November 2017

Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30    
Blog powered by Typepad

« What Austrian Economics IS and What Austrian Economics Is NOT | Main | The Continuing Crisis of Economics --- John Nye on Doug North, Theorists vs. Modelers, and Giants and Pygmies in Scientific Progress »

Comments

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

This also feeds into your point in the last post that you take the side of action, and not constrained optimization.

When I read this post, though, I think "yes I agree but doesn't constrained optimization do that too?".

What does constrained optimization say? Well, first the "constrained" part - that just notes that we all have budget constraints. We're faced with scarcity. I hope this doesn't need defense, and I don't think it does. But what of the optimization part?

All optimization says is "if the cost of one more is lower than the benefit of one more, choose one more; if the cost of one more is equal to the benefit of one more, stop; if the cost of one more is higher than the benefit of one more, choose less". Add a budget constraint, and that's all "constrained optimization" is. That's the math in plain English.

This never seemed to me to be any different than the implications of the action axiom. I'm no expert in praxeology by a long shot, but the implications seemed to be exactly the same. Both say "acting man will choose his prefered choice". The constrained optimization in plain English that I provided above doesn't even require cardinality! You can have ordinal preferences and still use that (obviously it becomes a problem if you try to translate it from English back into math).

When we actually mathematize it, I can see how some devotees of ordinal preferences could reject constrained optimization. If that's the objection, I don't share it but I understand and accept it. But that seems to be the only real obstacle.

The other obstacle is of course that we deal with a market process and not a Walrasian auction, but that's not an insight exclusive to any school of thought. The search process has been central to mainstream economics.

If the arugment is that by moving methodologically away from praxeology and towards formal optimization (let's say with search) there is some artificialization of the science that is unacceptable, then OK. Everyone has to decide where they fall on the artificialization/clarity trade-off spectrum. But as far as the fundamental insights of human action, I've never understood exactly what fundamental insight you suppose is lost.

Kirzner's INTERPRETATION of the first part of Human Action is much better than the reality of that text. Mises is stuck in a philosophy of science that is essentially one grand tautology. I do not want to get into all that here. It is at best static subjectivism, but it is more likely what Popper calls an immunizing strategy. I spent years being tied up in intellectual knots by that stuff.

Daniel wrote:

"The other obstacle is of course that we deal with a market process and not a Walrasian auction, but that's not an insight exclusive to any school of thought. The search process has been central to mainstream economics."

If you think "market process" is about "search," you have some reading to do, and particularly if you are thinking "search" in the Walrasian tatonnment sense.

I think the question of how AE differs from the mainstream is one that needs to be asked, and that Austrian types should be always willing to answer. But I also think that many who believe the differences are small to non-existent have not really grasped the Austrian argument.

"Mises is stuck in a philosophy of science that is essentially one grand tautology."

But of course. That's all that deductive thoughts are - grand tautology. Mathematics is another grand tautology because all the conclusions can be logically derived from A=A.

I really like this post, and I agree with Per on the previous. Stressing action per se is of paramount importance. Focusing on action, not choice, gives you time preference, and other extremely important deductions that other avenues don't. Like it was said before me, mathematical equations don't take place in time (or don't account for time) and so they neglect a necessary component of action.

Steve - I don't think the market process is search strictly speaking, but I do think they both have their root in similar dissatisfaction with equilibrium in economics and with answering the question of how we move from equilibrium to disequilibrium. In market process thinking, the entrepreneur is the prime mover in the economic system and as such he is alert to gaps in the market and enticed by the lure of pure profit closes those gaps through arbitrage. What is this if not entrepreneurial search? All search theory does is note that profit is not the only thing that is sought after - that employers, workers, demanders, and suppliers also take action to seek out their own respective surpluses, which does the work of closing the gap between disequilibrium and what we think of as "equilibrium".

So sure, there are differences - but I don't think it's wrong to make the connection. Peter made a similar connection when he blogged on the 2010 prize a couple weeks back, and I thought he was right for it (come to think of it I should comment more when I think you guys are right... this tendancy to comment when I have concerns probably makes me look less sympathetic to the blog than I really am).

I would not say that the differences between the mainstream and the Austrian school are "small to non-existent", but I would say that there are some differences that are overemphasized. Another one that bugs me is the aggregation point, for example. I'm in a Prices and Production reading group right now, and it's been amazing to read Hayek so freely use and speak of various aggregates in Prices and Production IN THE SAME YEAR that he penned his critique of Keynes for doing the same thing! The difference is not in aggregation, although this is often harped upon. The real, driving difference is that Hayek disaggregates along the time structure of the production process and Keynes disaggregates along the spectrum of the liquidity of an asset (and of course later Keynesians disaggregated along other lines at various points). Both Hayek and Keynes use their versions of disaggregation to good effect, and both abstract away from other contending disaggregations. And yet we hear this clarion call of "aggregation vs. disaggregation" over and over again as a driving difference. It's not that there aren't differences - there are. But they're often misdiagnosed.

All that having been said, I would not argue with you at all that I have reading to do.

*from disequilibrium to equilibrium I meant to say

Daniel wrote: "In market process thinking, the entrepreneur is the prime mover in the economic system and as such he is alert to gaps in the market and enticed by the lure of pure profit closes those gaps through arbitrage. What is this if not entrepreneurial search?"

I would suggest that this elides the important distinction between "search" and "discovery" that is at the heart of Kirzner's work on entrepreneurship. Now people like Caplan might deny that distinction, but I think they're wrong. Search theory is about finding something we know we don't know. Entrepreneurial discovery is about what we don't know we don't know. I think that distinction is at the heart of what AE is about and at the heart of the distinction between "action in an uncertain world" and "constrained maximization."

Pete,

Did you really mean to say, “some methods are ruled out of court explicitly”? Is that statement consistent with your recent response to Garnett on methodological pluralism? Is Gode & Sunder JPE 1993 “ruled out of court explicitly” because they have “zero-intelligence traders” rather than homo agens? Presumably, you will express appreciation for Gode & Sunder, but then I don’t know what methods are “ruled out of court” by a recognition of the difference between human and robot action.

My own view, as you know, is that Misesian action is indeed different from Robbinsian maximizing, the this distinction matters, but that the recognition of the distinction does not rule any method out of court and, finally, that methods that neglect entrepreneurship or Misesian action may nevertheless be useful and appropriate in economic analysis.

The first 100 pages of Human Action are quite unsatisfactory, in my opinion. The pure logic of human action is important, no doubt, but Mises draws in all kinds of irrelevant epistemological issues.

Economics _explains_ design like order in the overall coordination of economic plans -- that order is the _problem_ that demands explanation.


Economics is NOT a belief / desire "science" of human action -- this is the mistake the Mises picture inspires.

I get your point Pete, but you never seem to really get this big picture insight, which is all important in Hayek's picture of the science of economics.

Steve -
First, I'm not sure search is not also about what we don't know we don't know. Is there some requirement that an agent knows what he is searching for that I'm not aware of? Anyway, this sounds like Knightian/Keynesian concerns that may be more or less incorporated into any given search theory or market process theory, but it's not an insight that I would say makes a sharp distinction between schools, and it's not an insight that is foreign to search theory (even if we may argue that market process emphasizes it more).

This seems like a good place to ask: I found Bryan Caplan's criticism of the Austrian approach quite convincing, what do you guys think is the best rebuttal to it? I've watched the Bottke / Caplan debate and Caplan seemed to me to win that decisively as well.

I mean 'best' in the sense of most likely to convince a non-austrian, of course.

There is a considerable difference between search and discovery. If I am searching for something, I know what it is I am searching for (a job, for instance); but if I am discovering, I do not know what I am looking for when I set out (though I may mistakenly believe I am). When Columbus "discovered" America, he did not know it was there. Rather, he was searching for a better passage to the Indies. He failed in his search, but succeeded in making a discovery. On the other hand, we have all kinds of searches in science. Quantum physics makes predictions of particles that physicists then search for -- the know it should exist, so search for it. But the moons of Jupiter were discovered -- there was no theory that they should exist, yet observation resulted in their discovery. The entrepreneur makes discoveries, finding what he did not know he was searching for. A job-seeker makes searches, finding what he knew to exist (the kind of job he could do). One does need search theory, but one also needs discovery theory. The two are quite difference, even if they appear to have similarities.

I'm in love with the first 100 pages of Human Action.

I'm on my third listen through (I have the audio download). Each time I feel like I gain some new insight. His logic is very organized and clear, but you do have to follow it very carefully or sometimes you'll think he's saying something different than he means.

Troy - I think you're investing too much in the word "search" and the word "discovery". What in search theory requires that agents know what they're searching for? Can you tell me? I'm not aware of anything in search theory that says anything like that.

Yes, job-seekers search for jobs. That's practically definitional. Is that the only concern? Entrepreneurs search for investment opportunities or profits too. That alone doesn't mean they are searching for "known unknowns".

Anyway - they are definitely different, I agree. But they raise - broadly speaking - the same criticism of equilibrium economics and provide - broadly speaking - a similar answer.

Search and similar theories largely do the same task as the market process. Cosntrained optimization largely does what human action does ("do what is best and most subjectively prefereable for you"). The only differences I'm seeing are in any given economist's prefered point on the artificialization/clarity tradeoff spectrum. I prefer to write and describe things on the low-artificialization/low-clarity end of the spectrum but analyze and theorize on the high-artificialization/high-clarity end of the spectrum. I feel like Austrians tend to both write and do analysis on the low-artificialization/low-clarity end of the spectrum that is richer but a little fuzzier. But the concerns with "equilibrium economics" aren't especially different on this front... some of the language and citations are different of course.

Entrepreneurial learning is not an action.

The logic of action explanation as a causal explanatory system is very different than the character of kearning as a causal explanation for design like social plan coordination.

This matters.

From Dictionary.com

Search:

–verb (used with object)
1. to go or look through (a place, area, etc.) carefully in order to find something missing or lost: They searched the woods for the missing child. I searched the desk for the letter.
2. to look at or examine (a person, object, etc.) carefully in order to find something concealed: He searched the vase for signs of a crack. The police searched the suspect for weapons.
3. to explore or examine in order to discover: They searched the hills for gold.
4. to look at, read, or examine (a record, writing, collection, repository, etc.) for information: to search a property title; He searched the courthouse for a record of the deed to the land.
5. to look at or beneath the superficial aspects of to discover a motive, reaction, feeling, basic truth, etc.: He searched her face for a clue to her true feelings.
6. to look into, question, or scrutinize: She searched her conscience.
7. (of natural elements) to pierce or penetrate: The sunlight searched the room's dark corners.
8. to uncover or find by examination or exploration (often fol. by out ): to search out all the facts.

–verb (used without object)
to inquire, investigate, examine, or seek; conduct an examination or investigation.


Discover:

–verb (used with object)
1. to see, get knowledge of, learn of, find, or find out; gain sight or knowledge of (something previously unseen or unknown): to discover America; to discover electricity.
2. to notice or realize: I discovered I didn't have my credit card with me when I went to pay my bill.
3. Archaic . to make known; reveal; disclose

One can engage in search to discover, but one does not have to search to discover.

There is a reason there are two word: they have two different meanings. Two very different meanings.

jsalvatier, I read Caplan's critique (haven't seen the video) and thought it was childish. Sorry. I would really be interested in what you thought was so convincing from it, though.

PS, the reason I found it childish is that he debates straw men for the most part. Austrian economists do not say what he claims they say in much of the paper.

Troy - you miss my point. I know they have two different meanings. My point is that you are giving me a vocabulary lesson instead of an economics lesson.

I would prefer to know what requires that "searchers" in search theory know exactly what they're looking for.

My whole point is that Diamond, Pissardes, Petrongolo, and the rest of them could have probably just as legitimately called their theory "discovery theory". Defining the words as you do doesn't tell me much.

The words as used in economics have the same meanings as given above. To search for something means you know what you are searching for. To discover something is to find something you did not know what there. In the latter instance, new knowledge is created. In the former, one only confirms what is known -- no knew knowledge is created. That makes a huge difference.

Troy -
OK, my point is if I swiped all of search theory (adequately cited, of course), added a few bells and whistles to make it original, and re-published it in the AER calling it "Discovery Theory", what would you critique about my presentation of the problem?

That's what I'm still fuzzy on. I'm asking you to entertain the possibility that search theorists might have been sloppy in choosing the words they use, so I don't want to restrict the discussion to that.

Search theory is all based on constrained optimization where the means and ends are known and where agents maximize.

Austrian market process theory does not assume that the actor knows his/her means-end framework and makes no assumption that the actor is optimizing or maximizing, other than in the trivial sense of doing what he/she thinks best. That is a far cry from the neoclassical idea of maximization/optimization as normally construed.

Maybe the first 100 pages of Human Action would read better if von Mises had understood that positivism/empiricism does not work in the natural sciences, and fallible or conjectural apriorism plus the critical approach is appropriate for both the natural and the human sciences?

Hayek vs Boettke 1-4:

1H. Only individuals plan. The logic of choice helps us understand the design like structure of globally coordinated individual plans.

1B. Only individuals choose.

2H. The study of the market order begins with the empirical problem of design-like structural patterns in the global inter-coordination of countless individual plans. And understanding of the causal effects of individual entrepreneurial learning with an inherited social structure traditional laws and patterns of behavior explains the problem raised by the design-like structure of the global economic order.

2B. The study of the market order is fundamentally about exchange behavior and the institutions within which exchanges take place.

3H. The most significant "facts" of the social sciences are (a) that knowledge is local and divided, and (2) that learning in the context of changing local conditions and changing global price relations is a causal explanans capable of producing and explaining systematic social patterns with "design like" features.

3B. The “facts” of the social sciences are what people believe and think.

4H. Learning and re-ordering one's affairs involves (a) judgments of local conditions and (b) valuational relations that are unique to each individual. The logic of (a) and (b) are related but different, and neither can be reduced to constant conjunctions of physically or chemically defined categories.

4B. Utility and costs are subjective.

Hayek vs Boettke 5-6:

5H. The price system allows people to cooperate with people unknown to them, and to economize on the inputs to production, by allowing them to make use of socially generated knowledge of relative scarcities in the form of prices which in the absence of markets would be unavailable to them.

5B. The price system economizes on the information that people need to process in making their decisions.

6H. Private property, the rule of law, freely changing relative prices, and a functioning signal system of profits and losses are a necessary condition for the mostly successful coordination of individual economic plans in an extended economic order

6B. Private property in the means of production is a necessary condition for rational economic calculation.

If I am suggest, part of the confusion about a clear or relevant distinction between "search" and "discovery" is due to the fact that in the "real world" that are often going on at the same time in the same person.

For example, when I was a student and people actually went "looking" in the library stacks, I would often be looking for books on a particular topic or theme that I "knew" where or might be there, but which I did not have knowledge of (author, title, content).

But one of the pleasures I often experienced while do this "search" was to be entrepreneurially "alter" to what I might "discover." I would allow my eyes to roam over the shelves knowing only that there were probably a lot of books on a lot of interesting topics or by authors I had no knowledge about. By only "motive" in being alter to what I might "discover" while I was also doing my specific "search" is that I might come across one or more volumes the "discovery" of which I would find to be intellectually highly profitable.

And in hindsight, my "discovered" intellectual profits were often far greater than the rewards from my specific book "search" on a particular day. (This sometimes even included "discovering" works on Austrian Economics I did know existed and that I would have never thought of "searching" for.)

This can happen to the scientist in the laboratory. He is busy in his experiments consciously "searching" for a cure for "x." But in the process of his experimentation he "discovers" something that he had not thought of and imagined could result, say, as a byproduct of some test he was doing. "Searching" for a more tasty chewing gum, he "discovers" a cure of cancer, for example.

One has to be alert even during one's "search" so as not to miss something of value. But the alertness to what might be or is "discovered" still remains conceptually different.

I'm walking down a street "searching" for a particular store, and I'm being "alert" to reading the signs on the stores on both sides of the street. But, "unplanned," my eye catches something that I did not expect or know about on that street that has usefulness or value to me.

This is the type of distinction about which Kirzner and other Austrians are trying to make.

This is different that a standard "constrained maximization" problem, in which I know how much time I have to search, I have so much money to spend if I find it walking down the street. And, I finally stop searching if the "cost" of more search time is less than the expected "gain" from a marginally longer amount of time looking on one more block.

I didn't mean to be tedious about this. But I think these types of everyday examples help to highlight the conceptual difference between the two.

Richard Ebeling

Not tedious Richard! The man who discovered penicillin was not looking for something that blew in the window, if they knew things were coming in the window the laboratory manager probably would have shut them. Otherwise they would have put the dishes no the window sill.

Some of the most important books are the ones you find on the shelf near the one you were looking for, especially if you are being taught by people who don't know about the best books (or would tell you not to waste your time reading them, like that dreadful book "Human Action"). BTW I still think he should have called it "The Structure of Human Action".

Richard - No, I think that's great.

Let me put it this way - yes there are constraints and tradeoffs in search models. But that doesn't mean that agents know or have to know the distributions that they're searching through. At least I don't think they have to. They could have expectations about it, of course. Expectations of this are probably even necessary for decisions about investment in search time. And their search will be constrained by various costs and other constraints just as the activities of an entrepreneur is constrained by various costs and other constraints.

But the difference between a searcher and a discoverer is really just their knowledge of the distribution of "matches" they're searching through. Is this knowledge required in any search models? I don't think the knowledge of the distribution is required. And if it's not, I really think this boils down to a disagreement over relative formalization.

Because even your library "discoverer" knows he's looking for a book. There has to be some sense of looking for something. We think of entrepreneurs looking for a profit after all.

If there weren't some broad identification of the goal, they wouldn't be entrepreneur's they'd be lottery winners. Which is great - don't get me wrong. I might rather be a lottery winner than an entrepreneur myself :)

Richard said exactly what I was saying, just more clearly.

Oh, I get it! Thanks Richard.

Formal modeling is deductive: the solution to the problem must be assumed by given conditions, otherwise the solution cannot be deduced. Logically, the assumptions and implications of a set of premises are the same thing. There is no room in such a model for unexpected discoveries of alternative options during the search, nor the opportunism of individuals who are ever alert for such a possibility. This may be impossible to model formally, because such a search, by definition, ends with conditions not assumed (or implied) in the original problem-situation.

Is this right? I haven't read Kirzner. Perhaps I should, but books are expensive.

How to sort out the fascinating history of entrepreneurial activity from the important theoretical questions?

My jetsetting friend Dave got talking to a middle-aged to elderly American in the bar of the airport Hilton in Rome. When he found that Dave is an Australian he asked if he would like to buy two vinyards in Western Australia. Dave is a big wine drinker but he said "No thanks".

Clearly the man was loaded and Dave is an upfront guy so he asked the man how he made his money. The answer "Plastic coat hangers".

Several decades ago the young man had no job, no money and no prospects. He washed his shirt, put it on a wooden coat hanger to dry and the wood stained the shirt. So he decided that he wanted a coat hanger that would not stain his shirts.

Possibly the key to success was selling to dry cleaners who use huge numbers of hangers, and eventually he sold the distribution business for some tens of millions of dollars.

So what are the theoretical and policy issues raised by these kinds of stories?

Looking at the Hayek vs Boettke posts, it appears that Boettke is much more concise than Hayek.

Looking at Darwin vs. the Bible on the causal origin of species, it's clear that the Bible is much more concise than Darwin ....

Lee Keely,

"Formal modeling is deductive: the solution to the problem must be assumed by given conditions, otherwise the solution cannot be deduced. Logically, the assumptions and implications of a set of premises are the same thing. There is no room in such a model for unexpected discoveries of alternative options during the search, nor the opportunism of individuals who are ever alert for such a possibility. This may be impossible to model formally, because such a search, by definition, ends with conditions not assumed (or implied) in the original problem-situation."

As I see it, given my limited reading, there is no room for discovery and uncertainty in modern mathematical models. But I think that is not a universal property of mathematical models, it is only that the people that constructed modern mathematical economics, since the middle of the 20th century, didn't have an understanding of discovery and uncertainty, so they constructed a mathematical structure that cannot accommodate these basic elements of human action.

A formal model that integrates human action would need a different mathematical structure. Modern computation economics also has some elements of market process that classical mathematical economics cannot accommodate.

Rafael writes,

"As I see it .. there is no room for discovery and uncertainty in modern mathematical models."

Note well.

These are issues in the domain of the logic of LEARNING (something different from the logic of actions).

They are not inherent in the logic of "action" -- i.e. raising my hand is an action. We don't have to think about discovery or uncertainty to "model" an action. We simply refer to beliefs and desires which we can posit at "givens" -- just like in formal construction made out of "givens" which can be "taken" as shared between minds.

Mises wants to use language different from everyone else regarding "action" because he wants to create a "science" of actions for philosophical reasons based in the work of Kant and the justificational demands of the philosophical tradition, and as a way to defeat the historicists.

Hayek counters Mises by showing that economics is an empirical science which provides a causal mechanism which explains design-like global economics order -- solving a problem-raising issue in our experience sparked by empirical observations of order in the social domain.

Hayek further counter Mises by showing that the empirical domain of learning is different from the logical domain of choice theory, and furthermore, that the central causal components which provided the explanatory explanans in economics involves this empirical domain of learning -- e.g. changing judgments in the context of changing relative prices and changing local conditions (and also the empirical "institutional" domain of imitated and learned rules of morality and law, etc.).

Now, one can say that this is "all in Mises" -- but Hayek's point is that you get the picture of the science all confused if you muddle up all this and put in front what is false in Mises, and hide from view what is true in Mises.

Boettke puts the muddled and often false "science of action" stuff from Mises in the forefront, and he leaves buried or muddled Hayek's central points about the starting point of economic science -- the EMPIRICAL problem of design-like social order -- and the two core causal explanatory elements in the explanation of order provided by economic science -- entrepreneurial learning and common rule following "institutional" traditions.

The key ..

We need more of Hayek's sound picture of the structure of the empirical science of economics, and less of the Mises' muddles and mistakes derived from his failed effort to adopt Kant and defeat the historicists.

The history of the philosophy of science across the last 100 years is the history of showing that discovery and uncertainty cannot be formalized or "modeled" or captured in a logical or mathematical construct .. that's the whole point of Kuhn.

Rafael said,

"there is no room for discovery and uncertainty in modern mathematical models. But I think that is not a universal property of mathematical models, it is only that the people that constructed modern mathematical economics, since the middle of the 20th century, didn't have an understanding of discovery and uncertainty, so they constructed a mathematical structure that cannot accommodate these basic elements of human action."

Greg,

I think it's useful to mention here something Pietro M. reminded us about over on Thinkmarkets:
"Mises's argument was much less rationalistic than most think, both critics (like Hayek) and supporters (like Rothbard) of radical rationalism."

I don't think the differences are as great as you make out.

Should we not differentiate between the two forms of rationalism, as Hayek did?

Current, as I said above,

"one can say that this is "all in Mises""

The issue isn't "rationalistie" or "not rationaistic", the issue is laying out logical and explanatory structure of the thing correctly, and correctly identifying the logical explanatory nature of the elements involved.

Hayek sets it up non-pathologically.

The science begins with an empirical problem in our experience.

The solution to this problem is an empirical mechanism involving entrepreneurial learning and common rule following.

And Hayek correctly points out that learning is logically different than "action" explanations or "choice theory".

Now, you can find elements of all this in Mises, but he mischaracterizes everything, and he misidentifies the problem to be explained, and he misidentifies the logical character of the explanatory elements involved -- learning is an empirical / causal component within an empirical explanatory mechanism, it's not a Kantian category within a "science" of "action".

If you get the logical character of the elements wrong, you block explanatory progress.

Compare Erst Mayr on how the mistaken logical categorization of biological entities by Platonic and Aristotelian thought blocked progress in biology until the time of Darwin -- and long after on the Continent.

If you take species and biological characteristics as Platonic essences or Aristotelian natural kinds, then they can't evolve or adapt or change form or speciate.

And you can't really imagine any empirical problem to be explained or any possible causal explanatory elements to make up a mechanism solving that explanatory problem.

Reference:

Ernst Mayr, The Growth of Biological Thought.

The comments to this entry are closed.

Our Books