I cannot help sympathizing with Steve Horwitz’s desire to defend Austrian economics generally, and what Mises called “praxeology” in particular, from its detractors, and particularly from those of its detractors who insist, as Josh Barro does, “that Austrian economists reject empirical analysis, and instead believe that you can reach conclusions about correct economic policies from a priori principles.” Years ago I myself wrote a defense of praxeology, or rather of Mises’s views regarding the scope and method of praxeology; and although my essay’s intended targets were themselves members of the Austrian school, rather than outside critics of that school, I also believed that Mises had been misunderstood by those who took him to be saying that economists need never trouble themselves with history generally and hence with “empirical analysis.”[1]
Today, though, notwithstanding Steve’s counter-arguments, I’m afraid that I’m inclined to think that Mr. Barro has a point. He 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.
In defending Mises himself, and what Mises called praxeology, against the charge of being “anti-empirical,” Steve observes that for Mises, “‘praxeology’ was not intended to be a ‘method’ for economists,” and that while “[s]ome Austrians argue as if one can deduce all of economics in one’s armchair …Mises was pretty clear that this core of economics was fairly limited.” But while it is true that Mises generally employed the term “praxeology” to refer to a field of study rather than a method or set of methods, he also wrote things like
Praxeology is a priori. It starts from the a priori category of action and develops out of it all that it contains.
and
Praxeology is a priori. All its theorems are products of deductive reasoning that starts from the category of action…. What praxeology asserts with regard to human action in general is strictly valid without any exception for every action…. Every theorem of praxeology is deduced by logical reasoning from the category of action. It partakes of the apodictic certainty provided by logical reasoning that starts from an a priori category.
and
They [praxeologists] refer to experience only in order to separate those problems that are of interest for the study of man as he really is and acts from other problems that offer a merely academic interest.[2]
It is, for me at least, impossible to read such statements—and one can find many more like them—as not suggesting, first of all, that the term “praxeology” denotes not merely a field or fields of study but the one and only correct method for arriving at, not just a small part of, but “all” of the “valid” conclusions that the study is capable of yielding, and, second, that this method consists of deductive reasoning from a priori premises. It is therefore not simply owing to careless misreading of Mises that Murray Rothbard could write a highly influential essay titled “Praxeology: The Methodology of Austrian Economics,” devoted to arguing that praxeology consists of nothing other than “the logical implications of the universal formal fact that people act, that they employ means to try to attain chosen end.”[3]
Does this mean that Steve himself misunderstands Mises in insisting that his view of economics was not “radically anti-empirical”? It doesn’t, because at bottom the difference between what Steve is saying and what Mises appears to be saying boils down to a matter of semantics. For Mises and Rothbard, and for many other Austrian economists, the term “economics” means what is elsewhere referred to as theoretical economics, at least when they use it in the course of a methodological pronunciamiento. That is, it means (as Mises indicates in one of the passages quoted above) the set of extant economic theorems. Understood this way, the claim that “economics” is a purely deductive undertaking ought not to strike even the most mainstream of economists, or anyone who has worked through a standard graduate microeconomics text, as particularly controversial, let alone absurd.
But for most of us, Steve included, “economics” doesn’t just mean pure economic theory or analytical economics. It also means applied economics, which includes everything from economic history to economic policy appraisal to econometrics. In other words, it includes all the economics into which empirics—the facts of experience—enter. So when Steve says that Mises wasn’t anti-empirical, he means that Mises did not think empirics irrelevant for what we call applied economics, whereas when Rothbard defends Mises’s claim that economics is a purely deductive science, he means that purely theoretical economics is so. There need not, in other words, be any contradiction between the two understandings of the nature of economics, Austrian or otherwise.
It doesn’t follow, though, that an overwhelming majority of self-styled Austrian economists, or even of those possessing bona fide academic status, understand or write as if they understand that the “economics” Mises and Rothbard have in mind when speaking of economics as a body of deductions from a priori premises refers only to the “core” of the discipline and not to the broader enterprise of applied economics. Steve, it is true, supplies several examples of research that he claims uses “Austrian theory to offer better explanations of real-world phenomena,” and he could no doubt supply many others. Still he hasn’t convinced me that Austrian economics as such is not guilty of being at least relatively, if not resolutely, anti-empirical.
For one thing, I note that almost all of the examples he supplies of empirical Austrian economics come from George Mason graduates who, like himself, were heavily influenced, directly or indirectly, by the work of the late Don Lavoie. (The one exception—the article on the Fed written by myself, Bill Lastrapes, and Larry White is at most only one-third Austrian.) But just how “Austrian” are these examples? 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.
Discerning those broader tendencies requires, for one thing, that one consider the large body of writings by those Austrians who with some justice claim to be the strictest of all adherents to Mises’s methodological views—writings that are by no means confined to Internet forums. This includes work by Israel Kirzner, by Murray Rothbard, and finally by those Austrians whose books are published by the Ludwig von Mises Institute or who publish in its Quarterly Review of Austrian Economics.
Although I do not hesitate to include myself among the admirers of Professor Kirzner’s writings, and of his writings on entrepreneurship in particular, I do not think it can be denied that those writings are, to put it mildly, short on empirics, by which I mean not simply short on regressions and that sort of thing but short even on historical illustrations. Roger Koppl, himself a highly regarded Austrian school economist, finds that this has generally been true of Austrian writings on entrepreneurship—a major area of Austrian research:
Austrian economists have probably devoted too little effort to empirical work in the past. This trend has changed radically in the last few years…. Nevertheless, Austrian economics is still not very well endowed with empirical findings. [Other S]cholars of entrepreneurship, by contrast, have been very energetic in confronting theory with the facts of history.[4]
Murray Rothbard’s case is particularly revealing. Although he was an uncompromising praxeologist, far from thinking “empirics” unimportant, he wrote extensively on economic history. Yet even Rothbard’s historical writings, or some of them at any rate, only serve to underscore the often aggressively anti-empirical nature of his economics. America’s Great Depression, for example, is largely devoted to showing how the Hoover administration contributed to the Great Depression by persuading businessmen to resist cutting prices. 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.
Both Rothbard and too many other Austrian economists to count are also led to conclude, praxeologically as it were, that despite the evidently depressing effects of monetary contraction when combined with policies aimed at preventing prices from falling, the Great Depression was mainly the inevitable result, not of such monetary contraction after 1929, but of excessive monetary expansion beforehand. Empirics, and statistics especially, are neither needed to reach this conclusion nor capable of refuting it:
[S]tatistics can prove nothing because they reflect the operation of numerous causal forces. To “refute” the Austrian theory of the inception of the boom because interest rates might not have been lowered in a certain instance, for example, is beside the mark. It simply means that other forces—perhaps an increase in risk, perhaps expectation of rising prices—were strong enough to raise interest rates. But the Austrian analysis of the business cycle continues to operate regardless of the effects of other forces. For the important thing is that interest rates are lower than they would have been without the credit expansion. From theoretical analysis we know that this is the effect of every credit expansion by the banks; but statistically we are helpless—we cannot use statistics to estimate what the interest rate would have been. Statistics can only record past events; they cannot describe possible but unrealized events.[5]
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?
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. Empirics then serve only to determine which theories are applicable to which events:
Into the chain of praxeological reasoning the praxeologist introduces certain assumptions concerning the conditions of the environment in which an action takes place. Then he tries to find out how these special conditions affect the result to which his reasoning must lead. The question whether or not the real conditions of the external world correspond to these assumptions is to be answered by experience. But if the answer is in the affirmative, all the conclusions drawn by logically correct praxeological reasoning strictly describe what is going on in reality.[6]
As for statistics, “As far as there is discernible regularity in the succession of phenomena, no recourse to [them] is needed” either for constructing theories or for determining their applicability to particular historical instances. Those who imagine otherwise overlook the fact that there are no such things as “statistical laws” (with their implied, underlying numerical constants) governing economic events.
But Mises’s belittlement of statistics, and econometrics especially, as mere “childish play with figures that does not contribute anything to the elucidation of the problems of economic reality,”[7] true as it may be for a very large share of econometric research, overlooks the fact that 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. 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? To be sure, some self-styled Austrians have attempted to answer these and related questions. But in doing so, they can hardly be said to strictly heed Mises’s strictures concerning the limited relevance of statistics and econometrics for elucidating economic reality. And who will not suppose that, had it not been for those strictures, many more such attempts—and probably some more successful ones—might have been undertaken?
In short, while Mises and other Austrians have a point when they insist that “historical experience… can never falsify any [economic] theorem in the way a laboratory experiment can do with regard to the statements of the natural sciences,” 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. And it is precisely when it comes to demonstrating the explanatory oomph of their theories that Austrians lag behind other sorts of economists.
With regard to this last observation, I feel compelled to say that I think Steve comes dangerously close to missing the point of Deirdre McCloskey’s “cautions about the significance of statistical significance.” It is true that those cautions have mainly been meant for non-Austrian economists; but that only goes to show that Austrians run relatively few regressions: to judge from the cases I’ve seen, when Austrians do run regressions, they are as likely as any other economists to be guilty of the offense McCloskey cautions against—of confusing statistical with economic significance. But McCloskey’s point isn’t that economic regressions generally lack “scientific power,” and it is still less that they needn’t be employed at all so long as one can otherwise tell a good story. She means rather that magnitudes—many of which can only be estimated by running regressions of some sort—are precisely what matter most for establishing the “power” of a theory or hypothesis.
Notes
[1] Selgin, George. “Praxeology and Understanding: An Analysis of the Controversy in Austrian Economics,” Review of Austrian Economics 1988: 19–58.
[2] The quotations all happen to come from The Ultimate Foundations of Economic Science: An Essay on Method (Princeton, New Jersey: D. Van Nostrand Company, Inc., 1962), pp. 41 and 44. But many others like them can easily be found in Mises’s other works.
[3] Murray N. Rothbard, “Praxeology: the Methodology of Austrian Economics.” In Edwin Dolan, ed., The Foundations of Modern Austrian Economics (Kansas City: Sheed and Ward, 1976), pp. 19–39.
[4] Roger Koppl, “Gains from Trade between Austrian Economics and Entrepreneurial Studies: An Introduction to the Volume.” In idem, ed., Austrian Economics and Entrepreneurial Studies: Advances in Austrian Economics 6 (2003), p. 3.
[5] Murray N. Rothbard, America’s Great Depression, 5th ed. (Auburn, AL: The Ludwig von Mises Institute, 2000), p. 85.
[6] Mises, Ultimate Foundations, p 44.
[7] Ibid., p. 63.