Home > World Events > The efficient markets hypothesis goes berserk

The efficient markets hypothesis goes berserk

July 30th, 2003

Keneth Miles, Brad de Long and a lot of slashdotters have been all over this report that the Pentagon was on the verge of setting up a futures market in terror attacks.

Apparently, the genius behind this idea is Admiral Poindexter of Iran-contra fame.

Leaving aside the obvious points about moral hazard and insider trading that have already been made, I’m impressed that the most magical version of the efficient markets hypothesis, in which markets can divine the future better than any individual, still holds sway in Washington in the wake of the bubble. Perhaps these guys have been in some sort of bunker since 1999. I wouldn’t be surprised to learn that they planned to run the thing over the Internet, have an IPO and use the billions of dollars they made to fund the Defense budget.

Categories: World Events Tags:
  1. Brad DeLong
    July 30th, 2003 at 23:31 | #1

    On the other hand, what other social mechanisms are there for aggregating information? And how good are they?

    After all, if you took your list of authoritative commentators from http://www.truthlaidbear.com and its count of who the most-linked-to webloggers are, you would wind up with:

    1.Instapundit (1541) details*
    2.Daily Kos / Political State Report (1033) details
    3.The Volokh Conspiracy (971) details
    4.Andrew Sullivan (898) details
    5.Eschaton (Atrios) (889) details
    6.Talking Points Memo (879) details
    7.USS Clueless (796) details
    8.CalPundit (793) details
    9.James Lileks (754) details
    10.Little Green Footballs (734) details

    Could a market (even a market subject to periodic bubbles) do much worse?

  2. Daniel Lam
    July 31st, 2003 at 02:55 | #2

    How inefficient does a market have to be for it to be better for the market not to exist at all?

  3. Uncle Milton
    July 31st, 2003 at 08:52 | #3

    This market wouldn’t have to be efficient ie have prices reflect all the relevant information to be useful. Prices would only have to reflect some of the information.

    If a politician firms in the betting for next assassination from 8/1 to 5/1, that could be useful information that he is being targeted by terrorists, even if the true odds are 6/1 (so the market is not efficient).

    I admit the insider trading and moral hazard issues could be difficult.

  4. Mork
    July 31st, 2003 at 10:31 | #4

    I don’t see why insider trading would be a negative: in fact, it’s almost the whole idea, isn’t it? It seems to me that we have insider trading laws basically so that investors are not dissuaded by the probability of unequal information, not because they’re necessary for an effective price discovery mechanism. In this case, unequal information would be almost a given: the point is to measure its quality through its effect on prices.

    As for moral hazard, that could be addressed by limiting positions to a size at which you would only play for fun.

    Clearly the model for this is the Iowa Electronic Market: has anyone come across a serious study of its operation and predictive record?

  5. July 31st, 2003 at 10:49 | #5

    This is not something government should do. But isn’t Robert Shiller advocating much the same thing: using financial technology to better quantify, manage and share risk? Private betting shops already do a good trade in spread-betting on things like bin Laden captre futures and the fall of Saddam. Iowa’s political futures market covers various election outcomes. And private country-risk assessment firms routinely put numbers on these issues.

  6. Mork
    July 31st, 2003 at 10:59 | #6

    Why shouldn’t government be doing it? I could understand the fuss if they were proposing to shut up the CIA use this instead, but given that this is just a little, relatively inexpensive, experiment, where’s the downside?

    I’m also a little surprised that any academic economist would not be delighted to see a governmnet funding this cool little bit of research.

  7. Uncle Milton
    July 31st, 2003 at 11:19 | #7

    Mork

    The problem with your argument is that if the insider terrorists are allowed to bet on the outcomes, then, sure, prices will reflect that information – which is a good thing, if you’re interested in knowing the most likely terrorist target. But the outsider non terrorists, will know that the insider terrorists know (not just guess) who the next target will be – and not just that, the insiders will create the outcome they are betting on – and so the outsider non terrorists won’t participate in the market, and so there will be no market, and so we won’t know who the next target will be.

  8. Mork
    July 31st, 2003 at 11:27 | #8

    Right, Uncle Milton, insider trading is a bad thing if (and only if) it dissuades non-insiders from participating.

    But I don’t think that’s a real issue here, because (a) there’ll probably be position limits that make it difficult to try to make real money without being very obvious, (b) real terrorists won’t play because it would leave a trail for the U.S government: the “insiders” who will play are those with the sort of second- or third-hand information that they don’t mind explaining to an FBI agent, and (c) it sounds like fun: lots of people will play for that reason (and to have something to talk about at parties) and not particularly care if someone gets a tip that is not available to them.

  9. John
    July 31st, 2003 at 13:58 | #9

    The inside info point is crucial, since the supply of outside info is so limited. The only way that this would perform better than averaging the opinions of the bloggers listed by Brad is if those with inside info traded on it. But this would not produce good outcomes as everyone has noted.

    One thing I should have mentioned is that markets for this kind of thing already exist. British bookmakers are happy to give odds on almost any event. I don’t get the impression that the existence of this market has greatly improved prediction.

  10. July 31st, 2003 at 14:19 | #10

    I’m all for it: let’s also have betting on W. and Jack Howard getting assassinated, the Pentagon being bombed, the Opera House being blown-up … just the thing to cheer everybody up .. can’t understand what everybody’s complaining about … damn killjoys …

  11. Mork
    July 31st, 2003 at 14:22 | #11

    John – I think you’re missing the various levels of knowledge that might be out there: leaving aside actual terrorists, there are thousands of people who are professionally vastly more clued in to ME politics than a blogger ever would be: academics, journalists, government officials, etc. etc . . . including those who are in those jobs who are actually in the ME. Imagine the insights that a goverment aide in Jordan or Saudi Arabia might have, for example, or a political science professor in Egypt. Then there are all the people who know people like that who might call them up for a tip.

    I also think the comparison with blogs is a relatively poor one, because they tend to be so insular and self referential and are almost entirely dependent on secondary information that’s already been filtered through the world of journalism.

    It’s the primary information that a market could capture more effectively than conventional means: at least if the right people participated.

    C’mon, dude, tell me you’re not at least curious about whether it would work!!

  12. ano
    August 3rd, 2003 at 12:45 | #12

    The big problem with this idea is that a small open market is used to set the price for a big closed market or to put it in other words, Bush would buy “country X is going to attack us” and than claim the market says “country X is going to attack us” and that markets don’t lie so it is must be true so lets attack them.

Comments are closed.