LOOKING FORWARD TO IT. I always find it helpful to exit the world of progressive electoral prognosticators and head over to John McIntyre's RealClearPolitics which, while right-leaning, offers a rather dispassionate aggregation of poll information and political predictions. This week's forecast is particularly interesting:

while the national numbers have improved for the GOP over the last month, on balance the state polling has not. Looking at the RCP Averages in the contested Senate races, the Democrats are poised to pick up seats in Pennsylvania, Montana, Ohio and Rhode Island, while the Republicans look likely to win in New Jersey. That gives the Democrats a three-seat pick up with the need to pick up another three seats to win control. The problem for the GOP is Tennessee and Virginia have moved into toss up status, along with Missouri. The Democrats' odds of capturing the Senate have actually improved the last two months at the same time their national numbers vis-�-vis the Republicans have declined.

The better analogy politically for 2006 may be 1986 when the Democrats picked up 8 Senate seats and only 5 House seats. Because of Reagan's landslide in 1980 there were many weak GOP incumbents in 1986 that were taken out. Today Republicans have less of an issue in that regard as their 1994 weak incumbents were taken out in 2000 (Grams, Abraham, Ashcroft, Gordon, and Roth). The point of the '86 analogy is not that the Democrats are going to taking over the Senate, but rather that because of the inability to gerrymander states, Democrats might be headed for better success in the Senate than the House.

There is a reason 99% of incumbents win reelection in the House and right now even with the Democrats looking strong in the Senate they are only poised to pick up around 10 seats in the House. With the economy humming at 3%+ growth, unemployment below 5%, the Dow near all-time highs and gas prices back below $2.50, these are not exactly economic conditions associated with a "throw the bums out" type of election.

Much of that strikes me as plausible, if not correct. The House races tend to be harder to predict, if only because there's less polling and data flowing out of individual districts than whole states. Moreover, the GOP's GOTV advantage will likely make the difference in a number of marginal races where Democrats lack sufficient funds to power major turnout operations. That will likely be less true in the Senate, where there's more money, institutional support, and voter excitement.

--Ezra Klein