|In 17 mid-terms since 1944 the President's party's share of the House vote has dropped from the previous election 16 times. That's with Republican and Democratic Presidents. When something happens 94% of the time, it's a good sign it'll happen again.
I know people think 1998 was a good Democratic year and it was from the standpoint of Democrats gaining seats. What people don't realize is that the Democratic party's share of vote dropped from 1996 to 1998 from 50.2% to 49.5%. Gaining seats was a combination of good luck from retirements and 1994 winners who didn't lose in 1996.
So, in 1998 you have a Democratic President and a Republican party that did the unthinkable. Impeaching the President is one the few things a party can do to elevate the House above the President in people's minds. Despite this blunder, Republicans still got a higher percentage than 1996. Of course it shows that Democrats could drop in the vote and gain seats. I don't doubt that, but that's more of race-by-race analysis that I'll get into later.
The only time a President's party didn't drop in a mid-term was 2002, a year where George Bush's approval rating was 65% due to 9-11. So it might take an event like that for Democrats to win more votes. President Obama's approval rating is around 51%, a far cry from Bush's 65%. If Obama gets up to 65%, I'll revise my judgment.
Not only will conditions be less favorable to Democrats, but the electorate will be less favorable. Mid-terms have less voters. No matter what you, no matter how good the GOTV effort has been, this has been immutable. Every mid-term since 1986 has had between 71% and 76% of the votes in the previous general election. There's no reason to think that'll change.
While you'll get some new voters, most of the voters who vote in 2014 will be a sub-set of the 2012 electorate. The question is who will or won't show up. Minorities haven't shown up. Below are the percentage of Whites in the last four Presidential elections and the subsequent mid-terms.
There were 5.7 million less White voters in 2012. If all of them had voted, the Whites would've been 73%. The 2014 percentage of White voters will likely be somewhere around 73-75%. Democrats only managed to win 50.6% of the two party vote with minorities making up 28% of the electorate. Since Democrats clean up with minorities, but struggle with White voters more White voters is bad news. There were even less minorities in 2006, a good year for Democrats.
The Democratic coalition has become increasingly reliant on the 18-29 segment. They used to index no more than 110 against their actual vote but they indexed at 124 and 119 in 2010 and 2012. The problem is that this group has made up 12% of the electorate in the last three mid-terms, while being 16-19% in general elections.
So why did Democrats do well in 2006? They weren't as reliant on minority or young voters as they are today. They also did very well with swing voters, who were united against the President. Unfortunately, uniting against the President in 2014 won't be good for the Democrats.
Subjectively I can think of a bunch of reasons 2014 will be good for Democrats. The Republicans can blow it or become complacent. The problem with that is that the opposition party has surely done those things before and it hasn't resulted in a positive move for the President's party. There's a double whammy going on here, the Presidential party doing poorly in mid-terms as well as the Democratic coalition taking them off.
Objectively, I see the Democratic share of the vote declining by 1-3% in 2014 to 47.6%-49.6%. Circumstances will certainly change between now and the election to impact that prediction.