Saturday, May 17, 2008

US Senate Campaign Outlook

Can the Democrats get to 60.

Mainly I'm looking at the polls at Pollster.com which show two very likely Dem pick ups (Virginia and New Hampshire), 3 pretty likely pickups (Alaska, Colorado and New Mexico), two races which, by region, should be close but have the Republican ahead (Maine and Minnesota), two races which should be landslides but have the Republican incumbent in trouble (Texas and North Carolina) and only lonely incumbent Democrat with a possible (barely possible) loss in Louisiana. That gives Dem pickup of 1 to (very unlikely) 9 with 6 the posterior mode*.

However, there are two other races which aren't on the board at all, Oregon and Kentucky. They are both states in which the Democrats haven't had their primary yet, so Republican incumbents don't have a single oponent.

In Oregon a Republican cousin of the Udall cousins -- Gordon Smith -- will face Jeff Merkley or Steve Novick. In the only recent poll which I found Smith appears to be very vulnerable. Rasmussen reports

The latest Rasmussen Reports telephone survey of Oregon voters finds Smith leading Jeff Merkley by just three percentage points, 45% to 42%. In late March, he enjoyed a thirteen point lead. In February, he was ahead of Merkley by eighteen points.

When matched against Steve Novick, Smith leads by six percentage points, 47% to 41%. In the March poll, Novick trailed by eleven. In February, the gap was thirteen points.


I have no idea why the Oregon senate race is rarely discussed. Smith is clearly vulnerable. He is well under 50% with small leads over two hopefuls. It is well known (and also true) that when a hopeful ties up the nomination his or her support increases.


In Kentucky, Senate minority leader Mitch McConnell has "double digit leads over
both both Democratic primary frontrunner Bruce Lunsford and his nearest primary opponent in that race, Louisville businessman Greg Fischer...

Against Lunsford, McConnell garners 48 percent support among the 600 "likely voters" polled to 36 percent for the Democrat. In that contest, 16 percent of respondents were undecided.

Against Fischer, the results are similar. McConnell grabs 47 percent to Fischer's 35 percent, with 18 percent undecided."

McConnell has less than 50% support even though the Democrats don't even have a candidate yet. This poll does not suggest that McConnell is the solid favorite.

Now it looks like the Democrats will nominate Obama and that Obama is unpopular in Kentucky. This might help McConnell. Still, I'm surprised I haven't heard more about the race (except from the always optimistic Kos).

So I add two lean Republican seats bringing Democratic gains from 1 to 11 with a gain of at least 5 reasonably likely and 7 the posterior mode.

I am not cautious about giving probabilities (can't be refuted with one observation) so I will guess that the probability that the Democrats get to 60 (counting Lieberman) is around 1% (really I mean it).

My calculation to get to a pickup of 9 seats. Start with 2 (Virginia and NH which seem quite likely). Say chance somewhat less than 75% in Alaska, N.M. and Colorado and 25% in Minn, Maine, N.C, Texas, Oregon and Kentucky. This is generous to the Dems. Assume each race is independent. This markedly reduces the probability that the Democrats get to 60.

Hmm first chance to sweep AK,NM and CO is 27/64. chance to win at least 4 of the others is about 3.7%. Chance of winning 2 of 3 and 5 or 6 of 6 is 27/64 times .47 % chance of one of three then all 6 is 9/64 times two to the minus 12 (rougly zero). So I calculate around 1.8 % but I was being really optimistic about some of those states so I'll just round it down to about 1%.









*that's deliberately obscure as I don't want to make a comprehensible prediction.

No comments: