July 2014

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

« The Unintended Consequences of Ultra Easy Monetary Policy | Main | Quotes of the day inspired by recent ECB and Fed actions »


TrackBack URL for this entry:

Listed below are links to weblogs that reference Is Austrian Economics Empirical? This Month's Cato Unbound:


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

Professor Horwitz what do you think of this paper?


Pace, I don't think Prof. Horwitz is cutting to the issue here. Yes, discussing the relationship between theory and empirical work is necessary for pushing forward our understanding of them, but I do not think this captures enough of what is going on in Austrian apriorism to do so. Focusing on Mises, which if speaking about Austrian apriorism is going to be a fecund affair it is really only going to be so with Mises, if we really want to understand his notion of economics in Human Action and how that influences his view of the social sciences, we are only going to advance our understanding by employing some heavy Kantian philosophy.

Of course, Horwtiz certainly does mention Kant and Mises' indebtedness to him for supplying the notion of aprioristic analysis that permeates Human Action. He even does this to some extent, even noting Mises' own evolutionary twist on Kant, writing in the article:
"We all have 'a set of tools for grasping reality' that comes to us from our evolutionary heritage. The commonality of those tools allows us to engage in the reflection on action and the development of that core of economics as a set of necessary insights about how humans act. This core economic knowledge is not contingent but part of the very structure of human minds and is something that we can come to know."
Even though this along with other arguments in the essay certainly do provide the Kantian argument in a nut-shell, I do think he goes far enough. I do not think that the reader will come from this piece with a strong enough idea of the mode of argument found in Kantian aprioristic reasoning to really understand the issue in depth. This in turn is a problem because the debate about whether the fundamental laws of economics are aprioristic because if the reader does not know exactly what it means to be aprioristic in a Kantian fashion, they cannot address the truth-value of that statement.

If we are really to get across what Mises meant, I think we have to understand Praxeology as the architectonic of the social sciences unified by the synthetic a priori statements that could be made about human action just as Kant at the end of his Critique of Pure Reason tried to establish an architectonic of pure reason held together by his previous synthetic a priori principles like space and time. Then by understanding Praxeology as a Architectonic can we then clearly understand how Mises understood the the interplay between empirical work and theoretical work.

To start, consider Kant's own description of what an “Architectonic” was:
"Under the government of reason our cognition cannot at all constitute a rhapsody but must constitute a system, in which alone they can support and advance its essential ends. I understand by a system, however, the unity of the manifold cognitions under one idea. This is the rational concept of the form of a whole, insofar as through this the domain of the manifold as well as the position of the parts with respect to each other is determined a priori." (The Critique of Pure Reason, A831/B859)
Within Kant's system (like Kant, Mises was also a system-builder), the Architectonic served the role of being the most formal organization of human knowledge. Within each Architectonic, there is a single rational (i.e. a priori) statement that unifies an organized branch of knowledge that now rotates around.

Such systems, though, need a set of procedural rules, that will be necessitated by the character of the Architectonic, by which the representations of objects in the world are organized. Kant provides the schema as these procedural rules:
"For we have seen there that the concepts are entirely impossible, and cannot have any significance, where an object is not given either for themselves or at least for the elements of which they consist, consequently they cannot pertain to things in themselves (without regard to how and whether they may be given to us) at all; that, further, the modification of our sensibility is the only way in which objects are given to us; and that, finally, the pure concepts a priori, in addition to the function of the understanding in the category, must contain a priori formal conditions of sensibility (namely of the inner sense) that contain the general condition under which alone the category can be applied to any object. We call this formal and pure condition of sensibility, to which the use of the concept of the understanding is restricted, the schema of this concept of the understanding..." (The Critique of Pure Reason, A139-40/ B178-9)
Beneath the sheer density of it all, what Kant is getting to is that once we have our aprioristic concept, that concept must shape the way that the human being can experience the world, can have mental images of it, and the way that the concepts do so is through the “Schemata.” The human being then can only know about the world via the schemata and all true statements about it are thus conditioned by the way the schemata limit human experience.

Thus we are left with a Architectonic=>Schemata=>World process by which the statements that are true a priori limit and demarcate the possible experiences human beings can have. Indeed, within an orthodox Kantian epistemology, it is the way by which aprioristic statements limit human experience that they gain their truth-value.

However, do we indeed find the Architectonic=>Schemata=>World notion of human knowledge within Mises' own works. To that question I think a proper reading of a very few pages of Human Action yield a resounding yes so much so that they cannot be understood any other way.

First, compare Kant's notion of the Architectonic with Mises' definition of “Praexeology” in Human Action:
"Praxeology is a theoretical and systematic, not a historical, science. Its scope is human action as such, irrespective of all environmental, accidental, and individual circumstances of the concrete acts. Its cognition is purely formal and general without reference to the material content and particular features of the actual case." (Human Action, pg. 32)
By comparing the two, I argue that we must come to the conclusion that Praexeology is a Kantian Architectonic held together by what Mises considered the aprioristic idea of action. Action in turn was an a priori statement in Kantian terms because it limits and demarcates the possibilities of human experience, so it is necessarily true by the existence of human experience. Action was the pure idea and Praxeology in turn assembled as its Architectonic.

Now, we finally get to the meat of this discussion: the Schema of Praxeology. How do human beings come into contact with action in the world? They do so by its Schema! What are the Schema? Well, the Schema are what Mises spends most of Human Action trying to deduce: the necessary laws of economics that are antecedent to human experience within the economy.

So, I might have gone on too long already, so let me provide a quick summary. I do not think Prof. Horwitz has provided a study of apriorism within the Austrian tradition that can really provide a proper spring-board for debate. The reason I think so is that he does not provide enough time to explain why it is that Mises thought the laws of economics were a priori and if we are to suitably do so we must truly delve into the Kantian metaphysics that underlies it all. Only by doing that, only by emphasizing: Architectonic=>Schemata=>World can we truly place Misesian aprioristic reasoning in the form that not only captures why Mises had to theorize in such terms, but also provides a structure that is easy to compare with other visions of human knowledge thereby enabling the fecund debate we would like to have about this.

If we want to understand Misesian apriorism, we must grab the Kantian bull by its Kantian horns. To do so certainly is difficult and it certainly has a high intellectual opportunity cost, but to do otherwise would be to obfuscate the deep underpinnings of the issue.

Thanks for the link, Omar.

I definitely agree with Steve on all this, but one interesting omission from the list of empirical papers is the largish literature on ABCT. A lot of these find that ABCT is verified (contrary to the paper you provide, based on the abstract).

I think there is a lacuna in the current ABCT empirical literature - namely, it's very hard to determine from the existing literature which state of the capital structure is the "sustainable" state. Usually the analysis is concerned with showing how the capital structure and relative prices shift over the business cycle. They do shift in a way that confirms ABCT. The question is, is this a case of (1.) a sustainable capital structure being distorted by a recession or is this a case of (2.) an unsustainable capital structure being corrected in a recession. ABCT is not indifferent between the two options, but much of the empirical literature doesn't tackle this.

The one I know of that tries to is Carrili and Dempster (2008, RAE), but I'm not entirely satisfied with how they do it.

Very intersting! Thanks!

I do not think one can see all of this is a balanced way without taking into account the "apriorism" of Debreu's axiomatic method. That is true apriorism. In fact, Debreu made it clear that we can and should separate the generation of theory from ALL applications -- axiomatic theory is separate from content.

For further details on my point see:


Debreu's views are actually quite extraordinary and little appreciated by neoclassical economists who think naively about their field.

Hayek was right.

LEARNING is the principle causal / empirical / contingent element in economic science.

Hayek was right.

Economics is an empirical science because it begins with PROBLEM RAISING PATTERNS IN OUR EXPERIENCE.

Hayek is right.

Economics is an empirical science because it identifies the CAUSAL CONSEQUENCES of the existence of empirical patterns in the negative rules of just conduction we follow, eg property right rules & rules of honest and just dealings in contracts, etc.

Hayek is right.

Economics is empirical because is traces out the patterned causal consequences of people adjusting their judgments & plans in the context of changing relative prices and local conditions, ie changing empirical conditions.

The is the CORE of the empirical status of economics as a science.

Economic science provides a contingent causal mechanism, and empirical explanation, for problem raising causal patterns in our experience.

Hayek solved this problem of the logical status of economic science in the late 1930s and early 1940s.

All other efforts are based on pathological understandings of "science", most of the derived from the failed traditions of philosophy, especially the bogus demand for justification, the legacy of the demonstrative picture of knowledge going back to Euclidean geometry, a picture explained in the 20th century by WIttenstein, Kuhn, Popper, Polanyi & Hayek, among others.

"Economic research" of all different kinds in all sorts of different domains adds to various kinds of our knowledge in all sorts of different ways.

But the CORE of any understanding of the empirical / scientific status of economics is understanding the nexus of problem raising patterns and contingent causal explanatory mechanisms laid out by Hayek -- just as the CORE of any understanding of the empirical / scientific status of Darwinian biology is understanding the nexus of problem raising patterns and contingent causal explanatory mechanisms laid out by Darwin (see in particular Ernst Mayr's _One Long Argument_.)

Physics had Netwonian theory and now quantum theory and relativity theory, with efforts to combine them in M-theory.

Chemistry has atomic theory.

Biology has theories of evolution, and is becoming even more unified under self-organization theory.

Oddly, it is the so-called heterodox economic schools that seem to have any sort of unifying theory of economics. Mainstream economics seems to mostly be a hodgepodge. Austrian economics is unified under Austrian theory. Marxists have a unified theory. Keynesians have a unified theory.

Empirics alone don't make what you're doing science. You have to have a theory. Without a theory from which to interpret your results, you just have a hodgepodge. But you don't have science.

Troy -
I think you're making something of a category error. And it's not your fault. Economists often (but not always) use the word "theory" to describe these things.

What we have in mainstream economics is a bunch of models. The models describe different mechanisms at work in the economy, which is a very complex subject to study with lots of mechanisms operating simultaneously. I think economics does have a unifying theory, and that's Smithian theory. But we have lots of models flying around to explain different parts of the elephant. Some models are better at getting a more complete picture of the elephant. some models are really designed to just explain one particular facet of it (often that's helpful, sometimes the specificity is problematic).

This is similar in a lot of ways to biology. Yes, biology has a unifying Darwinian theory just like economics has a unifying Smithian theory. But biologists use all sorts of different models when they're approaching different problems, and they also talk at different levels of aggregation depending on their research interests without obsessing about it (too much).

Alex Gheg has a new framework that will shock all economists, not just the Austrians. Quantity, quality, variety and convenience in one equation. You will also see a scale for utility across generations. http://www.youtube.com/watch?v=u6tFLGpcOpE

I think Austrians don’t understand the game being played. Mainstream econ defines the terms and sets the rules. By definition, their method is science and the Austrian is not because it contradicts mainstream econ on important points. And since science is based on empirical data by definition, therefore Austrian econ cannot be empirical.

I like Steve’s explanation of the empiricism behind Austrian econ, but it’s a different definition of “empirical” from that of mainstream econ, so no mainstream economist will ever accept it.

Following on Steve’s definition of empirical, it’s obvious that Austrian econ is more empirical than mainstream because mainstream ignores time, heterogeneous capital and money (except for admitting it can change prices in the long run).

re: "I like Steve’s explanation of the empiricism behind Austrian econ, but it’s a different definition of “empirical” from that of mainstream econ, so no mainstream economist will ever accept it."

Well, maybe people might differ on Kant, etc.

Practically speaking, that's irrelevant. The mainstream will publish Leeson, Storr, etc. That ain't bad.

You might want to look at Micheal Keane’s ‘Structural vs. atheoretic approaches to econometrics’ published in the journal Journal of Econometrics (2010).

Keane argues that it is not possible to learn anything of interest from data without theoretical assumptions. All statistical inference relies on some untestable assumptions

"this paper attempts to lay out the sources of conflict between the so-called "structural" and "experimentalist" camps in econometrics.

Critics of the structural approach often assert that it produces results that rely on too many assumptions to be credible, and that the experimentalist approach provides an alternative that relies on fewer assumptions.

Here, I argue that this is a false dichotomy. All econometric work relies heavily on a priori assumptions.

The main difference between structural and experimental (or "atheoretic") approaches is not in the number of assumptions but the extent to which they are made explicit."

Also look at Keane, Michael P. 2010. "A Structural Perspective on the Experimentalist School." Journal of Economic Perspectives where he argues that we cannot begin a systematic assembly of facts and empirical regularities without a pre-existing theoretical framework that gives the facts meaning and tells us which facts we should establish.

Omar, Thanks for the link to the Lester-Wolff paper. It was very interesting, but as with most criticisms of the ABCT, they fail to understand the theory. Here are some examples:

“ABCT predicts PPI and IP to be negatively correlated with the federal funds rate.” P. 7.

Not necessarily. The authors are assuming that the ABCT understand the quantity theory of money working mechanically so that a reduction in interest rates mechanically increases the money supply and mechanically raises prices. That’s not the case. Also, econometric studies of changes in money have a lag of as much as four years before the effect shows up in prices. The link between interest rates and prices is much more complex than what the authors present. In addition, prices may not rise at all if productivity is growing fast enough, as happened in the 90’s.

“Theory predicts that entreprenuers are \surprised" by changes in base rates, otherwise indicating an independence of FFR from contemporaneous
values of prices and production activity.” P. 7.

I don’t think that is a part of any variety of the ABCT.

“The results here show that the components of the aggregates, decomposed by their stage of process, have the same qualitative pattern and similar quantitative magnitudes. In other words, the credit expansion looks as if it is uniformly distributed over the various stages- of-process. If anything, production expands the most in the middle order sectors, whereas the theory predicts resources should be leaving the middle and moving towards production of higher and lower order goods. Consequently, the resulting sectoral reallocation following
a monetary policy shock is not supportive of ABCT.” P. 10

The problem here may be aggregation. The ABCT doesn’t say that every last producer goods business will mechanically expand the instant interest rates drop. Many don’t. But history shows that every expansion has a leading industry that expands into a bubble. In the 80’s it was fiber optic cable. In the 90’s it was dot.com and computer companies. In the 2000’s it was housing.

And the ABCT doesn’t say that resources should leave the higher and middle order goods for the lower order goods. That’s what would happen in an expansion financed out of savings. In an expansion financed by credit, lower order production competes with higher order production for resources and drives the prices of both upward, but the prices of lower order goods rise faster.

“In addition to predicting a boom following an expansion of credit, ABCT predicts that a bust is sewn in the seeds of the boom. The results unequivocally dispute this. There is no evidence of a monetary contraction leading to a subsequent reduction in output. Even lengthening the time horizon to a period of ten years as in figure 4, shows no evidence of a contraction in prices or quantities.”

In the ABCT, a monetary contraction is not necessary, only a slowing of monetary growth.

“The case could be made that these aggregate indexes are still to coarse to study relative price and quantity changes seriously. We are dubious about this claim.”

They should take it more seriously.

“Finally, and as a consequence of the first two results, the predictions of ABCT are not validated in the empirical analysis.”

No. One flawed study that does not understand the ABCT proves nothing. There is plenty of empirical data that supports the ABCT. The recent paper by White is just one of the latest.

A lot of Austrians fixate on the interest rate, and it is important, but I think Hayek’s emphasis on the Ricardo Effect makes the ABCT much more robust. In “Prices, Interest and Investment” Hayek shows that the actual interest rates is far less important than profits in determining investment. If the Fed keeps interest rates too low and never changes them, the boom and bust cycle will happen regardless. The Ricardo Effect is supported by mainstream micro econ theory, too, in the PPF of the trade offs between capital and labor.

If only mainstream economics *were* Smithian. Lipservice is hardly adherence to a theory. Otherwise, economics would not have been so slow to embrace spontaneous order theory and, now, complex adaptive systems, emergence, self-organization, etc.

In biology, all models conform to evolution. Darwinist biology -- specifically, the New Synthesis -- IS biology. Period. And while there are now rumblings about creating a new New Synthesis that would include self-organization, network theory, etc. more explicitly, this is only working to expand on the already-existing unifying theory with which all biologists consciously interpret their results.

That is not the case in economics. It could be. And when it is, it will, broadly, become a real science. The few who actually do economics the way biology is done (something about which I do have experience) are real economic scientists rather than muddlers. Of course, each science when through it's muddling stage. And often things look more muddling in retrospect than when one is in it. Biology is a good example -- the per-Darwin Aristotlean biologists hardly thought themselves muddlers, though they certainly look that way from our perspective.

Economics will similarly be viewed in 100 years. Or less, if we can admit it's mostly a muddled mess. That will happen when mainstream economics returns to being mainline economics.

I recall Alfred Kahn saying at a conference that "introspection was his favourite form of empiricism." This seems to play a role both in Austrian apriorism and subjectivism. I would add that, subject to its limitations of scope, it is not only my favourite method but the method in which I usually have the most confidence.

I am wondering whether there isn't a fundamental difference in the way that Austrians use non-introspective experience or empiricism. The mainstream sees it most importantly as a test of a conjecture's/theory's/model's truth and secondly perhaps as a sign of or guide to the existence of an as yet unexplained underlying causality. In the second case, an opportunity presents itself to apply new or existing theory to explain events.

I see the Austrians as generally rejecting the first and accepting the second, subject to an additional concern in both cases regarding the difficulty or impossibility of determining the "facts" with great precision. However, due to the latter concern, as noted by Steve Horwitz, the Austrians often prefer qualitative empiricism to quantitative.

Hayek's concern in his Nobel speech was that mainstream econ had limited the definition of empirical to available data that would fit an econometric model. That type of data might not exist, and limiting all economic knowledge to just that emaciated definition of empirical rules out most of what we know about how economies work.

"...advocates of Austrian economics on blogs and Internet forums, rather than engaging with the professional research being published in the peer-reviewed journals by practicing Austrians."

Who are you referring to in the former category, Steve?

Hayek invented the word "scientism" to refer to fake attempt at "science" and this bogus & fraudulent conception of "the empirical":

"Hayek's concern in his Nobel speech was that mainstream econ had limited the definition of empirical to available data that would fit an econometric model. That type of data might not exist, and limiting all economic knowledge to just that emaciated definition of empirical rules out most of what we know about how economies work."

Hayek's attack on the fraudulent understanding of science & the empirical with his 1920 paper on Mach which eventually because _The Sensory Order_.

Very nice essay.

As far as Hayek is concerned, economics was an empirical science for him. See his "Economics and Knowledge" for example:

1) In short, I shall contend that the empirical element in economic theory--the only part which is concerned not merely with implications but with causes and effects and which leads therefore to conclusions which, at any rate in principle, are capable of verification -- consists of propositions about the acquisition of knowledge.

2) Whatever may occasionally have been said by overpure economists, there seems to be no possible doubt that the only justification for this is the supposed existence of a tendency toward equilibrium. It is only by this assertion that such a tendency exists that economics ceases to be an exercise in pure logic and becomes an empirical science; and it is to economics as an empirical science that we must now turn.

"I invite the reader to try and identify a single instance in which a "deep structural parameter" has been estimated in a way that has affected the profession's beliefs about the nature of preferences or production technologies or to identify a meaningful hypothesis about economic behavior that has fallen into disrepute because of a formal statistical test."

"The Scientific Illusion in Empirical Macroeconomics", Lawrence H. Summers
Scandinavian Journal of Economics, Proceedings of a Conference on New Approaches to Empirical Macroeconomics. (June 1991), pp. 129-148

George Selgin's reply is rather disappointing.

Selgin implies that Mises = Austrian economics.

And, then, later in the piece he implies that Hayek was a Misesian on the question of the logical status of economics, when Hayek tells us explicitly that his 1937 paper was directed (in part) at Mises to show him how he was wrong about the logical status and explanatory strategy of economic science.

I'm always stunned by those who grew up on Mises as very young men always seem unable to shake those Misesian lens when the pick up a Hayek text.


It is possible to argue that Hayek misunderstood Mises, which may explain why Mises liked Hayek's article.

Richard Ebeling

Whether or not Hayek understood Mises correctly in every particular is an issue independent of whether Hayek got the logical status of economic science right and Mises got is wrong.

There is a huge overlap between Hayek & Mises, and many, many elements of Mises picture fit into Hayek causal, contingent, empirical picture of the explanatory science of economics.

My sense is that Hayek was well aware of this.

In one of his video interviews, Hayek said that he agreed with Mises but didn't like the way Mises got to his conclusions. I'm not sure what he meant, but he never seemed to disagree with Mises much.

Hayek in the 1978 UCLA interviews says that his 1937 paper "Economics and Knowledge" -- laying out logical status of economics, clarifying the causal/empirical element vs the tautological/formal element in economics -- was directed at Mises (delicately so, Mises had ended relationships over lesser matters).

One thing that is rarely thought about is how Hayek's writings informed work Mises produced after the fact of Hayek's efforts on important topics.

Thinking through and articulating logical status & explanatory strategy of economic science has proved very difficult and extremely complex, particularly given the pathological traditions dominating our pictures of knowledge & science & social theory.

It isn't surprising that anyone should get a lot right and still get core issues wrong.

I think doing is usually more powerful than talking, and the younger Austrians are doing good empirical work. I do wish Austrians and their boosters would have more respect for and interest in "statistical correlation." My Big Players work included lots of statistical correlation. As far as I can, however, that part of it was pretty much ignored. I hope Andrew Young's very excellent work in macroeconomics enjoys a better fate. He produces some "statistical correlations" to show that ABCT fits recent facts pretty well. Click here: http://www.springerlink.com/content/68h2307p03352751/

The Austrian taboo on econometrics has become anachronistic in the wake of the huge declines in the cost of computing, which has made it easier to use nonparametric techniques and, more generally, to adopt tools suited to your data and your theory.

When the behaviorists blackballed the cognitivists in psych, only behaviorist empirical work got done because cognitivists didn't exist & weren't replicated in the grad schools.

Behaviorist 'empiricism' became _defined_ as empiricism in psychology, and all other empirical work was berated as non-scientific and non-empirical.

These parallel cases -- you can find others in biology & linguistics, etc -- highlight and spotlight the sort of phenomena we are interested in in the economic case.

Use parallel cases if you want people to see what you want them to see, without the clear examples outside of their field they are likely to be blinded to the phenomena in their own field.

The problem you will have with this strategy is that economists are so ignorantly science and the history of science & ideas they won' be able to follow the examples.

Thank you for linking Young, Roger. I agree, his paper is excellent. It (and other ABCT empirics, often from young Austrians) will definitely get attention from me in a forthcoming Critical Review issue on Hayek.

Roger, great paper by Young! Thanks for the link. And a nice response to the Lester/Wolff paper linked to by Omar.

Young and Lester/Wolff use essentially the same data. But that data was configured with Keynesian macro in mind and can't be used in its raw form for Austrian econ. Young shows how to correctly configure it.

But I can already hear critics saying that he manipulated the data to get the results he wanted.

I read the responses of Selgin and Caplin. Caplin doesn't even address the question of whether or not Austrian econ is empirical. He just vents.

Selgin had some good points. It cleared up for me some confusion I had reading Mises. On the one hand he emphasized the a priori nature of econ while on the other offered many real world examples. Hayek offered even more.

Like Koppl I would like to see more econometric analyses like what Young did among Austrians. It doesn't prove anything and won't change anyone's mind, but it is good applied economics. I would like to see Austrian econ be less about political economy and more applied economics.

Theory is the necessary foundation for empirical research, but even non-economists want to see how the theory has worked out in the real world. The real world brings the theory to life by giving the skeleton flesh and skin.

Greg Ransom badly misreads both Steve's opening essay and my reply to it in finding fault with me for implying "that Mises = Austrian economics" and that "Hayek was a Misesian on the question of the logical status of economics, when Hayek tells us explicitly that his 1937 paper was directed (in part) at Mises to show him how he was wrong about the logical status and explanatory strategy of economic science."

It should be perfectly obvious that I focus on Mises and "praxeology" because that is for the most part what Steve is talking about (he devotes only a single parenthetical statement to Hayek), and because I take it that when one is asked to respond to someone's arguments one is not supposed to respond to some other arguments instead!

As for my implying that Hayek was a Misesian, far from doing any such thing, I never so much as mention Hayek (see criticism #1); nor do I think I can be justly accused of not knowing about the different views he and Mises held on the subject of economic method.

To be sure, Hayek's distinct views on the proper role of empirics in economics are themselves worthy of attention. But for better or worse the issue, as it is developed in Steve's opening essay, has become one concerning whether "praxeology," meaning Mises' particular understanding of what others call "Austrian" economics, is or is not "anti-empirical."

I think the correct dichotomy is - deductive reasoning (e.g. Mises' Human Action framework) v.s. inductive reasoning (that's empiricism).

My take is that inductive reasoning is less rigorous than deductive reasoning - 'coz you have to make assumptions and so forth, but that's part of the process of interpreting data.

Deductive reasoning is most water-tight, subject to the axioms of your reasoning framework. But those axioms are usually not the end-all, be-all of things... they rarely conform perfectly to reality.

"they err in taking this to mean that experience—and experience captured in statistics especially—cannot tell us that some theories that are both logically valid and potentially pertinent are nevertheless false in the (common) sense meaning that they do not really explain what someone claims they explain, or explain far less than some alternative."

Problem is, it doesn't work that way. Never has. Look at the debates over the the effectiveness of TARP. Each side present models based on their own assumptions and results that reinforce those assumptions.

Arnold Kling has some great posts on the circular reasoning behind most econometric studies.

That mainstream economists are blind to the circular reasoning doesn't change the fact that it is circular reasoning.

Sorry, the quote above was from Selgin's rebuttal. Here is Leamer on econometrics from one of Klings old posts:

"...Let's face it. The evolving, innovating, self-organizing, self-healing human system we call the economy is not well described by a fictional "data-generating process." The point of the sensitivity analyses that I have been advocating begins with the admission that the historical data are compatible with countless alternative data-generating models." http://econlog.econlib.org/archives/2010/05/the_econometric.html

And here's Kling: "I think if you step back and look at econometric history, you see the rise and fall of fads: simultaneous equation estimators, distributed lags, VAR's, instrumental variables, regression discontinuity, and so forth. If you jump through the right hoops at the right time, you get your paper published. But what you publish is not reliable.

"Most economists eventually see through the older fads. But while the technique is going through its boom phase, woe be it to the young economists who fails to jump on the bandwagon. It shouldn't have to be that way."

Kling worked in government for decades running the latest and greatest econometric models. He has no confidence in them.

Here is Kling in his "Macroeconomics: the lost history":

"We badly want macroeconometrics to work. If it did, we could resolve bitter theoretical disputes with evidence. We could achieve better forecasting and control of the economy. Unfortunately, the world is not set up to enable macroeconometrics to work. Instead, all macroeconometric models are basically simulation models that use data for calibration purposes. People judge these models based on their priors for how the economy works. Imposing priors related to rational expectations does not change the fact that macroeconometrics provides no empirical information to anyone except those who happen to share all of the priors of the model-builder."

George in his discussion of "Austrian economics" and "anti-empiricism" directly - and at the beginning of his piece - invokes Josh Barro's piece, which in the first instance was an attack on Hayek, an attack in which Barro characterized Hayek as rejecting "empirical analysis".

George writes,

"I’m afraid that I’m inclined to think that Mr. Barro has a point."

George then writes,

"He [Barro] has a point, both because his claim seems to be in perfect agreement with many of Mises’s own claims about praxeology, and because in practice, notwithstanding the exceptions Steve mentions, many self-styled Austrians, including some who have perhaps the best claim to being consistent practitioners of the methods of which Mises approved, do in fact “reject empirical analysis” as a basis for drawing conclusions about the consequences of alternative economic policies either by condemning such analysis outright or by simply not resorting to it."

I think I can be forgiven for inferring that George is saying that "Mr. Barro has a point", when Barro's central point is to dismiss Hayek - author of the book Barro is out to ridicule - as anti-empiricist.

George himself makes Barro's remarks part of the focus of his discussion -- I can be forgiven for allowing myself to be misled, I would suggest.

Contrary to George's recollection, he does mention Hayek in his comments:

"Professor Lavoie was nothing if not eclectic in his approach to economics, and so left his students with an appreciation, for praxeology certainly, but also for hermeneutics and, in his last years especially, economic history, among many other things. It is, I believe, that eclectic approach rather than Austrian economics as expounded by Mises, or even by Hayek, that is reflected in the empirical work to which Steve refers. That of course doesn’t mean that one shouldn’t refer to such work as Austrian economics, or to its authors as Austrian economists. But it does mean that the work may not be representative of broader Austrian-school tendencies."

Let me suggest that the fact that Mises work is freed open when conceived in light of Hayek's critique & Hayek's cross-disciplinary unification, is a liberating reason why the door is open to Lavoie's more "eclectic" Austrian endeavors.

This whole label "Austrian" has a problematic & sometimes pathological role in the functioning of contemporary science & popular political economics.

I like George's own choice not to adopt the label.

I no longer use the label for Hayek's work, as far as I can help it.

The term "Austrian" today causes more confusion & false understanding than it offers clarity of significance in use. A fact exploited by journalist/trolls like Barro intentionally aiming to ridicule and marginalize scientific work via broken conceptual categories and false stereotypes.

Great post! Thanks for sharing!

McKinney: It is of course true that plenty of econometrics is baloney--I say as much in my post--but it hardly follows that we are therefore justified in dispensing with attempts to estimate magnitudes or to employ statistics in attempting to determine whether theories actually account for particular observed phenomena.

Greg Ransom: Let me say for the record that Barro's essay is itself as ignorant as it is arrogant. (Having seen the senior Barro perform at Prague last week, I can say with confidence that this is one apple that didn't fall far from its tree.) And, yes, in it he goes after Hayek rather than Mises. But it is Mises views on Austrian method that Steve chooses to defend: look at his essay again and you will see eight of the dozen or so paragraphs dealing with Mises, several showing work that supposedly illustrates the point that Mises' approach doesn't rule out empirics, and nothing at all on Hayek.

Once again: I was asked to respond to Steve's piece, not to Barro's.

George, Kling isn’t saying that much econometrics is baloney or poorly done. He’s saying the whole scheme is faulty because it is based on circular reasoning, especially macro modeling.

Have you ever known someone to be unable to find support for a theory using econometrics? Or have anyone change their mind about a theory because of econometrics?

Historical data is that is so vast and complex that one would have to be a dunce to be unable to find some kind of confirmation in it for one’s theory. Keynesians, monetarists, Austrians all find statistical support, but they can’t all be right about areas in which they disagree fundamentally.

I love econometric studies, like the one by Young linked to above. But all econometric models do is flesh out the details and provide illustration of a point. If good they can help with forecasting. But I have never seen econometrics “determine whether theories actually account for particular observed phenomena.”

For example, I tried at one company I worked for to estimate the price elasticity of demand and couldn’t do it. No matter what I did, the results showed a positive sign for price. I didn’t assume that the theory of demand is wrong, or didn’t apply to my company. I assumed that there were factors involved for which I had no data.

Again, the shortcomings of many econometric models do not supply an adequate reason for condemning econometrics altogether as a means for estimating magnitudes when several influencing factors have been in play simultaneously. It isn't a question of a regression settling things: it is one of several ways in which to argue the plausibility of a particular hypothesis. It is, in short, part of the story. And the fact that GIGO applies doesn't mean that there aren't cases in which one has more than G to work with.

Bill Lastrapes and I once wrote a paper arguing that the 2% check tax included in the '32 Revenue Act contributed substantially to the Great Monetary Contraction by raising the demand for currency. To establish that it was substantial we ran regressions. Might someone make the case, with different econometrics, that our estimates were not correct? Of course they might. Was our attempt to employ econometrics to make our case therefore nothing more than so much window dressing? I don't think so. It was but one piece of a body of evidence and argument; perhaps we got the magnitude wrong, but we surely would not have made our effort more convincing by merely asserting that the check tax mattered and leaving it at that.

Sweeping Hayek under the "Austrian economics" label, Barro smears Hayek as being against "empirical analysis", Steve defends "Austrian economics" focusing mostly on Mises, and George responds to Steve, and say that Barro has a point about "Austrian economics".

This episode suggests - perhaps -that the "Austrian" label should either be retired, or should be used by everyone with explicit care.

But we can't count on that outside a very tiny community.

So are the benefits worth the costs of chronic miscommunication, false stereotypes, and the general damage done to those not toeing the Mises Institute line?

For those advancing some aspect or other of the Hayek paradigm, likely not.

Has anyone written a good book or article on Henry Schultz's effort to develop a statistical science of demand functions?

Friedman from early on seems to have at least partly understood that the logical conditions for using statistical analysis did not hold (beyond a short period, as Friedman thought).

What other good books or articles would folks recommend on the problem of applying statical analysis to demand functions?

I think many Austrians take the right approach to regressions: they don't offer proof, but they offer support. They are more than window dressing.

Take your example of the check tax. Because of the problem of high correlations between historical variables, the addition of a single additional variable might reduce the magnitude of the effect substantially.

I have done a lot of regressions in business and as you know the addition of variables can cause the sign of original variables to change.

Then there is the problem of the error of the estimate. With historical data it is so large as to be almost meaningless.

I'm somewhere in between the attitude that econometrics offers proof and that it is useless. I think it adds support, but is not as powerful as reason.

McKinney: All true. But one is still entitled to ask, what is the best estimate one can offer of the magnitude of X? T-statistics and robustness checks are some ways to get to the question of how confident one can be in any proferred estimate derived econometrically. It remains the case, in any event, that non-econometric alternatives are quite often less rather than more credible than their (admittedly faulty) econometric counterparts.

It was good to be reminded of the Roger Koppl work on big players because that was precisely the kind of work that could have been expected when the Austrians had enough numbers and resources to do empirical work, rather than whatever the busload of Austrians in 1974 regarded as their top priorities. The point is to have theories that are worth testing, deduce some hypotheses, then get some numbers and apply appropriate tests to see if the hypotheses stand up. Of course most of the statistics that are readily available were collected for use in Keynesian macro models but if the underlying theory is incoherent then no amount of data or statistical expertise will help.

Contra Roger in one small way, the advance of computers has a downside, you can do too much without thinking what you are trying to achieve, like a hundred runs before lunchtime. I was into and out of econometrics in the 1970s when we had to punch cards and we only managed one run a day, so we had a brief team meeting before each run to decide what do to. I kept minutes of these meetings and so it is quite likely the best documented project in the history of econometics. When they disbanded the research unit and started throwing things out I put the records into the State archive for posterity. The project was a waste of time but it demonstrated that doing regressions is about as much fun as you can have with your clothes on. This is the report, but don't waste a lot of time reading it!


STOP PRESS. Interesting project using regression analysis picked up by Pete Klein and O&M.

As someone pointed out, it is all about understanding what the method can and can't do, especially when there are problems with data quality and also a lot of correlation between the variables which can make the coefficients practically meaningless.

There was some nice regression work coming out of GMU a few years ago, can't recall the details, possibly on the effect of foreign aid in promoting democracy or social stability. There has been a mountain of bad regression work on aid to the third world, I blogged on this and people laughed but a few months later Dani Rodrik had a paper in a refereed journal making the same points.

McKinney writes:

"Like Koppl I would like to see more econometric analyses like what Young did among Austrians. It doesn't prove anything and won't change anyone's mind, but it is good applied economics."

This is not true, at least from my sample of one. I started actively collecting what empirical studies there were for ABCT about two years ago and reading them has changed my mind about it. I went from being skeptical about changes in the capital structure to being pretty much convinced of the Austrian story. Young's paper was particularly important in swaying me - I first read it about a year ago, before it was released.

Empirical evidence matters a lot and theories have risen and fallen on empirical evidence. Don't discount that.

I'm obviously not an Austrian, though. I do think there are some real concerns about how prominent ABCT mechanisms are for explaining actual business cycles. But my read of the evidence is that those processes absolutely do happen - they're just not as important as some proponents suggest.

One of the problems is that a lot of the biggest proponents of the theory exert little effort in actually looking for it in the data.

In a lot of Austrian papers, we often get a formula of:

(1.) Noting that the Fed did something other than maintaining a constant MV before a recession.

(2.) Pointing out that there was a recession.

(3.) A detailed theoretical restatement of ABCT without any empirical effort of showing that that was what connected (1.) and (2.).

My view is that a lot of these papers can dispense with (3.). We're all aware of the contours of the theory.

Larry White says that blaming recessions on greed is like blaming plane crashes on gravity. It's always there - you need more than that. The same goes with a Fed that has been pursuing modest inflation for decades on end. It's always there. If you want to tie it to recessions you need to detail a more specific mechanism and actually provide evidence for it.

Daniel, I'm glad that the paper influenced you. The young are more open minded and receptive to change. I have not seen a regression change any older economist's mind. The obsession with data that our culture fosters means that we need to provide more in order to persuade the younger generation.

Your point about the Fed and inflation being like gravity is important, too. That's why Hayek's Ricardo Effect and PII are so important. The Ricardo Effect is nothing but the labor/capital PPF taught in most econ textbooks applied to business cycles.

Albert Einstein : As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality.

Do you think?

I see the history of economic thought (old guys, largely) rising and falling on empirical evidence.

Real wage cyclicality was critical in the rise of RBC. 1970s was critical in establishing expectations augmentation. Empirical evidence guaranteed that the Lucas critique remained but the Lucas supply curve did not survive (at least as an explanation of the Phillip's curve - you still see it hanging around in models where agents are sorting out signal and noise). Empirical evidence from Bob Hall and Fred Mishkin (I think Mishkin was involved in this???) has been critical in squashing rational expectations as traditionally stated and in introducing behavioral wrinkles to our thinking about rationality.

From what I can tell, empirical evidence is tremendously important in shifting opinion and arbitrating between competing theories.

Empirical evidence influenced the rise and demise of monetarism too.

I think you're being a little too pessimistic here.

Certainly no one regression is going to move a literature, but that's why we run lots of regressions on lots of data!

I was referring to specific econometric studies, or even groups of studies, like the many regression run on the effectiveness of TARP.

You're talking about the experience of reality and crises. Yes, some people change their minds when crisis makes it too obvious that their models are wrong, but it takes a major, prolonged crisis. I don't see the latest crisis changing the economics of anyone but young people.

The long stagflation crisis of the 70's changed a few people's minds.

I've often said that many Austrians are Keynesians who were mugged by reality.

Only changed a few people's minds?

Can you name me one person - a single person - after the 1970s who thought it was OK to use a Phillip's curve that did not have expectations augmentation built in?

But I'm not just talking about that. I'm talking about the econometric studies on intergenerational altruism. I'm talking about the econometric studies on real wage cyclicality. I'm talking about intergenerational studies on rational expectations and lagged market information. These aren't learning from crises! These are exactly the sorts of studies you were saying doesn't get people to change their minds.

I don't know any of the literature on TARP so I couldn't say. What do you have in mind there?

I find your last line interesting, since the reality of the last decade seems to have bolstered the Keynesian camp. I'm not saying it hasn't bolstered the Austrian camp too, but it seems odd to think that Keynesians moved toward the Austrian school. Got any examples?

Daniel, you're talking about minor points within the same camp. People change their minds often on minor points. I'm talking about changing camps.

On TARP is was thinking of the debates between the Taylor and Romer camps.

The latest crisis brought the paleo-Keynesians into the sun for a while because the neo-classicals had nothing to say. They had grown bored of business cycles in the 90's. And they were completely surprised by the severity of the recent crisis. But I didn't see anyone jumping camps because of it.

When the Queen visited the LSE and asked why they had missed predicting such a severe crisis, they wrote back to her that such crises are random events, so they successfully predicted that no one could predict it. Then they went back to sleep.

PS, Robert Higg's and Deirdre McCloskey are two that come to mind immediately as examples of converts through mugging by reality.

Minor points? Same camp? These are the defining macro debates of the 20th century between the major opposing camps of 20th century macro, McKinney!!

I don't know what a paleo-Keynesian is.

The Queen's question was pretty ignorant in my opinion (I thought this was the sort of thing that got us to elect Parliaments and Congresses in the first place).

It's pretty sad if those were the defining debates of the entire century!

You think it an ignorant to question why the UK's top economists were blind sided by the greatest depression since 1929? What good is economics if it can't see an economic meteor the size of Mars headed for the planet?

I can guarantee you people would be upset if astronomers didn't see an asteroid headed this way.

PS, perhaps if mainstream macro spent a little more time on business cycles and less on intergenerational altruism they could have seen the disaster coming.

Very good, I would like to learn from you and continue to work hard.

What McKinney said.

"it seems odd to think that Keynesians moved toward the Austrian school. Got any examples?"

Mark Blaug late in life became much more hospitable to the Austrians, especially Hayek but he drew the line at von Mises. I suppose he was spooked by strong apriorism which added no value to Austrian economics. You could say we are paying a bitter price for the distraction that piece of methodology created, especially when you see how dogmatic some of the von Mises people are about it.

The ignorance of the Queen would appear to be dwarfed by the failure of the economists. At least she asked the right question. The problem with demand analysis is that it attempts to answer the wrong questions, and the rest is history.

As we speak another Titanic (aka QE II) is being launched. Best of British luck to all who sail...

Rafe, thanks but I think mainstream econ isn't nearly as empirical as it claims to be, especially in macro. It's easier to be empirical in micro. But most macro is built on Keynes and Keynes never provided any empirical evidence for his nonsense.

There is little empirical evidence that any Keynesian policies have worked. Sure, there are models that use circular reasoning, as Kling has written. They assume a Keynesian model and then, lo and behold, the model works!

Mainstream econ could not be less empirical when it completely ignores the heterogeneity of capital and the destructive impact of money pumping on that capital structure.

Mainstream econ accepts a lot of things a priori but refuses to admit it while being dogmatic about empiricism.

The problem with mainstreams dogmatism on empiricism is that it forces them to do research on the available data and that leads them to trivial results. I think that is one reason behavioral econ is so popular. It's easy to set up controlled experiments on trivial things.

Look at the popularity of Freekanomics before the crash. It was mostly trivial or non-economic stuff. Mainstream econ has painted itself into a corner with its obsession over a very limited idea of empirical research that includes nothing but regression on available data. That's why they were totally blind sided by the crisis.

@Rafe -
Ha, OK, I guess I wasn't thinking of being hospitable to Austrians when he said that! Yes, I am a strong proponent of hospitability toward Austrians myself!

And my whole concern is that she is asking the wrong question and so many people seem intent on validating her on it. Economists are not fortune tellers. The right question isn't really whether we could predict it but whether we can understand and deal with it.

McKinney -
re: "I can guarantee you people would be upset if astronomers didn't see an asteroid headed this way."

Economics is not physics. People get in trouble when they think like that.

Did anyone see the mises.org article poking fun at Prof. Horwitz?

Daniel, it was predictable as a result of policies based on Keynesian macro plus other dirigist and interventionist policies and it was predicted circa 1999, it was just uncertain when it was going to actually happen.

Sure enough, on cue, strong apriorism turned up on Mises.org.

The US has enjoyed about 60 business cycles since 1790, an average of one every four years. If economists did nothing but tell people to look for a recession every four years they would be right at least once in a while. As it is, mainstream econ has never predicted a recession.

Now hold on - we're garbling a bunch of things here. People did predict recessions, McKinney. What was less predicted was a recession of this magnitude. And that's because forecasting isn't fortune telling and nobody should expect it to be and macro forecasters aren't privy to the transactions conducted in the bowels of investment banks.

Sorry, no miracles.

As I emphasized earlier - economics is not physics. It's more like biology.

Can we understand what went on? Yes - we have a good sense of it.

Can we propose good responses? Yes - unfortunately politicians don't seem interested.

Other than that forecasts are going to be short-term and OK-guides, not miracles. There are too many forces in play for what the Queen was asking for, and validating her question is tantamount to overstating the abilities of economists.

Daniel, other than Austrians, no one saw even a recession coming. Evidence is in the fact that so many people are complaining about it, even some economists.

Yes, economics is more like biology than physics, but so what? Do you think biologists can't predict anything?

How do you know the response is good when you don't know what caused the problem? No medical doctor would start prescribing medicine for an illness he didn't understand, but mainstream econ has no problem with that.

We are now embarking on QEIII after the massive failure of the first two plus the largest Keynesian stimulus in peacetime history. The failure is obvious to everyone but mainstream economists. But what is the response from mainstream econ? It could have been worse without the QEx's and multiple stimuli.

How do they know? We might have been better off without QEx's and stimuli. They make an unfounded assumption that such interventions work and then declare them a success with almost no evidence.

The history of recessions proves that economies recover relatively quickly (compared to recessions since 1929) when left alone. With that fact under our belts, it's not hard to assume that any recovery following stimuli or QE's is due more to the economies natural ability to recover than to the stimuli or the QE.

Mainstream econ reminds me of a guy in Times Square blowing a trumpet. A policeman stopped him and asked him what he was doing and the guy said I'm keeping elephants away. The policeman said there are no elephants in Time Square. The guy replied "I'm doing a good job, then."

The comments to this entry are closed.