Earlier in the day I had a nice (2 hour!) meeting with a group (including Lee Smolin and Sabine Hossenfelder) at Perimeter who are thinking about complex systems, agent simulations and economics. They referred me to the following paper, which is excellent, and again expresses many thoughts I've had over the years in thinking about markets, financial economics, etc.
Geanakoplos is a "real" economist (James Tobin Professor at Yale) and Farmer is a Sante Fe guy who ran a hedge fund called the Prediction Company. If you are a physicist trying to understand the thinking of traditional economists, or an economist who wants to understand why physicists are often dubious about neoclassical economics, read this paper.
The virtues and vices of equilibrium and the future of financial economics
J. Doyne Farmer, John Geanakoplos
http://arxiv.org/abs/0803.2996
The use of equilibrium models in economics springs from the desire for parsimonious models of economic phenomena that take human reasoning into account. This approach has been the cornerstone of modern economic theory. We explain why this is so, extolling the virtues of equilibrium theory; then we present a critique and describe why this approach is inherently limited, and why economics needs to move in new directions if it is to continue to make progress. We stress that this shouldn't be a question of dogma, but should be resolved empirically. There are situations where equilibrium models provide useful predictions and there are situations where they can never provide useful predictions. There are also many situations where the jury is still out, i.e., where so far they fail to provide a good description of the world, but where proper extensions might change this. Our goal is to convince the skeptics that equilibrium models can be useful, but also to make traditional economists more aware of the limitations of equilibrium models. We sketch some alternative approaches and discuss why they should play an important role in future research in economics.
Farmer wrote a paper this summer, How markets slowly digest changes in supply and demand, which analyzes order book data. It's been fascinating to watch what's going on in light of various of the numbers in the paper.
ReplyDeleteIt sounds pretty interesting. Geanakoplos was the thesis advisor of a good friend of mine who now teaches Econ at Brown. I can only assume he's very mathematical; my friend's thesis looks more like a math paper (theorem, proof, corollary, etc).
ReplyDeleteWhich is why I'm surprised this paper has no equations. I wish it had at least *some*; talk about leaving room for ambiguity, amid so many verbose definitions of mathematical objects!
Steve, you should read this article by CMU Statistics Professor Cosma Shalizi, who ridicules IQ, its genetic heritability and malleability as much as you pump it up.
ReplyDeleteSee the following, if you care to reevaluate your position:
http://cscs.umich.edu/~crshalizi/weblog/494.html
http://cscs.umich.edu/~crshalizi/weblog/495.html
http://cscs.umich.edu/~crshalizi/weblog/520.html
http://cscs.umich.edu/~crshalizi/weblog/cat_iq.html
--
V
And I forgot to mention: Perhaps he is better at finding flaws in the way "those people" (ab)use statistics than you or I.
ReplyDelete--
V
Ironically, Geanakoplos (who is indeed a great mathematical economist) is a partner at a hedge fund, Ellington Capital, which was a pioneer in (and recently ran into troubles) due to... mortgage securities.
ReplyDeleteAndrew: thanks for the link, I've been meaning to read that paper.
ReplyDeleteArtur: the article is pretty clear, even without many equations, at least in my opinion.
TC: what's going on with Ellington? I know some of the people there.
Anonymous: Your comment is off topic for this post!
Briefly: I've read Cosma's stuff (some time ago), and don't think we disagree very much. My interest in IQ is mainly pragmatic. To find supporting data for the claims below just look under the label IQ on this blog.
1) Are there simple tests of cognitive ability that have predictive power? The answer is clearly yes -- i.e., SATs DO predict college performance, that's why Harvard, MIT, Caltech and every other elite university uses them. There is no serious debate about this question, only about the precise degree. Don't you find it remarkable that a short exam is as predictive of college success as high school grades, which are the summary of years of evaluation by many teachers? Because of the high correlation between SAT, 12 minute Wonderlic and Raven's matrices, I could even predict college performance with a *very* brief test -- isn't that amazing?
2) Is IQ stable? Yes, it's the most stable of psychometric quantities, more so in adult life and less so in childhood. (See comments under my post on Roe's study of eminent scientists, in which I discuss the Terman study results.)
3) How important is the specific construction of "g" via factor analysis? (Cosma is particularly critical on this point and I agree with him) Not very. I don't ascribe any special significance to the existence of a large primary factor -- it's clear to me there are multiple factors to intelligence. You'll note I don't spend a lot of time talking about the amazing general factor construct, g, as opposed to other factors.
Finally, and this is where we might disagree the most, although only on the specific numerical value:
4) How heritable is IQ? I suspect I might have more confidence in the h deduced from twin studies than he does, although see my posts on Turkheimer's work and on regression to the mean.
My impression is that most people who cite Cosma's writing don't understand what he is saying. At the end of one post he basically comes clean with a "if you put a gun to my head and made me guess" discussion in which he reveals that his beliefs are pretty much like every other reasonable, smart person who has looked carefully at psychometrics. The uncomprehending nurture over nature crowd love to cite him because he seems to be on their side (I doubt they really understand the details), but I think they overlook this part where he comes clean.
By the way, I am being particularly nice to answer you in such detail as: (1) you comment anonymously and (2) don't even take the effort to identify the specific issues on which you think Cosma and I might disagree.
I suggest you now pester Cosma about the questions I've enumerated above and see what his response is. If he were looking at two candidates for admission to his PhD program, who were identical on paper except one had +2SD in IQ over the other, which would he admit?
A more pragmatic reason as to why standardized tests scores are used for college/university admissions (independent of IQ or anything "intelligence" related), is that nobody wants to read carefully through tens of thousands of applications. Test scores (such as SAT scores) and grades are an easy way of weeding out and significantly reducing the size of the pile of applications to a more manageable size. Precise numerical metrics such as gpa, test scores, etc ... make this process relatively easy. (Without precise numerical metrics, this weeding out process is a lot harder).
ReplyDeleteHarvard, MIT, Caltech, etc ... prefer to deal with the folks who put serious thought into their applicantion, and weed out all the other jokers and slackers relatively quickly (ie. lower GPA's, lower SAT scores, etc ...). If the next Einstein-like genius ended up in the joker/slacker category due to crappy grades/scores, laziness, and/or a bad attitude, so be it.
Hi Steve,
ReplyDeleteI didn't know the bags of gold go back to Wheeler either! Just in case you don't know it, check fig 4 of this paper. More bags :-) Best,
B.
Steve,
ReplyDeleteThis is a great blog. Just what I needed!
You might also want to follow the blogs over at Wilmott on derivatives, particularly those of Das and Taleb. You may not care for Taleb (I find him incredibly funny), but Das (that's what he goes by) is a physicist by training, and began to raise serious public doubts about the CDS market and other derivative instruments years before the rest of his field. He's since begun writing for Roubini and for the Washington Independent as well.
www.wilmott.com/blogs.cfm
I found the Janeway interview to be stellar (and so did my dad) -- thanks for the pointer. What I make of Janeway's argument is that the underlying Bachelier (fundamentally Brownian) school of financial complexity is that randomness = trendlessness. Dynamic hedging is placing bets in all possible directions to neutralize the randomness except along one vector you believe is predictable. But because complexity = trendlessness is a myopic assumption, you've done yourself in. The tethering down of uncertainty that your bets have supposedly caused has instead overleveraged you to the point where those tethers are bound to draw-and-quarter you. Am I understanding the flaw correctly?
When you often find others have been thinking along the same lines, that just means you're not a very original thinker :)
ReplyDelete