Michigan meshugas

Googling "turnout". Photo by dyingtobeadyingswan.

So, well. All the newspapers have been awfully coy about discussing the significance of the Democratic turnout in Michigan yesterday, which was apparently around 1.2 million voters (the Detroit News reports 1.3 million Republican votes out of 2.5 million total). Which is a subject I'm always particularly interested in, and especially now in the context of the Sanders campaign and the "people's revolution" (as he called it twice last night before correcting it to "political revolution"—he was clearly very tired): are the voters he calls to, the young and the ripped-off casualties of economic stagnation, unable to find decent jobs or pay off their school loans or in so many cases finish school at all—are they showing up for the revolution?

As you might think they are, from the hugely unexpected character of these results. As opposed to all the previous contests, where the Democratic turnout seems really weak while the Republican turnout is really scarily strong, this one seems like an enormous upset in Sanders's favor, and a fulfillment of his prophecy, that the normal nonvoters, people usually too cynical or alienated to do their citizen duty, are going to come out this time because the revolution is on offer.

The Detroit News claims that the turnout was huge—34% as opposed to just 19% of the registered voters in 2012! But this is a pretty misleading comparison, because in 2012 there weren't any Democrats running. That primary was for Republican candidates only. Moreover it attracted 1.5 million voters for Romney, Santorum, Paul, and Gingrich, or 200,000 more than voted in this year's Republican primary. For Republicans, turnout this year is actually down (though it's still nearly as much as the 1.6 million votes Rick Snyder won with in the 2014 gubernatorial election!).

As for the Democrats, it's really hard to tell. A big part of it is that it's impossible to compare with the 2008 primary, which was an idiotic mistake (the state party scheduled it for January 15 in defiance of the national party rules against having it so early, and the national party decided Michigan would have no convention delegates at all, and candidates Obama, Richardson, Biden, and Edwards all withdrew, and the turnout was exceptionally low that year), or the 2012 and 2004 contests, where they had caucuses instead (of course in 2012 there was only one Democratic candidate anyway). I also can't get any coherent picture of party registration in Michigan, and it's not clear how relevant it would be, since it's an open primary.

But I thought of an approach based on the 2012 general election. You can think of this year's Republican primary vote as a function of the population that voted the Republican ticket in that November (which included Democrats and independents, just as Michigan's primary does), as representing a kind of maximal picture of the Michigan Republican vote, and you can do the same thing for the Democrats.

So at that rate, this year's Michigan Republican primary vote as a proportion of the vote for Romney in 2012  (2.1 million) is about 71%; this year's Democratic primary vote as a proportion of the Obama vote in 2012 (2.6 million) is only 33%.

By that measure (incredibly crude, but it has novelty value!), recognizing that the Republican turnout itself is down from four years ago, the Democratic turnout looks as low as it's been elsewhere this year or lower, and the revolution hasn't started yet. Any thoughts?

Update: More useful (if horserace-oriented) analysis from Nancy Le Tourneau.