Speaking of the capacity of real-money betting markets to predict actual events, no one seems to have noticed that the Iowa elections market (*) has Bush’s re-election at roughly even money. People with their own money at stake don’t seem to agree with the conventional wisdom that he’s nearly unbeatable.
Pollkatz has a useful graphic (*), easier to interpret than day-to-day poll results. Right now, Bush seems to be just slightly above where he was just before invading Iraq, which in turn was just slightly above where he was just before 9-11. So far, he’s had two peaks: right after 9-11, and right after the invasion of Iraq. And the slopes down from those peaks seem to be equal and fairly consistent, at about 1% per month.
There’s no way to guess whether that trend will continue now that Bush’s numbers have come back to earth. Most of the people who are for him now were for him before 9-11, and you’d expect that support to be harder and less likely to erode than the rally-round-the-flag support. On the other hand, it’s striking how much less the Iraq invasion did for Bush’s popularity than Desert Storm did for his father’s.
In any case, the Pollkatz graph is more consistent with the Iowa numbers than either is with the conventional wisdom.