Over at the Volokhs, Jim Lindgren gets “upset”:http://www.volokh.com/archives/archive_2005_07_17-2005_07_23.shtml#1121820182 “twice”:http://www.volokh.com/archives/archive_2005_07_17-2005_07_23.shtml#1121843543 at co-blogger Orin Kerr for “claiming”:http://www.volokh.com/archives/archive_2005_07_17-2005_07_23.shtml#1121819891 and “repeating the claim”:http://www.volokh.com/archives/archive_2005_07_17-2005_07_23.shtml#1121843543 that prediction markets did a bad job of prediction the Roberts nomination. For Jim, the issue is whether markets are in general better than experts at aggregating publicly available information; he believes that “markets (however “good” or bad they are in absolute terms) should be better than experts on balance, or at least better than experts who lack actual first-hand knowledge of the forthcoming decision.” For Orin, Tradesports seems to be no more than a “way of monitoring what a few newspapers and blogs are saying,” and, on the whole, it “seems easier to just scan the headlines at How Appealing.”There’s an obvious alternative hypothesis which neither considers. Roberts futures shot up in value from 1% to near-certainty in the few hours before the decision was “officially” leaked. One highly plausible interpretation of this is that word had already leaked to a privileged few with good contacts in the Administration. Then, some of those people with insider knowledge took advantage of their privileged position by betting Roberts and fleecing the rubes. As Steve Bainbridge has “noted”:http://www.professorbainbridge.com/2005/07/inside_informat.html, Tradesports doesn’t seem to have any rules against insider trading. On this interpretation, Lindgren is right in saying that markets like Tradesports can provide more information on executive decisions than scanning the blogs – but only because they’re being used by those who have insider information to take advantage of the less-informed (who naively assume that they’re playing a fair game). In other words, there’s strong reason to suspect that this case doesn’t support Lindgren’s more general claims about the superiority of prediction markets vis-a-vis experts; in this case the markets are arguably being manipulated by people with insider knowledge that isn’t available to the experts. The reason that markets are doing better than experts “without first-hand knowledge” is most likely that they’re being used by experts _with_ first hand knowledge to make money from those who don’t have such knowledge. This is a very bad case to test the efficacy (or lack of same) of prediction markets in aggregating dispersed public knowledge into a usable metric; it seems to me rather unlikely that this sort of aggregation is what is in fact happening here.
Update: Orin Kerr says in comments. “You claim in your post that “Roberts futures shot up in value from 1% to near-certainty in the few hours before the decision was “officially” leaked.” That is incorrect. As best I can tell, Roberts futures shot up to near certainty only after every news website started posting that Bush had picked Roberts.” In which case, it seems to me that Kerr is right on this, and Lindgren and I are wrong, for different reasons.
Update 2: “Brayden King”:http://pubsociology.typepad.com/pub/2005/07/bad_versus_good.html has a very good post on the topic.
{ 20 comments }
Steve 07.20.05 at 10:22 am
Aren’t the “experts” in this sort of market the insiders? I suppose an outside expert can narrow the field to say 6 candidates, but to significantly bet on one of them would take insider information.
Now, I find it interesting that the insider trades waited until the last couple hours. Seems like it could have gotten as good a deal earlier in the day.
Could the greedy insiders only have gotten their info in the last couple hours? Or does it take a bit of time to open a Tradesports account?
Orin Kerr 07.20.05 at 10:27 am
Henry,
Thanks for your interest in this issue. Actually, though, both Jim and I are quite aware of the possibility you mention. In any market, it’s possible that some will have inside information and trade on it; the interesting question is, is there any evidence that this happened here?
As best I can tell, the answer is no. You claim in your post that “Roberts futures shot up in value from 1% to near-certainty in the few hours before the decision was “officially†leaked.” That is incorrect. As best I can tell, Roberts futures shot up to near certainty only *after* every news website started posting that Bush had picked Roberts.
SomeCallMeTim 07.20.05 at 10:33 am
I thought EMH assumed insider trading, and that predictive markets depended on EMH. Is this wrong?
ed 07.20.05 at 10:49 am
As best I can tell, Roberts futures shot up to near certainty only after every news website started posting that Bush had picked Roberts.
Yeah, but in the couple hours before, Roberts futures went up from negligible to substantial in a very short time period, which seems to me entirely consistent with Henry’s suggestion (a relatively small number of insiders could presumably move the futures only so far).
Or, given that Henry conceded in his update, am I missing something?
Cranky Observer 07.20.05 at 10:59 am
> markets like Tradesports can provide more
> information on executive decisions than scanning
> the blogs – but only because they’re being used by
> those who have insider information to take
> advantage of the less-informed (who naively assume
> that they’re playing a fair game)
This differs from the NYSE and Nasdaq exactly how?
Cranky
JR 07.20.05 at 11:34 am
Markets and experts both had equal access to all the relevant “objective” information about the appointment as of yesterday morning or earlier: all the information about the prospective likely nominees, the considerations regarding a woman or Hispanic nominee, the debt Bush owes to the religious right, the need to have a relatively uncontroversial nominee in light of the Karl Rove scandal, etc., etc. The only information that could possibly have developed in the last few hours would have been an actual insider tip, or observed events (like a neighbor seeing Mrs. Roberts and the children leaving the house in dress-up clothes). Most likely there were several events of both kinds. Markets have no ability to interpret each specific piece of such information better than experts, but they are capable of aggregating it quickly and adjusting the odds accordingly.
Bertrand 07.20.05 at 11:59 am
It will be a long time before I believe that the extension of futures markets past the stock market is anything other than a cutesie fad. Sort of a hula hoop for social scienctists. It’s not quite as bad as the systems for winning at the horse races or roulette, but almost.
abb1 07.20.05 at 12:04 pm
So, Mr. Bush and his cronies could (or maybe did) make a quick few bucks at the expense of everybody else by betting on a dark horse and then nominating it. What the heck kinda market is it?
Well, come to think of it, it’s just like in real life: they collect your money and distribute it to their cronies.
mkl 07.20.05 at 12:23 pm
Specualting in a trading market and publishing are just alternative means to seek to capitalize on insight one believes one has as to some future event. If one is right, one will earn cash for the right bet or greater esteem from your readers / tippees from successful punditry (which may in itself be for profit, or yield valuable information in return).
Markets can have two structural advantages over pundits.
First, there are two sides to every trade. For the market to print a 70 trade on Edith Clement, you need both a buyer thinking she has at least those odds of selection, and a seller who likes the 30% she wouldn’t be tapped. A pundit does not need a counterparty. So a trading price contains more information than a recommendation.
Second, markets are typically anonymous and give participants no exposure but financial gain or loss. This can attract insiders in some circumstances — I recall recently market disruptions in Harry Potter plot bets coming from bettors in towns near the printers — where the anonymity and financial gain are uniquely attractive. On the other hand, pundits have less cost to a wrong projection (or to change their position right up to the bell) and can in fact benefit from publishing wrong guesses, just to stay in the discussion rather than withdraw and be silent.
Further, as broadly noted, markets can quickly aggregate many opinions and weight their strength, which is not easy to do with pundits.
Orin Kerr’s correct that the markets generally trailed and tracked the pundits in this case. I would guess that this has to do primarily with the relative relationships of the insiders, pundits and bettors in the case. The insiders (say, WH staff) might well be better rewarded by leaking to pundits than betting on Tradesports (or leaking to bettors).
Zaoem 07.20.05 at 12:30 pm
I think Henry concedes too quickly. Roberts did become the favorite from almost zero odds about an hour before the reports appeared on-line. What’s the alternative hypothesis to insider trading for this? Did the markets suddenly aggregate publicly available information correctly?
Now, potentially markets are quite useful if they would reveal insider information reliably. The problem is that there is no good way to discriminate false rumors (about the Ediths) from credible ones until the reports are confirmed. Thus, people could pretend to be insiders and benefit from this.
Arm 07.20.05 at 12:40 pm
Could it be people hedging their bets? I.E. everyone buys stock i n Clement, and then–before Roberts is announced, but after the rumors start flying that it’s NOT Clement–start buying stock in the other candidates to even out their risk?
Barry 07.20.05 at 1:34 pm
“On the other hand, pundits have less cost to a wrong projection (or to change their position right up to the bell) and can in fact benefit from publishing wrong guesses, just to stay in the discussion rather than withdraw and be silent.”
Posted by mkl · July 20th, 2005 at 12:23 pm
Is there an actual cost to pundits being wrong? I haven’t noticed the absence of too many pro-Iraq war pundits, for example. Shame, of course, it beyond even laughing about, by now.
nikolai 07.20.05 at 2:35 pm
How do prediction markets operate? Are they like stockmarkets (where you trade something you own), or like a bet (where you agree with a bookmaker that if X happens you get a return)?
If they’re like a stockmarket, then insider trading (dishonestly buying or selling something) should be not allowed, but this obviously conflicts with making the best prediction. If they work like a bookmaker, then insider trading isn’t a problem, but the person setting them up stands to lose money if things go wrong. Is this a sensible distinction to make?
Chris. F. Masse 07.20.05 at 2:47 pm
Examples of recent prediction markets failures:
* SCOTUS nomination futures markets (note that the SCOTUS confirmation futures markets will certainly work finely);
* 2012 Olympic city futures markets (the markets saw Paris as the winner);
* papacy futures markets (the Pope would come from Europe, said the markets, but they failed to divine Ratzinger and Germany as country of origin);
* Michael Jackson futures markets (like the commentariat, the markets had him behind bars);
* Purcell resignation futures market (the market said he would not resign);
* George Tenet resignation futures market (idem).
Mrs Tilton 07.20.05 at 3:01 pm
Somecallmetim,
that depends on which EMH you’re talking about. ‘Strong-form’ EMH claims that, yes, market prices have already factored in all material information including inside information. But SFAICT strong EMH is a decidedly minority opinion even among EMH adherents (though I suppose that Daniel Fischel would need to believe in it, as he has written that insider trading is a good way to get information to the market).
Semi-strong EMH, which is what most EMH adherents seem to be talking about (and is the version of EMH underpinning the ‘fraud on the markets’ theory developed in the Texas Gulf Sulphur and Basic v. Levinson decisions) claims that prices reflect all publicly available information. Weak EMH claims only that you can’t extrapolate from past to future prices (in other words, technical analysis is voodoo). This last form of EMH would seem pretty uncontroversial, though I believe dsquared has said he has a mate who does in fact use technical analysis with a fair degree of success.
John Quiggin 07.20.05 at 5:17 pm
Following up Mrs T, the case for these speculative markets (such as the terrorism market) rests on strong EMH. The idea is exactly that inside information will be revealed.
Your summary is pretty much correct, though a big problem for weak form EMH comes with speculative bubbles like that of the dotcom boom, and even more the associated boom in the S&P 500.
Most of the sceptics (including me) said that the markets were wildly overvalued. In relation to the dotcoms, we could argue on fundamentals: most of these companies had no prospect of making any money, so were worth zero.
But as regards the S&P 500 we said that P/E ratios were way out of line with past experience, which is pretty close to saying weak EMH was violated, especially as you can replace E with a long-term trend assuming constant profits share.
Having said all that, I still think technical analysis is voodoo.
Darren 07.20.05 at 5:25 pm
Is anyone familliar with the mathematics behind percolation?
dsquared 07.21.05 at 2:36 am
Technical analysis isn’t voodoo. Many things done in the name of technical analysis are fairly voodooistic, but the literature is now, IMO, reaching the point at which it has been established beyond reasonable doubt that there are consistent time-series properties of stock price series. Andrew Lo has established this pretty conclusively as far as I can see.
In particular, there are lead-lag effects between large and small-cap stocks and a measurable Granger-causation relationship between them. You can make, over most time periods, a consistent return by using the returns on large-cap stocks as an element in forecasting the returns on small-cap stocks. Note that this is not a risk-free profit (although it is a more or less zero-beta profit in the sense of the standard CAPM) because what you are doing is providing liquidity to the small-cap stock market; your profit is the reward to this service, and you run the standard risks of a liquidity provider.
I think Lo’s smallcap/largecap result is the most solidly grounded violation of weak EMH, but there are others.
mkl 07.21.05 at 8:27 am
Of course, we assume the EMH is applicable at the macro level (or, as a starting point, in regard to any price we do not have specific information on), thanks to the efforts of a large number of well-compensated people who are busily deploying huge amounts of capital to wring out any inefficiencies they can exploit.
(time to stop blog commenting, back to bond trading)
jruspini 07.24.05 at 12:08 pm
In a way, what I have been saying over at my blog (http://riskmarkets.blogspot.com/) is that prediction markets will become more than a “cutsie fad” when they allow people to hedge real risks, from home prices to legislative risks like tax-code changes. (The Tradesports social security contract is the first example of this latter category)
For me the predictive ability of these markets is secondary to their usefulness in hedging risks. If price does not represent probability due to a supply/demand imbalance, longshot effect, etc, this simply represents an opportunity for speculators to profit.
Lastly, it shouldn’t be a surprise that these markets may tend to fail especially in cases where the candidate list is semi-open.
Comments on this entry are closed.