The news today is that Olympia Snowe will not be seeking re-election and that Bob Kerrey might try to return to the Senate. A couple of times already this morning on MSNBC — MSNBC, I hasten to reiterate — I heard people frame questions about the Senate and this year's election in terms that assumed that the Republicans will re-take the Senate this fall. Talking about prospects in Nevada, Nebraska and Maine, I heard Chris Cilizza say that if the Democrats win those seats it "increases the margin" the Republicans need to re-take. Similarly, while talking to Snowe herself, Andrea Mitchell twice talked about Democrats "holding on" to the Senate.

Where does this come from? Why are they assuming that the Republicans have that good a chance to re-take the Senate? Why aren't the questions framed in terms of the challenge to the Republicans to "hang on" to what seats they have? Or in terms of the Democrats increasing their majority, instead of increasing the Republicans' "margin of re-take"? My only thought is that it's left over from the assumption that took hold last year, and is only now loosening up, that economic troubles jeopardized Pres. Obama's re-election, and that if he did lose, the Republicans had a real shot at getting a Senate majority. 

Recent developments have begun to cast doubt on that assumption. (Not to get complacent, tho.)

But here's my point. I haven't been able to find exactly the right table of data to cite, but it's my impression that the party of a re-elected President always gains in the Congress.

A little more clearly, here's my assertion. You have a sitting President who wins (re-)election. That would include Reagan in '84, Clinton in '96, but also Johnson in '64 and Truman in '48, neither of which was technically a re-election, and Nixon in '72. As far as I've been able to confirm, in all cases going back at least to 1900, the winning President's party has gained seats in both houses. Not necessarily achieved a majority, but increased its representation. And I emphasize I'm talking only about the Presidential election year, not the mid-term. The Democrats gained seats in '32 and '36, but lost seats in '34 and '38.

So if we assume — as we have increasing justification to do — that President Obama will be re-elected, and statistically that predicts Democratic gains in Congress, why are we framing our questions exactly the opposite way?