Is neoclassical economics a Mafia?

Here is my short essay on that topic, riffing on Christopher Hayes’s recent article.  You’ll find other related essays and links here (David Warsh and Paul Krugman will be joining in at some point).  Here is my bottom line:

There’s much talk in Hayes’ article about discrimination against heterodox economists, but he gives surprisingly little attention to which of their valid propositions have been neglected. I’d like to see a simple list and start the debate there.

Addendum: James Galbraith is mad at me and responds.  And Max Sawicky has much to say about me.



You write "The door is open to anybody with a better approach, and it is circular to claim a gap exists because heterodox approaches have been kept down."

I think there is evidence that this was more true of the top journals thirty years ago or so than it is today; the evidence is clear that critical commentary has declined remarkably over this period. Of course it is true that EJW has stepped into the void, but this is a different door. In the long run the reputation of EJW may well come into parity with the top journals, but it is new and does not currently have the same impact on the profession. This comment is generic; that is, it does not apply only to "heterdox" papers, but to all papers challenging existing ones that have been published at the journals "top of the food chain."

As an outsider, what I observe is that economists tend to fall closer on the logic-facts cognitive dimension to mathematicians than to historians. Economists like logic a lot; facts, not so much. They like to believe that having a good theoretical model allows them to economize on knowing facts, which they aren't all that enthusiastic about in the first place.

This helps explain to outsiders why a fact-oriented economist like Thomas Sowell is treated condescendingly by the economics profession.


You would be correct except that the cost of manipulating data and
running empirical tests have fallen dramatically over the past 20
years or so.

But you ARE correct about some pure theorists who
puruse beautiful abstractions unrelated to ugly facts.

That is, there appear to be contrasting strands of literature: one that
pursues "pure" theory, and another that is leveraging new statistical
tests and databases made possible (and usable at low cost) by the
information revolution.

Much of the uneasiness with neoclassical econ seems to come from it's mathematical intensity. The heterodox economists tend to purely conceptual explanations, and view such explanations as better than ones that invoke a mathematical model. The question is, why should economics be purely (or mostly) conceptual? Mathematical models trump conceptual explanations in plenty of other sciences (physics being the obvious one, but fields like population genetics and computational neuroscience being others), why should economics be any different? The world is a complicated place, and our cognitive mechanisms may not be able to parse all the incoming data and patterns into neat little conceptual categories.

Ah, Tyler playing the contrarian Whig. Everything is fine, the Notre Dame department deserved to have their grad program shut down, they publish too many books and not enough AER articles. Move along, folks, nothing to pay attention to here.

I will grant that there has been a lot of loosening up in the micro area. Approaches that were not published much in very top journals because of the methods they used, e.g. experimental, are now OK to publish in those top journals, although QJE and AER seem a bit more open than JPE and Econometrica, despite the presence of Levitt and List at Chicago (and they seem to favor field experiments over lab ones anyway, at least somewhat).

The more problematic area seems to be in macro, where there has been the emergence of this neo-orthodoxy with the dynamic stochastic general equilibrium (DSGE) model. It embodies many of the worst sorts of ideas left over from the old, now getting discredited, neoclassical orthodoxy, representative agents with rational expectations operating in intertemporal general equilibrium. And, please Tyler, do not tell us that it is easy to publish a macro paper in the JPE if one does not do the appropriate hand kissing to this approach or its close relatives, even as its underlying assumptions are simply being shown to be baloney in general at the micro level.

I think this is why Hayes found the kvetching by the people at the Friedman do about George Akerlof's talk on macro. He showed what happens when one dispenses with these micro neoclassical fantasies that the experimentalists have been undermining for years, but which a lot of macro people are quite enamored of. I would say that there is still quite a bit of methodological repression going in parts of the profession.

So, the answer to notsneaky's question is that everyone has simply decided to ignore the Cambridge Capital theory controversy (pretend it never happened and publish lots of articles and textbooks using Cobb-Douglas or CES production functions with aggregate capital). But, my answer to your query, Tyler, is alternatives to DSGE. Are journals like the JPE really willing to even give the time of day to papers that pose alternatives to this approach?

Every time I questioned whether you can really add up captial like so many apples, or whether people really behave like optimising agents, I got told no of course you can't

IIRC, Hayek did develop a theory of disaggregated capital. The issue is extremely complex and not really amenable to statical analysis, so the claim that valid approaches are being neglected is unsupported.

I think the new reality is that disciplines are increasingly defined by their methods not their "topics". There are plenty of outlets for heterodox economics, but these are economic sociology, econ. geography etc.. Economics has become defined by its (neoclassical) methods, and just because you have a paper about an economic topic does not make it economics, so seems to be the implication.
As Econometron says, alternatives to neoclassical economics are situation-specific, again more like sociology or anthropology.

Thomas Sowell is treated poorly by economists? Please elaborate!

Luis, you wrote "if Card really did get ostracised for daring to suggest that the minimum wage might not actually destroy jobs (he says he did and I've no reason to doubt him) then those economists who behaved like that toward him should be very ashamed, and it's really troubling to hear that happened."

Quite the contrary, I believe he got the best-economist-under-40-award, not ostracised. It took a long time for the AER to publish a critique of his paper, some would say WAY, WAY too long; but then the AER was officially "hostile" toward critical commentary (the publication of comments, replies, and rejoinders) at the time that it published Card's co-authored paper. So albeit it after substantial delay, the AER allowed the science of that paper's analylsis to be challenged; this is anything but shameful. The publication of critical commentary is one of the checks and balances that long kept those at the top of the food chain from behaving like the mafia. Thankfully the EJW has stepped into the void left by the drastic downsizing of critical commentary that has taken place at the top general-interest journals. But again, Tyler, this is, given the newness of EJW, still a "different door."

Maybe Card thinks that his wage is too low? Just kidding; I actually have no idea why he would be upset about the open exchange of ideas that his co-authored paper aroused (if that is in fact what he is upset about).

If Robert Murphy is saying reswitching (more generally, recurrence of techniques) depends on there being "only two ways of producing a particular good", he just has gotten his arithmetic wrong. Anyways, it is wrong to say that economists know that, approximately, roundabout methods are more productive. Not that that approximation would justify the use of an aggregate production function in which the interest rate is equal to the marginal product of capital. (I don't know of any empirical work using Champernowne's chain index for measuring capital.)

I don't see what's heterodox about any of my links above, other than that I get my arithmetic correct. Nor do I understand the unwillingness, seen above, of mainstream economists to learn price theory.

There's plenty of neglected and ignored ideas in heterodox economics.

Louis Enrique,
Ridiculing a great person is not a good quality, you neo-classical (nonsense) admirer.Galbraith stands tall in economics.

Louis Enrique,
Ridiculing a great person is not a good quality, you neo-classical (nonsense) admirer.Galbraith stands tall in economics.

Cross-posted from TPM:

It is probably not polite to point it out, but given the anger displayed by James Galbraith and others at the refereeing process and at being shut out from top journals, I feel the need to offer the observation that the vast majority of the work in heterodox economics is of truly miserable quality.

What do I mean by that? I am a behavioral-ish economist and referee several times each year for the AER, JPE and QJE. Once or twice in a year I receive papers for review that are truly heterodox, and so far every single one of them was really bad -- logically flawed, non-identified econometrics (without realizing that to be the case), ignorance of the existing literature, ignorance of the available evidence, and so on.

My own work combines elements from neoclassical economics with elements from psychology and sociology. I am perfectly open to enriching economics with all kinds of new ideas, but I have a strong aversion against bad research. Since I don't want to be rude in my referee reports and hence don't like to write "you have no idea how to do empirical research, so please stop doing it", I often make a few substantive points and then end up saying something to the effect that "this paper goes against current trends in empirical economics" and recommend it for rejection. I am fully aware that this probably re-enforces the heterodox economist’s conviction that s/he is the victim of the bloody mainstream’s conspiracy to ignore the brilliance of his / her work, as so beautifully expressed by Galbraith at TPM.

Since James Galbraith likes to paint mainstream economics with a very broad and pretty offensive brush, let me apply the same brush to heterodox economics, where I believe it represents a closer approximation to the truth: Most heterodox economists are simply not good enough at what they are doing to break into the mainstream. Those that are good enough get absorbed into the ever-changing mainstream and end up re-defining what the mainstream is. The heterodox economists that remain outside the mainstream convince themselves that they are way smarter than all those fools working at Harvard / MIT / Chicago / Princeton / Yale / Stanford and without any doubt the victims of a conspiracy to silence them. Reality is that most of their work is either simply wrong or badly done or both.

1) Galbraith and Sawicky mostly use ad hominem arguments in their posts, which implies they have no substantial points or criticisms to make.

2) 99% of economists don't get to publish more than once, say, in the top journals. Does that make them victims of bias? If you take 1% of the truly heterodox economists, you're not left with many data points. The few heterodox papers that do get published, once in a blue moon, probably match the number you'd expect.

Rejecting a theory by saying that it is not founded in micro-optimisation is not good.Why should one adhere to optimisation always?


The problems with capital aggregation are so deep that it is not even ok as an approximation.

The problems with aggregating the representative consumer are even deeper... These are serious problems that cannot be waved away.

If you think it's just a matter of "discrediting the existing program", why don't you try building a disaggregated model and coming up with some useful science? At the very least, heterodox economists should make sense of these anomalies and explain how they matter.

"1) Galbraith and Sawicky mostly use ad hominem arguments in their posts, which implies they have no substantial points or criticisms to make."

your statement rests upon the unproven and idiotic assumption "ad hominem arguments imply insubstantial criticisms".

when you find a paper that you consider "fundamentally flawed" then it is all the more important to describe to the author(s) (and yourself) what you believe this flaw to be. Your discovery of such a fundamental flaw might be an indication of a fundamental assumption on your part that needs reexamination.

Heterodox papers are difficult for those editors steeped in the neoclassical tradition to read critically (in the sense of critique not criticize). The economic-theoretical world in which the editor works and thinks is usually significantly different from that in which the heterodox author works and thinks. Heterodox papers than serve the valuable purpose of providing "external" criticism to the mainstream method. Top journals should host such a forum for discussion, vitriol-laden attacks on the orthodoxy can only be improving if it is worth doing so. On the other hand, weakness and failings can be scrutinized and jettisoned. If as economists we truly believe in the prudence of competition, then we should have a marketplace for ideas that is not carelessly pre-screened for permission to enter. Let the letters to the editor and responses from other researchers do the job of tearing to shreds faulty arguments. Knowing rejection as the dominant strategy given the existing neoclassical game, the editor and the referees take a debate off the table before the community has had a chance to engage in it.

The Tsunami,

You wrote: "Top journals should host such a forum for discussion, vitriol-laden attacks on the orthodoxy can only be improving if it is worth doing so. On the other hand, weakness and failings can be scrutinized and jettisoned." I agree with you in part; I don't think that "vitriol-laden" challenges are superior to simply straightforward challenges. As I said elsewhere (above) this is a generic point that is not limited to the heterodox/orthodox debates that are the focus of this blog line.

Beyond our opinions, there is evidence that those in charge of the top journals once assigned a much greater value to discussion in the form of comments, replies and rejoinders than they do at present; for a discussion of the decline and debate about its causes see the following(hat tip to Barkley Rosser for persuading me to be less lazy about providing links to what I post):


I would say a bit too shy, but at least you are honest.
I am aware that at least one very heterodox economist,
also very famous, advocated closing the access to the
archives of a heterodox internet discussion list by
outsiders for pretty much the same reason.

Speaking of mafia-like behavior, when is the economics profession going to punish superstar economist and editor Andre Shleifer for his corruption in Russia in the 1990s?

I have endless problems with Mr. Sailer, but he is right to raise the Shleifer case in this context. It has been thoroughly ventilated on my own gigantic site, mostly by Barkley.

Galbraith mentioned high interest rates. Where are the interest rates high? Inquiring investors want to know.

He says "in the world economy" so presumably not here in US. Here in US it's about unemployment.

1) That's a non-answer.

2) I still want to know where in the world economy I can get a high interest rate on my investments. Oh wait, there aren't any.

On day 1 in our PhD macro class, the instructor essentially said, the Keynesian framework is plausible, sensible, empirically grounded but it does not enable us to answer questions. The DSGE/RBC framework is just the opposite. Somehow it works pretty well, even though it shouldn't. That seems to go along with the position of Commenterlein's former classmates.

sorry for the typos: I was too excited, I guess.


Huh? What article?

Mankiw's work fits into the DSGE mold. It is supposedly a synthesis.
New classical in its rational expectations/representative agent approach,
Nee Keynesian in allowing for sticky wages/prices, and neoclassical in
assuming general equilibrium.

Yes, there are lots of heterogeneous agents models out there, and I for
one see them as the big rival to the DSGE model. That is what I challenged
Tyler on when he asked his big question as to a rival that is being kept out.
And it is. You do not see such models in AER, Econometrica, QJE, or JPE,
except on very rare occasions (Brock and Hommes, Econometrica, 1997, "A
Rational Route to Randomness").

They are indeed all over the place in finance because people realize they
work. But this is Santa Fe complexity and econophysics, naughty naughty in
the clean JPE room of the DSGEers.


You wrote "Of course improving things means something was "wrong". You have to provide me with a better tool if you want to take away my hammer. This is how science works."

My question here is: A better tool for what?

When a tool being used to explain observational reality is replaced by a tool that is better in this regard, then yes "this is how science works [advances].

But what if a mathematically elegant model that is non-operational (non-falsifiable on the basis of observable reality) is generalized (say a calculus based model is reworked in terms of set theory)? Would you say that science has advanced if the generalized model, like the initial one, is non-operational?


Do any of these decade-old papers count as heterogeneous agent macro models published in top journals?

Aiyagari "Uninsured Idiosyncratic Risk and Aggregate Savings," QJE 1994, 500+ cites
Constantinides and Duffie, "Asset Pricing with Heterogeneous Consumers," JPE 1996, 300+ cites
Krusell and Smith "Income and Wealth Heterogeneity in the Macroeconomy," JPE 1998, 400+ cites

etc etc etc (cite counts from Google scholar).

These are not obscure articles, they're covered in first year graduate macro in many programs. I've taught Aiyagari's paper to advanced undergraduates. As should be obvious from the cite counts, there are hundreds of articles that build on these contributions.

I think it's fair to say Barkley that you don't know much about mainstream macroeconomics. Why do you pretend that you do?

Chris Edmond

Posted by: indiana jim at May 31, 2007 12:34:13 PM:

"But what if a mathematically elegant model that is non-operational (non-falsifiable on the basis of observable reality) is generalized (say a calculus based model is reworked in terms of set theory)? Would you say that science has advanced if the generalized model, like the initial one, is non-operational?"
First of all I am not sure what a non-falsifiable model is. I guess I don't think of models this way. I think of (mathematical) models as ways of laying out a theory in a clear, precise and "all assumptions on the table" way. The model can then be used to generate a falsifiable hypothesis. So I am not sure what you mean, and so I can't really respond. As two the second part of your question. I think that science for the sake of science is important too, and even if someone is just generalizing an existing model that there is value in that. Think of all the discoveries in the hard sciences that were not seen as impotent in their day but decades (centuries?) later these things were picked up by some grad student and were an important part of some new discovery.

Posted by: DarrenMc at May 31, 2007 4:33:03 PM:

"If you dislike the status quo, you won't get hired (or will be hired with extreme difficulty), and along the way will be forced to learn and do much that you disagree with before you are even in a position to be listened to. People who are capable of challenging the status quo and have the ambition to do so are pretty smart by definition; there are a lot of things they could be doing with their lives. Beating their heads against a dogmatic neoclassical wall is not an attractive option, and only the truly fanatical would pursue it."
I think you are wrong for the most part. As I said before I think that those who rail against the discipline will find it hard, but those who work within the framework of the discipline and bring in "heterodox" tools or theory into their work are often rewarded (or at least not "punished").

I am sorry you have been "forced to learn and do much that you disagree with before you are even in a position to be listened to." Thats just the way it goes in almost all aspects of life, even outside the academy. If you do not understand the current state of the art I say you are in no position to critique or even improve upon it.


you wrote:"I think that science for the sake of science is important too, and even if someone is just generalizing an existing model that there is value in that."

I agree that "science" for the sake of "science" is important (because I think of "science" as the pursuit of explanations having to do with observational reality), but what I was trying to explore was whether or not you think that unambiguously inapplicable models are worth making, and subsequently generalizing.

Comments for this post are closed