June 2022

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    
Blog powered by Typepad

« The Modified Friedman Dictum | Main | The Future of Austrian Economics as a Progressive Research Program »


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

Right, the subjective account is much more sensible, but psychology is speculative. Furthermore, the kinds of assumptions made in the psychology have been of a universal nature. You can't have extreme a priori positions concerning something so speculative. There is certainly room for disagreement about the nature of man.

I don't know how Mises got dragged off on these tangents. You can call them economics, if you like, but surely not in any Misesian sense of the term. Or, again, is Human Action not a reliable guide to what he really believed?

Mr. Lesvic,

Please re-read the Kirzner quote, and think before you write. Seriously.


This is excellent post on many counts, not the least for the appreciation of Emily's work. Well done and well-deserved praise for her.

Another quote on point also from Hayek's Counter-Revolution: "Unless we can understand what the acting people mean by their actions any attempt to explain them, i.e., to subsume them under rules which connect similar situations with similar actions, are bound to fail."

Much is packed into that quote. First, there is subjectivism. Second, the centrality of actions. And, third, an implicit theory of causality. The linkage with Mises, brought up by Pete, is apparent.

Go into any book store and look in the history section. There you will find many volumes on cultural, social, political and economic history.

Most of them use what is the traditional methodology of historical research. The first assumption is that you are studying some aspect of the human condition. He is concerned with human actions and their outcomes (intended and unintended), and he traces out the meanings of the actors, explains what ends they had in view and the means they attempted to apply.

The traditional historian analyzes the results of men attempting to interact with others for various political, social, and economic reasons.

For this, the historian devotes time to understanding the institutional context in which actions and interactions were occurring. How the institutions restrained, facilitated, or modified what the actors were trying to do. And how the institutional setting was, perhaps, changing through time as a result of the actors were doing.

The traditional economic historian would utilize his theory of economics to serve as an interpretive schema to arrange, orient, and analyze the human activities he is focusing on.

If this traditional economic historian was, say, doing a history of an industry or of a development of a region or country, he would (at least implicitly) explain how and why the division of labor evolved in the way it did, how market forces created various incentives and opportunities (or disadvantages) for the development and industrial evolution he is attempting to explain.

I could go on-and-on. But this is the way a methodological subjectivist (as Hayek was saying in the quote that Peter includes in his comment) would and has traditionally gone about his applied work.

It is the modern, mainstream economist who, still under the influence of 19th century positivism, has turned his back on the study of human action in all of its rich details and interconnections.

If one wishes to read such a traditional economic history as written by a recent writer, I would suggest looking at Tyler Cowen's two books on cultural capitalism and on creative destruction.

Tyler may not want to align himself with the "Austrian" approach. But, in fact, his two books are examples of this method applied.

Another example would be Thomas Sowell's work on culture, race, and social development in different institutional situations around the world, and how they influence the course of human events through time.

When a mainstream economist looks at such works as Cowen's and hSowell's he throws up his hands and says, "But were is the 'model'? Where is the data'?

He just does not see that the interpretive schema that the applied subjectivist economist uses IS is model, and that the relevant "data" is there, but is understood to be more than and deeper than merely a series of statistical aggregates that have been classified and correlated to determine "degrees of significance."

It is the mainstream applied economist who is out-of-step with reality and the richness of the actual empirical world, not the subjectivist or Austrian applied economist.

Richard Ebeling


Pete: FYI Brian and I got that book chapter proposal accepted. The Austrians are making moves in sociology too!

Richard's comment is excellent on all counts, I appreciate his recognition of Tom Sowell's work.

At the 1975 IHS Austrian seminar in Menlo Park, Sowell came over to present a talk on the research that became Knowledge and Decisions. As Roger Garrison observed, it was the Austrian textbook on economics.

For the sake of argument, let me ask this.

Mainstream economists want to use statistical measures and models because if you don't have data, you can't "control" for everything else that may be going on. Yes, that itself assumes that positivist statistical methodology is in a position to do so, but I don't understand how that somehow then justifies the use of forming a historical narrative without that backing.

For example, Professor Ebeling brought up Thomas Sowell's work in race and culture. I would agree that, from an informal and non-economic perspective, his work detailing the relationship between institutions and race, and that certain cultures have exhibited behaviors more highly valued on the market across time and national borders, is very persuasive. The point is, however, that unless you systematically aggregate this data into a regression (or whatever), you are unable to say whether or not the observed behavior was the result of some other factor, or that the number of observations is insufficient to say that the true effect of the variable is greater than zero.

I am in full agreement that believing we CAN control for all relevant and important data is certainly a pretense of knowledge. What I question is whether, even if we have the correct theory on our side, we are able to say anything at all about definite events at definite times. While methodological subjectivism may be fruitful and incredibly useful in dealing with the abstract questions of economic decision-making, I do not see why or how it gets around the concerns of mainstream economics.

The problem here is that the words "subjective" and "meaning" and "interpretation" have multiple unique uses -- and conflating one use with another is the equivalent and making a pun without being aware that you are punning.

I know of no work that has carefully distinguished and cashed out each unique and different sense of "subjective" used in economics, esp. in Austrian economics.

The "subjectivism" of the logic of marginal utility is different than the subjectivism of entrepreneurial "alertness", and social "meaning" subjectivism is yet another different and unique thing.

Conceptual clarity in distinguishing between these is called for -- and maybe newly coined words to keep track of different and unique ideas.


I'm in total agreement with you. This is made worse by the further distinction that needs to be made by those you mentioned and the literal meaning, "pertaining to the subject". This might not be very Austrian of me to say at all, but I think "subjective" in respect to marginal utility or entrepreneurial activity is perhaps better described as "Bayesian".

Subjectivism and other terms have different meanings in different contexts. Subjectivism means different things in epistimology, ontology and methodology. Sloppy authors conflate them, and careful authors don't.

There is no way to do what Greg Ransom wants: an exhaustive detailing of all the meanings of a word without regard to context. It is an attempt to compress connotation into denotation. Language is open-ended as is thought.

Unless Austrians get over this endless carping over the "meaning" of meaning, they will remain in an intellectual backwater. I recommend a jump over to ThinkMarkets for Mario's posting on what is Austrian economics. Let's move on.

Whether Prof Kirzner thinks so or not, economics proper deals not with the motivations of action, but with action itself, asking not why men act as they do, but whether or not their actions are suitable means for the attainment of their ends.

That is not to say that an economist must limit himself to economics proper. He is as free as anyone, and better equipped than most, to explain the phenomena of the real world. But he must not confuse what constantly changes with what never changes, the ephemeral phenomena of the real world with the eternal and immutable laws of economics.

As Mises put it,

“Economics…adopts for the organized presentation of its results a form in which aprioristic theory and the interpretation of historical phenomena are intertwined…this…procedure…has given proof of its expediency. However…uncritical and superficial minds have again and again been led astray by careless confusion of the two epistemologically different methods implied.”

So, let us use them both, but keep them straight.

And, by the way, I'm not so sure that there's any conflict between what Kirzner said and Mises said, and I say.

I would suggest that McCloskey made an important argument in his/her "The Rhetoric of Economics" when he/she pointed out that an "R-square" is not a conclusion but an argument.

It is an interpretive schema that needs to justify itself as reasonable in offering a satisfactory explanation of the subject that is being analyzed.

It is ONE WAY of looking at and organizing a selected group of data, with the presumption that by organizing it and studying it in this way some valuable information and insight is obtained.

For myself, I admit that such analysis can sometimes be suggestive, but I usually do not find it very satisfactory. My mind always tends to think, but what has been left out in terms of the complex "micro" interrelated processes that are the "real" unique events out of which the statistical averages are constructed, and which the aggregates have submerged beneath the statistical "surface"?

On the scatter diagram over which a trend line is superimposed, it is the individual points of real historical moments that represent actual actions, events and interactions. The task of real applied economics and history, in my view, is to attempt to explain how one of those points became another one of those other points in real historical time.
That is what REALLY happens in history.

It is too often forgotten that such things as a "trend line" does not exist. It is a statistical creation superimposed on the field of real unique events based on certain statistical techniques, which are created and chosen because the user believes something of interest or value may be gained by looking at the data in this particular way.

To think otherwise, that the statistical relationships are "really there" rather than simply a way of thinking about the historical events in a certain way, is to suffer from conceptual realism.

Richard Ebeling

Kirzner said,

"Economics has to make the world intelligible in terms of human motives."

That's exactly right, but you're drawing too much from it. The axiom of action, that men act purposively, is the essential starting point of economics, but not its subject matter. We do not have to analyze the axiom of action, for it is a given. It is the theorems that follow from it that we must analyze, and that are the subject matter of economics.

I am not arguing that the statistical R^2 are "really there" or that sufficient data is ever available in such a way that we can legitimately test it. What I mean to say is that the problems statistical analysis attempts (and almost always fails) to do is to separate which effects are real, which belong to something else, and which are indistinguishable from randomness. All caveats of how regression doesn't actually do this and that you need to have some a priori theory I'm not arguing against. But these problems don't go away when we adopt a methodologically subjectivist understanding.

How does Sowell know the fact that certain ethnic groups have succeeded financially isn't a coincidence? He implies that those with a Chinese cultural tradition (for example) have done well for themselves because of it and shows that this is true regardless of whether they ended up in Malaysia, America, or the other places where they constitute a significant ethnic minority. He implies strongly that this is true BECAUSE we've observed this behavior in so many place across so many times. But when does this relationship become historically (?) significant? Sowell draws the line and he expects the reader to as well because it makes intuitive sense. It's a persuasive argument, but it isn't science and ultimately arbitrary.

Theoretically speaking, at least, a crude of a tool as the normal distribution at least gives a starting point for whether or not those observations were a coincidence. I am NOT arguing that the hypothesis testing can actually perform such a function, but without such a theoretical tool, it means we cannot interpret history at all. The failures of logical positivism doesn't give us carte blanche to apply our theories to history without a tremendous deal of skepticism.

I think Richard Ebeling had it about right. The way I see it, if an economic theory can be perfectly formalized, that is great. But the formalization shouldn't be confused with the theory. And if the formalization is imperfect, that should always be kept in mind. The formal model isn't better because it is the formal model. If a formal model has implication regarding statistical data, then testing is great. It would seem to me that if statistical tests suggest the model is flawed, then a careful rethinking of everything from theory on is sensible. The theory could be logically correct but emphasing relatively unimportant things. Or, there is a failure in formalization. Or there is a problem with the data. or, it is that 1% fluke.

To me, what is important, is that economics isn't limited to things that can be tested using statistics. And if no formal model is appropriate to a theory, or formal models don't quite capture the argument, that doesn't mean the argument is unimportant or mistaken.

To me, data mining and pure mathematics are not economics.

The use of supply and demand curves is second nature to me, and I like some simple, partial mathematical models, like money supply and demand. They provide insight to me, anyway. And I very much like looking a macroeconomic data. But still, my understanding of what all this is about comes from Hazlitt and Rothbard, really. And so, indirectly from their teachers.

Perhaps enterpreneurship and creative destruction have been modeled effectively and I just don't know it. But if it hasn't, it still remains central to my understanding of the market economy. Poor formal models of entrepreneurship shouldn't be confused with the logical implications of entrepreneurship. On the other hand, formal models might provide some insight.

I agree with Pete and others in appreciating Emily's field work and her presidential address to the SDAE. She wants us to be rigorous and disciplined in the hermeneutical dimension of economics just as we try to be rigorous and disciplined in the "formal" dimension of our subject. It's hard to see how you could object to that stance, and yet some members of even the SDAE audience seemed resistant.

The formalist school of mathematics emphasized that you need to interpret your formal calculus. In economics such interpretation is interpretive. In one way or another (and there are different ways) it involves one person interpreting the meaning of another person's actions. No matter how "formal" or mathematical you might want to be as an economist, you can't practice your trade without appealing in some way to such interpretive acts. It would thus seem sensible to believe that we should be rigorous in our understanding of understanding, just as we esteem rigor in our our understanding of mathematics. Some enthusiasts for mathematical or statistical rigor and discipline in economics disparage hermeneutical rigor. But when they do, they are saying we should be careful, disciplined, and rigorous in only a part of our efforts. They are saying that we should be loose, sloppy, and inexact in the part of our discipline that links "formal" methods to social applications.

I'd like to point out, for Ryan M's benefit, that not all Austrians take the position being expressed here about historical methods. I think we all agree that econometrics is not an appropriate tool for "testing" theories. But, done correctly, econometric analysis may be useful for doing economic history (what Mises called "thymology"). Mises himself was skeptical about the use of regression analysis in historical investigation, but Rothbard and some modern Austrians have been more sympathetic. I'd characterize Bob Higgs's work, for example, as an excellent example of quantitative, econometric, thymological research. Of course, a good quantitative economic historian (i.e., a good econometrician) does not ignore problems of interpretation; quite the contrary, choices about data, variables, specifications, etc. are fundamentally subjective, interpretive acts. But it doesn't follow that econometrics is per se illegitimate in thymological research. As you point out, it can be a useful complementary tool for teasing out relationships within complex systems, for assessing quantitative magnitudes, and the like.

I met Emily at the SEA conference for the first time and she is really impressive. Since I have only a very limited consumer-knowledge of methods, I like to judge approaches by their outcomes. To the extent that David Sharbek and Emily Schaeffer are 'representative agents' of Emily's approach, I really really like it. Amazingly interesting and productive. Yet, while listening to her (the big Emily), I felt the sentiments expressed by George Selgin (to which Roger seems to refer). Certainly this is because I react allergic against economics-without-theory, which is not what Emily stands for, but what her message seems to allow for in less enlightend circles. Anyway, Emily's approach seems to be less useful for us macro-guys.

Emily's approach is useful for field studies, amv. It is useful when you're doing open-ended interviews, which is something some of us should be doing, especially when dealing with the sort of terra incognata Emily was exploring post-Katrina. She never said everyone else should use only the methods she described in her talk. On the contrary, she explicitly endorsed the notion that we need a portfolio of methods including math and stats. I share Pete Boettke's puzzlement: Why is this so hard?

Speaking of methodological diversity, I would express mild disagreement with Peter Klein on one point. (It is always dangerous to disagree with Peter Klein!) I think econometrics *is* an appropriate tool for testing theories. I don't really buy into simple demarcation criteria. I recognize that theory may test the evidence just as surely as the evidence may test the theory. And I generally recognized that falsification does not work at all like some folks represent in simplistic pseudo-Popperian stories. But testing, with all its ambiguities, is one part of what we do and it generally a good thing when you can contrive a plausible statistical test of your ideas.

This is a canard, Jerry.

There is a clear problem here. "Austrians" conflate the different contexts and connotations of "subjectivism" routinely. And no one is asking anyone to be Carnap -- the suggestion is imitate the practice of Darwinian biologist in their conceptual clarity and sophistication about hard concepts.

The truth of the matter is that conceptual problems at the heart of economic science aren't scientific problems just one for "Austrian economists" -- they are core problems for the science of economics full stop.

The deepest scientific problem with economics as a science is the unwillingness of economists to work to get right the hardest problem of all -- showing how the scientific explanation works in economics, now _economics_ works as an explanatory enterprise.

Economic professionals have demonstrably failed in this regard.

You can't find an economist who can explain how an equilibrium construct plays a coherent and sound role in a causal explanation -- indeed, you can't find an economist who has a coherent and worked out sound idea of how economics provides scientific explanations.

There is no professional payoff or incentive in other words for doing what, say, Ernst Mayr did in Darwinian biology.

And it's significant that Darwinian biologist find getting conceptual distinctions clarify to be a valuable part of their science -- they don't belittle this vital part of science as "mere philosophy".

But then, Darwinian biology is a sound and solid science, rightly without hangups and insecurities about what they are doing.

One reason everyone knows that "macroeconomics" is in massive crisis is because everyone knows that it is a joke as a coherent or even plausible scientific enterprise -- it's a failure on multiple grounds, conceptual, logical, explanatory, predictive, etc. I.e. the public fully right to identify macroeconomists as naked emperors without legitimacy in the "Republic of Science."

If work done to achieve explanatory coherence and conceptual clarity in economics makes one a "backwater" so much for worse for economics as a legitimate member of the "Republic of Science".

Jerry wrote:

"There is no way to do what Greg Ransom wants: an exhaustive detailing of all the meanings of a word without regard to context."

[Strawman alert -- the is the exact opposite of what I suggested. I suggested clarifying meanings by distiguishing between contexts -- if you can't do that in economics, then maybe you don't have meaningful scientific terms.]

Jerry wrote:

"It is an attempt to compress connotation into denotation. Language is open-ended as is thought.

Unless Austrians get over this endless carping over the "meaning" of meaning, they will remain in an intellectual backwater. I recommend a jump over to ThinkMarkets for Mario's posting on what is Austrian economics. Let's move on."

When it comes to laying out the explanatory problem and explanatory strategy of economics -- how economics works as a coherent explanatory enterprise, very few economists have every spent much time working to get this successfully achieved -- economists can't "move on" from this because almost none of them have ever spent a day working on the problems.

Hayek and Mises are memorable and commendable scientists because they worked on this hard problem right at its core -- and they advanced the ball. But Hayek's work is spread all over his hundreds of publications -- collected in no single coherent place, and Mises work is flawed in several significant ways.

But despite all of their helpful special studies on special topics related to this problem, has any Austrian work on this problem in a fashion as broad based and as successful as Hayek and Mises?

In my judgment, no.

Once again, there is nothing to "move on" from here. No "Austrian" has really tackled the problem -- most especially in light of the various advancements of the last 50 years, in all sorts of different areas.

"I recommend a jump over to ThinkMarkets for Mario's posting on what is Austrian economics. Let's move on.""


to the question of 'why it is so hard' to accept methodological pluralism and diversity I answer that it isn't, at least not for me. Isn't my post positive by and large? I also accept your remarks on econometrics. The sentiments I expressed didn't point to Emily. And let me say that I really like the GMU crowd for their openness and their Boettkesque 'look out of the window'. What I see in the profession, however, is that work in pure theory is more and more sourced out to a small group of experts, whose output is used as a toolset for applied economics. That is of course necessary, but the theory choice of applied economists is usually guided by their pragmatic interest and is not advised by what Colander calls 'consumer-knowledge of theory'. I don't worry about theory in Austrian circles. I travel among the so-called 'mainstream' and I see little interest in UNDERSTANDING theory and its limitations.

Greg thinks biology is paradigm of science, and economics falls short. Particularly so, Austrian economics. All economists since Mises and Hayek have fallen short.

Two obvious paths are open to him. He could become a biologist himself. Or he could redirect his energies in economics from constant criticism of what everyone else has done, and produce the great work that others have failed to do.

Emily gave a great talk. I think going further into ethnographic research would be a nice way to go(Pete and I have discussed this since our days with Peter Berger and Dan Rose) but it comes at a huge cost. Ethnography is a more complete way to get into "the native's point of view," but it is less efficient than Emily's approach.

George's point was that Emily said nothing new, and wrapped her language in interpretive philosophy of science. What she is really appealing to, George thought, was good old journalism. But journalists don't work with sophisticated theory, most have no ability to, for example, apply theories about the transmission of knowledge within and across specific cultural groups to something like the events surrounding Hurricane Katrina.

In an age where econometrics almost completely crowds out field research of the kind that Emily and others do, it is sensible to offer a scientific/philosophical defense of her approach. More than that, it is fantastic that she can demonstrate the fruits of her research so clearly and convincingly. In the end, that's what can win the day.

Let me add just a few thoughts:

1. Emily's talk was excellent, in both content and delivery, as Pete has noted. She is among the very best teachers of economics we have as Austrians and her ability to convey ideas eloquently and effectively was on full display Sunday night.

2. As others have said, she most certainly did NOT say her way was the only way. Quantitative work of all sorts is appropriate in the right context for answering particular sorts of questions.

3. What she DID say, and what I took to be the real message from her talk, was that IF we are going to do economics as if humans and their meanings matter, we need to do so *rigorously* by training ourselves and our students to use qualitative methods that have been validated by their use in the other social sciences. Put another way, Emily's talk could be read as the Austrian version of "let's take the con out of econometrics": if we're going to do this work, let's do it right.

And that's why her response to George's suggestion (which, I'll note, even though George may well be reading, was inappropriate in tone and language for a celebratory address - it wasn't an academic seminar) that her talk was not much more than advice on doing "good journalism" is so important. Qualitative social science has a set of methods that has been refined over decades and demonstrated to be effective at answering the questions it is appropriate for. This is not journalism, it is social science, even if economists don't know it or don't want to admit it. A sneer is not an argument against decades of refined methods.

You can't just make up a quick list of survey questions and send out some untrained graduate students to ask them and expect to get reliable results. That's probably not the way to get the data you want and it's BAD qualitative methods. You need to take methodology seriously. (Hence the parallel to taking the con out of econometrics.) As I think Emily would say, doing bad qualitative research is easy, but doing GOOD qualitative is hard and maybe harder than doing good quantitative research.

4. And that segues to an issue that came up Sunday night that still mystifies me: the supposed reluctance of IRBs to approve this sort of work. Frankly, I don't believe it. My long-time teaching partner chaired our IRB for 7 or 8 years or so and I have a pretty good sense of how ours worked. I can tell you this much: If Emily came to our IRB with a proposal for the sort of work she has been doing it would be approved without hesitation *assuming*:

a. The research design was sound (e.g., the survey instrument was well-designed according to accepted methodologies and appropriate to the research question being asked, the interviewers were appropriately trained, etc.)
b. The participants gave informed consent.
c. The proposal allayed any concerns about the non-anonymous nature of the interviews and any sensitive topics that might be discussed.

If any of those conditions were NOT met, our IRB would simply send the proposal back to Emily and ask her to fix the problems before it was approved. Yes, IRB approval would be needed as it IS human subjects research, but there's NOTHING in this sort of work that would make it automatically suspect in any way. IRBs do have a role as a legal/ethical guardian of human subjects research, but they are also effective as peer feedback for improving the quality of human subjects research. The bottom line, though, is that Emily's work is perfectly acceptable human subjects research when done in the methodologically sound way she has done it.

So why are Pete (and others) so afraid of IRBs? I offer two not mutually exclusive hypotheses:

a. IRBs at state schools are screwed up. I've seen a few horror stories about other schools, but those IRBs bear little resemblance to what I've seen on my campus.

b. Researchers doing this sort of work have submitted crappy and non-rigorous proposals that IRBs have rejected not because this sort of research is problematic, but because the researchers didn't know what they were doing.

This is why Emily's call for rigorous qualitative methods is also important: not only to we have to get the methods right for our own sake in generating valid results, but it's the best way to avoid having problems with IRBs.

If my second hypothesis is right, it's ironic that Pete would be complaining about IRBs, given how he constantly tells us that we shouldn't lay the blame on others (e.g., journal editors or hiring committees) for our own failures and should instead ask ourselves if we are being as rigorous and scholarly as we need to be. The same argument should apply to IRBs.

To whom it may concern: How do Austrians who work according to Emily's standards insure against the risk to pet Hayekian priors by selective choice of antecdotical evidence?

No guarantees, amv, for qualitative or quantitive methods. Updating doesn't always eliminate the influence or your priors. That's why competing voices are so important in science as in life.

I'm pretty clueless on the particular methods Emily discussed. As I understand, however, they include tricks to reduce the role of one's own priors. The frequency of key terms my lead you in a different direction than you had been pointed previously, for example, and the binning she described helps with that. That's sort of the point, isn't it? To have methods that help you really get at the agents' models, as opposed to your own prior model.

yes roger, something like this is what I meant. and of course: there is no perfect routine to get rid of biases. even to try is dangerous, as I think Mises's method can show. I understand that the best can be the enemy of the good. you have to weigh, yes, but there is certainly an optimum and I see still some free-lunch in increasing theoretical awareness. Further, pure theory is not quantitative in the Lucas-sense. Isn't it?

To give an example: given you have filtered out a perfect set of each agent's models. what's next? if we want to say things about, for instance, the Hayek problem (which was what at least one GMU presentation at the conference wanted to do), or how orders emerge by the agents' INTERACTION, I do not see what kind of help any descriptive sample of agent's models can be WITHOUT ALSO CONSULTING some kind of agent-based model or the like.

Well, it looks like we may be talking past one another, amv. I mean, I like agent-based models too, but I'm thinkin' "So what?" You seem to say that Mises showed it's dangerous to try to get around your biases, which is rather perplexing to me. I'm afraid I'm just not connecting to your POV at all. Sorry about that.


Besides internal procedures to minimize unwarranted bias, there is also the discipline of the discourse community. I really think this discipline get understated too often.

See Michael Polanyi's Personal Knowledge, and also his The Republic of Science.


amv and others,

Why is it that people are missing the theoretical ideas Emily was working with? The coordination problem of moving back to New Orleans, and the regime uncertainty point about investing once back in New Orleans. They she was choosing the methods appropriate for the task -- to get at beliefs, expectations, strategies. Her question (informed by the theoretical concepts she was working with) is not ameniable to statistical analysis. I am sure, which is what she said, that if her questions were those that were answered by quantitative analysis she would have done that. But alas they are not, and to pretend as if they are is to commit the drunk under the lamp looking for his keys error.

To Peter Klein and others --

What is the objection to chosing methods appropriate to the task? And if the data of the social sciences is what people believe (as Hayek argues) then shouldn't the techniques be appropriate for that? How else do you makes sense of Searle's football example, Kirzner's Grand Central Station example (which he got from Mises)?



In my opinion, IRB should have NO SAY at all in social science research. I understand in medical research, but not in social science.

Also, if you think the concern with IRB is due to lack of rigorous proposals you are missing the issue. Science is about discovering things, IRB often wants to know in the social sciences what you are looking for in advance. If you knew that, they you wouldn't need to engage in research.

IRB isn't about making better science, it is about trying to protect the university -- not even to protect subjects.

I find the entire exercise to be stifiling -- not that it cannot be conformed to, and gotten around, but it doesn't work costlessly.


"IRB isn't about making better science, it is about trying to protect the university -- not even to protect subjects."

This is an assertion and one open to empirical evidence. My point was simply that as I've seen IRBs operate (at reasonably close distance), they see their goal as encompassing both. They have a legal responsibility to protect the university and an ethical responsibility toward human subjects.

And I'm just going to have to disagree with your first assertion. I can think of numerous examples, both hypothetical and real, of social scientific research that treated human subjects unethically and for which an IRB process was/would be appropriate.

Seeing IRBs as only about medical research assumes they are only about preventing physical harm. It ignores the other ethical obligations that researchers should have toward their human subjects. There's nothing, in my view, "unlibertarian" about recognizing such obligations. In fact, I'd say good libertarians would not want social science researchers to proceed without IRBs given the ways in which the state and social scientists have ganged up historically against subjects and used the data to expand the state's power. I don't think *I'm* the one missing the issue.

No one said it was costless. It has costs. All rules do. The question is the appropriateness of the rules. All I can say is that the evidence I know is that IRBs can and do work well without stifling legitimate social science research.

It would be very interesting to hear from Emily on this particular issue since she's in the trenches. I happen to know that she thought my previous comment was pretty much right on target.

Jerry, my point was simple.

Austrians use "subjectivism" to cover things that are not the same -- to cover very different contexts wheremthe word "subjective" does different work.

I've pointed this out before.

Then you criticized me on false grounds.

Jerry, I have written on Darwinian biology, applying limits of knowledge arguments to the problem of reduction, writing for Alex Rosenberg. Rosenberg has taken up some similar themes in his work written after reading my paper.

Jerry, the great work I have underway is a development of the argument contained in the second part of this paper:


I see De Grauwe just this month developing a central theme in that paper (e.g. the bottom-up econ explanation is like bottom up Darwinian explanation).


I'm glad to learn you have published a paper applying a Hayekian limits of knowledge argument to the problem of reduction in biology. This is great news. What is the cite for that?

Roger, I guess you misunderstood me (which is my fault). My pointer to Mises is not meant to say that it is dangerous to try to overcome one's biases (my entire discussion here is based on the assumption that it is important to do so), but that it is also dangerous to think that one has found a perfect method that insures against any such bias. Mises's praxeology seems to be that way. I think this is rather trivial. Further, I am not insisting on agent-based modeling. You did not answer my question: how can we say something SCIENTIFIC about emergent orders by asking people how they feel and filtering how they see the world without highlighting the dynamics of their interaction in some kind of model? How to insure against becoming the German historical school which produced loads of highly detailed studies on each detail of human life without ever grasping the operation of the economy as some kind of system?

Peter, you ask: "Why is it that people are missing the theoretical ideas Emily was working with?" Perhaps people outside GMU like me just don't have the possibility to refer to the hours of internal discussion at GMU since Lavoie and honestly try to put Emily's lecture into context. I also wonder why you react as if I did not also say good things about the approach. So perhaps the answer is: the ideas Emily is working with are not that obvious and self-evident as it looks to you experts at GMU. You ask for communication with the mainstream. Just wondering why we (the unblessed outside GMU) are unable to see the light is surely a bad communicating strategy. If even such outstanding and benign scholars like Selgin have problems understanding what it is all about, perhaps you should be less dismissive of people that seriously try to understand Emily's approach but are not there yet.


This isn't a GMU thing, it is ECONOMICS. She was talking about coordination problems (a problem that comes from Schelling), she was talking about regime uncertainty (a problem that comes from Higgs). And of course she is talking about expectation formation.

All of these ideas are not foreign to anyone knowledgable about mainstream/mainline economics. This is not some code language from Lavoie. Same is true of her work on development.

My confusion was (a) that people could not understand as I said in my post that theory and empirical analysis are two different acts and that Emily's post was about empirical analysis (and she mentioned explicitly the theoretical frame she was working with), and (b) that empirical analysis from an Austrian perspective is different than the quantitative analysis most economists are trained within.

That was the point of the long Kirzner quote --- what is the task of the social sciences from an Austrian perspective.

So perhaps we are having a communication problem, but I am not so sure where the breakdown is coming from. BTW, Emily has published several journal articles and books so she must be communicating with someone other than some private community of GMU folks.


BTW, Emily published work based on her field work on Katrina in Rationality and Society, Public Choice, and Journal of Urban Affairs among other outlets just in the past year. It might be good to look at those rather than thinking there is some sort of German Historical resurgence.



Pete said some of what I was going to say.

I'm not sure what you mean by "model," but the short answer to your question is "theory," of course. As you know, sometimes "model" just means some mathematical notation to illustrate a made-up story lacking explicit theoretical grounding. (Boettke has been eloquent on this point.) Thus, I'd say "theory" not "model." I thought you had already spotted this point to Emily when you said earlier, "Certainly this is because I react allergic against economics-without-theory, which is not what Emily stands for, but what her message seems to allow for in less enlightend circles." Now it seems you view her message as consistent with a sort of historicism. I guess I should not ask more consistency than the blog medium allows.

I think Mises gets a bum rap, BTW, on this point about perfect method. It is a rather difficult point however, in part because he sets himself up for misunderstanding with all his talk about "apodictic" and the like.

Thank you Pete. This is clarifying. I didn't know that you split theory and empirical studies ("theory and empirical analysis are two different acts "). In saying that the method is GMU specific, I did not mean that it isn't economics (even though it is not Robbins's economics what it does not have to be). Why so thin-skinned? However, I should consult Emily's paper before making further claims (even though I mostly posed questions and never intended to offend anybody).


yes 'theory' is better than 'model'. Just the way I am used to talk. Awful according to McCloskey-standards, I know. On inconsistency: yes, perhaps I am guilty. I really don't think that Emily stands for lack of theory, etc, but I saw some stuff by my fellow GMU grads who work under her guidance or at least work with that spirit that seems to go in that direction (even though, I repeat, I like their stuff). But as I said to Pete: The discussion has reached a point at which I should either consult Emily's work or shut up, or both. On Mises we disagree. His economic theory falls short by so many standards, and he is so damn sure that it doesn't, that something must be awfully wrong with his updating routine.


Why do you detect "thin skin"? I am not, just responding to your points.

The method isn't unique to GMU either --- though Lavoie did push it. But as I pointed out earlier, there is a long history of ethnographic work that was inspired in part by Alfred Schutz (Mises's student in Vienna).

And again, I am not sure what GMU students you might be referring to --- my students are doing analytic narrative work in political economy. The emphasis is on analytical clarity, and with the purpose of doing narrative empirical work --- either historical or contemporary.



you write: "The method isn't unique to GMU either --- though Lavoie did push it. But as I pointed out earlier, there is a long history of ethnographic work that was inspired in part by Alfred Schutz (Mises's student in Vienna)." do you admit that this view is still a minority view? - held by a subset of austrians that are a subset of the profession? do you agree that the economics department pushing most persistently in that direction is GMU? I went on some conferences and your approach is totally new to me. since I am curious about you guys and your work, I try to understand what you are doing (and I cannot read every article). your reaction lead me to think of "thin skin" because you gave my intentions little chance. for instance: where did I suggest that GMU folks are not doing econ? this would be such a miserable suggestion that I wonder how you can ever think of me saying this! I like you guys and GMU (just offer me a post-doc position and I prove my devotion by accepting it ;-)) and I just wonder how you treat people who don't.


There's the limits of blogs for you. You were sort of coming across as saying, like, "Oh, you nutty GMU types with your secret handshakes and private language." I completely get that you did not intend to cop that attitude, that you weren't really saying that at all. Still, the impression was sort of there. Like I said, that's blogging for you. And it's not just this thread on this blog; it's a general thing. It's a minor miracle the blogosphere has not produced open fist fights at academic meetings!

glad you see my intentions though.

Overall, I still feel some puzzlement over resistance to Emily's message evident the evening of her talk. Some have mentioned George's name, but I don't think we should hold him to anything he might have said in one rather off-hand comment in a setting far less public than this blog. He may or may not have been saying what some thought he was saying. I do think it's fair to say, that, that we see resistance to Emily's message, which is strong evidence that it needed delivering and requires reiteration. Long ago Pete drew my attention to the great Spencer quote to the effect that it takes varied iteration to impress alien ideas on reluctant minds.

"To Peter Klein and others --

"What is the objection to chosing methods appropriate to the task? And if the data of the social sciences is what people believe (as Hayek argues) then shouldn't the techniques be appropriate for that? How else do you makes sense of Searle's football example, Kirzner's Grand Central Station example (which he got from Mises)?"

Pete, I thought my comment was in fact a call for methodological pluralism in thymological research. Was I unclear?

BTW, I think you are misreading Kirzner's Grand Central Station example here. The point is that we cannot understand the general idea of a train station, a stock exchange, or most any social arrangement without invoking the purposes, subjective beliefs, knowledge, etc. of the actors. That is the hallmark of methodological individualism. But it doesn't follow that all empirical research on train stations or stock markets must use ethnographic methods. It depends on the particular questions asked. I was simply making the point, in response to Ryan M.'s concerns, that small-N fieldwork, like econometric analysis, has benefits and costs, which must be evaluated on a case-by-case basis. Once we have the general idea of a train station, for example, quantitative analysis may help us tease out certain relationships between train traffic and local-market characteristics, or whatever.

And Roger, what can I say, you have always been a maverick!


Nothing in what I said or did in my own research, or what Lavoie ever said or did, that would contradict what you wrote in your second paragraph above. What did I say -- choose the methods appropriate to the question being asked. Sometimes, those questions are about beliefs, expectations, and 'meanings'.

The work that my students and I have been engaged in for the past decade, including Ed, Ben, Virgil, Scott, Pete, Chris, Tony, etc. is 'analytic narratives' and you can see this in their work. Some of these analytic narratives employ ethnographic techniques, some more traditional historical techniques or contemporary historical analysis. And I encourage the readers to look at the work produced by these and other students in this line of research. The analytical lens is provided through Austrian market process theory, institutional analysis and spontaneous order theory. And the applications range from stock markets ancient and modern; development from Japan to Ireland; Caribbean development; Botswana as an African exception; pre-colonial Africa to social order in pirate 'societies'; post-war reconstruction; and post-communist societies.

And these works have been published in a variety of journals and book publishers from the JPE to the CATO Journal; from Princeton and Stanford to Edward Elgar, etc. In other words the track record is evident on what people do with their work, the argument they make, and where they are able to place it and what sort of professional hearing they are getting.

One of the concerns raised was "historicism" -- nope, not going on; work is grounded in rational choice, institutional analysis, and spontaneous order theory (and everything in between). One of the concerns raised was "approach which cannot get a professional hearing" -- nope, not going on; work is able to be published in journals and with publishers and positions are being secured and promotion and tenure achieved, etc.

So ultimately, I came back to a complete and utter confusion over why the resistance to the message. In Emily's case, she was able to publish both in Economic Development and Cultural Change, and in Rationality and Society --- two very well known outlets for her work both in Africa and with regard to Katrina.

Let me end by repeating --- it is about choosing the methods appropriate for the questions being pursued.


Very interesting topic and discussion, ladies and gentlemen. Thanks!

I'm not an academic and I don't want to come across as a salesman, but I really like structural equation modeling and think the technique has a lot to offer Austrian econ. I use it regularly in my work as an analyst for an HMO, mainly for analyzing customer satisfaction surveys. SEM is widely used in the social sciences. It just seems to me that the technique was tailor made for Austrian econ.

PS, is Emily's speech on line anywhere?

It's not Fundamentalist. It will be available in the RAE next year. We could certainly ask her if she has a text she'd like to put online somewhere, but my guess is that she'll want to wait til the final version is out.


I think what you're calling SEM is basically just econometrics. Linear regression and factor analysis are subsets of SEM and quotidian tools of econometrics. If I have not misunderstood, some SEM techniques judge goodness of fit by comparing estimated and hypothetical covariance matrices, which if nifty IMHO. Perhaps the SEM trick you find particularly meritorious?

In any event, SEM is parametric if I have not misunderstood. You need to specify the model in SEM, right? But economic theory does not often indicate which of an infinite number of possible models is the true model. That's why I tend to push for nonparametric models. Thus, SEM seems fine and all, but, honestly, it doesn't seem "tailor made" for AE to me. It seems to me that SEM is a fine tool, suitable for many purposes. It would be rather silly to declare to taboo for "Austrian" empirics. But, again, I'm not seeing where it's "tailor made." Please let me know if I'm missing the boat here, which is certainly always a possibility I'm sorry to say!

Roger, Let me pull two quotes from the main article to illustrate the advantages of SEM:

“Emily wants to understand the meanings that economic actors attribute to their actions, the behaviors of others, and the circumstances they find themselves acting within.”

“…the subjectivist perspective focuses the analyst to take into account purpose, intentionality, and meaning.”

People in marketing, advertising, psychology and sociology use SEM to accomplish just those things. The data usually come from surveys of people’s attitudes and intentions. The factor analysis establishes latent factors that are regressed on each other to help understand the “meanings that economic actors attribute to their actions.” The survey data can be combined with longitudinal economic data. The whole point of SEM with latent factors is to take into account purpose, intentionality, and meaning.

In my case, I use survey data to understand member satisfaction with various aspects of the HMO. I use SEM with latent factors to determine interactions between attitudes to the various elements as well as the drivers of overall satisfaction and loyalty. I could extend the analysis to determine the impact of satisfaction on sales.

Roger: “But economic theory does not often indicate which of an infinite number of possible models is the true model.”

One of the main purposes of SEM with latent factors was to provide a means to objectively compare models. Surely economic theory can limit possible models to at least a couple dozen possible models. If so, comparing them is fairly easy in SEM. Essentially, the computer program creates a covariance matrix of the data and a predicted covariance matrix based on the specified model. Then you use a number of fit stats (one as chi-square) to determine how closely the predicted matrix matches the actual matrix.

The advantage of SEM with latent factors over all forms of plain regression is that factor analysis reduces measurement error. For example, you wouldn’t have to rely on a single measure of money, but could combine five different measures (maybe M1, M2, gold prices, commodity prices, etc.) into a factor called money. And instead of relying just on gdp as a measure of economic performance, combine several measures into one. Then you regress the latent factors on each other. Also, unlike other forms of regression, you can have a lot of correlated variables and separate out their effects.

The only empirical work I have seen done in Austrian econ has been done with cointegration or regular regression. SEM offers a lot of flexibility, the ability to add abstract concepts such as attitude and meaning, and the ability to quickly test different models of the data. If Austrian econ intends to be a social science, and not just an historical or math science, then it needs to use the main tool of social sciences—SEM with latent variables.

"there is also the discipline of the discourse community. I really think this discipline get understated too often."

A point made clearly in chapter 23 of The Open Society and its Enemies (1945) and people thought that Kuhn discovered the social context of science. This chapter contains a critique of central planning and also a pre-emptive strike against the strong program in the sociology of science. Not a bad effort for an amateur in the social sciences:) http://www.the-rathouse.com/OpenSocietyOnLIne/Chapter-23-Sociology-of-Knowledge.html

This chapter signals what Ian Jarvie later called Popper’s “social turn”, his recognition that whatever objectivity and rationality we can achieve cannot be attributed to special qualities of mind but to the give and take of criticism in a community.

He identified two dangerous ideas that were emerging in progressive intellectual circles during the 1930s. One was the idea of controlling social change by means of largescale central planning, the other was the theory of the social determination of scientific knowledge. One of the products of this impulse (some decades later) is the "strong program in the sociology of science".

Today the range of fields of study exploring, registering and describing the economy or a part of it, include social sciences such as economics, as well as branches of history (economic history) or geography (economic geography). Practical fields directly related to the human activities involving production, distribution, exchange, and consumption of goods and services as a whole, range from engineering to management and business administration to applied science to finance. All kind of professions, occupations, economic agents or economic activities, contribute to the economy. Consumption, saving and investment are core variable components in the economy and determine market equilibrium. There are three main sectors of economic activity: primary, secondary and tertiary.

But the training they received that fuels those pretensions to the mantle of science.

The comments to this entry are closed.

Our Books