Did Unopposed Seats Cause The Democrats' Gains To Be Understated?

Fivethirtyeight is always an interesting and entertaining read, but I think Nate Silver overreaches in his piece from last week entitled "Are the Republicans Still a National Party?"  My disagreement here isn't so much with his tentative conclusion, which I think is indeterminate at this point.  I tend to agree more with his initial answer to the question, which was in essence "we'll have a better idea if this is a re-alignment in 2016."  After all, when I took David Mayhew's elections course in 1992, we began the semester reading articles about the permanent Democratic majority in the House and the Republican lock on the electoral college.  Both theories were toast within two cycles.  Many of us will remember 2004, when Republicans were crowing about how their strong position in the South and Mountain West, how the Upper Midwest was trending their way, how they had won 97 of the fastest growing 100 counties, and how the Democrats were being reduced to a coastal party relegated to decaying, dying urban cores.  Heck, I remember 1998, when Republicans held every district in the Mountain West and the Great Plains except for five.

For that matter In the non-wave year of 2004, Democrats came within single digits of House Republicans in only about ten races, and within twenty points of House Republicans in about forty.  This before they picked up fifty seats over the course of the next two elections.  In other words, things can change very, very quickly in politics, I think the whole concept of re-alignments and permanent majorities is largely bunk in the modern world of personality- and media-driven elections.  In other words, I think the Democrats' position may be a lot more precarious than many seem to think it is.  But time will tell on that point, and it may well be that we're seeing the first thing that resembles a truly permanent majority in a long, long time.  It's just too early to tell.

But there was one thing that I pretty strongly disagreed with Silver on.  He writes that:

What happens when a candidate runs unopposed varies from state to state. Some states automatically award him the seat without putting him on the ballot or tallying his votes. The House popular vote total, therefore, will not give any credit to the Democrats when this happens. Other states will put the candidate on the ballot by himself (or against third-party opposition), let the voter punch the card for him, and then tally the results. When this happens, though, there is often a rather severe undervote, since there's not much reason to vote in a contest where you have only one choice.The upshot of this is that the House popular vote tends to discount those areas where a party is so dominant as to discourage competition, because less competition in a district also means fewer votes in that district. Congressional Districts have roughly the same number of people as one another, and so a fairer way to evaluate the House might simply be take the average of the vote share received by each party across all 435 districts (giving a candidate credit for 100 percent of the vote when he runs unopposed). If we do things this way, then we find that the Democrats won, on average, 56.0 percent of the vote on November 4th, and the Republicans 41.3 percent. That's a difference of 14.7 points, far more formidable than nominal 8.9 point advantage that the popular vote total gave them.

It's true that unopposed candidates result in an undervote, but in this case the undervote is probably for the Republicans, though its very difficult to measure.  First off, there were only eight congressional districts where Republicans failed to field a candidate where no votes were recorded:  AR-01, FL-03, FL-17, FL-20, GA-04, GA-05, LA-03, and WV-01.  This certainly hurt Democrats to varying degrees.

But what about the remaining thirty-three seats?  All of them recorded some vote for members of Congress, and many of these members who didn't have Republican opposition still had some type of opposition (in many of these districts, we'll see a Libertarian or a Green candidate get an uncharacteristicly high 20%+ of the vote). 

More importantly, it appears that the Democrats by-and-large still vote for the Democrat in this situation, and that the undervote comes from Republicans.  I've taken these thirty-three seats, recorded the Democrats' vote in 2008, and then put in the next columns their margins when they had Republican opposition.  As you can see, there is really only one instance -- MI-14 -- where the Democrat came even close to his unopposed margin in the face of major-party opposition.  [ED NOTE:  For some reason this table looks right in "edit" mode, but doesn't publish correctly]

You may argue that the increased turnout in 2008 makes any comparison to 2006 or 2004 irrelevant.  Perhaps, but we would probably expect to see at least one district where the unopposed Democrat won with a small margin than when he faced major-party opposition.

But to the extent we can compare like districts (we can't, for example, compare Watson's district to Waxman's district) where there was major party opposition to those where there was not, we see similar results.  For example, with the exception of MA-08, the Mass districts are all pretty similar (as far as Congressional districts go) in terms of partisan bent and demographics.  We see that the unopposed Democrats got between 225,000 and 272,000 votes (In MA-08 the Democrat got  185,000 votes).  The opposed Democrats didn't fare too much better, receiving between 203,000 and 226,000.  The difference was that in the districts where a Republican was on the ballot, the Republican tended to get about 80,000 votes, reducing the Democrats' margin in those districts to about 120,000.  Perhaps its possible that there weren't a similar number of Republicans in the unopposed Massachusetts districts, and perhaps its possible that there were even more Democrats waiting to vote against a Republican in these districts, but that strikes me as improbable given the overall similarity (again, relatively speaking for Congressional districts) of these districts.

We see similar results in Tennessee, where unopposed Democrats in fairly diverse districts TN-06, TN-08, and TN-09 got 193,000, 180,000, and 198,000 votes respectively, while Democrats in TN-04 and TN-05 got 140,000 and 180,000 votes.  The Republican in those latter districts got 94,000 and 85,000 votes respectively, reducing the margin of the winner substantially vis-a-vis the opposed Democrat. 

Finally, let's note that there is a reason why you might get a better result by averaging winning margins.  The reason is that, while Congressional districts in states that aren't single-member are generally pretty similar with respect to the number of persons in their districts, they aren't with respect to the number of voters. Typically, urban Democratic districts have averaged lower turnout than Republican districts, especially in urban minority districts, althought with Barack Obama at the top of the ticket, the difference may not be as pronounced this time around.