Redistricting Matters

It appears Republican gains in the House will wind up at around 64 or 65 seats. Looking quickly at the casualties, it appears the vast majority were either veterans in heavily Republican territory or Class of 2006-2008 "Democratic wave" members. Six wins were in southern open districts that were all but conceded months ago. There were virtually no out-of-the-blue upsets; as Nate Silver put it early this morning, it was a very "orderly wave."

But Republicans did seem to enjoy some luck at the margins. They won the national House popular vote by between 6% and 7% (which means the final Gallup generic poll, predicting a 15-point margin, was indeed an outlier, along with Rasmussen, which predicted a 12-point margin). This margin would in theory normally produce a gain of about 55 seats. The excess peformance will be attributed to superior Republican vote "efficiency," which is another way of saying that the advantage they obtained during the last round of redistricting endured to the end.

Speaking of redistricting, the worst news of the night for Democrats was in state legislative races. Republicans appear to have gained control of 15 state legislative chambers. In conjunction with gubernatorial wins, they obtained control of the redistricting process in several big states which will lose House seats (alway an opportunity for gerrymandering mischief), including Ohio, Michigan and Pennsylvania...

Finally, something must be said about the electorate that produced these results. According to national exit polls, 2010 voters broke almost evenly in terms of their 2008 presidential votes; indeed, given the normal tendency of voters to "misremember" past ballots as being in favor of the winner, this may have been an electorate that would have made John McCain president by a significant margin. Voters under 30 dropped from 18% of the electorate to 11%; African-Americans from 13% to 10%, and Hispanics from 9% to 8%. Meanwhile, voters over 65, the one age category carried by John McCain, increased from 16% of the electorate to 23%.

These are all normal midterm numbers. But because of the unusual alignment of voters by age and race in 2008, they produced a very different outcome, independently of any changes in public opinion. Indeed, sorting out the "structural" from the "discretionary" factors in 2008-2010 trends will be one of the most important tasks of post-election analysis, since the 2012 electorate will be much closer to that of 2008. That's also true of the factor we will hear most about in post-election talk: the "swing" of independents from favoring Obama decisively in 2008 to favoring Republicans decisively this year. Are these the same people (short answer: not as much as you'd think), or a significantly different group of voters who happened to self-identify as independents and turned out to vote?