orgtheory.net

boundary control in economics and physics

The following fact was recently told to me: apparently, the leading physics journals no longer accept social network analysis, even though these articles are now among the most cited in these journals. I found this odd. From a “pure” perspective, it makes sense: why should a journal dedicated to stuff like quantum mechanics accept articles on the Internet just because a physicist wrote it? On the other hand, it seems like disciplinary suicide. Why ditch the folks who’ve done such great work? Obviously, physicists have made wonderful contributions to the study of social networks and that’s a highlight of contemporary physics, then so be it. Furthermore, I was informed that socio-physics is a tough sell on the physics job market. Apparently, most tenure track socio-physics people teach in social science/professional units & most physics programs don’t hire new PhDs in this area.

In contrast, the economics profession seems to strongly support boundary crossing, long as you stick to some variant of the neo-classical framework. It’s hard to find a topic in the social sciences where economists – often highly regarded ones – haven’t treaded. Also, I have rarely heard of an economist who was punished for publishing in non-social science areas like math, statistics, and the sciences. Diedre McCloskey, if she wanted, could probably get away with an article or two in literary journals. Sure, there’s a core to the economics profession, but there are lots of well supported outposts.

So we have an interesting puzzle: two highly regarded fields with an abundance of clever people. Yet, one has very tight internal control, while the other is quite pleased with imperialism. The flippant answer is that, well, economists are self-interested. By requiring people to use a neo-classical framework, you maintain the brand. By allowing imperialism, you increase the value of the brand. But why don’t physicists do the same? Why don’t they say (like economists do), “we’ll just come in and clean up?” As a field with some serious internal problems (small market/running out of easy problems), conquering other disciplines would seem like a good solution. Both economists and physicists are encouraged to comment.

Advertisements

Written by fabiorojas

October 14, 2009 at 4:05 am

Posted in academia, economics, fabio

23 Responses

Subscribe to comments with RSS.

  1. Economists engage in self interest with guile ( http://bobsutton.typepad.com/my_weblog/2009/04/do-economists-breed-greed-and-guile.html ). Perhaps physicists are just guileless…

    Like

    REW

    October 14, 2009 at 4:55 am

  2. That’s a really interesting question Fabio. I’ve been to a couple of econophysics conferences, and the physicists have outnumbered the economists by a fair margin both times. My memory is that the physicists were predominantly located in physics departments too. In these settings, I found the physicists to be at least as imperialistic as the economists (though not as insufferable about it). So based on those experiences, I find this turn of events surprising. Obviously it’s just anecdotal evidence rather than a systematic study, but still…

    Like

    Tim Kastelle

    October 14, 2009 at 5:53 am

  3. Hm. I’m not sure I’m convinced of the phenomenon. I was looking at disciplinary citation data the other data – the extent to which works in a field cite works from the same field – and economics was very much a stand-out for the proportion of citations within the field rather than to anything outside of it. Econ was something like 80% own-discipline citations, higher than all other social sciences and, I think, natural sciences (Sociology was more like 60%). Perhaps that supports your point in a way though – economists go out in search of new territory to dominate, and do so by bringing the main apparatus of the field with them, rather than by “going native”. Physicists are happier to be interdisciplinary and accept the legitimacy of other specialties (math, chemistry, etc.), but are more tight in their control over what it takes to count as “doing physics”. In other words, an openness to interdisciplinarity can go hand-in-hand with tight internal boundaries.

    Like

    Dan Hirschman

    October 14, 2009 at 10:53 am

  4. It’s an interesting puzzle. It really brings out how scientific disciplines can be treated as cultures, with the usual themes of acculturation, elite dominance, “pots not people”, etc. Physics and Economics are both “apex” cultures which influence others with little back influence.

    My take is that physics is more strongly associated with subject matter. From Newton on down, things like matter, energy, etc. have been at the core of the discipline, and the set of tools keeps expanding.

    Economics (at least modern economics) is far more defined by a methodological approach, based to a certain extent on “rational” behavior, but on agent maximization and empirical research more broadly. While physicists are more inclined to treat their toolbox as a useful set of tools, I think economists really internalize their models as useful descriptions of reality.

    Like

    Thorfinn

    October 14, 2009 at 3:05 pm

  5. I do not know the answer to your question. What I do know is that this is what people I hang out with would call a market opportunity for sociologists! Go get those physicists who are getting read in your department!!

    Like

    Michael F. Martin

    October 14, 2009 at 9:52 pm

  6. Thorfinn seems right to me; econ is more defined by method, physics more by subject. I have an MS in physics and a PhD in econ.

    Like

    Robin Hanson

    October 15, 2009 at 5:24 pm

  7. First off, I think Thornfin is basically right about the core of this problem, but one thing I might add is that physicists are more interested in empirically testable results. Untestable results, like the ones coming out of String Theory, tend to get branded as Metaphysics, which is just about the nastiest thing you can say to a physicist.

    The sort of predictions made in economics tend to be hard to test, at least, not at the levels of precision physicists are accustomed to.

    Finally, it’s worth noting that this is not a new phenomenon. Edgeworth’s and Jevons’ treatment of “Mathematical Psychics” was basically the methods of mechanics applied to economic phenomenon. While such subjects in the history of economic thought are falling out of favor in economics departments, they were certainly never of interest to physicists.

    Like

    Bradley Spahn

    October 15, 2009 at 5:55 pm

  8. One datum: I currently work as an economics RA; but my previous job was in a neuroscience (birdsong) laboratory. At least one prominent researcher in that field (Fee) has a physics PhD and if I recall there were others with similar backgrounds.

    Like

    Sam Henly

    October 15, 2009 at 6:08 pm

  9. Can you please tell us what you mean by “The leading journals in physics” ? There are awfully few journals that are that broad–phys rev comes to mind, and that’s it. The rest of the “leading journals” are domain specific.

    you said On the other hand, it seems like disciplinary suicide. Why ditch the folks who’ve done such great work?

    Huh? No one is suggesting physicists can’t publish in these fields in other journals. To suggest that the “discipline” of physics isn’t supporting interesting physics is without evidence. You’ve jumped several steps in your reasoning.

    Yet, one has very tight internal control, while the other is quite pleased with imperialism.

    The notion that physics has very tight internal control is silly. It’s without evidence again. You don’t seem to understand physics journals, or where the majority of preprints are.

    I have a better question: why don’t the economists have a preprint archive, like the physics community does?

    But to the question of why physicists might not think social networking etc. is interesting: one possible answer is that they’ve been doing this stuff for 40 years now and they’ve not gotten anywhere for it.

    Like

    Allison

    October 15, 2009 at 6:16 pm

  10. You write as if it is a similar leap from physics to social science as from economics to social science.

    To a physicist, economics is social science. Physics is not. This seems like comparing the interaction between political science and economics to that between chemistry and economics.

    I don’t think physicists are discouraged from studying other subjects. My PhD adviser spent a significant fraction of his time studying finance. But to be promoted in physics you need to do interesting physics research, and at the same time physics is competitive, so it’s hard to advance if you’re splitting your attention. It seems silly for a physics department to consider research outside of physics when deciding promotion — it does not have the expertise to evaluate the value of said research.

    Like

    Mike

    October 15, 2009 at 6:19 pm

  11. It seems silly for a physics department to consider research outside of physics when deciding promotion — it does not have the expertise to evaluate the value of said research.

    Mike’s comment illustrates what seems to me to be, on average, a difference between how new hires are evaluated by physics departments and other departments or schools. Others seem to have no problem passing judgment on a new hire without understanding the work.

    Like

    Michael F. Martin

    October 15, 2009 at 6:25 pm

  12. Allison: I was told that Phy Rev letters E has stopped accepting social networks papers. Was I told something that was not true?

    Also, I never said people can’t publish in other journals, but the norms seem much stronger in physics. And regarding tight internal control, I’d say physics has extremely tight control: it’s hard to become a physicist; they don’t accept as faculty people with PhDs in other field, except math; their journals publish overwhlemingly people with physics PhDs. Contrast with another field, like public policy. I’ve seen people in policy schools with all kinds of back grounds: social science people, people with physical science degrees, statisticians. It’s a very loose system in comparison with physics.

    Like

    fabiorojas

    October 15, 2009 at 6:34 pm

  13. Actually, economists have this problem too. If we publish in Science (which environmental economists often do) it counts as nothing to your tenure review.

    Like

    Amy Googenspa

    October 15, 2009 at 6:38 pm

  14. Allison: I just clicked through the websites of leading programs (MIT, Berkeley, Cal Tech) in physics and I don’t see many assistant professors who list social networks as a research interest.

    Like

    fabiorojas

    October 15, 2009 at 6:40 pm

  15. I have to disgree with Robin and Thorfinn here. I think the research in physics is defined much better by method than by subject matter. If you look at the top general physics journals (Physical Review Letters, Reviews of Modern Physics, Nature Physics, etc) you can find a very broad array of subject matter, including biology, chemistry, networks, applied math, econophysics, neuroscience, computation, and so on. In fact, I think you can find a paper in PRL addressing pretty much any scientific field that lends itself to mathematical or quantitative analysis. This kind of diverisity is not there in generalist journals of other hard sciences, like biology, chemistry, or math.

    The common thread among these papers is their methods, which tend to be physical in their emphasis on simple models, symmetries, general lack of rigor, etc.

    @Fabio: I think the main reason you don’t see social network research in physics journals or in physics departments is that most physicists (rightly or wrongly) view this particular type of research as dubious and not very physical in spirit. This isn’t evidence of any type of disciplinary narrowness in physics. In fact, my impression is that physics is growing more and more excepting of cross-disciplinary work. For example, the annual APS March meeting now has sections devoted to econophysics and (I think) other areas of social science research.

    Like

    Phil

    October 15, 2009 at 7:27 pm

  16. Socio-physics, much of which consists of social network studies, although some is traffic modeling, is much more limited than econophysics, although some network papers are showing up in econophysics as well, for example regarding links in financial systems.

    Physical Review Letters E has been an outlet for econophysics papers, which are written by physicists on economics issues, sometimes in conjunction with an economist or two (I have one forthcoming with a physicists in Reviews of Modern Physics). Other physics journals that publish econophysics papers have been European Physical Journal B, Phsica D, and especially Physica A, whose editor, Eugene Stanley, is the man who coined the term.

    There are also some economics journals that publish econophysics papers, including the one I edit, JEBO.

    Like

    Barkley Rosser

    October 15, 2009 at 7:35 pm

  17. The following fact was recently told to me: apparently, the leading physics journals no longer accept social network analysis, even though these articles are now among the most cited in these journals

    This is emphatically not true. The leading journal for physics is Physical Review Letters; everything below that in the prestige hierarchy is field-specific, but probably the most impressive one that’s relevant is Physical Review E>. I referee for both and can assure you that they are certainly still getting networks papers (and other “sociophysical” papers), which are being refereed and some of which are being published. In fact, turning to the table of contents of the latest issue of PRE and just looking for titles containing the word “network”, I count eleven papers which are relevant (plus a few containing “network” from other contexts); I didn’t bother to look for other relevant words or phrases.

    Since I don’t, thank God, have to worry about the physics job market any more, I don’t keep up with it, and for all I know that claim might be true, but I doubt it.

    Like

    Cosma Shalizi

    October 16, 2009 at 2:01 am

  18. Your above commenter dealt with the issue of PRL and PR-E. To your issue of tight control: I misunderstood your notion to mean tight control of what physics faculty DO for research, not how they get to be physics faculty in the first place.

    You contrast it to a field that is not a real liberal arts and sciences field. You pick an example that makes the comparison to physics apples and shoes. Public policy? What does it even mean? It’s not a science; its principles can be defined by what, exactly? They may need to have science IN public policy, and that may be why they “allow” different backgrounds in to the field, but physicists don’t need economists to do physics.

    So yes, tight control is needed so that the only people actually doing science in their discipline actually know how to do science. how this is disciplinary suicide is beyond me.

    to the specific issue of social networks: the “physics” or stat mech of simulated annealing, MCMC, condensed matter, or whatever other model you want to apply to demonstrate there is some analogous predictable behavior of a social network to a known physical process has been around a long time. The social network side of that stuff hasn’t gone anywhere, even though it has been studying for over a decade now. So maybe it’s a dead end for physics, and that’s why young profs are not doing it. Certainly young profs chase real money. Is there money in that these days in physics?

    Like

    Allison

    October 16, 2009 at 4:29 am

  19. […] 1. Boundary control in economics and physics. […]

    Like

  20. Well, given all the talk about physicists not needing economists to help them out, just for the record, certain ideas used in physics first appeared in economics, with the irony being that some of those ideas then filtered back into economics later from physics, at least to some extent.

    Thus, while Einstein got his Nobel Prize officially for his study of Brownian motion, it was a math Ph.D. thesis about the stock market in 1900 by Louis Bachelier that beat him to the punch (apparently Einstein was unaware of it though). In 1959, the physicist Osborne proposed in a paper in Operations Research that brownian motion was relevant to the study of stock markets, and then became the foudation for the efficient markets hypothesis and the random walk model.

    Also, the idea of power law distributions was first initiated in the 1890s by the socio-economist Vilfredo Pareto, although he was originally trained as an engineer. One of the big ideas of recent econophysicists has been using physics models to explain how power law distributions might arise in economics, especially in financial market returns (in violation of the brownian motion hypothesis), although it was the mathematician Benoit Mandelbrot who first argued for their relevance in asset return markets in 1961 (others were using them in such areas as city size distributions for a long time without any input from physicists or mathematicians, such as Zipf of Zipf’s Law).

    Like

    Barkley Rosser

    October 16, 2009 at 5:34 pm

  21. Since Barkeley Rosser brings them up, I’d like to point out an excellent review article by Mark Newman of Michigan on the mechanisms behind power laws. As Professor Newman does a good job of explaining, there are many mechanisms by which power law-dependence may arise, and the prescriptions for any social economic theory would be quite different for different mechanisms.

    http://arxiv.org/abs/cond-mat/0412004

    Like

    Michael F. Martin

    October 16, 2009 at 5:47 pm

  22. […] Controlling the boundary between physics and economics. Interesting problem of who researches what. […]

    Like


Comments are closed.

%d bloggers like this: