, hey, you know, intermediate rough isn't so bad; let's call it a hit fairway. Now the "hit the fairway" category averaged -0.16 shots to par, and the "missed the fairway" group averaged +0.4! Both categories saw their scores get worse! This must mean the hole got harder, right? Well, of course not; I simply drew the lines differently. Specifically, while both groups got worse, the group with the better scores got bigger, to an extent that exactly cancels the first change.

The analogy is fairly obvious with political parties. Suppose that an electorate contains 21 liberals, 40 moderates, and 39 conservatives. One classification scheme 13 liberals, 18 moderates, and 9 conservatives identifying as Democrats; 6 liberals, 12 moderates, and 9 conservatives identifying as independents; and 2 liberals, 10 moderates, and 21 conservatives identifying as Republicans. The Democratic Party is then 33% liberal and 23% conservative, while the Republican party is 64% conservative and 6% liberal and independents are 22% liberal and 33% conservative.

But now suppose we have a different classification scheme in which the Democratic Party contains 17 liberals, 17 conservatives, and just 5 conservatives; the Republican Party has just 1 liberal, 9 moderates, and all of 22 conservatives; and the independents are 3 liberals, 14 moderates, and 12 conservatives. Now the Democrats are 44% liberal and just 13% conservative and the Republicans are 69% conservative to 3% liberal, and independents are 41% conservative and just 10% liberal. It looks like the Democrats have become alienated from the moderates, right? Well, obviously this electorate has the same 21 liberals, 40 moderates, and 39 conservatives as the last one, so the electorate hasn't changed at all. If being more liberal makes Democrats more likely to choose policies that alienate the median voter, they'll be worse off, but the identity of the category "independents" is almost meaningless. (Moreover, the shift toward ideologically pure parties has been going on for a while, and is, you know, natural. It's just the Great Sorting Out of the confusion about racial issues of the mid-20th century. Parties are supposed to be networks for advancing the cause of certain political ideologies, so it makes sense for them to be relatively ideologically consistent.)

As to the second claim, which is really necessary to claiming that Part 1 is a problem: the idea that we liberals show a tendency to eat our own. In the 2010 elections, in the year when a shocking 40% of Democrats identified as liberals (and 70% of Republicans identified as conservatives), there were 37 Senate races. Of those 37 elections, I count all of 4 primaries in which liberal Democrats opposed a more centrist, mainstream Democrat: AR, PA, KY, and CO. Colorado's inclusion is debatable: liberals did prefer Bennet's opponent, but not strongly. In any event, the point is that in none of these races was the liberal candidate less electable. Specter would have lost by a lot more than Sestak lost by. Lincoln lost by a lot more than Halter would have. Conway ended up losing pretty big, but that's largely due to his unforced and unforseeable error with his late campaign ad; at the time of the primary he looked like the stronger candidate. And in Colorado there was essentially no difference between Bennet and Romanoff.

As for Republicans, I count 11 right-wing primary challenges to establishment candidates: AK, AZ, CO, CT, DE, FL, KY, NV, NH, PA, and UT. Of these, nine were successful, and only one, I think, can be argued as having nominated a candidate who was stronger in the general election than the establishment alternative would have been, namely Pennsylvania. I'm not certain of that, though. Meanwhile, in Utah it was a wash, being a ridiculously conservative state; in Florida, Rubio ended up winning big anyway though in a one-on-one against Meek he would probably have done much worse than Crist would have; and the Tea Party candidates in Alaska, Arizona, Colorado, Connecticut, Delaware, Kentucky, Nevada, and New Hampshire would all have been, or were, much weaker than the establishment candidates in the general election. Indeed, these Tea Party primaries cost the Republicans Delaware, Colorado, and Nevada, pretty clearly, and also caused Alaska's Republican to no longer be a nominal Republican.

So, to recap, liberal Democrats applied primary challenges sparingly and only when the liberal candidate was no less electable than the centrist one, while conservative Republicans brought primary challenges early and often and almost always in cases where doing so cost them electability. Who eats who, again? (That liberals have not tended to fare worse than moderates over the last decade is known; see here.)

And notice that liberals are not actually clamoring for an Obama primary challenge; nor are liberal Democrats, 69 percent of whom support the tax deal, uniformly "foaming at the mouth" in opposition to the deal. And I saw a previous Gallup poll showing Obama's approval from liberal Democrats hadn't slipped, though there does seem to be one now showing that it has; it's somewhat murky. And there's no indication that liberal Democrats are going to vote for anyone other than Obama in 2012; if those of us who are malcontent do anything, we'll just be apathetic and stay home. We've learned the lesson of Ralph Nader; it's the Tea Partiers who will not yet learn the lesson of Christine O'Donnell.