Bush, Kerry, Partisanship and Intrinsicism

Who to believe on the debate last night? Peter thinks Bush won it. Abiola finds it hard to locate non-partisan assessments but is adamant that Kerry won it. Brad De Long declares that "Kerry is presidential; Bush isn't" and posts a list of polls which seem to support his conclusion. I didn't actually watch the debate but it seems to me that all of the above display an intrinsicist assumption: That the debate itself has intrinsic value and that winning the debate is some kind of end in itself. Yet, from the point of view of each candidate, the purpose of the debate is to seek votes, so that they can win the presidential election. It doesn't really matter, even if there is some widespread consensus that candidate A 'won the debate', if that candidates likelihood of election is not improved after that debate. A hollow 'victory' indeed. It takes an economist to deliver that elusive objective assessment: Tyler Cowen, by comparing trading futures for a Bush victory before and after the debate, declares that Bush won. It seems the smart money is still betting on a Bush victory and no less robustly than before.

Comments

[Abiola Lapite]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've seen the Tradesports numbers, as well as those on the IEM, but you have to remember that these are shallow markets that aren't open 24 hours a day (and in the case of the IEM at least, Dsquared has shown that it isn't efficient). I'd wait a while before assuming that they are an accurate reflection of how things stand.

I don't get how you derive the conclusion that I see the debates as ends in themselves from anything I've written. When I say Kerry won them, I mean that he won them in terms of how he came across on the big screen, and how that is likely to affect voters. Had I meant otherwise, I wouldn't have concentrated on the appearance of the two men, but on the content of their responses. Kerry looked presidential, confident and alert, while Bush looked petulant, tired and confused, and the polls accurately reflect that perception.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (MingW32) - GPGshell v3.10
Comment: My Public Key is at the following URL:
Comment: http://www.alapite.net/pgp/AbiolaLapite.txt

iD8DBQFBXS4fOgWD1ZKzuwkRAv5OAJ9abSI+bgzP1J14Y1+9fbPjg52iDgCfd/iQ
rPQ1a6dDterVt4vHA38FcXI=
=r2t4
-----END PGP SIGNATURE-----

[Frank McGahon]

(and in the case of the IEM at least, Dsquared has shown that it isn't efficient)

Well, dsquared showed that these aren't efficient but crucially he didn't show that they were less efficient than opinion polls which is the relevant comparison.

I'd wait a while before assuming that they are an accurate reflection of how things stand.

Of course, my point is more to frame how 'victory' ought to be assessed - likelihood of election improved - than to declare the winner per se (remember also, I didn't see the debates).

When I say Kerry won them, I mean that he won them in terms of how he came across on the big screen, and how that is likely to affect voters.

But this is exactly what I mean: according to your judgement he appeared better than Bush. It seems plausible to you that this performance would sway voters but the best test for whether it actually did sway voters is whether his likelihood of election was improved by the debate. It may well turn out that the trading futures exhibit a "lag" effect and you could see "Bush victory" numbers go down. Tyler Cowen left that possibility open. It is really only if any bounce in Kerry's numbers is discerned that you could say he "won" it. Until then, the sober (provisional) assessment has to be that, however shambolic you might have thought his performance, Bush won it.

[Abiola Lapite]

"Well, dsquared showed that these aren't efficient but crucially he didn't show that they were less efficient than opinion polls which is the relevant comparison."

Nor has it been shown that it is *more* accurate than all the polls in the aggregate, even if it is a more accurate barometer of who is eventually likely to win the election - and even that is a dubious assertion to make unless it can be shown that it displays less long-term volatility than all the polling series. All that has been shown for the IEM is that it does a better job of calling the winner *at the closing date*, not that it is an accurate barometer of public opinion at any given instant.

You have no evidence to support the view that either market is efficient in the sense that either adjusts immediately and accurately in taking on any news, and contrary to your claim, *that* is what is required to assume that the present figures are more "objective" than the polls.

"however shambolic you might have thought his performance, Bush won it."

Really? Where are the numbers to support it? Even on the assumption that "winning" is to be assessed as a matter of whether Bush retains his lead or not, let's wait and see how both markets close tonight to assess how their participants viewed the outcome of the debates. There isn't a scrap of evidence that these markets are so efficient that they instantaneously reflect all new information in an accurate manner.

[Frank McGahon]

All that has been shown for the IEM is that it does a better job of calling the winner *at the closing date*, not that it is an accurate barometer of public opinion at any given instant.

But the winner at the closing date is what matters and not public opinion per se. It's of no use to Kerry if a majority think he's a great candidate but this opinion doesn't translate into votes on election day. The thing about trading futures is that people put their money where their mouths are. I'd trust smart money over expressed preferences any day.

Really? Where are the numbers to support it? Even on the assumption that "winning" is to be assessed as a matter of whether Bush retains his lead or not,

Note my qualification: If you take trading futures, as I do, to be the worst method of predicting the winner apart from all the others, then the best way assessing 'victory' in the debate is whether your predicted chance of victory is improved by the debate. At the moment Bush's numbers are up, therefore it appears that he 'won' the debate. If Kerry's numbers increase and it is reasonable to associate those numbers with the debate, then he 'won' it.

Note that this is not exactly to do with retention of Bush's lead: It would be possible to declare Kerry winner, even while Bush retained a lead, so long as that lead was cut. The relevant question is: was there an effect from the debate on likelihood of election and was that effect positive?

[Brian]

Thomas Jefferson said that people tend to get the kind of government they deserve. If we give Bush four more years, then we certainly deserve him. Whether the rest of the world does is another question.

[Derek Scruggs]

"Smart money" on Bush? It just means *money* is betting on Bush, period. Was it "smart money" that bet on VA Linux at $120? How about taking the Jacksonville Jaguars over the Titans last week?

[Jimmy Sands]

Your intrincisicism point was the one which had been bugging me. I regard Kerry as having run a staggeringly incompetent campaign but he did very well last night, albeit in circumstances that clearly favoured him (not too many supplementaries pointing out his rather illogical positions, and a 40 watt opponent). He looked the part whereas I half expected to see Bush put his hands in his pockets or carve his initials into the podium. I doubt it will be of more than marginal help to him though.

There seems to me to be a very obvious flaw in the IEM model, however, and that is that betting on politics, like betting on sports, will be for many an expression of support as much as a coldly rational calculation and this is likely to skew the results.

[Frank McGahon]

I doubt it will be of more than marginal help to him though.

This is really my point: if we define "winning the debate" as meaning "looking more impressive to pundits", then it is not too difficult to declare Kerry a winner, but a more useful definition of "winning the debate" is "helping your election campaign at the expense of the other guy" and it seems as if, however counterintuitive it may seem, Kerry didn't succeed in this aim.

As for betting and "smart money" - I doubt very much that anyone is betting in this obscure market with the aim of influencing the result. This would be a tenuous proposition in actual global financial markets for anyone less rich than George Soros. It is easily more plausible that punters are betting on what they think is the likely outcome. These markets aren't perfect, but the fact that someone is prepared to risk some of their hard-earned on the outcome suggests a higher level of confidence than a cost-free prediction. The aggregate of these bets tends to produce more accurate predictions. Such markets can absorb new information very quickly. My feeling is that the expressions of support are more likely to be found in cost-free predictions.

To give an example. Rooney's hat-trick the other night has slashed the odds on United winning the premiership. This is because people rushed to put money on at the bookies rates. They faced a risk of payment so they cut back the odds. They don't need to make United so attractive to punters (all of this, by the way, while Arsenal are still favourites). There are probably still the same number of people before and after Roo's debut - including me - who hope that United do win. The difference is that while we were all prepared to declare our support for this outcome. It is only after Tuesday night that some of us were prepared to plonk down some cash on it.

[Abiola Lapite]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Some Tradesports numbers, and the latest IEM graph. Looks like Kerry did "win" after all, though Bush supporters are trying hard to spin this data as saying something else.

By the way, I don't know if you noticed, but Jane Galt has also noted that the IEM is systematically overvaluing Bush's chances, though she doesn't go into it in the detail Dsquared did. The argument that Tradesports and the IEM are individually more "objective" than the polls is a plausible one, but it isn't an established fact.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (MingW32) - GPGshell v3.10
Comment: My Public Key is at the following URL:
Comment: http://www.alapite.net/pgp/AbiolaLapite.txt

iD8DBQFBXdy7OgWD1ZKzuwkRAkQIAJ4/VjkFS2SQ7jYt9+RAeZSFJmXufgCePfmX
d9ZWzNPJgFYQWAgEmLNa14c=
=swSi
-----END PGP SIGNATURE-----

[Peter Nolan]

I did try and make it clear my estimate was very subjective, as this is the first such debate I've ever watched, and I very watch TV news. Arguing over who won seems very much a matter of de gustibus reactions, and both men seem to inspire quite irrational loathing that I don't share.

[Jimmy Sands]

What I suspected was the sort of overvaluation that Jane Galt notices, and that is I believe a factor of support. To pursue your football comparison, your current belief that United can win led you to back them, but did your previous belief (presumably) that Arsenal would win lead to a similar reaction? Football betting is not a purely rational decision. Few I think will bet against the team they want to win, however favourable the price may seem.

[Frank McGahon]

Jimmy, the point is not that individual actors have perfect information and act perfectly rationally but rather that the aggregate of those decisions tends to prove more accurate than cost-free opinions: there will be plenty of irrational, hopeful and foolish bets but so long as the market is big enough, those bets will tend to get punished. The smart money is still on Arsenal to win, they are (regrettably!) deserved favourites. There are two ways of figuring out what is the precise chance that they will win. a) The bookies' odds, reflecting the pattern of betting or b) the aggregate view of the experts in the papers. My guess is that the bookies will be right more times than the "experts" because there is more dispersed information and hopeful predictions will tend to get punished.