Saturday, May 23, 2009 30 Comments

Professor Hanson responds

Here. Further feedback is in the comments.


Blogger Daniel A. Nagy said...

DAMN! Once I fail to mirror UR at Thiblo and a discussion excellently suited for the medium breaks out elsewhere in a technically clumsy way, with Professor Hanson forced to insert MM quotes between his lines instead of just annotating the article...

I was on a business trip to (gasp!) Minsk, Belarus, and it was quite busy...

May 23, 2009 at 1:59 PM  
Blogger Malchus X said...

I have just finished reading Professor Hanson's response, and all the follow-up comments on his blog. That breezy sentence makes it sound like I know what any of this - including Mencius's original post that touched off the debate - is really all about; other than in the vaguest sense, I don't. Most of what I know about complex economic models or theories can be summed up thusly: "Life is funny; skies are sunny; bees make honey". Well, not entirely, but pretty close. The great thing about the cross-blog debate on this topic is that in the back and forth reply of the comments I actually learned something - not an everyday blogosphere occurrence. I kinda, sorta get what all the kerfluffle is about, even if not what the correct answers are (I defer to my intellectual betters in such subjects). Thanks to all the commentators on both blogs, whose replies helped clarify things for me a bit. And, of course, to our estimable host, Mr. Moldbug.

May 23, 2009 at 2:46 PM  
Blogger nazgulnarsil said...

i don't really see the fruit in arguing with professional economists who dont even understand how banking works. I'm currently trying to get the wikipedia article on fractional reserve banking changed.

getting to maturity transformation is hard when people refuse to even believe that banks dont lend demand deposits. hasn't anyone ever bothered to interview a bank accountant?

May 24, 2009 at 1:34 PM  
Blogger TGGP said...

Scott Sumner is probably a better partner for that sort of thing than Robin Hanson. Sumner even complimented Mencius' defense of the barbarous relic. I think one difference may be that Sumner has fully inculcated a sort of low-status identity free to be playful, while Hanson has achieved a higher status (albeit among a very select group of unusual folks).

Speaking of maturity transformation, I did a search for the phrase at and found that most of the results were nazgulnarsil talking about MM. Did the Austrians previously use some other phrase for it? Maybe whatever they called it has fallen by the wayside, like "apodictically certain synthetic a priori" talk. I don't read the Mises blog as much since I was banned there. Have you been tangling with George Selgin when he posts there, nazgulnarsil?

May 24, 2009 at 3:36 PM  
Anonymous JP said...

@TGGP - over there you said, " A Galton-esque "wisdom of crowds" aggregation consisting entirely of quite wrong (high standard deviation) sheep who are not even being rewarded for accuracy can result in a quite accurate prediction."

Is there a real-world example of this, or is it pure theory?

May 24, 2009 at 7:54 PM  
Anonymous Anonymous said...


I don't think maturity transformation is something that austrians have ever really talk about at all. Rather moldbug has said that it's just a generalization of the austrian criticism of fractional reserve banking.

May 25, 2009 at 4:25 AM  
Anonymous josh said...


"In 1906 Galton visited a livestock fair and stumbled upon an intriguing contest. An ox was on display, and the villagers were invited to guess the animal's weight after it was slaughtered and dressed. Nearly 800 gave it a go and, not surprisingly, not one hit the exact mark: 1,198 pounds. Astonishingly, however, the mean of those 800 guesses came close — very close indeed. It was 1,197 pounds.[10]."

May 25, 2009 at 7:54 AM  
Blogger georgeweinberg said...

Daniel, I think thiblo might get more use if there were a more obvious path from the thiblo homepage to articles.

BTW, there's also some discussion of the futarchy idea (including replies from Hanson) at Unenumerated.

May 25, 2009 at 9:55 AM  
Anonymous Eggs said...

This is hilarious:

Even the Orientals know what's up.

May 26, 2009 at 4:52 PM  
Blogger nazgulnarsil said...

I stopped posting anything substantial on mises when I realized how much of an echo chamber the forums are. they really don't like biology (cognitive psych) because it gives the socialists ammo (happiness research).

I was considering rewriting my unified banking thesis and posting it looking for criticism. Not that MM doesn't accurately portray FRB, just that a concise summary of the topic is useful.

May 26, 2009 at 6:03 PM  
Blogger TGGP said...

I didn't know you'd written a unified banking thesis. I'd gotten a similar impression of the echo chamber, where all knowledge comes from Rothbard. You'll find the dastardly neo-classicals are more interested in evo-psych or science more generally.

May 26, 2009 at 9:42 PM  
Anonymous Jim Simons said...

Mencius, did you see Hanson's latest post re the topic, titled "Reinventing Idea Futures"?

He of course doesn't respond in detail to any of your criticisms or make any further case for his silly ideas that have been thoroughly critiqued by you, Szabo, and others.

He just points to the fact that some other scientists are copying his idea futures idea. This is supposed to, you know, indicate his brilliance and deflect all your and any other person's criticisms. After all, other scientists, no doubt equally retarded as he is, agree with him!

The hubris!

Basically he just brags that these other professional scientists are lifting his idea.

Although I don't know why he doesn't consider the possibility of "convergence." Stupid people often end up doing the same stupid things.

May 27, 2009 at 8:41 AM  
Blogger nazgulnarsil said...

I've been poking at it for awhile. It's not original, just my take on Maturity Transformation Considered Harmful, A Straightforward Explanation of the Financial Crisis, and A quick Explanation of Fractional Reserve Banking in shortened, more digestible form. Hiding amongst these three lengthy diatribes are a few extremely devastating points that are not being addressed anywhere else as far as I can tell.

May 27, 2009 at 10:55 AM  
Blogger TGGP said...

Jim Simons, the post you reference was not a reply and there was no reason to discuss the actual merits of his idea. There is nothing hubristic in focusing on other aspects of a topic when they come up (as was the case with that article), although to some extent there is in saying "imitation is the sincerest form of flattery" about yourself.

May 27, 2009 at 11:11 AM  
Anonymous winterspeak said...

Scott Sumner may be new, but he got the ear of Paul Krugman pretty quickly. You don't really need anything more than that.

He's also, frankly, a joke. His schtick is that there is nothing that cannot be solved by monetary policy, and takes monetary policy to such extreme limits where it's fiscal policy.

For example, if Ben Bernanke was to actually get freshly minted greenbacks from the Mint (ahem) and throw them out of a helicopter in real life, Sumner would say this was "extremely unconventional monetary policy".

Whatever. At this point, it's clearly fiscal, where the Govt is paying people to be in the right place, at the right time. To call it monetary really is twisting the truth to destruction AND eliding over the fact that, at the limit, monetary and fiscal look pretty similar.

The good news is that it's easy to make Sumner jokes. For example:
"Why did the chicken cross the road?"
SS: Monetary policy!

"What's black and white and red all over?"
SS: Monetary policy!


The man has no idea about how any operational reality at the Treasury, Fed, or individual banks actually works. And he's a macroeconomist, which tells you all you need to know about that profession.

May 27, 2009 at 5:20 PM  
Blogger TGGP said...

And he's a macroeconomist, which tells you all you need to know about that profession Zing! Cowen & Tabarrok claimed they vowed not to stop working on their book until they could transform macro from the "ugly sister" to full-partner of micro. I'm not holding my breath.

It's tough writing monetary obsession jokes because Robert Solow already did it better: "Another difference between Milton and myself is that everything reminds Milton of the money supply. Well, everything reminds me of sex, but I keep it out of the paper."

he got the ear of Paul Krugman pretty quickly Maybe if by that you mean "barely acknowledged his existence". Krugman is somebody and Sumner isn't. That's why Krugman didn't bother to give a serious response to Scott and Sumner wastes time yakking with yahoos (such as yours truly). Whether or not he is a man worth killing, he'll stoop to grappling with aspiring dispatchers.

May 27, 2009 at 6:37 PM  
Anonymous Johnny Abacus said...


We should chat. I've been trying to throw together something as well.

May 28, 2009 at 5:09 AM  
Blogger winterspeak said...

TGGP: I hope that Cowen & Tabarrok stay focused on transforming macro from the "ugly sister" to full-partner of micro. It would spare us from their book.

Good point re: Solow. I didn't really see Sumner as a modern day Friedman because I think mostly of Friedman's micro work.

Krugman responded to Sumner in (web) print. That's more than I can say!

May 29, 2009 at 10:19 AM  
Anonymous Anonymous said...

What of the distinction between *prediction markets*, which in some cases when well set up and well tended can be useful, and *decision markets*, which I've seen Hanson promote, although I've never seen him explain understandably how they would actually make any sense?

May 29, 2009 at 8:01 PM  
Anonymous Anonymous said...

As near as I can gather, one of the sub rosa "beautiful" features of a typical proposed prediction market is that, because the market's "findings" are never tested against objective reality, the market conforms to and helps "realize" the dogma that "reality is socially constructed."

So it would seem that the most "enlightened", reality-denying university professors should be fond of it.

May 29, 2009 at 8:12 PM  
Anonymous Anonymous said...

Sorry, in preceding post I should have said *decisions* market, not predictions.

May 29, 2009 at 8:13 PM  
Blogger TGGP said...

because the market's "findings" are never tested against objective reality You can read Chris Masse's Midas Oracle site to find numerous examples of wrong prediction markets. Decision markets have payouts just like prediction markets and so they are tested against objective reality. It is the case that the counterfactual that didn't happen can't be evaluated. This is the case for both prediction markets (such as Intrade bets on a candidate winning the general election conditional on winning the primary) and decision markets.

I think the most reality-denying professors would be the post-modernists, who are more often found teaching English, law, philosophy or other such "soft" fields.

May 31, 2009 at 11:03 AM  
Blogger Jacob said...

I would like to distinguish between two different decision types - one can be aided by prediction markets, the other can't.
Prediction markets can predict who will be elected President, but they can't predict if he'll be a good president (which is what we are interested in, the important issue).
Decision questions are of two kinds: 1. "How many units of the new gizmo will be sold? Will people buy them?" 2. What will be the effect of some policy (say - the stimulus plan).
To answer question 1 all you need is a poll, asking people "would you buy it?"
Prediction markets are good at answering this (maybe better than pollsters) because they rely on a bigger sample. That's all, they have no inherent wisdom or wisdom aggregation - just a bigger poll sample.
The second kind of question cannot be answered, because there is no answer. The data and method of deducing the answer doesn't exist. You can't aggregate a wisdom that doesn't exist. You can aggregate guesses, but a guess isn't knowledge. A prediction (and decision) market is useless, if no prediction is possible.
No need to argue about a "trained" prediction market vs. an "initial" prediction market.
So, yes, futarchy is nonsense on stilts.

May 31, 2009 at 1:49 PM  
Blogger TGGP said...

"Good" is a normative term. As such it is inherently subjective and cannot be evaluated. What prediction markets can do is ask things like "What will GDP be at date D conditional on candidate C1 becoming President? What will GDP be at date D conditional on candidate C2 becoming President?" As I was perhaps inarticulately trying to explain, "reality" and hence prediction markets can only evaluate one of those two questions: if C2 wins all bets conditional on C1 are off. It doesn't establish causality either, though we tend to assume that bettors estimate different outcomes because the election has an effect rather than the election itself being caused by some third factor (whose value is unknown, and so cannot be taken into account, before the election) which in turn also causes a change in GDP.

I am not sure, but I think most prediction markets actually have SMALLER samples comparable polls. In that sense they are inferior.

because there is no answer Wrong. There is nothing theoretically preventing someone from doing a controlled experiment involving different policies (in fact such controlled experiments have been done on a number of occasions). There are just great practical difficulties. There is an answer, it just a difficult question.

May 31, 2009 at 3:26 PM  
Blogger Jacob said...

When I said a "good" president I meant the same as you: for example: if you define good=GDP up - prediction markets cannot predict that. They cannot predict any meaningful metric that would measure performance of a future president. They can (or have been seen to be able to) predict who will be elected.
They are a kind of poll, no more.
There is nothing theoretically preventing someone from doing a controlled experiment involving different policiesIf that were possible we wouldn't need prediction markets. We would do our calculations or experiments, or whatever method exists to solve our equations, and reach the desired prediction analytically.
Prediction markets are claimed to be useful in complicated policy questions were no analytical, engineering solution exists. Here my claim stands: if there is no solution - prediction markets are useless, if there is a solution - they are irrelevant.

May 31, 2009 at 10:08 PM  
Anonymous Mish said...

wait, didn't Mencius Moldbug argue in favor of prediction markets in his post titled 'statistics moons the bull' or something like that?

June 1, 2009 at 1:01 PM  
Blogger TGGP said...

prediction markets cannot predict that Sure they can estimate future GDP. What would make them unable to do so?

If that were possible we wouldn't need prediction markets. In practice, theory is different from practice so theoretical possibility doesn't mean practical. Also, randomly assigned controlled experiments are also not popular among policy-makers, who would like to believe they are applying the best policy to everyone.

Prediction markets are claimed to be useful in complicated policy questions were no analytical, engineering solution exists Comparing two different policies, either they will have different results or they won't. Either way there is an answer to the question.

Here my claim stands: if there is no solution - prediction markets are useless, if there is a solution - they are irrelevant. As I stated, there is always an answer. Prediction markets are believed to be useful because they aggregate the available information. Prediction markets have outperformed other sources of estimates, such as estimates, so they are not irrelevant.

June 1, 2009 at 9:07 PM  
Blogger Jacob said...

Sure they can estimate future GDP. What would make them unable to do so?Everybody can estimate GDP, me too. The problem is: making systematically accurate predictions (and not merely randomly accurate).
I can't say I've read all the literature, but I don't think prediction markets have given better estimates of (future) GDP than any random economist, or that there is any reason to believe they will.

I don't think you understood my idea of an engineering solution. In an engineering solution you don't estimate the outcome of a policy, you calculate it, you know in, with certainty, in advance. If that is the case you surely don't need prediction markets to tell you what the future result will be.
If no engineering solution exists - it means than the outcome (say - the GDP a year from now) cannot be calculated with certainty. In such a situation any prediction is a guess, and prediction market's guesses aren't any better than any other guess, so they're worthless.

June 2, 2009 at 3:25 PM  
Blogger TGGP said...

I can't say I've read all the literature Yet you seem quite confident in your conclusions.

better estimates of (future) GDP than any random economist If they gave the same result as polling economists, that would still be better than randomly accurate.

with certainty 0 and 1 are not probabilities. Certainty is a pipe-dream. Cue Steve Sailer on the nature of reality and why deductive philosophy has been so useless.

ny prediction is a guess, and prediction market's guesses aren't any better than any other guess You haven't established that. There are a great many ways of making poor guesses.

June 2, 2009 at 5:37 PM  
Blogger TGGP said...

Reality is not black and white. It is gray. The fallacy of gray is to say that all gray is the same.

June 2, 2009 at 5:48 PM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home