George A. Selgin's critique of praxeology

Economics Economics Questions
1

George Selgin has posted a critique of praxeology at Cato Unbound:

Here I would like to bring to light an important omission in George’s narrative, an omission that is common to many contemporary discussions about Mises and praxeology.

As George understands things, Mises advocated something he called ‘praxeology’ and insisted that this ‘praxeology’ is “a priori.” Rothbard then came along, defending the a prioristic nature of this ‘praxeology,’ and using it himself in support of various economic policy conclusions. Praxeology begins with Mises and ends with Rothbard. This is the picture George presents. Unfortunately, this conception of things is not historically (George might say “empirically”) accurate.

To understand why, we need only consider some elementary but very important passages from Carl Menger’s Investigations into the Method of the Social Sciences. Here are two:

  1. The types and typical relationships (the laws) of the world of phenomena are not equally strict in all cases. A glance at the theoretical sciences teaches us rather that the regularities in the coexistence and in the succession of phenomena are in part without exception; indeed they are such that the possibility of an exception seems quite out of the question. However, some are such that they do indeed exhibit exceptions, or that in their case exceptions seem possible. The first are called laws of nature, the latter empirical laws.

(Menger clarifies in a separate passage that by “laws of nature” he means “exact laws.”)

  1. The realistic-empirical orientation of theoretical research, as we saw, offers us in all realms of the world of phenomena results which are formally imperfect, however important and valuable they may be for human knowledge and practical life. They are theories which give us only a deficient understanding of the phenomena, only an uncertain prediction of them, and by no means an assured control of them. From the very beginning, too, the human mind has followed another orientation of theoretical research beside the one discussed above. It is different from the latter both in its aims and in its approaches to cognition.

The aim of this orientation, which in the future we will call the exact one, an aim which research pursues in the same way in all realms of the world of phenomena, is the determination of strict laws of phenomena, of regularities in the succession of phenomena which do not present themselves to us as absolute, but which in respect to the approaches to cognition by which we attain to them simply bear within themselves the guarantee of absoluteness. It is the determination of laws of phenomena which commonly are called “laws of nature,” but more correctly should be designated by the expression “exact laws.”

There can be no doubt that the theoretical orientation Menger refers to as the “exact one” is the same thing Mises refers to as “praxeology.” In the context of Austrian economics and Austrian theoretical social science, praxeology does not begin with Mises; it begins with Menger. Mises is simply continuing the “exact” orientation of theoretical research, while giving it a different name, and providing his own understanding of the nature of this orientation.

One wonders why critics of Mises and of praxeology only want to critique them from the Rothbard/Apriorism point of view and not from the Menger/Exact Law point of view. In my opinion, anyone who wants to advance a serious critique of praxeology needs to include, if not begin with, a critique of Menger’s writings on the subject of exact laws. If Menger’s notion of ‘exact laws’ is defective in important ways, what is needed for an educational discussion is an explicit statement to this effect including an explanation of the reasons Menger’s vision is defective and should be abandoned.

The critique of Rothbardian apriorism is by no means a substitute for a critique of Menger’s concept of exact laws. Any knowledgeable Austrian knows that Rothbard’s primary concern and passion was in the ethical and moral field. It is natural then to seek to demonstrate how Rothbard tried to support his normative conclusions with “aprioristic reasoning.” Lo and behold, this is what George does in his article:

“Yet the theoretical framework that takes up the book’s opening chapters is, as Rothbard himself indicates, one which supports “[t]he Austrian policy of refraining at all times from monetary inflation” (my emphasis). Thus Rothbard the Austrian praxeologist is led, by a chain of deductive reasoning starting from the action axiom, to deny the potential utility of monetary expansion even under circumstances which, according to Rothbard the historian, made severe unemployment inevitable in the absence of such expansion.”

“Praxeology” begins with Mises. Rothbard then tries to use praxeology to support his normative conclusions that money and credit ought not be expanded, and this concludes the “praxeological” chapter of economic history. This is the picture George presents.

Here is an alternate explanation:

As the passages from Menger indicate, Menger conceives that there are two types of laws or regularities; empirical and exact.

When I walk toward a given location, I usually arrive there. This is an example of an empirical regularity. If I attempt to walk to a location tomorrow, I may arrive there but I may not.

When I walk toward a given location, I always walk away from one. This is an example of an exact law. If I walk to a location tomorrow, I know I will also walk away from one.

The fundamental concern of Menger and Mises is the relationship between something an actor may do or observe, X, (and observing is an instance of doing), and something else Y, that will follow or will be coexistent with X. Theoretical exact science (Menger’s term for praxeology) is only concerned with that aspect of “human phenomena” (human action) for which the relationship between X and Y is “exact,” “absolute,” “necessary,” or “apodictically certain.” Relationships between X and Y that are not exact or certain define the boundary between praxeology or exact science on the one hand, and the empirical or realistic approach on the other hand. Praxeology is only concerned with relationships that are necessary because the relevant supposition X, “simply bears within itself” (i.e., logically entails) Y.

“If the “given situation” conforms to a certain pattern, certain other features must also be present, for their presence is “deducible” from the pattern originally postulated. The analytic method is simply a way of discovering the necessary consequences of complex collocations of facts—consequences whose counterpart in reality is not so immediately discernible as the counter part of the original postulates. It is an instrument for “shaking out” all the implications of given suppositions. Granted the correspondence of its original assumptions and the facts, its conclusions are inevitable and inescapable.” (Lionel Robbins, An Essay on the Nature & Significance of Economic Science)

As Menger sees things, empirical laws and empirical research may be important, but they are only one part, or one side of the phenomena we experience:

“Thus we arrive at a series of sciences which teach us strict types and typical relationships (exact laws) of phenomena, and indeed not only in respect to their nature, but also to their measure. We attain to sciences, no single one of which teaches us to understand full empirical reality, but only particular sides, and therefore must not be judged rationally from the point of view of one-sided empirical realism. But the totality of these sciences conveys to us an understanding of the real world which is just as distinctive as it is profound.”

“Not just any one theory of human phenomena, only the totality of such theories, when they are once pursued, will reveal to us in combination with the results of the realistic orientation of theoretical research the deepest theoretical understanding attainable by the human mind of social phenomena in their full empirical reality.”

“Realism in theoretical research is not something higher than exact orientation, but something different.”

(chapters 4 & 5)

The distinction between the empirical and the exact approach, and the advocacy of praxeology as an exact science that stands in relation to human phenomena as mathematics does to physical phenomena, does not begin with Mises. Nor does it end with Rothbard and his ethical concerns.

An earnest critique of praxeology and the distinction between exact and empirical laws would have to begin with Menger, since Mises is largely extending and revising Menger’s vision of the exact orientation. One would have to show the sense in which Menger is wrong in conceiving exact laws, and how therefore Mises headed down the wrong path in trying to extend and revise the exact orientation.

Instead, George critiques praxeology and Mises via Rothbard’s ethics.

This may be effective in the short term. But it does nothing to address the more substantive issues that are present in the work of Menger, and following him, Mises.

2

  1. Yep, Rothbard makes a good heavy.

  2. I like this line “…self-styled Austrian economists, or even of those possessing bona fide academic status”.

So you can’t be anything but a self styled Austrian until you get bona fide academic status.

Henry Hazlitt was thus merely a self styled Austrian, as opposed to the bona fide George Selgin.

By the same standard, Euclid was a mere self styled Geometer, as opposed to all present day grad students in any university.

This is the atitude typical of small minds. Great minds are more interested in the idea itself, as opposed to the bona fide academic status of the person. George Selgin may find it instructive to read about Ramanujan, and Hardy’s response when first reading his letters. Oddly enough both were Selgin’s intellectual superiors by far [proof: Selgin believes in econometrics], are respected in the intellectual community in a way Selgin would die for, and yet their atitude is exactly the opposite of Selgin’s.

  1. I like his distinction between theoretical economics and applied economics, and think he got that part right. Unlike him, I would place econometrics in a third category, the one where alchemy and voodoo and witchdoctory belong.

  2. Selgin thinks that he is one the few who understands that economics is supposed to tell you something about the real world. Take a look at my humble blog, George, and you will see this “self styled” internet Austrian showing how AE predicts the death of the Euro, with Mises predicting it 50 years before the Euro even existed. You will also see my humble prediction about the death of bitcoin, based on Mises regression theorem.

You can’t get more self styled than the Wild West, Biker bar types on the mises.org forum [as Bob Murphy so graciously complimented us]. And it’s chock full of discussions about the real world.

  1. OK, here’s the hatchet job on Rothbard:

Yet the theoretical framework that takes up the book’s opening chapters is, as Rothbard himself indicates, one which supports “[t]he Austrian policy of refraining at all times from monetary inflation” (my emphasis). Thus Rothbard the Austrian praxeologist is led, by a chain of deductive reasoning starting from the action axiom, to deny the potential utility of monetary expansion even under circumstances which, according to Rothbard the historian, made severe unemployment inevitable in the absence of such expansion.

George, my good man, this is the kind of muddle headed thinking you fall prey to because you have contempt for that “self styled Austrian” with no bona fide academic status, the great Henry Hazlitt. [See Failure of the New Economics, Chapter 26]. George, my son, have you heard, perhaps in a dream, perhaps in a vision, that there are two kinds of jobs, productive jobs and parasitic jobs? Do you know the difference between them? Do you grasp that one is good for the economy, and one is bad? [I leave it to your econometrics researches to figure out which is the good one and which is the bad one]. Do you know how one can derive from first principles that jobs created by monetary expansion are all parasitic?

If you answered yes to all these questions, then shame on you for critcizing Rothbard when you should know better. If you answered no to any of them, then shame on you for criticizing Rothbard when you obviously have no clue about the very basics of economics.

  1. The attack on Rothbard continues:

But is the fact that, when credit expansion takes place, interest rates are lower than they would otherwise be, really the only “important” thing? Allowing, as Rothbard would certainly allow, that the Austrian boom-bust theory is only applicable to circumstances in which credit expansion has indeed taken place, and that the Austrian theory is logically valid, does the mere fact that credit expansion did indeed take place during the latter 1920s suffice to establish that the theory “explains” the Great Depression, either wholly or in part? If we cannot, as Rothbard insists, employ statistics to tell us how much difference the expansion made, then how can we know that the difference it made wasn’t trivial, and that the true causes of the depression must consequently be sought elsewhere? How, in other words, can we tell that the theory is not just logically valid but actually useful in explaining any particular episode for which it might be relevant?

Time is short, so I’ll cut to the chase. Thou art barking up the wrong tree, the interest rate tree. What counts is not the interest rate, but the amount of increase of the money supply. Rothbard goes on and on, in boring page after boring page [All of Chapter 4], to establish beyond any doubt that the increase in money supply was HUGE in the twenties, by any metric.

So now your argument is this:

I. AE predicts that huge increase in money supply will create a businees cycle. It explains logically, step by step, why this must happen. No one has ever refuted the logical chain of reasoning employed.

II. In the twenties, there was a huge increase in the money supply, and a business cycle happened, just as AE predicted.

III. But maybe it’s a coincidence.

Is that what they teach those with bona fide academic status in George Mason U? Because at UC Berkeley, in the Department of Math and Logic, when they granted me bona fide academic status in those areas, they taught me that such reasoning is more to be pitied than censured.

  1. Now comes a real howler:

The praxeological—which is to say, Misesian and therefore “Austrian”—view of empirics suffers from its implicit assumption that quantities are irrelevant, not merely for constructing economic theories, but for doing applied economics.

What an absurd thing to have the gall to say! And right after you read AGD. which goes on and on and on [in Chapter 4] about, you guessed it, quantities. That should have given you some clue. I mean, it’s hard to declare to the world ones ignorance of AE in a clearer fashion than that stupid statement.

But I will grant you one thing. You admitted you never found any Austrian whom you can actually quote as saying such nonsense. You admitted that it’s “implicit”, a polite way of saying “a figment of my, George Selgin’s, imagination.”

  1. George then defends econometrics by saying that there are important questions it can answer. That econometrics is not as useless as mises made it out to be.

By how much were interest rates driven below their “natural” levels as a result of the open-market purchases and discount rate reductions that took place between 1922 and 1928? How much additional investment activity can be attributed to the difference? How great was the substitution of more capital intensive or “roundabout” investment activities for less roundabout ones? (How, indeed, might one quantify roundaboutness?) How, finally, does the scale of consequences statistically attributable to the mechanism described by the Austrian theory compare to that attributable to, say, the monetarist theory, which blames the depression on monetary contraction?

Yes, it certainly would be nice to know all these things. But it can’t be done. Tell us all the simplifying assumptions you make when you do your econometrics, so we can get a good laugh.

BTW, that last line of his shows great ignorance of AE. According to AE, the bust always starts with a monetary contraction, just like the monetarists say. That is not an issue where monetarists and Austrians differ.

  1. He then makes the absurd statement that statitistics can disprove logically valid statements. In other words, statistics can disprove the Pythagorean theorem. goodluck with that one.

  2. Finally, he shows he misunderstood McCloskey. But that’s a whole 'nother kettle of fish.

3

That’s not a very good critique.

4

Most of Selgin’s confusions would be dismissed were he to read and absorb this chapter of Ultimate Foundations. The problem is that he doesn’t understand the dividing line between praxeology (the study of acting beings, whether their specific historical acts or the general laws governing their natures) and natural science.

The problem with statistical studies of historical human behavior is that you can never get from there to general laws of human nature. As acting beings, the telic behavior of humans only makes sense in a thymological framework, that is, by asking answering the questions, “Why? To what end? For what gain?”

A proton does not move in order to gain something, it moves, mechanically, in response to electromagnetic stimuli. And there is a lot of confusion over the role of causality in relating the behavior of protons and the behavior of the human being. If, after all, our motions are fully determined by physical laws in exactly the same sense that the motion of a proton is fully determined by physical laws, then we are mechanical, too.

On any theory of the world that preserves causality, this must be true but it misses the point. The human brain cannot analyze and comprehend fantastically complex machines - such as living things - at the same level of analysis at which we can analyze and comprehend a proton. This should go without saying. There is an out-of-favor interpretation of quantum randomness called the “hidden variables” interpretation, which posits that quantum randomness does not violate causality but, rather, that there are “hidden variables” that are so fantastically complex as to make the behavior of a quantum particle appear random. This is the sense in which human behavior is non-deterministic. The variables that go into determining human behavior are so fantastically complex that human behavior might as well be non-deterministic*.

The way I like to illustrate this is to show how the second law of thermodynamics interacts with the economy. We can think of the entire biosphere as a gigantic machine. This machine is, ultimately, converting energy incident the Earth from the Sun and energy from the Earth’s core and the orbit of the Moon into mechanical work. The biosphere is a gigantic heat engine. As such, it is subject to the second law of thermodynamics, the same as any engine. Hence, we can state - with the same certainty that a physicist can state a law regarding the motion of a proton - that the biosphere is increasing the entropy of the troposphere, the oceans and the Earth’s surface, insomuch as we can think of these as a closed system (which is appropriate since Earth is not ejecting high entropy waste out into space). Since humans are a part of the biosphere and since the economy is part of human behavior, we see that pollution is the inevitable byproduct of any economic activity by the laws of physics. (It also shows how artificial the “green” distinction is.)

Hence, we can apply the laws of physics directly to human behavior but only at the appropriate level of analysis. In order to apply the second law of thermodynamics to human behavior, you have to “zoom out” - so to speak - on human beings. You can learn absolutely nothing about the stock market from the study of thermodynamics, except to say that it must be contributing to the entropy of the environment, as well. But what Selgin and other economists talk about (using “regression” to derive economic laws) is like trying to derive the laws of thermodynamics from observing the stock-market. There’s just no connection. Yes, the stock market is subject to the laws of thermodynamics but so what? This can’t tell us anything interesting because it’s simply not the appropriate level of analysis.

Finally, to put the final nail in the coffin of this idea of deriving economic laws from regression analysis, we need look no further than game theory. We know from game theory that statistical methods simply are not suitable for modeling the behavior of acting agents. As soon as you introduce an acting agent, the probability distributions are not what is called in mathematics “stationary”, i.e. eventually converging to a definite probability distribution. Hence, if an agent in a game naively applies probabilistic methods against opposing agents, he will be defeated because they will be able to learn what probability distributions he is imputing to them, incorporate that into their strategy, then make moves which the naive agent will guess wrong every time.

In summary: The human brain is not a gas. It doesn’t act like a gas, it doesn’t look like a gas and it can’t be modeled like a gas. We even know why it’s not like a gas, because it’s actually a neural-net computing device. The generalized behavior of a computer can’t be modeled by probabilitistic methods because the long-run behavior of a computer does not give rise to a stationary probability distribution. This is why the brain cannot be modeled by probabilistic methods. In addition, as agents in a “live game” called life, naive application of statistical methods to modeling the behavior of human beings guarantees failure.

Clayton -

*Dennett has a lecture on this topic and he analyzes human free-agency from what is (unwittingly) a praxeological perspective - the use of free-agency is to force others to “guess” our motives and, thus, force them to treat us as “acting agents”, that is, peers. That’s why we can conceal out motives through language and body gestures.

5

Smiling Dave, I hope I’m not the only one here who is capable of concluding, from your item #2, that you are not a very good reader. I also hope others here will have the decency to actually read my essay, as well as Steve Horwitz’s, before they make the mistake of believing any of the other claims on your list–that is, the few that aren’t self-evidently foolish.

6

Smiling Dave…

Oh, I read them both, rest assured. Very amusing.

In any case, I stand by all I wrote, including statement 2.

Let the readers decide.

7

Adam Knott neglects one not minor detail: my assignment was, not to review Austrian perspectives on empirics generally, but to respond to Steve Horwitz’s essay, which was itsellf almost entirely devoted to the claim that praxeology, and praxeology as Mises understood it, is not anti-empirical. Those who have read Steve’s piece can readily affirm this.

8

That suits me, SD: I would not think to insult the others by explaining English syntax to them.

9

Hi George,

This is not related to the thread, but now that you’re here, I was wondering if you as a free banker could comment on the issue of Mises’ regression theorem. I’m guessing that you stand by it, and that you would thus reject Menger on money?

See: http://nielsio.tumblr.com/post/25583537960/menger-versus-mises-and-rothbard-on-how-money-works .

10

+1 Clayton. An interesting point about determinism and free will.

11

Hi Nielsio. I am a bit puzzled by your question, as I do not at all think Mises regression theorem inconsistent with Menger’s. On the contrary: I believe it perfectly consistent. Some of my thoughts concerning it are in my JMCB paper “On ensuring the acceptability of a new fiat money.” I confess, though, that my views there are challened by certain episodes, including the Massachussetts colonial paper money and bitcoin.

12

Massachets paper money had a declared value, and was accepted in payment of taxes.

Bitcoin is not generally accepted.

My humble blog [and especially the comments] spells it all out: http://smilingdavesblog.wordpress.com/2012/08/03/bitcoin-all-in-one-place/

13

To the OP.

Selgin’s attack on Rothbard has nothing to do with ethics. In the very excerpt you quoted, he writes about “the potential utility” of printing money that Rothbard denies. Meaning even if we all agree on the ethics, Rothbard is saying printing money will not help the economy as a whole, and Selgin thinks it will.

If he’s lurking here still, you could ask him.

14

Clayton, nowhere in my article to I suggest that I disagree with Mises view that you can’t derive “economic laws” using regressions. Like SmilingDave and Adam Knott, you seem to have done a poor job of reading what I wrote, assuming that you actually read it at all. Here for the record is the relevant passage from my article:

“there is something between merely checking to see whether an occurrence satisfies all of the “contingent claims” needed to apply a particular theory to it, and pretending that you can construct economic theories using regression coefficients. What’s in between is trying to arrive at an informed estimate of just how much of any observed phenomenon an applicable theory explains and, when there are several equally applicable theories, their relative worth.”

Do you see that word “pretending”? Do you not understand that that is agreeing with Mises on the matter of not being able to use regressions to construct theoretical laws?

I could go on with many other examples, but the main point I wish to make, to those following this forum who are innocent of the ways of certain “hard-core” Rothbardo-Misesians, is that they should heed the principle of caveat emptor, and note that much of what is passed of as fair criticism here is just plain lying.

P.S. to SmilinDave: I’m a great Hazlitt fan, and I never disparaged non-academic Austrians as such. Had you really read Horwitz (I do not believe your claim to have done so) you’d know why I referred to “bona fide” academics.

15

P.S. to SmilinDave: I’m a great Hazlitt fan, and I never disparaged non-academic Austrians as such. Had you really read Horwitz (I do not believe your claim to have done so) you’d know why I referred to “bona fide” academics.

In that case, apology accepted.

And, as Ripley is my witness, I read that Hurwitz article.

16

That’s Horwitz.

17

…we’re not talking about this Horwitz essay, are we? Cuz that would be kind of funny.

18

Yep, that’s the one.

The underlying tone is “Mainstream, don’t laugh at us. We are the good guys. We know math, too.”

19

Smiling Dave is exactly why we can’t have nice things.

20

Easy SD. You don’t want another month. Just something to consider.