Permutations 01/20/08

At the end of today's "Meet the Press", Tim Russert threw out a challenge to the panel: who would predict the name of the person to be inaugurated on Jan. 20, 2009--exactly one year from today?

It was a bit of a red herring (there were only seconds left in the telecast), and of course no one dared to put out a name, proving Russert's implication that predicting the winner today would be impossible.

Of course, it was a challenge we dared take up.

To do our analysis, we relied on one BTH factoid--that Mike Bloomberg's best scenario, according to one, would be the Clinton-Huckabee matchup. This confirmed my hypothesis that Bloomberg's analysis of his chances looks at 1) a perception of managerial competence, and 2) ability to appeal to independents--these are the things he would bring to the table, and he will run only if he feels they are lacking among the major party nominees. I decided that a nomination for Obama (appeals to independents), Romney (also managerially competent), or McCain (possibly both?) would each tend to reduce the likelihood he would win, and thus that he would run.

I came up with this matrix of "Bloomberg In" probabilities:
Clinton-McCain 20%
Clinton-Huckabee 75%
Clinton-Romney 40%
Obama-McCain 5%
Obama-Huckabee 40%
Obama-Romney 20%


I focused on the 12 "canonical variations" of Clinton or Obama, vs. McCain, Huckabee, or Romney, and whether Bloomberg would run in each case. Then I factored in a subjective (but carefully considered) estimate of the probabilities of victory for each candidate in each of the matchup scenarios. The probabilities I assigned to nomination outcomes were as follows:
-- 50% Clinton, 40% Obama, 10% no decision indicated from the UNP outcomes;
--40% McCain, 30% Huckabee, 20% Romney, 5% Giuliani, and 5% "X"--someone else. (Here I'm still sticking with my age-old calculation of 40% McCain, 35% anti-Bushite Right-Winger, 25% "GPR"--Giuliani/Pataki/Romney.)
In the case of the first, I completed the probabilities recursively, i.e., the 10% remaining probability were divided up 50% Clinton* (late victory), 40% Obama late victory, and 10% Edwards or someone else (combined probability of 1%). In the case of the Republicans, I made no assumption as to when the decision would become clear. I reduced the probability of the Democrat winning if they didn't win decisively, and I also increased the probability of Bloomberg running for those outcomes.

I assumed that Bloomberg's decision was after the Democrats' and Republicans' nomination decisions, and that it would be informed by his chances of winning, consideration of whether his running would improve the chances of a "suitable" President winning, and a guess as to how much he valued the two outcomes (vs. the cost of his participation). His participation probability would rise sharply as the chances of winning increased, as long as his running didn't hurt the chances of a "suitable" outcome (I actually fixed the probabilities so they were never hurt by his running).

Finally, I looked at most variations of the two 10% extreme cases, going down to cases that had as low as 0.1% likelihood (e.g., Edwards, or some other Democrat vs. Giuliani or X)

Bottom line: I came up with about a 64% probability for two outcomes:

1. That a Democrat would win; and
   
2. That someone "suitable to Bloomberg" would win (i.e., Bloomberg himself, Obama, McCain, or Romney).

The probabilities for individual candidates being elected (covering 98.15% ; let's say the other 1.85% is of a Bushite coup preventing someone being nominated, or some other unlikely event such as a nominee assassination, accidental death, or Bloomberg defying his calculations and not running or something) were, in order, as follows:

• Obama 33.58%;
• Clinton 28.32% (Bloomberg took more of her chances away than he did from Obama, plus I like Obama's chances better in a straight-up race against McCain);
• McCain 15.00%;
• Bloomberg 10.39%;
• Huckabee 6.45%;
• Romney 3.17%;
• Giuliani 0.65%;
• X (some other Republican) 0.34%; and
  
• Edwards (or some other Democrat) 0.26%.

You can take this to the bank--I'm going to take it to the Rasmussen Markets trading floor.