Behind the numbers

Tuesday, February 12, 2008

I've been following the Democrat race more closely than the Republican for the last few weeks, and I've become fascinated by the political trading markets. The most interesting question for me has been: given that Obama (or Clinton) wins the nomination, what are the markets saying about their chances of winning the election.

As of right now (12:48 AM EST, February 12), it's clear that Obama won the MD, DC, and VA primaries handily. The markets reflect that - he's now at 74% to win the nomination. That's his biggest lead yet.

But he's also at 49% to win the election. That implies that if he wins the nomination, the markets are estimating a probability of .49/.74 = .662 that he wins the Presidency. Meanwhile, Clinton, though given only a 28% probability of winning the nomination, is at 19.5% to win the Presidency - implying a .195/.28 - .696 probability that she wins the Presidency if she wins the nomination. Apparently the markets think Clinton is a unlikely to win the nomination, but would be a slightly stronger candidate in the general election. Interesting that this has held up even as Obama has taken a commanding lead (at least probabilistically) in the race for the nomination.

These probabilities have been fairly consistent for the last few weeks, with Clinton generally leading, Obama generally close, and both ranging between 60% and 69% implied probabilities of winning the election given that they win the nomination. The markets have enforced this consistency, even though no one is actually betting on it (in the market in question, anyway).

Another interesting thought - if the race goes to the super-delegates, will the perception that Clinton is a better candidate in the general election be a swing consideration?