December 2021

Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31  
Blog powered by Typepad

« Austrians in New Jersey | Main | Sachs on Easterly (again) »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Why do economists still insist on calling fundamental differences of this kind "methodology" -- a term paradigmatically tied to 1950-1990s neoclassical economics.

What you have are alternative explanatory strategies -- explanatory paradigms are prior to and dominate over any explication of a "methodology" derived from successful explanatory paradigms.

The "methodology" talk comes from a false / scientistic conception of 19th century physical science -- it comes from Mill's false vision of science and knowledge, a false vision directly tied to the false / scientistic 20th century vision of "scientific economics".

"Meta" descriptive progress in the science / philosophy of economics will catch up with explanatory advance in economics when it dumps this term which in part gets its meaning from a false 19th century understanding of science and economics.

Greg,

Um, http://en.wikipedia.org/wiki/Methodology. Methodology (or, more precisely, methods) is (are) how we do the stuff we do. It is the appropriate term here.

Prof Boettke,

I did a quick silly search of "methodology" in EconLit; most of the hits are totally unrelated, but this (below) looked promising. It's at least worth skimming?

Fine, Ben. 2009. The Economics of Identity and the Identity of Economics? Cambridge Journal of Economics, vol. 33, no. 2, March 2009, pp. 175-91

Drawing upon individual and social choice theory, the economic approach to identity associated with Akerlof and Kranton is critically assessed in terms of its own definitional and technical conundrums. This leads on to an external critique in terms of the limited extent to which identity can be addressed by mainstream economic methods. Indeed, the latter essentially sets aside the insights into the nature of identity that emerged with modernism, let alone postmodernism. As a result, the enterprise of constructing an economics of identity is set in the context of recent changes in, and disagreement over interpretation of, the identity of contemporary economics.


Pete,

The Stanford Encyclopedia entry isn't that good. There's a whole lot of stuff an realist vs. constructivist approaches, *modern* a priorist methodology, problems with data interpretation, the philosophy of mind issues that impinge on economic theory, and so that are either left out or largely unaddressed.

Just check the bibliography - no Hayek, no Mises, no Robbins, no Kirzner, no Menger ...

I used Daniel Hausman, “The Philosophy of Economics: An Anthology” as my first textbook in economic methodology. I also used extensively Roger Backhouse. Another author I like is Deborah Redman (“Economics and the Philosophy of Science”), she has a view of the methodology of economics in the context of the latest developments in nowadays science and she is also a historian of political economy. I’ve read Mark Blaug, of course, but personally I think he is biased towards modern positivism. Some other “methodologists” that I have read and come to my mind now are Thomas Hall, John Elliott and Frank Hahn.
About modern methods into economics, I can only say a few things about agent based modeling. A few authors that wrote on the methodology of agent based modeling into economics are Robert Axtell, Joshua Epstein and Matthew Hoffmann.
More new there is the cognitive agent based modeling, but I don’t know about any implications of it into economics.
I’ve also read David Colander a few years back and he cites, if I remember correctly, Brian Arthur as the first one to employ complex adaptive systems notions into economics, although I don’t think Brian Arthur is truly a “methodologist”.
I also think that prof. Richard Wagner uses a new methodological approach into political economy, but I don’t think anybody has written about his methodological approach. Yet.

I agree with Anamaria that more of Richard Wagner's methodology should be described and presented to a wider audience.

Um, Jared, read some Kuhn.

Jared, Carl Menger gave us an explanatory exemplar with his explanation of the origin of money.

Economic "methodologists" have given us a at least dozen different and conceptually non-compatible stipulations for how to do the "scientific method" of "methodological individualism".

The Menger explanatory exemplar is the explanatory strategy that produces powerful explanatory results -- all of the various and incompatible stipulated rules of economic "methodology" tell us how to do "methodological individualims" have mostly given us confusused and often utterly failed explanations.

Go with the explanatory exemplars, forget the "methodology".

E.g in "experimental economics", or "behavioral economics", or "neural economics" make sense of the paradigms of powerful explanation -- forget the gassing about stipulated "scientific method".

Jared and Greg,

Fritz Machlup taught that methods and methodology are different things. "Method" refers to a technique, e.g., constrained maximization or praxeological reasoning. "Methodology" is the study of methods. Pete asked about the "methods" we like, not the methodology we like, although I think he might have been playing a bit fast and loose with the difference between "methods" and "methodology."

Anyway, there is no taboo on studying methods, so "methodology" is just fine, thank you very much. An important methodological view (or is it meta-methodological?) decries *prescriptivism* and invokes Percy Bridgeman's remark that you should apply your mind to the problem, no holds barred. If we accept such methodological tolerance, it still makes sense to study methods! Economists should understand, for example, Duhem-Quine, which makes you wiser about the connetions between theory and history.

Greg,

What Roger said. Or from the wiki link (which is just good ol' Merriam-Webster):

"the systematic study of methods that are, can be, or have been applied within a discipline"

So it sounds like Prof Boettke is writing a piece on the 21st century study of the tools of the trade. There are lots of tools; we should have an article about 'em! Why are you so upset about a word meaning "the study of the tools those people use to do their stuff"?

Roger, you're missing my point. Let me try this. Here's something of what I'm driving at.

When Carl Menger gave us an explanatory exemplar with his explanation of the origin of money, that was helpful and represented progress. When Mill, Mises, and others dozens of others sought to fit this and other explanatory strategies within the demands of the philosophical tradition defining "demonstrative knowledge" or "science", then we got confused and false stipulated accounts of the "method" of economic science.

The lesson is the one taught by Thomas Kuhn -- what is learned from an explanatory exemplar is much richer than any "method" which meets the stipulated requirements for "knowledge" or "science" mostly derived from textbooks and the philosophic tradition, e.g. Aristotle, Kant, Hume, Mill, Carnap, Popper, Nagel, Hempel, and in economics all of the folks discussed in Caldwell's book on the history of 20th century economic "methodology".

So, I'm saying that most of the whole of the study of "methodology" is the history of a mistake -- a mistake picture of science constructed as a response to the demands of a false picture of knowledge given to us by the philosophical tradition (the Euclidean paradigm of demonstrative knowledge / "justified true belief").

And the same is true in economics -- and tragically looking what this mistake produced in economics, i.e. the 1950-1990s neoclassical "mainstream", the pathologies of which Caldwell among others has located in the first instance in a false view of science given to it by the stipulated methods that have more to do with "scientism" and the philosophical tradition than they have do to with real science. (Listen to Calwell's FEE lecture of last year or to his bloggingheads interview with Will Willkinson.)

I enjoy and advocate the study of "methods" and "methodology".

My larger point is Kuhn's -- explanatory exemplars are primary and primitive, "methods" are derivative and secondary, and often mislead and block the way of progress and understanding.

If we want to understand the new domains of economics, go looking to identify the explanatory exemplars in behavioral economics, in neuroeconomics, in finance, in "experimental economics". Any "methods" cribbed from those exemplars will be pale substitutes as conveyors on understanding and information.

Roger writes:

"Fritz Machlup taught that methods and methodology are different things. "Method" refers to a technique, e.g., constrained maximization or praxeological reasoning. "Methodology" is the study of methods. Pete asked about the "methods" we like, not the methodology we like, although I think he might have been playing a bit fast and loose with the difference between "methods" and "methodology.""

Greg,

I am completely on-board with Caldwell on the false view of science and the negative impact on progress in economics due to it.

But I am following the Machlup distinction precisely because the survey I am writing has to be non-partisian by construction, and I genuinely believe that most work-a-day economists are not that philosophically reflective on the discipline. Economics is what economists do Frank Knight once argued.

So while the philosophical justification of methodology has not really changed in economics, the methods employed are changing. But many of these methods implicitly challenge the underlying "folk-positivism" that governs most economists judgements about model and measure techniques.

Anyway, Caldwell's defense of "basic economics" will be represented, but so will North's institutional analysis, Peyton Young's use of evolutionary game theory, Axtel and Epstein's simulations, Smith's experimental economics and ecological rationality, Levitt's natural experiments, etc. etc.

Economics on the one hand is more open now than it has been at any time since I earned my PhD, but also less open to fundamental challenges to "folk positivism".

To Roger --- how would you write this entry if you were tasked? I think your SDAE address provides an excellent survey, what other areas do you see emerging since you wrote that piece.

Pete

"Economics is what economists do" -- right, look at the _explanatory exemplars_ to make sense of their "method", don't look at there explicit methodological stipulations. Kuhn, not Mill or Mises.

Peter writes:

"I am completely on-board with Caldwell on the false view of science and the negative impact on progress in economics due to it.

But I am following the Machlup distinction precisely because the survey I am writing has to be non-partisian by construction, and I genuinely believe that most work-a-day economists are not that philosophically reflective on the discipline. Economics is what economists do Frank Knight once argued.

So while the philosophical justification of methodology has not really changed in economics, the methods employed are changing. But many of these methods implicitly challenge the underlying "folk-positivism" that governs most economists judgements about model and measure techniques."

Greg,

I said I was against prescriptive methodology and you said you advocate the study of methods. So I don't see how you're correcting me. Your Kuhnian point seems about right to me. Indeed, prescriptive methodologists are in the position of the Randian villain from "The Fountainhead" who reduced innovations of modern architecture to stripping out ornaments and throwing in corner windows here and there.

Pete,

Thanks for the kind remark on the SDAE address. It was meant for a different purpose than yours so I would have different emphases. I would give more emphasis to experiments and neuroeconomics than I did in my SDAE address. Also I don't recall whether I mentioned analytical narratives. Anyway, I hardly need to remind you of these methods.

As you know, I would give some weight to the methods of complexity theory, including the methods of computable economics. That last one brings in the methods of constructive mathematics. On computability, I think a little subtlety is required. Velupillai has a paper coming out in New Mathematics and Natural Computation, “Taming the Incomputable, Reconstructing the Nonconstructive and Deciding the Undecidable in Mathematical Economics.” He blames nonconstructive math for certain problems he perceives in economic theory and asks us to study nonconstructive methods. Now I think we need to distinguish between 1) an economist using a nonconstructive prove to establish, for example, the existence of Walrasian equilibrium and 2) an agent within an economist’s model computing the uncomputable. We can agree to reject 2) even if we completely reject every version of a constructivist philosophy of math and thus do not agree to 1). I don’t want to have an opinion on constructive math per se, but I sure don’t think our models should have super-rational agents. Now Velupillai uses computability issues in a way very agreeable to me in his important 2007 paper, “The impossibility of an effective theory of policy in a complex economy” (in Salzano, M. & Colander D. (eds.), Complexity hints for economic policy. Berlin: Springer).

I personally, would give a fair bit of attention to these issues, which also bring in Mirowski’s stuff on Marketomata. I don’t really buy Mirowski’s story and I think he waffles in a crucial way: Are markets finite automata or are they the phenotypic results of genotypic automata? Complicating the picture is the possibility that markets may be able to reach noncomputable equilibria, as noted by da Costa and Doria (“Computing the future,” in Velupillai, K.V. (ed.), 2005, Computability, complexity and constructivity in economic analysis, Oxford: Blackwell Publishing, 15–50.). The “determination of equilibrium prices in a competitive market” is “formally equivalent” to “determining equilibrium in finite non-cooperative Nash games,” and will sometimes be, therefore, formally impossible. “So, the main argument in favor of a planned economy clearly breaks down.” (Thus, they defend the “Austrian” side of the socialist calculation debate.) Yet they report, “the equilibrium point of the market is eventually reached while we cannot in general compute it beforehand” (pp. 38–39). Because markets do not have to know what they are doing they can reach equilibria that cannot be computed ahead of time. The market’s achievement might well be described as a computation, but not the computation of a Turing machine. (See also their recent paper in JEBO.)

I think those issues are worth fleshing out in the context, probably, of Colander, Rosser, & Holt.

I might also give some attention to how different econometrics has become. We have much more use of computation-rich techniques and nonparametric statistics. That’s old news, but new since we were in school. It is, nevertheless, an important issue in a survey of economic methods, IMHO. A big cite in that connection is John Sutton’s Marshall’s Tendencies: What Can Economists Know?

That’s about all I got, Pete. Plus, it’s time to fire up the grill!

All this reminds me of how interested I used to be in methodology. Ultimately, I found it so unsatisfying. There were two reasons. First, 99 percent of economists don't care one bit about it and don't respect it. Second, there is a methodology "mafia" that seems to control the field and is extraordinarily intolerant of different approaches to the subject. I remember the trouble I had publishing my genetic-causal piece until Kylkos rescued Robin Cowan and me. I still remain "methodologically aware" and hope I will always be so. I think my current work displays this. Good luck to those of you who continue in the area.

Words, meanings, ideas are historical individuals (see the work of David Hull), and the historical individuals which are "method" and "methodology" in economics are bound up with a deeply false view of knowledge and science -- and a false picture of sound economic explanation.

A significant consequence of this false view are the professional pathologies Mario only begins to list.

I suggest a new word and a new field for the careful study of sound and successful explanation in economics.

Mario writes.

"All this reminds me of how interested I used to be in methodology. Ultimately, I found it so unsatisfying. There were two reasons. First, 99 percent of economists don't care one bit about it and don't respect it. Second, there is a methodology "mafia" that seems to control the field and is extraordinarily intolerant of different approaches to the subject."

Pete,

I have just been reading your "What went wrong with economics" paper, and I think it is the best paper of yours I have read. For me, it answers the methodology question pretty well, and succinctly.

As to this question:

"What modern methods do you find most promising and attractive for your own research and teaching interests?"

I do recommend agent based and other computational complexity approaches. For a discussion of my reasons, you can check out my forthcoming paper in RAE which connects market process and the computational approach.

Could have a precis, liberty? It doesn't seem to be up yet on the RAE site.

Hi Roger, it is on Pete's website, 1997
http://economics.gmu.edu/pboettke/pubs/articles/boettke-cr.pdf

Several people wrote comments and there is a rejoinder
http://economics.gmu.edu/pboettke/pubs/formalism_in_contemporary.pdf

It is a beautifully nuanced critique of two abuses of formalism, on the right, the claim that equilibrium theory is near enough to the truth and on the "left" the claim that departures from equilibrium call for intervention to correct market failure.

Mark Blaug should have cited this paper when he issued a critique of formalism.

Disturbing Currents in Modern Economics Challenge, Vol. 41, 1998.

The Problems with Formalism: Interview with Mark Blaugh; Challenge, Vol. 41, 1998.

Oh, the perils of ambiguous pronoun reference! I meant the "it" to refer to liberty's paper, Rafe, not Pete's. You and liberty are right: Pete's "What went wrong" paper is great. I'd like to know why liberty digs agent-based economics and computational complexity. Does she like them for the same reasons I do? Thus, I'm hoping she'll give us a precis.

Sure!

You can grab it from here:

http://economicliberty.net/Competition_MarketProgress_paper_Submit_rev.doc

Nice paper, liberty! I agree with your main points, namely, that GE is not really a model of competition and that agent-based modeling is simpatico with Austrian economics. Walras called himself a "scientific socialist," though without any intention of alluding to Marx. He was a cooperativist socialist for whom GE described the desirable results of policy, not the tendency of laissez faire. He derided "liberal" economists. So you are on solid historic grounds IMHO. I confess I'd allow a larger for GE than you seem to, but I agree GE is no substitute for process analysis. I also like the idea of taking agent-based modeling seriously. I think it's a great way to get at the self-organizing nature of market economies.

You did not argue that agent-based models are the be all and end all, but you did express optimism about what they might show us. That POV seems about right to me.

What examples can we give of new insights coming from such models? I can think of one right off, namely, Rob Axtel's model of firm size, which mimics the observed conformity to Zipf's law. What other examples might we point to?

Thanks Roger!

I think, sadly, that much of the work that is currently happening in the agent based community is ill conceived and misdirected. I have become somewhat disillusioned with the present state of agent based modeling, which has probably contributed to my stance on it now--that it is certainly not the be-all and end-all, but may prove useful.

The current work seems to be a mixture of (a) physics-minded work devoid of economic reasoning, (b) some interesting behavioral work, e.g. on social network effects and diffusion of information, (c) financial modeling of various sorts, and finally (d) interesting policy work and evolutionary work, of which there is relatively little.

However, this is only historical fact--it can change. Agent base models can and should be used to explore complex systems in economics such as the market process. One would expect it to change, given the excitement around complex systems--however the trend is the other way right now in economics. Agent based models are used in biology, and these models respect the complexity approach more, I think. So, what is needed is simply a handful of economists that have the economic way of thinking, and can relate it to complex systems and agent based modeling.

It does hold a lot of promise.

Also take a look at The Oxford Handbook of the Philosophy of Economics, ed. by Kincaid and Don Ross. The other work by Don Ross seems promising too.

Also take a look at The Oxford Handbook of the Philosophy of Economics, ed. by Kincaid and Don Ross. The other work by Don Ross seems promising too.

Roger wrote: "Now Velupillai uses computability issues in a way very agreeable to me in his important 2007 paper, “The impossibility of an effective theory of policy in a complex economy” (in Salzano, M. & Colander D. (eds.), Complexity hints for economic policy. Berlin: Springer)."

Certainly, very nice paper, but who wrote recently on this blog about the inadequacy of mathematics in economics?

The comments to this entry are closed.

Our Books