Monday, March 16, 2009

Physics and Economics

Jim Devine posted an article on Benoit Mandelbrot on my pen-l list condemning economists' appropriation of physics. I covered some of the humorous history of the economists' machinations at becoming physics like in The Invisible Handcuffs. You will especially appreciate the exchange between Macleod and James Clerk Maxwell.

Mandelbrot:

http://www.sciam.com/blog/60-second-science/post.cfm?id=benoit-mandelbrot-and-the-wildness-2009-03-13

The extract from my book:

http://michaelperelman.wordpress.com/2009/03/16/physics-and-economics/

11 comments:

Barkley Rosser said...

michael,

Hmmmm. A French political economist named Canard was influenced by physics models in discussing market equilibrium in 1801. The first to use just the term "economics" without the political were German proto-neoclassicals in the 1830s. Mirowski is the main source on the misuse of physics in standard neoclassical economics. Mandelbrot's remarks follow along these lines.

Of course the more recent odd development has been "econophysics," which has involved physicists, some at Santa Fe like Doyne Farmer, using more recent physics models, most often those from statistical mechanics, to model economic phenomena, most commonly trying to get at the fat tails that Mandelbrot worries about in financial markets, but also in such areas as income and wealth distribution. Most of this gets published in physics journals such as Physica A, which is edited by Gene Stanley, who coined the term "econophysics." I wrote the New Palgrave entry on this, and some other papers on it as well, most available on my website.

I will note that part of the issue here is different kinds of physics. I also note that the relationship between physics and economics is more complicated than it seems, with the random walk being developed in 1900 by Bachelier, a mathematician studying stock markets, prior to Einstein and Brownian motion, with a physicist named Osborne reintroducing it back into economics in 1959, while power laws, the favorite tool of the current econophysicists, was initially developed by Pareto.

Michael Perelman said...

Barkley, I don't know about the Germans. Who were they? Are they a canard?

Obviously, there was some spillover between physics and economics, but I was concerned with the conscious attempt to emulate physics.

rosserjb@jmu.edu said...

michael,

The best source on them is a paper by Erich Streissler, "Rau, Hermann and Roscher: contributions of German economists around the middle of the nineteenth century," European Journal of the History of Economic Thought, September 2001, 8(3), 311-331. I think it is Streissler who coined the term "proto-neoclassical" for this bunch, although Wilhelm Roscher is widely viewed as the founder of the German Historical School.

The most interesting and important of the three was Karl Heinrich Rau, born in 1792, whose famous book was first published in 1826, but went through many revisions up to 1848, and contained many standard ideas of neoclassical economics in a "primitive" form. The 1841 edition contains the first figure showing supply and demand curves with price on the vertical axis and quantity on the horizontal axis as it was done by Jenkins and Marshall in the English language tradition, in contrast with the French tradition of Cournot and Walras, who put them the other way. Streissler shows fairly convincingly that Marshall read Rau and was influenced by these guys, without ever citing them.

BTW, this puts a completely different light on Carl Menger. Rather than an inventor of marginal utility, which had been around since Rau and Gossens for decades in the German literature, he can be seen as the culmination of this proto-neoclassical school that then got into the Methodenstreit with the German Historical School in the 1870s. Much of this was political, with the earlier school being scattered about in the independent German city-states, with all of that being suppressed after the unification of Germany in 1870, and the Historical School then exercising a dominant and centralizing tendency based in Berlin, with the older view surviving in still independent Vienna as the Austrian School, a label originally applied by Schmoller as a derisive term to indicate the supposedly inferior Viennese who were not part of Germany.

Anonymous said...

i wonder about the physiocrats, and comte. weren't they after newton, and thought you could model 'social atoms' and reach some sort of utopian equilibrium? maybe i'll look at the extract to see if they're there. (there were also the technocrats, of the 30's, but while energy conservation may have played a role, not so other symmetries like the currently hip gauge symmetry (or the ever popular goldstone boson---also applied to the phenomena of hallucinations in the 70's)).

i've never been much of a fan of Mirowski's 'technical' critiques though his history is somewhat interesting (though i'm not sure if its accurate any more than 'shock doctrine' or 'confessions of a corporate hit man' (or whatever, by perkins). and 'mirkomauta' or whatever really lose me, even more than category theory. and i don't really have a problem with samuelson's 'ergodic' approach (as davidson---who seems pretty on point retrospectively---) as a technical tool (everything is ergodic in some context, is my view, and others too from what i gather.)

rosserjb@jmu.edu said...

media,

The physiocrats were not on the list. They were clearly a group that viewed the economy as based on physical reality and nature, with their essentially land theory of value. Quite a few modern ecological economists, such as Paul Christensen at Hofstra, have cited Quesnay's _Tableau Economique_ as a foundational work for ecological economics (one of the first papers in the journal Ecological Economics was by him making this argument). So, much of this is more from biology than physics, although clearly involving some inputs from it and chemistry as well.

Comte is the viewed by many as the father of both "positivism" and also sociology. Clearly there was a rationalistic tradition in France, dating from Descartes, if not all the way from Thomas Aquinas, and including Saint-Simon as well as Cournot and so on, that took a strong "social science" perspective of fitting economics and society into a scientific mold. Hayek has pinpointed this as "the fatal conceit," the idea of a rationalistically planned society and economy.

I tend to agree with Davidson against Samuelson on the ergodicity issue, and I am also a fan of "markomata," the main paper about which was published in the journal I edit, JEBO, although I would say that sometimes Mirowski is off on some details.

Michael Perelman said...

Regarding Barkley's first comment, I do not think that the economists you mention were attempting to emulate physics -- which was a goal of many economists since the time of Adam Smith.

rosserjb@jmu.edu said...

micheal,

You mean the German proto-neoclassicals? I would agree. Most of them had strong links to the sorts of traditions that would lead to the German Historical School, even if much of their work foreshadowed later neoclassicism. Heck, by today's standards, Menger was not much of a neoclassical, certainly less so than either Jevons or Walras, both of whom were definitely strongly influenced by either physics or engineering (which is mostly physics, if applied).

Anonymous said...

jbr--on my post, somehow i missed the qualifier---davidson seem's 'pretiscient' (if that's the word) or 'spot on'. steven keene also seemed to have the data, though alot of stuff (including your own papers) show how various bubbles and crashes arise. (keynes or h simon had it---imperfect information or sticky prices). i still stick with my view of ergodicity. 'ergodic on an attractor'. samuelson's case was all iid, like boltzmann. but now its beyond that.
as for markomata, sure; complexity theory. 'show me the money'. 0{n}. i'd rather find the effective hamiltonian; or messier (the real one). and actually, there are pleny of pasper showing that a dynamical system is an algorithm, so you can classify them complexity classes (eg pour el-richards). not that i'm jumping on the politics of notre dame. (i forget who was banned from there).

as for hayek, i actually read some of his debate with lange/lerner and a huge book on cultural evolution, which heavily borrowed from the biological lit(t)erature. already knew it. i did think it interesting that he shared his noble with gunnyr myrdal, who seems to be of a diffrent stripe (i read that yesterday on wikipedia).

Barkley Rosser said...

Well, while ergodicity is certainly attractive for doing all kinds of theory, it is far from clear that the economic world is ergodic; Davidson may be right.

The work of Steve Keen is most interesting, especially on Minsky dynamics. Ironically, he is coauthor with Mauro Gallegati, Thomas Lux, and Paul Ormerod of a piece that appeared in Physica A entitled "Worrying trends in econophysics" that triggered quite a firestorm in the econophysics bailiwicks.

The department Phil Mirowski is in at Notre Dame is the one that was suppressed for ideological and methodological reasons, having its grad program taken away and given to a new, openly named "neoclassical" department that so far seems pretty mediocre. Besides Mirowski, some other prominent members of that department are Amitava Dutt, who has written much on growth and development falling a more or less Post Keynesian approach, and David Ruccio, the post-modernist economist who edits the journal, Rethinking Marxism.

Many think that the Nobel committee was having fun when it gave one to Hayek and Myrdal at the same time, given their rather contrasting ideological positions. Hayek moved more towards an evolutionary approach later in his career.

Anonymous said...

yeah, i saw that 'worrying trends' on arxiv. worry about what? i noted ormerod was complaining he couldnt get his quite acceptable paper in a journaL. it aint look new.

in the long run we'se all ded; ergodic. i stick with that. (500 police today; they couldnt do sh-t)).

rosserjb@jmu.edu said...

Well, in the long run, "we" may all be dead, but there will (hopefully) be new people who are not and can behave non-ergodically. As for the really, really long run, well, the ultimate end of the Cosmos remains a matter of some debate.

Of the four coauthors of the "worrying trends" article, Ormerod is the one who has probably published the fewest articles in journals, either economics or physics, although he has published some fairly big selling books that have gotten a lot of press, making him probably the most widely known of the four. He also owns and runs a major consulting firm based in UK an Australia, Volterra.

Steve Keen is very much a heterodox character who has published econophysics papers and has a widely read book critiquing economics in general, the title of which is escaping me at the moment. It is quite well written and very witty, although I disagree with some of his arguments. He has published some very well done mathematical papers showing Minsky dynamics, in the Journal of Post Keynesian Economics, and also in some other multi-disciplinary nonlinear dynamics journals, a lively and strongly heterodox economist from Australia.

Mauro Gallegati is in Ancona, has published numerous papers in physics journals as well as in economics and is a cofounder of WEHIA/ESHIA, the Workshop on Economic Heterogeneous Interacting Agents with Alan Kirman. He has produced an interesting group of students, and several of them with him published an excellent book with Springer last year called Emergent Macroeconomics.

Thomas Lux is director of a major research at the University of Kiel in Germany. He may be the closest to being a conventional economist of this group, but is still pretty non-mainstream. He has published the most in physics journals of the four, having written many papers with physicists, a very capapble and cutting edge economists. He is the one who showed why Didier Sornette, a prominent econopysicist, was making errors in his model of forecasting crashes of speculative bubbles. Ironically, for all the anti-conventionality of much of the econophysics lit, Sornette was falling back on one of the sillier ideas of conventional economics, namely that of rational speculative bubbles, such foolishness!

I comment on this paper and the replies by others, such as Joe McCauley in my paper from last year in Nonlinear Dyhamics, Psychology, and Life Sciences, "Debating the Role of Econophysics," which is available on my website also.