July 29, 2003

Moral Hazard

DARPA's ideas to research whether derivative markets will produce good forecasts of future politico-military events is an interesting idea in how to aggregate information (Robin Hanson has worked on this, and produced some cool stuff). Make people put their money where their mouth is, and perhaps you wind up with better forecasts. However, there is the... moral hazard problem. I mean, people who offer fire insurance have a moral hazard problem. This might create a much bigger such problem.

The UnAustralian collects some comments on this problem from Slashdot:

The UnAustralian: Predicting Terrorism

Apparently, DARPA is trying to create a idea futures market for events in the Middle East. Naturally this leads to many opportunities to take the piss out of this project. Because it's late, I'm not going to think up any myself, but will rather quote from some slashdot posts.

you hit three or four correct terrorist acts and the next thing you know you're in an orange jumpsuit overlooking guantanamo bay.

--sweeney37

I think any organization bright enough to pull off a major terrorist act would also be bright enough not to make a bet with the pentagon about when and where it would happen.

--Fred IV

"Woot! Car Bomb in Riyadh! That pays 100-1, guess I shouldn't have sat on that report about Saudi terrorists."

--typical geek

If al-quieda sign up will they get taken to court for insider trading?

--Yaruar

So this is what they mean when they talk about a "peace dividend". I never realised it was quite so literal.

--iainl

Posted by DeLong at July 29, 2003 07:45 PM | TrackBack

Comments

This program was an easy target to lampoon. However, the description provided by the proponents does not provoke immediate laughter:

"The overview of the plan said the market would focus on the economic, civil and military futures of Egypt, Jordan, Iran, Iraq, Israel, Saudi Arabia, Syria and Turkey and the consequences of United States involvement with those nations. The creators of the market envision other trappings of existing markets like derivatives."

Now, it is not at all clear what that means. If there are credible stats for things like Egypt's GDP, and we think that predicting that number is useful, well, full speed ahead. Maybe there are also credible figures which measure their military capacity - number of tanks, or some such.

But the program certainly lends itself to caricature. And I don't know how, if the stakes are meaningful and the participants are involved in the intelligence community, you solve the problem that the folks betting on terror events to happen are also supposed to prevent them from happening. (I am setting aside the notion of terrorists betting as well, since the NY Times suggests a limited number of (presumably) screened participants).

Can you imagine the 9/11 investigation if, in addition to the other lapses in communication, Congress had to puzzle over the job performance of folks who won big bets when the WTC was struck?

If you were long the WTC, and failed to circulate some bit of info that was deemed to be relevant after the fact, people would never get over it.

But be of good cheer! The future is now, at TradeSports.com

They have propositions on obvious stuff (Dem nominee, Pres election) and not so obvious stuff - Supreme Court vacancies, Gray Davis, Kobe Bryant (will it go to trial - market is roughly 65% that it will).

And for DARPA and the terror crowd, you can bet on the Homeland Security Threat level at different dates. They also have a Finding Saddam contract, a Finding Osama contract, and a Finding WMDs in Iraq contract.

Also, sports, movies, European elections, economic figures - basically, it looks like anything a bored bond trader could bet on is here.

Posted by: Tom Maguire on July 29, 2003 09:38 PM

____

I don't really see where the moral hazard problem is. Fire insurers have a moral hazard problem because giving people fire insurance makes them more careless about burning themselves up. What's the equivalent here?

It can't be, as the one slashdot poster suggests, that it gives people an incentive to hide information that would be useful. On the contrary, the whole point of the market would be that the only way for people to profit would be to disclose information (by wagering that something would happen).

Similarly, it can't -- contrary to the caricatures presented by the senators -- be that it gives terrorists an incentive to commit actions in order to profit from them. In the first place, the sums of money involved in the pilot program that was just killed were going to be very small, certainly much smaller than the cost of mounting a successful terrorist action. More to the point, in order to profit from an impending action, a terrorist would effectively have to inform on himself (since betting on the action occurring would make it more likely that DARPA would notice and do something to stop it). I think it's unlikely that terrorists would inform on themselves in this fashion, but if they did, certainly our intelligence would be helped, not hurt.

I guess there is a minor danger of the opposite possibility: that terrorists would bet against themselves in order to throw DARPA off the scent. But unless they were investing huge sums of money their bets would be unlikely to move the market all on their own.

In any case, if what's wrong with the program is that we're afraid terrorists will try to manipulate it, it seems easy enough simply to require all traders to get a national security clearance or the like, as Tom suggested.

The one substantive problem DARPA's plan did have was that if the U.S. acted on the information the market aggregated and stopped events from happening, it would take away the incentive for traders to be right (since making the right bet would make it more likely that the event you were betting on would never happen). I don't think the problem was insoluble, but it was real. The rest of the concerns the senators raised struck me as very poorly thought out.

Posted by: James Surowiecki on July 29, 2003 11:22 PM

____

Brad, I was intrigued by this idea when I first heard it. And I do think that the debate should be one of serious positive economics and not one of slogans like "Oh they're trading on death." Could you write something about the goods and the bads of this plan and expand on your present commentary?

Posted by: Bobby on July 30, 2003 02:04 AM

____

My take on the terror markets thing: http://www.demog.berkeley.edu/~gabriel/weblog/2003_07_01_archive.html#105948868666134845

Bill at rational explications critiques my take:
http://www.rationalexplications.com/blog/archives/000112.html

My overall reply: terrorist attacks are not the same as presidential elections.

He seems to think that pointing out that I live in Berkeley is a slur against me. In fact, I live in Oakland.

Posted by: Gabriel on July 30, 2003 02:15 AM

____

The point seems to be that there is a short-coming in analysis as it is now conducted, one that could be at least partly overcome by aggregating assessments of the odds of future events from those who think they have a handle on a particular future event. The very first thing that comes to my mind is that this is an effort to overcome the more poisonous aspects of competition between analytic institutions. They don't trade information, but they may be willing to compete with each other in a way that reflects that information, without revealing what it is.

Here's an idea. Why don't we just fire the next 100 intelligence analysts and managers who say "I'll get back to you with that" to a colleague from another agency, but never do.

Posted by: K Harris on July 30, 2003 04:29 AM

____

This scheme, Terrorism Futures, reminds me of Neutron Bombs, the ultimate capitalist weapon. Kill all the people, save all the capital. It's another way to make Americans seem obsessed with the dollar. Whatever it's intelligence gathering potential it's a horrible way to represent our culture, unfortunately a *small*, albeit in charge, piece of our culture is horrible.

Posted by: dennis on July 30, 2003 05:16 AM

____

DARPA was willing to try a new approach to analyzing not just terrorism but politcial instability. It might not have worked, or it might have turned out to be a pleasant surprise. Now we'll never know.

What this incident really demonstrates is how easily politics and smart alecks can stifle experimentation. Too bad for all of us.

I hope at least that DARPA management stands by its analysts and shows some guts.

Posted by: Jim Harris on July 30, 2003 05:24 AM

____

The success of any such endeavor is likely to hinge on the level of noise in the market. Just how much information about political and economic events is held outside of western intelligence agencies, valuable information not already in the intelligence mix, relative to the amount of noise from folks who think they know something, but don't. Now, who do we know who has a good grasp of signal-to-noise sorts of issues in trading?

Posted by: K Harris on July 30, 2003 05:32 AM

____

Trying very hard to be sympathetic to the people who dreamed this up, I think they probably said to themselves:'Hey! We know that the market is a means of processing information that beats anything a bureaucracy can do (Bohm-Bawerk, Mises, Hayek, etc). We want to process information on terrorism; we're a bureaucracy; let's set up a market, which will do the job better than we can. Result!' (I imagine they said 'Hey!' etc because these guys really do sound like the Grad Students who Never Grew Up.)

But there are some rather obvious potential sources of market failure, among them the fact that confidential information gathered by intelligence agencies won't be in the public realm; and, even more seriously, the fact that hte type of market that DARPA are planning to set up, a futures market, is one which is particularly prone to massive informational failure- remember the NASDAQ bubble? Apparently no-one at DARPA does, because it happened...oh..years ago. Informational failure- not what you're actually looking for in any plan to improve intelligence on terrorism.

Posted by: Dan Hardie on July 30, 2003 05:33 AM

____

Gabriel --

Terrorist acts are not like presidential elections. But if markets are good at predicting elections, box-office results, the fall of Saddam, and the outcome of sporting events -- all of which are the result of purposive actions by human beings trying to achieve a particular result -- it doesn't seem too outlandish an experiment to see how they'd be at predicting acts of terrorism or geopolitical events. Your objections to the market (with the exception of the fourth objection)are reasons why it would work imperfectly, not reasons why it shouldn't be tried. Yes, markets are noisy and messy and often driven by rumor. They're also incredibly powerful information-aggregating mechanisms. It seems strange for a country that's trying to aggregate information to dismiss them out of hand.

Posted by: James Surowiecki on July 30, 2003 07:18 AM

____

IIRC, MSNBC was reporting yesterday that the money to be deposited in the testing accounts was going to be $100.

Not a princely sum, and unlikely that anyone would make a substantial profit on $1 contracts.

To maintain the predictive integrity of the market, the system would have to be closed anyway.

It's a game, hopefully one that would have made the basis for new systems of information gathering. The issue here isn't whether or not people should be gambling on the lives of people overseas, it's whether or not DARPA should be funding a glorified Newsfutures game.

Personally, I think it was a good idea unnecessarily killed because of a fanatical bloodthirsty media that loves to shock the public, and our dullar congresspeople who love to posture for pretty media opportunities.

Everyone needs to play NewsFutures and chill out: http://www.newsfutures.com/

Posted by: J.Goodwin on July 30, 2003 08:38 AM

____

My take on the moral hazard problem: If this market did get up and running, and if it did start to predict ‘a coup in Turkey’ then if its predictions had any credibility at all there would be serious problems for the Turks (fall in share values, downward revision of their S&P’s and Moodie’s ratings, etc). This would make it well worth the Turkish govt’s while to manipulate the market in the other direction, which it could quite easily do. And people like the Saudis, with their huge resources, would have to be assumed to be manipulating the market in whichever direction most suited them.

I can’t see the big institutional funds getting involved in such a whacko idea. If the market does develop credibility, the Saudis or the big oil firms, or other players with plenty of currency, will want to finance big shifts in value to influence the Pentagon in their preferred direction. If it doesn’t develop credibility, the market will just become an another forum for the fantasies of Den Beste and LGF fans, and suchlike ignorant loons.

Finally, the point of this is presumably to provide intelligence analysis in time for policy to be changed. You didn’t get any kudos for predicting an Islamicist attack on the US on September 12th, 2001. What the US wants is better advance warning than it got last time.

So the underlying logic of this market is: markets process information better than bureaucracies; we set up a market to predict coups, wars, or terrorist attacks; when the market ‘predicts’ a coup, attack, etc, but before that event has taken place the government should change policy to forestall the event.
So- given the fact that markets do develop bubbles- we’re going to see, for example, air strikes on the Bekaa Valley because everyone in the Darpa Terrorist futures market has decided to buy the ‘Hezbollah to strike Chicago’ shares. Which might be caused by a classic market bubble, stoked perhaps by hot tips from the rightwing blogs, or perhaps because some Saudi princeling has cornered the market.

Madness.

Posted by: Dan Hardie on July 30, 2003 09:25 AM

____

Good grief. Think folks, this idea was and is a lunatic's notion. Why not have an options pool on the lives of American soldiers in Iraq? Absurd. These Administration lunatics trade in fear. Give up economics, if this lunatic's idea seems rationale.

Posted by: lise on July 30, 2003 09:29 AM

____

My take on the moral hazard problem:
If this market did get up and running, and if it did start to predict ‘a coup in Turkey’ then if its predictions had any credibility at all there would be serious problems for the Turks (fall in share values, downward revision of their S&P’s and Moodie’s ratings, etc). This would make it well worth the Turkish govt’s while to manipulate the market in the other direction, which it could quite easily do. And people like the Saudis, with their huge resources, would have to be assumed to be manipulating the market in whichever direction most suited them.

I can’t see the big institutional funds getting involved in such a whacko idea. If the market does develop credibility, the Saudis or the big oil firms, or other players with plenty of currency, will want to finance big shifts in value to influence the Pentagon in their preferred direction. If it doesn’t develop credibility, the market will just become an another forum for the fantasies of Den Beste and LGF fans, and suchlike ignorant loons.

Finally, a problem that Brad didn't mention was the use that was to be made of this market. I presume the reasoning behind DARPA's project was as follows: markets process information better than bureaucracies (Hayek, Bohm-Bawerk etc); we set up a market to predict coups, wars, or terrorist attacks; when the market ‘predicts’ a coup, attack, etc, but before that event has taken place the US government should change policy to forestall the event.

So- given the fact that markets do develop bubbles, or can suffer from any number of informational problems- policy is going to be made on the basis of an opaque and quite possibly wholly inaccurate market judgement. We could, for eexample, see air strikes on the Bekaa Valley because everyone in the Darpa Terrorist futures market has decided to buy the ‘Hezbollah to strike Chicago’ shares.

Madness.

Oh, and has anyone noticed that there really are pieces of information in national security that can't be given away? Does anyone think World War Two would have been won quicker if DARPA had been around to issue futures on 'German U-Boat codes to be completely cracked by April 1943?'

Posted by: Dan Hardie on July 30, 2003 09:39 AM

____

I'm sorry to see this thing go down the tubes: I'm in favor of the widest possible playing of games of skill for money. It redistributes capital away from the stupid and toward the intelligent.

China and the Tigers are growing great on the back of capital accumulation made possible by the mass playing of Mah Jong. America grew great when it played poker, faltered when it went to the track, and got itself ready for Vietnam by playing the slots.

Ah, but isn't gambling at the track a game of skill? Well yes and no. Nobody's gambling on the horses, because the outcome of the races is known: the favorite will win most often, the second on the Tote board will win second most often, and so monotonically on.

The gamble at the track is, are you or are you not smarter than the average of the betting crowd at picking which times the favorite will win, by a percentage greater than the sum of the State Betting tax and the Jockey Club's rake, typically about 15~17%. This is a hyper-recursive question: can you identify those occasions when your own judgement is 15% better than the crowd's at knowing whether the crowd, as shown on the Tote, is right? This is certainly a skilled question, but it is a skill which the public is known to not have, since the vast majority of the population believe themselves to be smarter than average.

Policy betting, University of Iowa style, http://www.biz.uiowa.edu/iem/, breaks this mold, and gets us back to poker, a pure game of skill. Play on, sez I.

Posted by: David Lloyd-Jones on July 30, 2003 09:41 AM

____

Besides the obvious problem that no one with any political instincts at all must have seen this program before it blew up on TV, what makes 10,000 half-assed guesses better than one certainty? If someone actually has hard evidence of an upcoming strike, besides having an incentive to sit on it, he's not going to have any more effect on the market than a guy who figures it's a better payoff than the Maryland State Lottery and picks a date out of the air?

OK, so it's limited to $100 so the incentive to sit on information is small. So is the incentive to research information. Might as well take a flyer on 10 Christmas contracts of Sharon Assassinated at $6.50.

Posted by: Rich on July 30, 2003 09:43 AM

____

Moral hazard in these information markets is a non-issue. First, these markets were intended for information gathering and aggregation and not for insurance and thus positions were to be limited to relatively small amounts. Second, if terrorists want to make money from terrorism there are far better ways than investing in the DARPA market - for example, by shorting airline stocks, taking various positions in oil, shorting nuclear energy companies, agricultural companies (e.g. shorting beef and then threatening to introduce mad cow disease) etc. These markets are more anonymous than the DARPA market and you can take much larger positions without calling attention to one's self.

The bubble issue is real but 1) the issue is comparative - how do information markets compare with say "experts" in forecasting? Evidence from the stock markets and betting markets indicate that markets perform better than the experts. HP used a market to predict future printer sales and found that it outperformed traditional methods (see Charlie Plott's Presidential address to the SEA.) Note that the experts are also subject to herd behaviour, irrationality etc. 2) There are things one can do in the design of information markets and in how one extracts the information in these markets to reduce bubbles.

Finally, one should note that these markets were primarily about pricing securites that were conditional on various policies - e.g. if the US pulls troop out of Saudia Arabia does the probability of terrorism/uprisings etc. go up or down? The conditionality of the bets precludes the reverse causality critique.

Alex Tabarrok
Department of Economics
George Mason University

Posted by: Alex Tabarrok on July 30, 2003 09:57 AM

____

"Make people put their money where their mouth is, and perhaps you wind up with better forecasts."

Yes, if the "scientists" at the Intergovernmental Panel on Climate Change (IPCC) had been required to put money on the crock that is known as their Third Assessment Report (TAR), you can bet your bippy they wouldn't have produced warming predictions ranging from a *minimum* of 1.4 degrees Celsius to a ridiculously high maximum of 5.8 degrees Celsius.

In fact, I hope to set up, on my own website, a mechanism by which people's official predictions can be recorded for:

1) surface temperature in the 21st century,

2) lower tropospheric temperature in the 21st century,

3) methane atmospheric concentrations in the 21st century,

4) carbon dioxide atmospheric concentrations in the 21st century, and

5) anthropogenic carbon dioxide emissions in the 21st century.

http://pages.prodigy.net/mark.bahner

My guess is none of the dishonest "scientists" who made the TAR not worth the paper it's printed on will submit predictions. (Robert Watson, this especially means you!)

P.S. My own prediction for surface temperature in the 21st century hasn't been absolutely nailed down yet, but a reasonable approximation might be:

5% probability: Warming will be less than -0.5 degrees Celsius (i.e., I think there is a 5% probability that Earth's average surface temperature will COOL by more than 0.5 degree Celsius, in the 21st century).

50% probability: Warming will be less than 0.5 degrees Celsius.

95% probability: Warming will be less than 1.4 degrees Celsius.

In other words, my guesstimate is that there's a 95% probability that warming will be less than the MINIMUM given by the IPCC's Third Assessment Report.

And I'm willing to put money on *my* predictions.

Posted by: Mark Bahner on July 30, 2003 09:57 AM

____

On NPR yesterday the commentator made a very good point, tangentially related to the moral hazard problem everyone is talking about: if you go long an incident you are giving the government information to prevent the incident.

As others have said above, set aside the possibility of terrorists participating in the market. Does the contract pay when the government successfully stops that incident? If not, why would you ever make such a dumb bet? The market is structured so you can't win.

Posted by: Dave on July 30, 2003 10:11 AM

____

If this market gains credibility, then any rise in value for, say, 'Coup in Turkey' futures will have serious problems for the Turkish economy (slump in share prices, downward revision of S&P and Moodie's credit ratings)- thereby giving the Turkish government, say, an incentive to buy into the market to push futures prices the other way.

The people who would seem to be best placed to rig the market are those who combine a deep interest in US policy in the Middle East with enormous financial resources. Like...oh, I dunno...the Saudi royal family. (Of course, they'd be detected the minute they tried to push share prices up or down, because they're way too unsophisticated to be able to buy through intermediaries. Ahem.)

I would disagree with Brad that the real problem in all of this is moral hazard, however. The real problem is the rationale for creating such a market. Being as kind to the DARPA people as I can be, that rationale seems to be: 'We in the US govt have a problem processing information on terrorism or the Middle East in time to act on it; We have read Hayek and know that markets process info far more efficiently than bureaucracies; Ergo we will establish a market for such information and, when it predicts a coup or a terrorist attack, act before this event has happened, to punish its would-be perpetrators.'

Now, moral hazard can distort futures markets, and so can a lot of other things, not least asymmetric information. (And a lot of the best information on terrorism or Middle East realpolitik will, by its very nature, be known only to small numbers of people.) But if this market has any purpose at all, it is to provide warnings for the US government to act on.

So if there is, say, a rush in demand for futures in 'Hezbollah to attack Chicago'- because of market manipulation, or herd behaviour by buyers, or whatever- the US govt can go ahead and bomb the Bekaa Valley. Not really a good basis for policy.

Still, World War Two would have gone better if DARPA had been around to issue futures in 'Bletchley Park to crack all U-Boat Ciphers by April 1943....oh, damn, forget we mentioned that.'

Posted by: Dan Hardie on July 30, 2003 10:16 AM

____

Why not think about what life and death really means from your cushy little retreats from reality....

Casualties:

American soldiers 107
British soldiers 11
---
118 Since May 1

American 246
British 44
---
290 Totals

Note: American forces have risen to 148,000
British forces have been cut from 10,000 to 5,000

Posted by: moen on July 30, 2003 10:21 AM

____

market gains credibility, then any rise in value for, say, 'Coup in Turkey' futures will have serious problems for the Turkish economy (slump in share prices, downward revision of S&P and Moodie's credit ratings)- thereby giving the Turkish government, say, an incentive to buy into the market to push futures prices the other way.

The people who would seem to be best placed to rig the market are those who combine a deep interest in US policy in the Middle East with enormous financial resources. Like...oh, I dunno...the Saudi royal family. (Of course, they'd be detected the minute they tried to push share prices up or down, because they're way too unsophisticated to be able to buy through intermediaries. Ahem.)

I would disagree with Brad that the real problem in all of this is moral hazard, however. The real problem is the rationale for creating such a market. Being as kind to the DARPA people as I can be, that rationale seems to be: 'We in the US govt have a problem processing information on terrorism or the Middle East in time to act on it; We have read Hayek and know that markets process info far more efficiently than bureaucracies; Ergo we will establish a market for such information and, when it predicts a coup or a terrorist attack, act before this event has happened, to punish its would-be perpetrators.'

Now, moral hazard can distort futures markets, and so can a lot of other things, not least asymmetric information. (And a lot of the best information on terrorism or Middle East realpolitik will, by its very nature, be known only to small numbers of people.) But if this market has any purpose at all, it is to provide warnings for the US government to act on.

So if there is, say, a rush in demand for futures in 'Hezbollah to attack Chicago'- because of market manipulation, or herd behaviour by buyers, or whatever- the US govt can go ahead and bomb the Bekaa Valley. Not really a good basis for policy.

Still, World War Two would have gone better if DARPA had been around to issue futures in 'Bletchley Park to crack all U-Boat Ciphers by April 1943....oh, damn, forget we mentioned that.'

Posted by: Dan Hardie on July 30, 2003 10:21 AM

____

And if we had a market for such events, and if it worked, providing the NSC or the Pentagon with far better assessments of the odds of certain events? Then the same bunch of Keystone Cops would be left to act on the data. Here are a few excerpts from a Jane Meyer piece in the New Yorker entitled “The Search for Osama.”

(http://www.newyorker.com/printable/?fact/030804fa_fact).

They smother any hope that the high level bureaucrats in the agencies responsible to protect us would take better intelligence and do much with it. The “Clark” who is quoted is Richard Clark, the top counter-terrorism guy in the Clinton and early Bush administrations.

Two or three times that fall (2000), intelligence analysts thought they might have spotted bin Laden himself. The man in question was unusually tall, like bin Laden, and drove the same model of truck that bin Laden preferred, the Toyota Land Cruiser. (The images weren’t clear enough, however, to allow analysts to discern facial features.) The C.I.A. rushed the surveillance tapes over to the White House, where the President, like everyone else, was stunned by their clarity. Later that fall, however, fierce winds in the Hindu Kush caused the Predator to crash. The accident led to recriminations inside the C.I.A. and the Air Force and quarrels about which part of the bureaucracy should pay for the damage.

“Every time we were ready to use it, the C.I.A. would change its mind. The real motivation within the C.I.A., I think, is that some senior people below Tenet were saying, ‘It’s fine to kill bin Laden, but we want to do it in a way that leaves no fingerprints. Otherwise, C.I.A. agents all over the world will be subject to assassination themselves.’ They also worried that something would go wrong—they’d blow up a convent and get blamed.”

As Clarke, who was there, recalled, “Tenet said he opposed using the armed Predator, because it wasn’t the C.I.A.’s job to fly airplanes that shot missiles. The Air Force said it wasn’t their job to fly planes to collect intelligence. No one around the table seemed to have a can-do attitude. Everyone seemed to have an excuse.”

Posted by: K Harris on July 30, 2003 10:24 AM

____

Amusingly, tradesports https://www.tradesports.com/ a UK betting website offers a line as of today which pays a dollar if admiral poindexter is still on the DARPA payroll at the end of August. It is .60 bid at .68 offered.

Posted by: John Eckstein on July 30, 2003 10:48 AM

____

If the market gains credibility, then any rise in value for, say, 'Coup in Turkey' futures will have serious problems for the Turkish economy (slump in share prices, downward revision of S&P and Moodie's credit ratings)- thereby giving the Turkish government, say, an incentive to buy into the market to push futures prices the other way.

The people who would seem to be best placed to rig the market are those who combine a deep interest in US policy in the Middle East with enormous financial resources. Like...oh, I dunno...the Saudi royal family. (Of course, they'd be detected the minute they tried to push share prices up or down, because they're way too unsophisticated to be able to buy through intermediaries. Ahem.)

I would disagree with Brad DeLong that the real problem in all of this is moral hazard, however. The real problem is the rationale for creating such a market. Being as kind to the DARPA people as I can be, that rationale seems to be: 'We in the US govt have a problem processing information on terrorism or the Middle East in time to act on it; We have read Hayek and know that markets process info far more efficiently than bureaucracies; Ergo we will establish a market for such information and, when it predicts a coup or a terrorist attack, act before this event has happened, to punish its would-be perpetrators.'

Now, moral hazard can distort futures markets, and so can a lot of other things, not least asymmetric information. (And a lot of the best information on terrorism or Middle East realpolitik will, by its very nature, be known only to small numbers of people.) But if this market has any purpose at all, it is to provide warnings for the US government to act on.

So if there is, say, a rush in demand for futures in 'Hezbollah to attack Chicago'- because of market manipulation, or herd behaviour by buyers, or whatever- the US govt can go ahead and bomb the Bekaa Valley. Not really a good basis for policy.

And the fact is that information on terrorist organisations and state intelligence agencies does rather have to be asymmetric. Unless you think World War Two would have gone better if DARPA had been around to issue futures in 'Bletchley Park to crack all U-Boat Ciphers by April 1943....oh, damn, forget we mentioned that.'

Posted by: Dan Hardie on July 30, 2003 10:49 AM

____

Why not keep the market secret, and limit its participants to qualified intelligence personnel -- perhaps to include the intelligence agencies of U.S. allies?

The idea would be to overcome bureaucratic barriers and other failures to communicate by allowing intelligence pros to play the market. Sure, your boss at the CIA may think your concern about terrorists taking flying lessons is overblown, but, by "investing" in the in-house terror market, you could signal your concerns to the broader intelligence community. If you're willing to bet enough of your allotted share of "terror dollars" (I'm assuming real money wouldn't be used; payoffs could be measured in "Starbucks" credits or something; predict the death of 3,000 innocent people, win a cup of coffee!) would prompt other spies to look more closely at the issue. It would also signal to higher-ups (hello, George and Condi) that there's something to worry about -- even if it's not reflected in the summaries their getting from their deputies.

Of course, the public wouldn't have access to any of this. It would be a purely internal, classified resource.

Just a thought.

Posted by: Charlie Robb on July 30, 2003 11:01 AM

____

Dan, we can all read your comments the first time you post them.

Predictive markets are closed, that's why they work. The Saudis can't just take their wealth and dump it into the market to affect the trade in contracts for "Saudi Government will not fall to a Coup by September 1st 2003."

Traders who make successful guesses are rewarded with the ability to control a greater portion of the total market capital (which is a fixed amount). With successive contract closings, the market should become more accurate, as those who are either just extremely lucky or who are knowledgeable about middle east politics begin to control more and more of the totla investable dollars.

Believe it or not, this Discover article [ http://www.discover.com/aug_03/gthere.html?article=feattech.html ] on automomous computer animation is related to the concept. The market becomes "smarter" over time because those who have made the best attempts in the past are enabled to make more attempts in the future.

Posted by: J.Goodwin on July 30, 2003 11:04 AM

____

Rich asks: "Besides the obvious problem that no one with any political instincts at all must have seen this program before it blew up on TV, what makes 10,000 half-assed guesses better than one certainty?"

Markets don't average all the guesses. They assign a value to the margin where the most aggressive buyers meet the most aggressive sellers.

When the commodity being bought and sold is opinion about reality, the margin is where the people who have worked the hardest at figuring out reality get to meet.

There they get to take money away from people who haven't done their homework as well.

This is a strong Darwinian process for doing two things: transferring money from dumb lazy people to smart hard-working people, and incidentally getting a pretty good guess at what the truth of a situation is.

Posted by: David Lloyd-Jones on July 30, 2003 12:52 PM

____

Rich asks: "Besides the obvious problem that no one with any political instincts at all must have seen this program before it blew up on TV, what makes 10,000 half-assed guesses better than one certainty?"

Markets don't average all the guesses. They assign a value to the margin where the most aggressive buyers meet the most aggressive sellers.

When the commodity being bought and sold is opinion about reality, the margin is where the people who have worked the hardest at figuring out reality get to meet.

There they get to take money away from people who haven't done their homework as well.

This is a strong Darwinian process for doing two things: transferring money from dumb lazy people to smart hard-working people, and incidentally getting a pretty good guess at what the truth of a situation is.

Posted by: David Lloyd-Jones on July 30, 2003 12:57 PM

____

Fundamentally, the function of this market is to provide the US Government with advance warning of adverse events *in order that the US Government can act before these events occur to forestall them*. (Unless J Goodwin foresees Don Rumsfeld testifying thus to us Senate Committees: 'Senators, if I get credible prior notice of a terrorist attack on the US or on states friendly to the US, my policy will be to do absolutely nothing'. 'Why thank you, Secretary Rumsfeld, excellent plan.')

What is wrong with this? Predictive markets work, says J Goodwin, because they are closed. Actually, before that, they work because they reward accurate predictions. This 'market' will in practice do precisely the opposite.

If the market trend is to buy futures in, say, 'US base in Bahrain to be attacked by terrorists in October 2003', then, if the US Government takes the market seriously as a predictor, it will *take action to ensure that US bases in Bahrain are not attacked in October 2003 *. (Plenty of things you can do to forestall such an attack, believe me.)

Thus: if I and most other traders in the market guess right and terrorists do plan to bomb GIs in Bahrain in Oct '03, then the US govt will take precautions against any such attack. If no attack takes place in Oct '03, it could be because no such attack was planned in the first place; it could be because it was called off for reasons unconnected to the DARPA market; but it could be because the DARPA market predicted it and the US Govt acted to forestall it.

In other words, if I guessed right about terrorist attacks in Bahrain, I would lose money if my warning were taken seriously. That seems like a fairly classic case of 'Perverse incentives'.

So while I agree with Brad Delong that moral hazard is a problem in this market (and so is herding behaviour, and so is asymmetric information) the most fundamental problem seems to be the one I've outlined above.

Posted by: Dan Hardie on July 31, 2003 05:58 AM

____

Cultivated people foster what is good in others, not what is bad. Petty people do the opposite.

Posted by: Cohen Norman on December 10, 2003 10:20 PM

____

In his errors a man is true to type. Observe the errors and you will know the man.

Posted by: Maroney Terry on January 10, 2004 03:01 AM

____

Post a comment
















__