March 2017

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

« MPS 2010 in Sydney Australia | Main | Search and Market Frictions »


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


"Economic analysis is about economic forces at work , not the analysis of situations after those forces have done their job."

Nicely put.

Excellent post.

Spot on, Pete: "Cleanonomics" is the invention of mathematicians dabbling in econ, and often getting it all wrong; _real_ economists have always been messyconomists; and the real case for markets has always been the case that they handle messy reality better than government regulation can. Nothing could be more absurd or idiotic, for instance, than the claim, so popular among persons who've obviously never read Smith (and who usually havan't read Debreu either) that Debreu polished-up and thereby improved upon Smith's rough-hewn ideas. Klein's "frictionless world" bugbear--a favorite among apologists for big government--just compounds this foolishness, by rejecting the notion of a clean economy only in order to replace it with the equally absurd one of "clean" government.

I enjoyed Steve's "Why Inflation Matters". It doesn't really say anything new, and I am not sure anyone would disagree, but it was still good. The opening explanation of "price inflation" vs. "monetary inflation" was good. It really needs a companion piece: "Why Deflation Matters". With appropriate substitutions of key words, the first quarter could be almost identical.

Pete is probably right too. I really have no idea. The notion of cleanonomics strikes me as rather absurd and irrelevant, but that's just because I don't move in the same circles. For you guys it really is relevant, because it's your lives, careers, and passion. Sometimes you make me glad that I am not a part of that world, though I realise that is probably not your intent!

We've got a problem, though, don't we.

Stiglitz and Diamond and others believe that their own "God's eye view" math construct adequately captures everything that matters in "messynomics" -- and they use their own construct to argue for "market failure" -- and "prove" what government policies are required to fix the "market failure".

And I don't see that there has been any effective way to get these guys to see that they are missing anything.

There needs to be a knock down / plain as day example or story or illustration or formal result to establish that they have utterly failed to do what they claim to do.

But where is it?


A grand post by Pete.

I learned my formal economics at UCLA very much in the Alchian tradition. Complete markets, perfect competition and full information were counterfactuals against which to analyze real markets. Real markets deal with the absence of all these assumptions.

I'd call all these straw men except some of our friends bought into the counterfactual. The New Chicago School wanted to be mainstream. Friedman knew he was a Marshallian, not Walrasian, but his successors didn't understand or care. I blame Stigler a lot for sowing confusion.

I am right now grappling with how to get Pete's points across on non-neutrality of money for the Cato monetary conference. I will look at Steve's paper. Insights welcome.


Two responses.

(1) You want a definitive proof, a "knock down." Such does not exist in economics, perhaps not in science at all.

(2) Ludwig Lachmann was the master sociologist of the economics profession. He had this wonderful phrase, "the neoclassical top brass." In his mind, they were the elite who ran the profession and determined what is permissable dialog. No one takes them on directly. His thesis remains unrefuted.

Can anyone think of a mathematical economics proposition that was accepted that was not consistent with what Boulding called the literary vagueness of classical economics and economic sociology?

If mathematical economics came up with a result that was not reproducible through economic intuition, did the result become popular or was it ignored?

Until this threshold is passed, mathematics will be a shorthand language rather than an engine of enquiry, as Marshall argued long ago.

According to Pigou, Marshall "saw that excessive reliance on this instrument [mathematics] might lead us astray in pursuit of intellectual toys, imaginary problems not conforming to the conditions of real life."

Beed and Kane has a nice paper on the history of the critique of mathematics in economics in Kyklos 2001, 44 (4).

One problem is that the MIT / top 10 folks simply dismiss "messynomics" as "narrative" economics -- and they don't see that as anything but confused economics that hasn't made its assumptions precise.

Compare Roberto Caballero's remarks on Hayek's Nobel Address in the introduction to his new paper on how to fix macroeconomics:

Caballero implicitly dismisses Hayek's economics as "narrative" economics.

You've got to come up with some simply illustrations or examples or arguments to make it clear that this isn't simply an issue of vague and sloppy "narrative" vs precise and mathematically sound "modeling".

@Jim Rose,

Good point.

Math is not a language. A language does not need to be transalted into another language (e.g., English) to be understood. You can't talk "math." But the failure to confront the mainstream on this issue exemplifies Lachmann's point.


"Narrative economics" is economics. Period.

When I started at UCLA in graduate school, the major professors held an assembly to introduce the incoming students to the UCLA approach. Leijonhufvud explained that we would be taught verbal economics, or "economics in other words." No concessions. What is now called "theory," mathematical economics, was a field that few took.

The rest is covered by Lachmann's sociology (see above). His analysis was seminal.

Great post, that I hope will be elaborated into a paper (or has that already been done?).

As a layperson, I want to agree that this is a fantastic post. Very nicely explains the seemingly complicated Austrian view of the market - or, as it should be called, the market process.

I'm gonna have to bookmark this. Thanks, Professor Boettke.

And this is why my economics is Austrian. When studying molecular and cell biology, I learned about self-organizing complex processes/systems. I saw this repeated in the emergence of organisms, in the emergence of ecosystems, in the organization of the brain -- and thus in social systems as well. The fact is that the world is complex and emergent -- and self-organized at each of those various levels of complexity. If this is true at the level of cellular organization, organisms, ecosystems, and brains, it makes little sense that, all of a sudden, it's not true at the level of human social organization. If it's true at that level as with every other level of reality, then it makes sense to find the people whose theories actually match that reality. That is the Austrian economists.

Could you imagine using the same arguments for ecosystems as are used for economies? Imagine creating a mathematical model of an ecosystem, then declaring that the real ecosystem was a failure because it didn't match the model. Yet, this is what passes for "science" in economics. I certainly doesn't pass for science in biology.

Peter Boettke:

Great post! I love the idea of the 2x2 matrix to describe schools of thought.

Jerry O'Driscoll;

I'm not so sure math is not a language and that it absolutely needs to be transcribed in another language to mean anything. I've had classmates that could solve intricate microeconomics math exercises, and in the NCE sense they were really doing economics not just math, but were incapable of transcribing the results or the process into prose. (Nevermind the fact that transcribing it into prose or explaining its economic significance was not rewarded in those classes...)

That is interesting - see, I've always thought of Mortensen and Pissarides (I was less familiar with Diamond before the prize) as complex/order types, and I guess I thought of Stiglitz and Keynes in that way as well. Search and matching models are nothing if not market process models, right?

The only way you can really call this gang "complex/disorder" is if you call the inability to achieve full employment "disorder". But that position - that we can operate at below full employment - doesn't deny the fundamental order of the economic activity that does go on.

I suppose your typology strikes me as a little self-serving in that sense... perhaps a good way of illustrating what you see as the advantages of the Austrian school, but has anyone who you wouldn't consider "complex/order" consider this to be an especially useful typology?

I don't know - if you're interested in that market process and markets dealing with complications and messiness, I think the search and matching literature is where I'd go. That is the pre-eminent market process theory on the table right now. I wouldn't shove that off into "complex/disorder" just because it presents a view of the economy that isn't the best of all possible worlds. It's whole purpose is to demonstrate how markets bring about order after all.

Keynes and other interventionists all see the economy as naturally disorganized and, thus, needing an organizer. The Austrians see the economy as naturally self-organizing. This is why Boettke's division is exactly right.

Macroeconomics will continue to be a series of disastrous experiments until we come to understand macroeconomic behavior as emergent features of a complex system. Understanding the economy as a complex system that evolves -- and can evolve into a new kind of complex system -- is what is necessary. One way of understanding recessions are seeing them as transition states from one stable economy to another (catastrophe theory models these kinds of recessions). As changes and errors accumulate in a stable economy, it becomes unstable, a shift occurs, and the economy leaps into a new stable order. In natural ecosystems (to which an economy can be fruitfully compared), mass extinctions are often needed to make space for new forms of life. Sure, some mass extinctions are caused by asteroids -- but that's a rare event. In the case of cambrian explosion. there was a sudden increase in number and kinds of species after a period of relative stability.

At the ecological level, we have species coming and going off and on, here and there all the time. But sometimes, every so often, there will be a conjunction of species going extinct, opening up new niches, new species emerging to help those extinctions along, and other new species, later, filling in the newly opened niches. New niches may even form, with new species emerging to fill them. A species that was just barely holding on may suddenly find the environment perfect for it. This is much like what happens in a typical business cycle. Some are as spectacular as the Cambrian explosion. Others are more subtle. Typically, when there is a mass extinction, though, there is some outside force which causes it -- one can think of the dinosaur extinction event caused by the asteroid. Such massive extinctions may find better explanations in such theories as the Austrian Business Cycle Theory (though obviously here the analogy fails, as the coming asteroid can hardly cause an overproduction of dinosaurs prior to the extinction event, as happens in the bubble preceding a recession).

I think Ormerod's "Butterfly Economics" sets us on the right path, but he still falls a little short, still missing the forest for the trees. His ideas need to be combined with Hayek's spontaneous order model of the economy, as well as Prigogine's dissipative structures, and Gould's punctuated equilibrium model of evolution. Indeed, the punctuated equilibrium model I think brings us much closer as to why there are business cycles in market economies. It also explains why we have so many scandals at the time of recessions -- as cheaters (those contributing to the errors accumulating in the system) are exposed. Cheaters have to be swept away for the next, healthy system to come into existence.

Some may object that things are in fact simpler than this. They could not be more wrong. All economies are nonlinear, complex, self-organizing, dissipative systems. Simple is not an option. Simple, linear thinking is what makes recessions in these cycles worse. The attempt to make such a system simple and linear -- through welfare statism or socialism -- stretches the recession out through time and at best flattens growth. What we are witnessing right now is the actions of people who see the economy as simple ,linear, and as a zero-sum game. They are wrong on all accounts. To not understand the need for systems theory in economics is like not understanding the need for it in biology. You can keep trying to apply Newtonian physics equations to cellular behavior, but you're not going to have the foggiest idea what's going on. Of course, it's actually worse in economics, since an economy is more complex than a living cell.

Accurate calculation, planning, and prediction are impossible, meaning good intentions cannot be matched with good outcomes in any sort of rational way. Chaos theory shows us that we have to have every single element exactly right to get the models to match reality. Bios theory shows us that creativity is even less predictable than is chaos. And Wolfram shows us that if you let a system run, you will find out sooner how it will turn out than if you try to model it mathematically. Isn't it time we applied these insights to macroeconomics?

Finally, in complex systems theory and network theory, we see that systems/networks/processes which are creative are far-from-equiibrium. Now, if economies are creative, and economies are complex systems/networks/processes, may it not also follow that they are far-from-equilibrum? May not the problem, then, be that our macroeconomic theories are and have been based on an erroeous theory -- equilibrum?

Prof. Boettke, this was a great article. I really enjoyed reading it. Thanks for sharing it with us.

RE: "Keynes and other interventionists all see the economy as naturally disorganized and, thus, needing an organizer. The Austrians see the economy as naturally self-organizing. This is why Boettke's division is exactly right."

Troy, I agree with you on the Austrians but that is simply not true about Keynesians. I think "order" is the wrong thing to identify a difference on. The Keynesian economy is an ordered economy - but it is sub-optimal. If the matrix were a simplicity/optimality matrix then I'd probably put Austrians and Keynesians in different cells. And what Keynesians do with that sub-optimality (to intervene or not to intervene) depends on the Keynesian. But is it an ordered vision? Of course it is. As I said above - what are Keynesian search and matching models (not all search and matching is Keynesian, of course) if not market process models!!! Mortensen and Pissarides IS a spontaneous order perspective.

I agree with you completely on the points of emergent, spontaneous order. I've been reading this literature for almost a decade now - most of my career as a student of economics, and long before I even heard of "the Austrian school". I think if you didn't take such a combative view of things you'd find a lot of people agree with you on that.

Jerry O'Driscoll -

You said: "I am right now grappling with how to get Pete's points across on non-neutrality of money for the Cato monetary conference. I will look at Steve's paper. Insights welcome."

Why not go straight to the source? I think Mises did a great job of explaining this in Human Action. It could use an update of language, but all the essential points are there. You could use markets for consumption goods (like movie tickets) compared to markets for essentials (like food) to illustrate how inflation impacts different markets differently.

Well, I should clarify - some of the earliest stuff I read about spontaneous order was Hayek - I just didn't realize at the time he was part of a wider school called "the Austrian school". But I was also reading Krugman, Mandelbrot, Arthur, Bowles, Gintis, etc. - not to mention more recent views of the market process I picked up in grad school (and was happy to see recognized by the Nobel committee) from Petrongolo, Pissarides, Mortensen, etc.. Troy, there are many roads to these insights, Austrians do not have to be in a war of all against all, and identifying sub-optimality or inefficiency is not the same as identifying disorder or failure.

The problem is precisely with thinking that there is such a thing as sub-optimaity and inefficiency. These things only exist if we compare what we have to a mathematically pure world with a utopian market economy. There is no such thing as a "market failure" except in comparison with such models. We shouldn't be surprised when reality falls far short of perfection -- but that doesn't mean there is a failure in reality.

Let me give an example. Imagine that one had a keynesian biologist. A Keynesian biologist looks at a real human body and is horrified at all the jury-rigging and inefficiencies in it. What to do? Well, first he sets about creating a model of an ideal body. With this in hand, he compares the real body and, low and behold, it's a disaster in comparison. Why, the lens of the eye is in backwards! So lets fix that -- ignoring the fact that the rest of the eye evolved in response to this, and "fixing" the eye now distorts the vision. Now the organism is worse off than it was before -- and all because it didn't fit with the orderly model the Keynesian developed.

This is not order in the sense Boettke is describing. This is a concern with disorder. We can see it in Krugman's take on self-organizing systems. He sees what happens, how the order comes about, and declares it's not to his liking and that something should be done about it.

The Austrian, on the other hand, looks at complex systems and recognizes that the patterns in the system are in fact orderly, and that they evolved to have the patterns they did for a reason. One does not set about changing those patterns lightly (this is Hayek's lesson on tradition and how change occurs in spontaneous orders). In fact, some things can emerge which one may not even understand, but which when you change or remove them, result in a collapse of the system as a whole. The Keynesian thinks he can know what those are; the Austrian refuses to have such a pretense of knowledge. Wisdom is knowing what you do not and cannot know.

Let me ask this: if Keynesians didn't think that the economy needed their organizing hand, why then do they recommend interventionist policies?

@Jerry: I recommend the more "academic" version of my stuff on the costs of inflation:

Good post, Pete. I wish economists would remember that the cleanonomics of Leon Walras was meant as a description of the ends of policy, not the real messy economy. He called himself a "scientific socialist," though not in a Marxian sense. Thus, GET describes the goal of the interventionist (and, in Walras's case, cooperativist) state. In other words, the originator of cleanomics was not even trying to describe the market process. He was criticizing it. He explicitly opposed French liberal economists. Only later was the myth invented that Walras was somehow describing Adam Smith's invisible hand with "rigorous" mathematics.

Is there a book where this has been explored at length?

I've often seen comparisons between monopolistic competition theory, perfect competition theory and the Austrian point of view. Hayek discusses that in a few papers. And there have been many discussions about the differences and similarities between Stiglitz's position and the Austrian one. But, has anyone tied all this work together into a book on the topic.

What we need is a theory of where the "God's eye view" should live in economic thinking. (I'm not persuaded that the answer is "nowhere".)

I wouldn't want to endorse an economics that calls itself a narrative. "Verbal," okay, but "narrative" has a connotation of not just lacking math but lacking a method of adapting or responding to any phenomena observed in the real world. Certainly that is inconsistent with Hayekian pattern prediction.

"Narrative" is very applicable to something like Marxism, in contrast. I would hope to believe that Austrian economics stands on slightly higher epistemic grounds than Marxism.

A purely narrative versus scientific distinction, where all that is verbal is narrative, is just boilerplate positivist BS.


inventions are often used in ways unintended by the inventor. Thus, to say that GET "describes the goal of the interventionist state" just because Walras "called himself a scientific socialist" is somewhat simplistic. Walker's "Walras's Market Models" shows that (the early) Walras was indeed interested in market processes. Further, in neo-Walrasian models general equilibrium is reached in the absence of any institution (bad enough), including the government intervening in the economy.

On Daniel's point about Keynes and order, there is much to be said on either side of the question. Some people interpret Keynes as having pointed to an unemployment equilibrium. Others emphasize that he was really talking about a disequilibrium position that continually left workers and others frustrated in efforts to implement their plans ("involuntary" unemployment). There are aspects of equilibrium and disequilibrium in the Keynes story. However,in either conceptualization the outcome is not good.

Machlup often reminded us that equilibrium is not necessary good and disequilibrium is not necessarily bad.

Therefore, I find the cell "order" to be in need of further explanation.

I want to propose a simple/complex and irrational/rational matrix.

Austrians (particularly those in the Hayekian tradition) tend to see rationality as a property of systems, conventions, rules of conduct or method. A system is "rational" because it adapts to feedback and criticism. Others tend to see rationality as a property of particular beliefs, goals, or persons; the system itself is inherently irrational, because it serve no explicit purpose or premeditated direction.

Roger -- Wieser was doing something of the same thing as Walras. Wieser called it a dictator model -- for a reason.

@ Lee Kelly,

good point.

Thanks Mario - and yes, I do think people think of the tradition in different ways. But what is really the fundamental problem - "disorder" and microeconomic inefficiency, or "sub-optimality" and macroeconomic inefficiency. I think it's quite clear that the latter is the quintessential Keynesian problem. Even where newer Keynesian do introduce microeconomic inefficiencies, they've elaborated on market processes (like search models) that identify how those inefficiencies are overcome. That is not a disordered understanding of the economy.

So who would I put in the complex/disordered box? I would have thought that is where the behavioral economists go, not the Keynesians.


there is a difference between dictator models, which are called social-planner models today, and Walrasian models where prices decentralize allocations. The social planner model is often used to define welfare criteria: if there would be a omniscient super-human, what would allocation look like (if only Pareto-improvements are allowed for)? And do prices decentralize alloactions that would be chosen by an omniscient agent? Another approach is the analysis of the core. Here you don't need the dictator. Walras, however, remains.

FWIW, I defend "literary economics" in a forthcoming JEBO paper now available on "Some epistemological implications of economic complexity." To go from your "formal" theory to application, or even to known what your mathematical model is saying you need to interpret human meanings. Thus, even if we restrict "theory" to math models, we can't really escape "literary economics." You can't have rigor in mathematical economics without rigor in literary economics.


Must run, but Walker's work shows *me* that Walras was forced to see that his basic model could not handle trading at false prices, which tells us that it's no good for market *process* theory. Well, of limited utility. I have an ancient JITE article defending a role for GET. But it's not telling us about processes.

Arash -- I don't buy it. The knowledge problem says you are just begging the question.

Have you read Wieser?

"there is a difference between dictator models, which are called social-planner models today, and Walrasian models where prices decentralize allocations. The social planner model is often used to define welfare criteria: if there would be a omniscient super-human, what would allocation look like (if only Pareto-improvements are allowed for)? And do prices decentralize alloactions that would be chosen by an omniscient agent? Another approach is the analysis of the core. Here you don't need the dictator. Walras, however, remains."

Peter, I suggest you have a conversation with Nick Rowe:

Rowe wants to discuss good and bad ways to put "money" into economics.

You also say you want to discuss good and bad ways to put "money" into economics.

Rowe thinks you can do this mathematically.

My suggesting is that the issues at stake here can be drawn out -- brought to light - best through conversation. People can gain an understanding of what matters here and what doesn't by finding out what unstated premises people are working with, and what the strength and weaknesses of those promises are. And a conversation builds a common store of "reminders" which people can go back to re-use to establish bench-marks of understanding.

I may be wrong, but I don't see that folks have really engaged the math guys directly on any of this.

You have guys out there ready to have a conversation -- one with an audience.

So what is stopping this conversation from happening?

Nick Rowe is great - have him to GMU. I'd show up to that.

I don't know if I'd agree with you on the math, though, Greg. Roger is right above that you can't have math without "literary economics" - but if the concern is laying bare our premises it seems to me math makes that easier and relying too exclusively on conversation gives too many opportunities to obscure premises and assumptions.


I did. Social Economics and Natural Value. Wieser is about imputation. I was thinking of Social Economics, esp. the definition of the simple economy:

"The single economy, as described in our introduction, is the economy of the single subject. [...] But we do not have in mind the scant economy of an isolated householder. Rather we envisage an economy that has the breadth of a national economy with all its wealth, technical knowledge and problems of economic calculus. But this broad economy is guided by a single mind. [...] This director foresees ends, weighs them without error or passion and maintains a discipline which ensures that all directions are executed with the utmost precision and skill and without loss of energy. We shall further assume that all requisite individual forces are placed at the disposal of this social management as cheerfully as though enlisted in their individual interest." (p. 19-20)

This is Wieser's "benevolent dictator model" (in contrast to his complex economy model). Excactly this "director" is what is called social planner today. Please note that imputation is not a market phenomena.

sorry, the quote begins with: "The simple economy ... ."

Thanks to Roger for demythologizing Walras.

Walras is a theory of centralized resource allocation handled by the auctioneer. As Roger said, no trade at false prices. Trading at false prices is what decentralized markets do, and then "correct" them.

That is the messiness in markets. There is no messiness in Walras.

Daniel -- the concern is to lay bare what is missing from any formalism or formalist understanding of formal systems like math or logic, i.e. see the work of Wittgenstein and Hayek on this topic.

Godel is also relevant.

Daniel wrote,

" if the concern is laying bare our premises it seems to me math makes that easier and relying too exclusively on conversation gives too many opportunities to obscure premises and assumptions."

The "premise" is that math can't "lay bare" the premise of open-ended learning or genuine uncertainty, among many other things.

Try "laying bare" those premises in a bit of mathematics ...

Daniel wrote,

"if the concern is laying bare our premises it seems to me math makes that easier and relying too exclusively on conversation gives too many opportunities to obscure premises and assumptions."

I think that Pete is thinking more about Post Keynesians (and post-Keynesians) than New Keynesians. The New Keynesian view is about sub-optimality, the Post Keynesian view is about disorder too.

It's also worth mentioning that the income-expenditure, AD-AS and IS-LM Keynesian views are quite simple.

Nice post. I am wondering where dynamic general equilibrium models might fall on this matrix. Although they are neo-classical, DSGE models can capture more complex indirect effects that static models can't. Also, from a market process perspective, DSGE models would not need to attain a steady state, but would continually adjust to new parameters in each time period. Maybe DSGE models might overlap the simple/complex border. If done well, DSGE models could help to tell the market process story. About Walrus, it's interesting that he was trained as an engineer.


yours is such a biased reading of Walrasian economics. Walrasian theory is fundamentally a theory of decentralized resource allocation! The artificial groping process, which no GE-theorists considers to be a real-world phenomenon, is of course problematic. I refer to Mirowski's Markets Comes to Bits paper:

All his critique is right to the point.

Nevertheless, the groping process leads to a price vector that decentralizes allocation! It's all about price signals that allow locally informed agents to behave in harmony with information outside their reach. Walrasian economics at least tells us that if a equilibrium price vector could ever be found, it would decentralize allocation. That's not the Holy Grail, because it tells us little about reasonable market processes (with trading at false prices), but still something.

@ Andrew

I would put NK-DSGE into the simple/order box. It's about equilibrium dynamics, so no disorder. It about representative agents, so no complexity. Further, NE-models reach stochastic steady-states, with stationary (and well-behaved) probability distributions. On the extent that DSGE can improve our understanding of market processes, check one of Alan Kirman's excellent papers. A good and easy intro is:


I'll ignore the ad hominem.

"The artificial groping ... is of course problematic." That's the issue and not the math. Since prices are not actually set that way, it is not a decentralized process.

Your "nevertheless" comment is a breathtaking contradiction of what you had just written. There is no groping process.

You want a theory of equilibrium price determination -- the price vector, which, if found, would bring about equilibrium. The folks here want a theory of how market (not equilibrium) prices are set (not mathematically determined).

Read some Alchian and we can all have an intelligent and civil discourse.

The "narrative" & history vs formalism / deduction debate over causal explanation in the 1960s was arguably the most important even in the philosophy of science in the last half century.

The results of that debate are still playing out in the profession.

The debate centers on whether narrative explanations can provide "scientific" causal explanations.

The two philosophers of science I studied on are on opposite sides of the debate -- Larry Wright and Alex Rosenberg. Their work engages with one another in the domain of biological explanation -- and a core topic in the philosophy of Darwinian biology has been the nature of explanation in that science, and workability of narrative or nomological / deductive models in the science, or what alternative model might be more appropriate.

Wright studied with two of the key figures in the original 1960s debate -- Hanson and Salmon.

Wesley Salmon gives a summary account in his _Four Decades of Scientific Explanation_ for the perspective of the formalist / deduction side of the debate.

In the last 15 years there have been some very good work arguing that all scientific explanation has something of a narrative structure.

But these issues aren't easy - or easy to bring others see things from the opposite view.

And note well -- there are deep institutional and incentive biases in the profession behind the "formalist" or deductive / inductive logic ideal. It's a filter device throughout the profession, from class work to publications to dissertation, etc. -- formalism is the easy and objectively clear way to identify "understanding" and intelligence, and there are massive incentive and self selection mechanism to fill the profession with folks with a formalist / deductive / inductive logic bias.

Similar to the mechanism in economics.


As far as I can see the only way to represent decentralised knowledge in GE is through introducing new goods.

Let's suppose that there are three people who know where to buy candles cheaply in Cork city centre, and one person who supplies them. If we have a good in GE "candles for sale in Cork city centre" then it would have one price. So, to represent different knowledge we would have to introduce another good "cheaper candles for sale in cork city centre known to Tim, Frank and Sheila, but not anyone else".

How would you deal with this?


I just said that your view on Walrasian theory is biased. That's all. Certainly, it does not qualify as ad hominem: it's just a statement about your perspective on Walras.

The point is that you use the term "decentralize" differently than GE-theorists. Their use is that the price system provides information such that locally informed agents adjust in harmony with each other's data. You use the term to what is going on in reality. That's fine. Given your use of the term, Walrasian theory is not on decentralization.

"Read some" Leonid Hurwicz "and we can" see that my use of the term is the more common and, therefore, my comments a logical consequence.

Finally, my "breathtaking" comment makes a lot of sense if you are willing to accept the language and purpose of Walrasian analysis. Please check Kirman's foreword to Colander's "Post-Walrasian Economics". Search for the footnote, where he makes exactly the same point against Colander's interpretation of Walrasian economics. Check Kirman and Hildenbrand textbook on GE, or any other textbook, like Starr or Arrow and Hahn.

Jerry - you do this a lot, to me as well. Not everyone who offers a counter-argument is making an attack or trolling or introducing an ad hominem. I found Arash's comments (all of them) very interesting and helpful. If you disagree, then disagree. Don't get upset over it unless he actually starts calling you names.

I should add -- the other key person at Indiana writing on the explanation, narrative and the nomological / deductive / formalist idea Michael Scriven.

The key problem is:

How do we provide scientific explanations?

And the affiliate problem is, How do we provide scientific causal explanations?

The formalist tradition has it that you can't do this with narrative, the model of causation and explanation is formalist -- explanations require laws or statistical regularities, and these need to be captured in deductive or inductive formal structures, usually involving math formulas.

The problem is to show why the formalist project is unsound -- and why the non-formalist project is both sound and helpful.

And the issues involved are merely replayed in economics.

The formalist project has taken massive hits -- and people like Kuhn and Wright have given prestige and explanatory power to the non-formalist project. The the sale has yet to be made to the die-hard formalist -- who publish or perish based on working within ever new variations on the formalist paradigm.


very good point. Indeed, the law of one price is a presumption in GE-analysis, that is, also in disequilibrium it is assumed that homogeneous stocks of commodities have the same price. These aspects of simple arbitrage are ruled out to see the more complicated (not complex) problem: does the price system guide the allocation of many different goods? Here, it does not suffice to think in terms of profit equalization (the classical or pre-Walrasian approach to competition). It comes as no surprise that the economy is defined in commodity space (with all nasty restrictions necessary). The introduction of even more commodities, namely state-contingent ones, allowed to apply the static framework to intertemporal aspects. BTW, Hayek introduced 'dated commodities' to economics. He used such frameworks even though he knew of their limitations.

Most economists think you need formalism and quantification to provide scientific explanations -- math models and econometrics.

The roots of this model of "science" are essentially identical in economics to the roots of that in the philosophy of science.


thank you! Big hug!

Arash, on Austrian blogs "illegitimi non carborundum" will go a long way. The vast majority are quite pleasant, of course.

Oh who am I kidding - that'll work well for you on ANY blog, and in life too.


So, from your explanation I suppose Walrasian economics is like this...

* The entrepreneur who deals with prices is modelled by the auctioneer, but ...
* The entrepreneur who introduces his own non-price knowledge falls outside the model.

Is that right?


I'm not sure I would use the word entrepreneur at all, but ...

#1 ... yes, only the auctioneer is allowed to deal with prices, and ...

#2 ... for each price vector arbitarily chosen by the auctioneer all agents express their excess demands and thereby(!) inject their local information into the groping process. (The process is build on recontracting) No knowledge falls outside the model. All information is used if economy is competitive (read: large enough; read: large but finite number of agents).

It is also important to keep in mind that, in contrast to the social planner, the super auctioneer has no other knowledge than to increase prices (proportionally or not) whenever market excess demands are positive, to decrease prices whenever excess demands are negative, and to leave prices unchanged whenever excess demand is zero. That's all. If all excess demands are zero, the equilibrium price vector is found. The process halts and agents are allowed to adjust according to their self-interest. This at least is the most prominent groping process called 'the law of supply and demand'. The process converges, if income effects are dominated by substitution effects. But we cannot find restrictions which we can impose on agents and which guarantee such property.


Here's the problem: If an economist's theory consists of "Alpha + beta = lambda" then you don't advance a theoretical conversation in which you assume there is no additive relationship between the three terms.

How does one "disprove" Y=C+I+G+NX? It is silly by assumption - the math is irrelevant. How does someone who espouses a Misesian framework engage in a debate of macro models with mathematical economists who believe that GDP is a meaningful number?

The debate you want to see won't ever happen because it CAN'T happen. It is a definition-stasis argument. Nobody wants to go back in time 100 years, so here we are.

Mario Rizzo's comment that "Machlup often reminded us that equilibrium is not necessary good and disequilibrium is not necessarily bad" is an important one. Its a sentiment not expressed often enough in economics. Mario - if you have the reference on hand for that Machlup quote, I would love to read it up. Thanks!

If we analyse the macroeconomy as a complex adaptive system, then "disorder" is an essential part of the system. For example, it is true of many such systems in ecology, biology and economics that fragility and disorder at lower levels in the system leads to resilience at higher levels in the system i.e. micro-fragility is consistent with macro-resilience, as I try and explain here . Ecologists understand this concept very well, for example in their concept of "Adaptive Cycles" .

Daniel Botkin explained this concept eloquently in his aptly titled book "Discordant Harmonies" as follows:
"The true idea of a harmony of nature, as Plotinus wrote long ago, is by its very essence discordant, created from the simultaneous movements of many tones, the combination of many processes flowing at the same time along various scales, leading not to a simple melody but to a symphony at some times harsh and at some times pleasing."

On the comparison between microeconomic inefficiencies and macroeconomic inefficiencies that Daniel Kuehn raises, as I understand it the Keynesian position is that put forward by James Tobin when he quipped: "It takes a heap of Harberger triangles to fill an Okun's gap." i.e. micro-inefficiencies are much smaller than the macro sub-optimality.

In the simple static world of neo-classical economics, James Tobin's assessment is probably correct. But in a dynamic conception of the economy, the nature of the micro-inefficiency matters. Mancur Olson's disagreement with Tobin in his book "Rise and Decline of Nations" arises from his dynamic conception of the damage done by rent-seeking. And in the post above, I refer to Burton Klein's work in his book "Dynamic Economics" who also analyses situations where micro-inefficiencies are a key cause of the macro sub-optimality. Klein incidentally is a Keynesian or to be more precise a Shackleian.

As a middle ground, Botkin's statement on ecosystem management again captures the essence: "The fact that some changes are natural and necessary does not imply that all changes, regardless of time, intensity and rate are desirable. There are both natural and unnatural changes, and there are natural and unnatural rates of change. To recognize that melodies and themes are made up of changing tones does not imply that any noise is music. The key to a new but wise management of nature is to accept changes that are natural in kind and in frequency, to pick out the melodies from the noise."

Even if you don't buy that some changes can be "unnatural", some micro-inefficiencies clearly do more damage than others.

Mario & Daniel,

Obviously cells need refinement in an academic presentation -- this is something I hope will be made explicit in my current book (based on 6 lectures I gave last year) tentatively titled The Science of Liberty: Economics, Politics, Policy and Prosperity.

But order is a different concept from equilibrium, though equilibrium is a concept embedded in order. Perhaps the right image is a corridor, or the banks of a river. The activity within the banks is dynamic, fluid and ceaseless yet ordered; whereas if the flow of water breaks over the river-banks, then disorder results.

To Daniel, here we have a difference of opinion on the interpretation of results in search/matching models. Note I am a fan --- see my reference to Alchian --- but I would separate Diamond from the other two. Diamond is more or less similar to Stiglitzian style --- imperfection = market failure. Not an invitation to inquiry into how markets adjust and institutions emerge to cope with the imperfections.

Now where we go from here on how to resolve our disagreement I believe (as my interactions with AMV) not possible through blogs, but only through more traditional academic styles of discourse.


Greg Ransom, re: Nick Rowe.

Greg, if you read through the comments section of Nick's post you'll see he is *not* saying that math is required to "put money into economics". His method is actually very Austrian.

Here is Nick: "Menger would be the original version of "the proper way", because he was trying to explain (succeeded in explaining, in an informal way) why people use money rather than barter."

Sometimes the "mainstream" is more Austrian than you think.

JP, I'd suggest that Rowe has been groping toward something non-mainstream for the last few weeks ...

Note these words, however, "succeeded in explaining, in an informal way" ... Rowe at least hints at a dissatisfaction with this "informal" element. And note also, this is merely an explanation of _why_ people use money, it's not a model of the economy with money within that model, as Rowe is after.

And if I recall, I was the one who introduced Menger into Rowe's conservation, in a comment two weeks ago.

"And if I recall, I was the one who introduced Menger into Rowe's conservation, in a comment two weeks ago."

Nick has been talking about Menger for ages. (ie. He was taught by Laidler who is very much aware of these things.

Lots of backslapping and self-referential bullshitting in these comments. All your "great papers" haven't amounted to anything in the profession.

And the "profession" has proven itself -- by giving us a depression ..

Backstabber said:

"All your "great papers" haven't amounted to anything in the profession."

Arash -

I find your comments about GE models very interesting, but I am not entirely convinced that you touch on the most important market process aspects. I have a paper in a recent RAE about equilibrium, Austrian market process theory, and the advantage of agent-based (complexity, evolutionary) models over DSGE models. You might find it interesting.

Greg - how has the profession given us a depression????

Daniel -- answering your question would be a long conversation.

Check out my blog:

and I'll let you browse around and count the ways yourself.

Daniel wrote:

"how has the profession given us a depression????"


Thanks for your explanation, I'll have to think about it more.

How has the profession given us a depression? By not understanding what causes a depression. But putting economists in power to don't have the least understanding of what causes recessions or that they cannot possibly know enough to understand the nature and outcomes of their decisions. By creating mathematical models that have nothing to do with reality - and using them to argue that there are market failures because the real world doesn't match their magical mathematical thinking. It's hard to say who is more at fault: the foolish politicians who believe foolish economists, or the foolish economists who create mathematical models that magically tell politicians everything they want to hear -- such as that without them, the economy wouldn't work well, that the economy would in fact collapse without their interventions, that they create jobs and wealth, that they can tax and deficit spend and inflate without impunity, that one can measure "Jobs saved" and nonsense such as that, that they can know things that they cannot possibly know, etc. I"m not sure who the bigger villains are, but I am certain bad economists who promulgate folk economics with bogus math is either first or second.

Machlup's article is called "Equilibrium and Disequilibrium" in his Essays in Economic Semantics. It is partly available on Google Books -- but the first few pages are where he discusses the point that equilibrium is not necessarily good and disequilibrium bad.

Professor Boettke,

Another great post. I'm curious however about your opinion on guys like Acemoglu, Shleifer, Stiglitz, Becker, etc. These are all people who, I think, would be considered neoclassical, but at the same time deal with the messy real world. We are studying Acemoglu's growth text in grad macro and although it's a very mathematically dense book, he talks about institutions, culture, and political economy issues. It seems to be that the most relevant economists today on the theory side use a lot of math but they also stress intuition.

Additionally, as an economist in training, I find the balance between messy real world issues and the neat precision of mathematical economics to be very difficult to find. Both approaches have produced great achievements. The insights of Coase, Buchanan, Alchain, Demsetz, and Williamson are profound. But, at the same time, the contributions from Arrow, Samuelson, Diamond and Lucas are just as important. To say they just ignore the real world and fall victim to the economist's caricature is an unfair criticism. I guess it comes down to a marginal argument though. Mathematical economics has been beat to death with hundreds and hundreds of irrelevant papers, whereas a lot of the messynomics has not been exploited.


d'accord. You mention Acemoglu's text book. It is really great, but he is representative in his treatment of aggregation problems. Check Ch. 5, pp. 150-2, Theorem 5.1, very he introduces SMD to raise a "severve warning against the use of representative houshold assumptions" (151). This is great, because most growth textbooks (or macro textbooks, in general) usually don't mention SMD at all. But he teaches a misrepresentation of SMD: He argues that the problem posed by it can be solved by "imposing further structure" on households like Gorman utility functions (homothetic preferences). This is opposed to the very meaning of SMD which was introduced in opposition to Gorman's aggregation theorem. So be aware of this grave mistake in an otherwise excellent textbook!


I'll check. Can you please provide the title?

read "... where he introduces SMD ..."

Arash: Competition as market progress: An Austrian rationale for agent-based modeling

If you want to understand today, you have to search yesterday.

If you want to understand today, you have to search yesterday.

The comments to this entry are closed.