Saturday, November 11, 2006

atrios: market expert

I've been meaning to comment on this from atrios, given that I was quoting the markets all week:

Well, the magic power of market predictions is not so magic.

The magic market was heavily predicting Senate control would remain with the Republicans. Oops.

And, hey, so was I. But the idea that there's something magical about market aggregated preferences seems to have infected the minds of too many people. They provide a cute distillation of conventional wisdom, but that's all.
He's wrong for a bunch of reasons, and he should know better, being a perfessor and all - (although as i mentioned the other day, understanding the betting market does take some familiarity.)

1. the market wasn't 'predicting' anything, 'heavily' or otherwise - the market was saying that there was a 25% chance of a Democratic win. The alternative way to frame/interpret the markets, to use a sports betting analogy, is to say 'if these two football teams play four games, we expect Team B (dems) to win once out of those 4 games.'

If Macaca Allen received 7000 more votes, would Duncan have said that the market was 'right'?

2. he is being cute when he says markets are a 'cute distillation of conventional wisdom' - firstly, this particular market was an aggregate of each of the individual races - each of which had it's own particular probabilities. If Duncan wants to say that the aggregate market was 'wrong' - then he'd need to tell us which of these were 'wrong':
Maryland - remaining Democratic - 65.3%
Missouri - turning Democratic - 60.0%
Montana - turning Democratic - 69.0%
Rhode Island - turning Democratic - 68.0%
Virginia - turning Democratic - 66.0%
(plus all the others)

The market was saying that everything had to go our way, and markets 'know' that there are surprises every now and again.

He's also wrong to suggest that markets are a distillation of conventional wisdom - market prices are determined by professionals, not by CW at the margin (assuming that there's reasonable volume). If he wants to argue that there aren't any professionals in political betting markets because these events are too rare (compared to say, horse races), then he's free to make that argument.

I'm not saying that markets are perfect, they're not, but they are notoriously difficult to 'beat' over time - otherwise we'd all be rich. And if they're notoriously difficult to beat, that means that they are 'right' (or, more accurately, not identifiably wrong)

Was the market 'right' or 'wrong' in this case? And if so, by how much? I have no way of knowing (in retrospect, i'd still probably argue that the market was pretty close) - but Duncan's argument is false, and his choice of graphic is cheap. Would he make the same argument with an expiring option on say, the Dow? He'd be laughed out of both of his professions.

And, no, I'm not defending all of my pre-election references to the markets - I just thought that atrios' post deserved a response.

so there.

No comments: