Robin Hanson on prediction markets

One common worry about NGDP futures markets is that no one would be interested in trading the contracts.  After all, if there was interest then why wouldn’t someone already have created such a market?

In fact, the lack of interest in NGDP futures would actually improve the efficiency of an NGDP targeting regime.  If lots of people wanted to hedge against NGDP risk, then a futures price might be biased, it might not reflect the expected future level of NGDP.  In other words, it might include a risk premium.

But the bigger problem with these critics is that they haven’t bothered to even read the proposals.  The plan is to have the Fed subsidize futures trading, at a level high enough to create a relatively efficient market.

Robin Hanson makes a similar point (when criticizing Casey Mulligan):

No, a market can single-task, with no other function than prediction, and no other trader motive than selfish financial profit, if someone who wants the info will pay to subsidize the market. It is only when you want people to answer your question for free that you’ll have to piggyback on their having some other reason to trade.

Of course there’s no guarantee that your willingness to pay for some info exceeds other folks’ cost to supply that info. Supply and demand curves need not intersect at a positive quantity. But that’s hardly a failure of an info exchange mechanism.

Later Robin criticizes Snowberg, Wolfers, and Zitzewitz on much the same grounds:

Noise traders are traders who subsidize your market for free, for reasons of their own, such as risk-hedging, idiocy, etc. If you fail to attract noise traders, you fail to get their free subsidy. But you can still offer to directly pay for your info, by subsidizing the market, as the Microsoft sentence in the quote indicates. Similarly, if employees find executive questions uninteresting, that just means they won’t answer such questions as freely in their spare time. But that hardly means firms can’t pay employees to address key firm questions. Here we are only talking about a “failure” of prediction markets to mooch stuff for free!

Given the cost of business cycles can run into the trillions of dollars, I don’t see any problem with the Fed spending a few millions, or even tens of millions, to insure interest in an NGDP futures market.

Michael Sankowski says the following:

Economically useful contracts apply to every possible futures contract design. If the contracts aren’t economically useful, nobody trades them. This means the Central Bank does not get private sector forecasts from NGDP level futures.

Michael still doesn’t seem willing to actually read the proposals; otherwise he would have known that the market would be subsidized, so there is zero risk of no one trading the contracts.  If no one else trades, I will, and I’ll gladly walk away pocketing the entire multi-million dollar subsidy.

Later he repeats the claim that Goldman Sachs would walk off with $500 billion despite the fact that the Fed could easily set up the market in such a way that it never took a net long or short position.  I think people might find it more useful to think of this as a “prediction market” rather than a “futures market.”  The point is to have the public predict the monetary policy instrument setting that is most likely to result in on target NGDP growth.  There are many ways of doing this, some allow the Fed to take a position (if it sees market inefficiency) and others work more automatically, setting the instrument at the point where the public’s net long and short positions exactly balance out.  It’s entirely up to the Fed how much risk it wants to take.  I could easily visualize a small prediction market, in the millions of dollars, not billions.   Studies suggest that even smaller prediction markets (in the thousands of dollars) can be highly efficient.

Here’s another common misconception about prediction markets, this time from Free Exchange:

FOR over a year, the prediction market site Intrade offered a contract on whether the Supreme Court would rule the Obamacare’s individual mandate unconstitutional by the end of 2012. For most of its life, the contract traded below $5; the collective wisdom of the market suggested the mandate would stand. In late March, however, a surge of public scepticism about the Court’s tolerance for the mandate led to an impressive jump in the price. By the eve of the Court’s ruling, the market put the odds that it would be struck down at nearly 80%. Then the fateful day arrived””and on word that Chief Justice Roberts voted to uphold the mandate as a tax the contract instantly plummeted to near zero. So much for the wisdom of the markets, right?

Not quite, says a new NBER working paper by Erik Snowberg, Justin Wolfers, and Eric Zitzewitz. Their research sets out to show how prediction markets can provide the best available estimates of future events and figures. Yet while the paper argues strongly for the utility of markets, it also offers plenty of reason to treat their conclusions cautiously.

.   .   .

In what ways do prediction markets fail? The paper provides some discouraging answers. First, they struggle when there is a high degree of insider information. On the question, “Will the mandate be struck down”, for instance, only the Chief Justice himself could say for sure, and so the market was likely to be wrong. There must be information to aggregate.

The market didn’t “fail” at all, the 80% forecast was probably the optimal forecast.  And it’s not true that the traders had no information—the questions asked by the conservatives made many people think they were leaning toward rejecting the individual mandate.  And they were leaning that way!  Kennedy was almost universally viewed as the swing vote, and even he ruled against Obamacare.  What few people expected was a last minute change of heart by Roberts.  Sure, there was always some uncertainty, that’s what 80% means.  That’s why the market didn’t price in a 100% chance of the law being overturned.  (Robin Hanson makes a similar argument.)

Consider the following analogy:  Two prediction markets are set up to predict the toss of the coin before the next Super Bowl.  One says 50% odds of heads and the other says 58% odds of heads.  Then the coin is tossed, and it’s heads.  Which market “failed?”  I’d say the market with the 58% forecast.  They made a bad forecast and simply got lucky.  I find people are way too mesmerized by predictive success.  The important issue is whether a market delivers an optimal forecast, not whether it turns out to be correct in any single case.



42 Responses to “Robin Hanson on prediction markets”

  1. Gravatar of JoeMac JoeMac
    20. July 2012 at 08:18

    I think I will ask the obvious question. How come no one in the financial markets has already created such a market?

  2. Gravatar of Saturos Saturos
    20. July 2012 at 08:19

    Yes, you should officially rename it NGDP prediction markets.
    Better yet (since prediction markets don’t have such a reputation yet), call it, “the NGDP standard).

  3. Gravatar of Josiah Josiah
    20. July 2012 at 08:19

    While I’m a fan of the prediction market concept in general, I’ve grown more skeptical about actually existing prediction markets such as Intrade. During the Republican primaries, for example, there were a lot of implied predictions that were wonky even compared to predictions being made by polling experts or based on a “look at the fundamentals” approach.

    No doubt the question will be: if this was so why I didn’t take the opposite position and get rich? Well, it’s not so easy. You can’t just log onto Intrade and start betting. You have to wire money to them at which point the money sits in an account not bearing interest until the contract expires. If a had a big pot of money to play with there are arbitrage opportunities you could exploit, but if you’re just an ordinary guy who thinks a single market is off the transaction costs involved in trying to rectify it can be prohibitive.

  4. Gravatar of Saturos Saturos
    20. July 2012 at 08:20

    JoeMac, it’s a public good, only the central bank would or should subsidize it.

  5. Gravatar of Michael Michael
    20. July 2012 at 08:27

    Who would be expected to participate in NGDP futures/prediction market, and why?

  6. Gravatar of JoeMac JoeMac
    20. July 2012 at 08:37


    But plenty of other derivatives markets exist, why doesn’t somebody create one?

    If people would e willing to buy and sell such derivatives, then somebody on wall street can make much profit by creating such an exchange.

  7. Gravatar of ssumner ssumner
    20. July 2012 at 08:39

    JoeMac, Fortunately, there is little interest in trading NGDP futures contracts.

    Saturos, You might be right. By the way, as I point out in one paper, it’s conceptually no different than tying the pay of FOMC members to how accurately they voted on instrument setting decisions. And then expanding the FOMC from 12 to 7 billion.

    Josiah, Studies have showed that prediction markets are pretty efficient. In any case, NGDP traders would earn interest on their margin accounts.

    Michael, People who follow the economy, and think they have insights into NGDP. In late 2008 I would have gone massively short, as it was obvious NGDP would fall short of target.

    It might be people who see information before it shows up in government data, like inventory controllers for Walmart.

  8. Gravatar of JoeMac JoeMac
    20. July 2012 at 09:08


    Why “fortunately”? I don’t understand. Shouldn’t you WANT people to be interested in this and have some entrepreneur set one up?

  9. Gravatar of Nick Nick
    20. July 2012 at 09:11

    How do you deal with the basic problem that setting policy based on a single NGDP futures market is not a robust system? That is, doesn’t it introduce a single point of failure, and if that one market is wrong it screws up the entire economy? While I tend to agree with EMH, there will be times when any market is wrong in that it doesn’t even give the optimal forecast. When that happens to say, the value of internet stocks or real estate, it doesn’t blow up the economy – assuming that the Fed does not sharply tighten as a policy response 🙂

    The current situation allows a central bank a certain amount of wiggle room (“charade” you called it the other day) where they can look at the various metrics and then decide to do whatever they want to do, and so if some market is out of whack (in their opinion), they can ignore it and look at the other data that is available. But if were an NGDP futures market in operation, and they believe that it is in a “bubble,” don’t they still have to tighten (or loosen) as that market prescribes, regardless of their opinion?

    Now I know you are going to say that such an autopilot is a feature, not a bug, but then you are guaranteeing a policy error whenever the NGDP futures market gets out of whack.

    You would argue that when the NGDP market is “wrong,” the chances that the experts at the Fed are (more) “right” is infinitesimal. But sometimes prices move not merely due to crazy or stupid herd behavior by participants, but rather by market actors squeezing others to increase returns. Even fairly well traded markets such as natural gas (Amaranth) and CDS (JP Morgan’s “Whale”) have seen such market reactions, and it is likely that an NGDP futures market will experience a bubble or a squeeze at some point, bringing the entire economy along with it.

    Do any of the proposals contemplate bringing a diversity of different markets to bear on the problem of NGDP expectations? Even at the expense of some clarity for policymakers, it seems better than to risk the state of the economy on a single set of NGDP futures prices.

  10. Gravatar of Bill Ellis Bill Ellis
    20. July 2012 at 09:29

    So often I think the biggest problem in discussing economics (as well as everything else ) lies in Semantics.

    In the case of the coin toss neither market “failed”.
    The 58% assumption was a flawed assumption based on a normal coin and normal toss conditions, But no market would have predicted 58% with out some information (true or not ) of a nonstandard toss/Coin.

    In the case of the mandate the Market DID fail. It was wrong. The 80% was based on flawed info.

    But none of this should have anyone saying… “So much for the wisdom of the markets, right?”
    Markets are the best way we have of making predictions, tho imperfect. It would be stupid to disregard the best tool we have because it is not perfect.

    Markets are not always right because information is not always right. Arguing that they always are only confuses things.

  11. Gravatar of StatsGuy StatsGuy
    20. July 2012 at 09:31

    “I think people might find it more useful to think of this as a “prediction market” rather than a “futures market.” ”

    Sorry Scott, I have to side with the critics here. I don’t see how any of these items – whether or not the Fed subsidizes participation or does not subsidize participation – affects the core issue. The market is inherently NOT independent. If it’s common knowledge that the results of this small market affect the Fed’s action vis-a-vis a much larger market (which in turn impacts every market drastically), then the prediction market cannot remain small…

    If the market is small, then it is too easily manipulated – someone could bid up or down futures to influence Fed action, taking a loss in the small market to make a much bigger gain in the larger market. These markets become inseparable.

    The only way this works is if the losses are comparable – and, indeed, if the market was sufficiently liquid this would happen. In other words, the prediction market would need to be HUGE (either in real terms, or in synthetic terms), and to serve as an effective hedge on the economy and other markets. I might thus take a position in the futures market to offset interest rate risk in the treasury market.

  12. Gravatar of Major_Freedom Major_Freedom
    20. July 2012 at 09:38


    In fact, the lack of interest in NGDP futures would actually improve the efficiency of an NGDP targeting regime. If lots of people wanted to hedge against NGDP risk, then a futures price might be biased, it might not reflect the expected future level of NGDP. In other words, it might include a risk premium.

    Why would an individual care about hedging NGDP risk, when the only risk that affects him would be the risk of his market activity? If I have a synthetically constructed risk-free security (combination of options, underlying, etc) that has payoff structure X in the future, then it doesn’t matter to me what NGDP becomes in the future. A seller of corn won’t care if the demand for computers fall such that NGDP falls. He would only care about the demand for his corn.

    If “NGDP” is claimed by MMs as something that is a risk to this farmer’s business, then it should show up as a risk in the future price/demand of his corn. Well, he can hedge against that risk by selling corn futures. What would he need an NGDP futures for? What would any individual investor need an NGDP futures for?

    I think JoeMac’s question is a reasonable one. If NGDP futures are indeed economically useful, then surely some trader at some financial institution in some country in the world at some point would have created them by now, no? NGDP has been around since 1776. You would think that at least ONE investor would have found them useful. But to my knowledge, not a single investor has ever created or traded these types of futures.

    Shouldn’t that tell you something? That maybe the concept of an NGDP futures market that implicitly requires central banks to create what no investor in the market has ever created on his own in the history of trading, is really nothing but arm-chair theorizing that is fundamentally flawed but you just don’t see it?

    Have you considered that maybe the flaw is your belief that NGDP targeting is inimical to free markets, which manifests itself in a total absence of NGDP futures in the market to begin with?

    I’ve been told many times on this blog that CEOs and CFOs and businessmen really do care about NGDP. Yet how can that be if not a single one is hedging against NGDP volatility? Why aren’t investors betting with each other on the future path of NGDP? Why haven’t NGDP futures markets appeared on any exchanges? If NGDP really was something that is important to an individual’s business, then surely at least SOME investors would be making NGDP bets with each other. I would submit that the total absence of NGDP futures is consistent with my position that nobody cares about NGDP except MMs.

    If it is the case that the more investors become interested in NGDP futures hedging, the less efficient NGDP targeting based on said futures becomes, then what you are inadvertently showing is Goodhart’s Law applying to NGDP targeting. Once NGDP is made a target for the purpose of conducting social or economic policy, then it will lose the information content that would qualify it to play that role.

    The path of NGDP growth, therefore, could only show what you think it shows (I’m not saying it does show it however), if a Reverse Goodhart’s Law applies to it, that is, only if enough investors believe it is not a credible target, can NGDP even in principle qualify as having the information content you say it has.


    If no one else trades, I will, and I’ll gladly walk away pocketing the entire multi-million dollar subsidy.

    Suppose a few investors do purchase these NGDP futures with you. Suppose that you form a cartel-like agreement to not spend the entire sum of money the Fed creates on anything that is tracked by NGDP, but only, say, $1 million a month. Wouldn’t the rate of money creation go to infinity, because the Fed would be stuck with having to target an NGDP that depends on you and the few other futures market traders actually spending the subsidies?

    In other words, if the NGDP futures market will work most efficiently if there are only a few investors trading in them, then isn’t Michael Sankowski right that the new “Goldman Sachs of the world” can walk away with $500 billion?

    You say this is not necessary, because:


    …the Fed could easily set up the market in such a way that it never took a net long or short position.

    That is impossible. For the Fed could not even inflate the money supply. The Fed is a non-market institution, and so it has to take an offsetting position to the market if the end result is going to be non-market inflation into the market! Just like an investor in a long position requires an offsetting short position investor to PAY UP at contract expiry, so too would the market that (allegedly) needs inflation have to take an offsetting position with the Fed so that the Fed PAYS UP, which is inflationary, which then generates 5% NGDP growth.

    The Fed is going to have to remain in a net short position the entire time if they are going to maintain inflation. For they have to buy futures in order to inflate. If they sold futures, that would remove money from the market. The rate of buying NGDP futures MUST be greater than the rate of selling NGDP futures, in order for 5% NGDP growth to take place. After all, the prices of these futures would be fixed by the Fed, would they not?

  13. Gravatar of Major_Freedom Major_Freedom
    20. July 2012 at 09:43

    typo: “not inimical.”

  14. Gravatar of Bill Ellis Bill Ellis
    20. July 2012 at 09:55

    Does targeting NGDP have to be based on an actual market ? Is it a mechanical necessity ?

    I know I just said that markets are the best predictive tools we have, but markets can be virtual.

    If the hang up is getting investors to participate …AND keeping investors from manipulating the market…why bother with actual investors ?

  15. Gravatar of Tomasz Wegrzanowski Tomasz Wegrzanowski
    20. July 2012 at 10:32

    Intrade has pretty mediocre track record on issues other than US elections. Using market efficiency arguments on Intrade is pretty ridiculous, they’re internally inconsistent, it’s just impossible to make enough money out of their inconsistencies due to structure of the market to cover your time.

    Anyway, why neither you nor Robin Hanson managed over all these years to get anybody interested in even toy money NGDP prediction market for major economies? (or inflation/unemployment/whatever) This sounds like such a simple project – you just need to find someone to setup a website like old play money site, and advertise it widely enough.

  16. Gravatar of Mike Sax Mike Sax
    20. July 2012 at 10:50

    Actually the one person who predicted right about Roberts flipping was Alberto Gonzalez:

    “I spent a great deal of time vetting Justice Roberts in making my recommendation to President [George W.] Bush that he appoint Chief Justice Roberts to the court,” Gonzales said on CNN. “One of the traits I most admired about him, and this is very consistent in his judicial decision making, is to decide decisions on the most narrow grounds possible, to not get to constitutional issues you don’t have to in order to dispose of a dispute….So in that respect, I expect Justice Roberts to follow that approach in deciding this case.”

    “During oral arguments in March, the justices appointed a lawyer independent from the case to argue that a law known as the Anti-Injunction Act bars the court from blocking collection of a tax, like the penalty mechanism the law would use to encourage Americans to get insurance. The Anti-Injunction Act says a tax must be collected before it can be challenged in court.”

    “At oral argument, none of the justices seemed to think much of that idea, but Gonzales suggested it would be the most consistent with Roberts’s judicial philosophy favoring something lawyers call “the doctrine of constitutional avoidance.”

    “That may mean, that he’s going to be pushing the court to perhaps not make a decision on this case, wait until 2015, when the penalties on individual mandate come into play….Perhaps the chief justice is not going to go that way, but I wouldn’t be surprised if he did,” the former attorney general and one-time Texas Supreme Court justice told CNN.

  17. Gravatar of Mike Sax Mike Sax
    20. July 2012 at 10:55

    Of course you don’t think predictive success is so important. If you’re in the market though that’s all you care about.

  18. Gravatar of Mike Sax Mike Sax
    20. July 2012 at 11:30

    Still Hanson seems to feel differntly about predictive success:

    “Finally Snowberg, Wolfers, & Zitzewitz end with this stunner:

    We believe the real promise of prediction markets comes not from their ability to predict particular events. Rather, the real promise lies in using these markets, often several at a time, to test particular economic models, and use these models to improve economic forecasts.

    “This seems like saying the real promise of democracy is that academics can study votes to refine their theories of human behavior. Really?!”

  19. Gravatar of dwb dwb
    20. July 2012 at 12:20

    i see i got featured in Michael Sankowski post. i just dont have time to get into a is-the-earth-flat-or-round debate with an MMT/MMR’er anymore (however, i will say this: even if he is right he is wrong, no risk manager on the planet would approve a risk limit where you could make or lose a 400 bn, and the margin call for the loser would be killer). no seriously, killer: a bankruptcy event.

  20. Gravatar of Becky Hargrove Becky Hargrove
    20. July 2012 at 12:35

    he sure wanted a fight! But I remember what you said about the bottle of scotch the other day: same here, not as easy to drink these days as to think about the fact that it’s pleasant.

  21. Gravatar of o. nate o. nate
    20. July 2012 at 13:01

    I agree with StatsGuy. If the Fed is committed to moving monetary policy on the basis of where these NGDP futures contracts trade, then they wouldn’t have to subsidize it – it would grow huge on its own. Otherwise, it would be too easy for any deep-pocketed player to game the market and then make a killing on side-bets on the future direction of Fed policy (for instance by betting on short-term interest rates).

  22. Gravatar of dwb dwb
    20. July 2012 at 15:20

    “same here, not as easy to drink these days as to think about the fact that it’s pleasant.

    im weak. if my nose gets anywhere near a peaty single malt im toast. we brought some back from Scotland from some distilleries we toured, didnt last long.

  23. Gravatar of Overcoming Bias : Prediction Markets “Fail” To Mooch Overcoming Bias : Prediction Markets “Fail” To Mooch
    20. July 2012 at 16:01

    […] 20July: Scott Sumner says people similarly misunderstand his proposal for NGDP futures. GD Star Ratingloading… Tagged […]

  24. Gravatar of Mike Sankowski Mike Sankowski
    20. July 2012 at 17:03


    That’s my exact next post. NGDP derivatives are a natural for wall street exotics desks, and as far as I know, there are no NGDP derivatives.

    I know the Inflation Trader, and he wrote inflation derivatives for years on Wall Street. We talked about inflation futures/derivatives at length.

    I also worked a with a well known company on possible inflation futures. You would not believe the pedigree if I told you. You would not believe the brainpower and reputation of the people involved.

    Inflation futures are close to NGDP futures, but are still quite far away. Inflation derivatives don’t trade that much on the street.

    Nobody currently trades NGDP derivatives because nobody wants NGDP derivatives to hedge existing risks. Basically, what is being proposed is the equivalent of an online survey.

    If the market is too small, it’s open to massive manipulation by China, Koch, Soros, or Dr. Evil. Basing Fed policy on an online survey is a bad idea.


    Once again, I support NGDP level targeting.

    You make a good point in that prediction markets offer the best possible estimate. The problem with this particular prediction market is the market size and national importance.

    NGDP is not a good enough hedge for big enough players to make the market large. Therefore it’s totally open to manipulation by Dr. Evil.

    I can only address one part of flawed NGDP futures at a time. I needed to individually knock down each plank with 100% certainty otherwise people would say “But what about contract structure X?”. I didn’t want there to be any question any possible contract structure for NGDP futures are flawed.

    You vastly underestimate how the weak correlation of NGDP to most corporate and institutional risk factors makes NGDP futures/derivatives a non-starter. It’s a problem for every participant in the market ecosystem.

    When I say “nobody” will trade them, it means the market would be very small. I don’t literally mean “nobody”.

    I’ve disposed with the idea the fed could be a market maker at the NGDP target level. If the price floats, the market will be very small because nobody needs to hedge NGDP. This leaves the market open to manipulation.

    In addition, I am not clear as to what you might possibly mean by subsidize the market. In any market, the lowest fees are the brokerage and execution fees. The highest fees are the bid/ask spread.

    I’d recommend using something like TIPS or other widely traded vehicles to estimate inflation. GDP – you’re on your own, because I don’t see any index anywhere which can help. The GDP data collection process is so subject to error, it makes me sick. But even worse, it dooms NGDP futures.

    I suspect as data collection becomes better, NGDP futures will become more viable. I estimate it will take another 30 years before using real time data will allow NGDP futures to be a possible solution to the Lucas Critique.

  25. Gravatar of Saturos Saturos
    20. July 2012 at 17:49

    JoeMac, Michael, did you not read the blog post on this page?

  26. Gravatar of Mike Sankowski Mike Sankowski
    20. July 2012 at 20:18

    I did read this post.

    How exactly will the NGDP prediction markets be subsidized?

    Will there be market maker incentives, based on exactly what criteria?
    Will participants get a cash bonus regardless of profit or loss?
    Will execution fees be waived – or perhaps execution credits?
    Do participants get a per contract payment against possible losses?

    Let us know the exact subsidy mechanism. I’ve actually written the legal verbiage on market subsidy contracts for market makers. I have a pretty good idea how market subsidies work.

    Subsidies can’t overcome poor hedge correlation. The bid/ask spread is the largest cost for every transaction. You’ll need to think of a way to make the bid/ask spread much smaller than the measurement error of GDP through subsidies. Otherwise market makers will not make markets smaller than a full percent wide.

    It’s easy to make markets when you can arb against something liquid. GDP does not have a liquid market to arb, therefore any incentives need to overcome possible losses for the market maker. Since the measurement of GDP is such a guesstimate, the fed will have to provide huge incentives to market makers.

    Basically, you need to give market makers something close to a guaranteed profit. That’s the subsidy. And a guaranteed profit is tough to promise given the imprecision of the GDP index itself and the lack of highly correlated, liquid hedges.

    As we know, subsidies distort markets. We will no longer be getting pure information from the NGDP prediction market if subsidies are used to incentivize trading.

    Then regarding small markets under many billions in size – How exactly will you justify to the CFTC the NGDP prediction market won’t be manipulated? When I’ve been part of the team launching futures contracts, the CFTC was extremely interested in how a market (and pricing) might be manipulated.

    Yes, the CFTC may not have jurisdiction over the NGDP prediction markets. The concerns of manipulation in small markets remain in force no matter who is overseeing the market.

    Players like Walmart may not have a direct interest because NGDP doesn’t hedge anything. Certainly China does care about the actions of the U.S. federal reserve.

    What would China do with a small market which dictates U.S. federal reserve action? Heh.

    Of course you already know the CIC and other sovereign wealth funds make many investments in hedge funds through difficult to trace foreign entities.

    Additionally, there are political reasons to force the fed to act in dumb ways. Does anyone doubt the Koch brothers would toss a $20-$30 million at NGDP prediction markets to compel the fed to tighten rates today? Certainly Soros would be accused of such manipulative behavior, even if he did not make trades in the NGDP prediction markets. There’s a PAC for that.

    dwb, It is true goldman alone would not get $500bn, and no risk manager on the planet would allow their firm to risk anywhere near the amount necessary to make $500bn. I exaggerated the title to get you to read the post.

    But as an industry, $500bn is a reasonable estimate of the profits. This is why I pointed out the entire hedge fund industry is $2.1tn in size. Allocating less than 5% of industry assets would make this trade worth hundreds of billions to the hedge fund industry. The headline would read something like “U.S. pays Goldman, Hedge Funds $500bn during recession.”

    You should copy and save my comments and posts, so when someone tries to launch NGDP prediction markets, the person designing the contracts won’t have so much work to do.

  27. Gravatar of Major_Freedom Major_Freedom
    20. July 2012 at 21:15

    Mark Sandowski:

    The headline would read something like “U.S. pays Goldman, Hedge Funds $500bn during recession.”

    Haha, joke’s on you, they already do that now. “Recessions” are already times when the Fed pays the primary dealers oodles of cash.

    The word “recession” in our lexicon has the same effect as “war.” It makes people believe that certain actions have the opposite ethical connotation. “You say 73,846 Americans have been killed in Iraq since Aug 1990? That’s horrific! What’s that? We’re…..“at war”? Oh, well, that’s still pretty bad, but it’s war, so, you know, the alternative would be worse…”

    “You say the Fed printed off trillions for the major banks? That is an outrage! What’s that? We’re…..“in a recession”? Oh, well, that’s still pretty bad, but it’s a recession, so, you know, the alternative would be worse…”

  28. Gravatar of ssumner ssumner
    21. July 2012 at 05:50

    JoeMac, No, you want no interest, so that there’d be no risk premium.

    Nick, Yes, they’s make mistakes, but far less than discretionary policymakers.

    Bill, I define failure as when the market price doesn’t equal the optimal forecast given available information.

    Statsguy, There are lots of options:

    Allow a huge market.

    limit the position of each trader.


    Sankowski, You said;

    “Nobody currently trades NGDP derivatives because nobody wants NGDP derivatives to hedge existing risks.”

    That’s a huge advantage to my proposal, as I indicated.

    You said:

    “How exactly will the NGDP prediction markets be subsidized?

    Will there be market maker incentives, based on exactly what criteria?
    Will participants get a cash bonus regardless of profit or loss?
    Will execution fees be waived – or perhaps execution credits?
    Do participants get a per contract payment against possible losses?”

    Have you ever considered actually reading the proposal before asking all these questions, and writing two posts?

    Everyone, Short of time, I’ll do the rest later.

  29. Gravatar of Mike Sax Mike Sax
    21. July 2012 at 07:27

    The idea that more guns means less crime is baloney. Here’s some actual sense being spoke:

    “this model of multiple equilibria applies to guns. There are societies with few guns and strict regs on them. Think Japan. Most of their gun deaths are socially approved suicides derived from the Samurai code of honor and Seppuku. But, in general, the rate of gun ownership and gun violence in Japan is extremely low. Many other nations on this planet resemble Japan in this generally, although varying in details on many important aspects. These societies are in the “good” equilibrium, where general social disapproval of guns combined with strict laws regulating them has led to few around, while organized crime and selected others can get them, but the rate of gun deaths is low by global standards, and certainly compared to the rate in the US.”

    “OTOH, the US is clearly stuck in the “bad” equilibrium, derived from a long individualistic frontier history of widespread gun ownership, reinforced by the Second Amendment with its two parts, one emphasizing the right to own guns, recently raised above the other part by the US Supreme Court against long established precedents, with the other emphasizing the need for state level militias to be supported by a gun-owning citizenry in a society without a federal national defense (uh huh, compare 1787 national military [basically zero] to current DOD, duh). The control variable is the number of guns out there per capita, and the long US history favoring guns has meant that the barn door was torn off long ago and guns are everywhere, with no chance to go to the low gun “good” equilibrium anytime in the near future. Local efforts to control guns are hopeless as they pour in from other parts of the country, such as Virginia supplying the gangs of New York, with newly revived relaxations against the pleas of law enforecement officials to the northeast, but, hey, here in Virginia, the NRA really has the legislature in its grip to kill kill kill.”

    “However, this most recent shocking event makes clear that even if the US has no hope of ever getting to that “good” equilibrium, maybe we are going too far into the bad equilibrium zone, and that more and more guns do not lead to less crime at all. What can we do?”

    “Well, there is one obvious move. Reinstate the previously existing ban on assault weapons. The evidence is clear that when that ban was in place, there were fewer deaths from such weapons. The main weapon that James Holmes used in his invokation of the Joker was an assault rifle banned under the previous ban on assault weapons, left to expire. This ban must be reinstated”

    Uh, yeah! Look I’m not against gun ownership though it does not reduce crime, that’s baloney. How is it that so many countries have few guns and little crime like Japan? Yet in Japan and Swistzerland and lots of other countries “only the criminals have guns”-that they have to import from out of the country-and yet they have lower crime rates than we’d know what to do with?

    By no means do I believe we should take people’s guns. However we should take their assault rifles tomorrow. This Joker guy was stocked to the gills. He had explosives all around his house.

    He had the most domestic casualties in US history with 70. Yeah, I know he would have still been “an evil guy” or certainly plain Loco without his heavy artillery. But no way would so many be dead and wounded.

  30. Gravatar of Mike Sax Mike Sax
    21. July 2012 at 07:28

    Oops the above post in in the wrong spot!

  31. Gravatar of Gene Callahan Gene Callahan
    21. July 2012 at 08:01

    What is an optimal forecast for a one-off event?

  32. Gravatar of Morgan Warstler Morgan Warstler
    21. July 2012 at 08:14

    Stats hits the EXACT nail on the head.

    And Scott repsonds in a way that leads to MY FRIGGIN PLAN.

    1. The market is HUGE. Big enough to take money from and put new money into the hands of traders – and have that be the PRIMARY benefit.

    2. LIMIT the traders to non-price setters, so Goldman and WalMArt are out.

    3. LIMIT the amount one US Citizen can deposit and trade.

    4. ONLY allow a bet on the under.

    5. Closest to the Under, without going under pays off the highest X return, reverse on losses.

    So if the NGDPLT run over, everyone loses some piece of the bet.

    If it runs under, print money and pay off.

    After a couple monthly rounds or iterative gameplay…

    a LARGE percentage of Stakeholders in the Economy – SMB EMPLOYERS are using the system as a hedge against economic downturn.

    You don’t mind losing a little bit when the economy is running hot, and you are doing well.

    You are rewarded when the economy cools off and new money is being injected.

    Think of it like real Futures for SMB owners.

    Plan WORKS.


  33. Gravatar of Morgan Warstler Morgan Warstler
    21. July 2012 at 08:17

    and oh yeah, Goldman gets bupkiss!

  34. Gravatar of Morgan Warstler Morgan Warstler
    21. July 2012 at 08:18

    And this is great news!

  35. Gravatar of ssumner ssumner
    21. July 2012 at 18:16

    Tomasz, I have no interest in a NGDP futures market that isn’t subsidized by the Fed.

    O. Nate, See my response to Statsguy.

    Gene, I suppose there’d be no way of knowing in that case.

  36. Gravatar of dwb dwb
    22. July 2012 at 07:20

    @Mike Sankowski
    But as an industry, $500bn is a reasonable estimate of the profits.

    No. you make some fine points about correlation but the $500 Bn analysis is still flawed in many ways. lets start with the fact that if the Fed were actually targeting NGDP, there would not be so much variation in it. The variation would be far less than 5%, closer to 1% in *unexpected* NGDP changes (annualized) because the fed would be targeting ngdp . TIPS or bond yields do not vary that much (even on the long end, 30 yr) over a quarter and in fact serve as a measure of Fed credibility insofar as inflation targeting. We can infer the volatility in unexpected ngdp from TIPS or PCE, which the Fed *is* targeting.

    Lets call it .75% ot 100 Bn over a quarter to be generous. Much more than that is made or lost in the commodities, stock and bond market every quarter, its distributed among thousands of players. Yes, *somebody* might try to manipulate it for their own personal gain. so what.

    Second, there is no way to consistently make money (because by definition, variations would be unexpected). So to make money requires a consistently good guess. Whats the sharpe ratio for a trade with zero expected return and volatility equivalent to PCE??

    This is where i get a little irritated, you seem to be building up a straw man in order to attack it on the grounds of market manipulation. Rather than attack it, please tell us how how would design it??

    /blockquote> there are political reasons to force the fed to act in dumb ways. Does anyone doubt the Koch brothers would toss a $20-$30 million at NGDP prediction markets to compel the fed to tighten rates today?

    Soros made billions betting against exchange rates in the 90s, Koch company is active in commodities. This is not an argument against ngdp futures or Schillers Trills. They can do all this manipulation now. The Fed has been accused of helping Romney, the ECB accused of forcing desired political outcomes and budget discipline on the Eurozone.

    The real question is whether its better with or without a vehicle like Sumners ngdp futures or Schiiler’s Trills. You have made a case these will be less than optimally designed, but have not made the case that the Fed’s toolkit is better off without either. Rather than building up a straw man to beat, please tell us how you would design it better.

  37. Gravatar of Mike Rulle Mike Rulle
    22. July 2012 at 09:44

    All markets are prediction markets. Most futures markets are priced at cash spot, plus or minus carry costs. Commodity markets are more complex, because of the difficulty of shorting real commodities, hence futures prices reflect more than cost of carry.

    VIX theoretically is a cost of carry market, except it is not really arbitrageable.

    NGDP is different all together although also a prediction market. It is more like a sports gambling market or as a few have stated, an intrade like product. I assume Scott means the Fed should be a market maker in NGDP, when he he says subsidize.

    I assume there would be many different contracts, not dissimilar to eurodollars, for example,(with forward markets and so called forward forward markets).

    I still maintain that it is unclear that a stated known target can in fact be hit. Actions in one environment (say the Fed tripling its balance sheet today) may not work in some other market. Will the Fed change its views if the market starts shorting the contract, or will it stick to its guns? It is difficult to imagine the Fed changing its views.

    How important is market feedback? Scott has always placed great emphasis on that. But we already have TIPs spreads predicting sub 2% inflation and the Fed does not appear to be responding the way it should if it cared about feedback.

    I don’t know. Maybe targeting NGDP is more clear for a variety of empirically observed reasons. I am open to that—-I think. But as I stare into “tax cliffs”, entitlement money recirculation and creation, regulatory mazes, anti-business rhetoric, entrenched government funded money losers, excess reserves larger than M1, it is difficult for me to believe that NGDP is our current problem. Fiscal insanity seems far more dominant.

  38. Gravatar of ssumner ssumner
    22. July 2012 at 10:02

    Mike, I’d suggest you stop talking about my proposal until you actually read it, as what you are talking about has no bearing at all on what I’m proposing. We are talking past each other.

  39. Gravatar of Mike Rulle Mike Rulle
    23. July 2012 at 14:29

    will do

  40. Gravatar of NGDP Targeting: Are Expectations Enough? | Last Men and OverMen NGDP Targeting: Are Expectations Enough? | Last Men and OverMen
    23. April 2017 at 07:18

    […]    “Michael still doesn’t seem willing to actually read the proposals; otherwise he would have known that the market would be subsidized, so there is zero risk of no one trading the contracts.  If no one else trades, I will, and I’ll gladly walk away pocketing the entire multi-million dollar subsidy.” . […]

  41. Gravatar of What is Bitcoin Hivemind? Complete Beginner’s Guide | BitcoInternet What is Bitcoin Hivemind? Complete Beginner’s Guide | BitcoInternet
    10. December 2018 at 06:05

    […] also cites an essay by Robin Hanson detailing how the public might be interested in paying for useful information. […]

  42. Gravatar of What is Bitcoin Hivemind? Complete Beginner's Guide – BTC Investor What is Bitcoin Hivemind? Complete Beginner's Guide - BTC Investor
    10. December 2018 at 22:01

    […] also cites an essay by Robin Hanson detailing how the public might be interested in paying for useful information. […]

Leave a Reply