Today, Ezra Klein pointed out that on Intrade, people are betting Democrats have a decent chance of winning back the House but are unlikely to hold the Senate:

I was tooling around InTrade last night and noticed that the betting markets are giving Democrats much better odds of retaking the House of Representatives in 2012 than of holding the Senate. As they see it, there’s about a 60 percent chance that Barack Obama will be reelected as president, a 40 percent chance that Democrats will win the 20-some seats they need to retake the House and only a 25 percent chance that they’ll win enough elections to retain control of the Senate.

I understand that the map of Senate seats up in 2012 is very unfavorable to Democrats, but looking at historical patterns, there is something seriously wrong with the Intrade market.

Going back 100 years, not once in the last 50 elections has control of the House switched to one party while control of the Senate switched the other way. For the most part, partisan movement in each chamber is highly correlated with the other. There just hasn’t been a modern example of a national party making large gains in one chamber while suffering big loses in the other.

In 2012, Democrats might gain back the House, or they might lose the Senate, but if history means anything, it is extraordinarily unlikely both will happen. And if you want to play the market, a bet that Democrats will gain the House should never be so high compared to a bet they will lose the Senate.