Complexity and Human Action - what if human beings were automata?

In Human Action and elsewhere, Mises emphasizes the difference between the natural and social sciences by pointing out that there are no constants in the latter. He means that there are no natural laws in social science equivalent to the law of gravity, or the constant speed of light. Human beings are not automata. Thus, our understanding of human action comes not from the empirical observation of constant relationships between the elements of the social system (the human actors), but rather from our “understanding” of their motives and the logic of action itself in terms of means and ends, categories that have no meaning or application in the natural sciences.

One might be tempted to conclude that the absence of such categories, as intention, or means and ends, is sufficient to render the subject matter equivalent to the natural sciences with its constant relationships, its natural laws – that if human beings were automata simple prediction would be possible. But, of course, as Hayek pointed out, natural systems displaying such natural law regularities, that enable precise quantitative prediction, are, most probably, in the minority in the natural world. Biological, chemical, astronomical and cybernetic systems for example, are "complex phenomena" for which prediction of a quantitative kind is impossible. A striking example is the work of Stephen Wolfram on systems of automata. Wolfram shows how some very simple systems of interacting automata produce complex outcomes that are not predictable. In fact, this suggests that, in one sense, Mises’s claim implies the opposite. If human beings were automata (given that human societies are characterized by huge numbers of multi-level interactions) we would be at a total loss to predict likely outcomes. It is because they are not automata that we have more knowledge than we otherwise would – through understanding the logic of human action and armed with empathizing insights into the motives of our subjects. It is an understanding of the incentivized behavior of human beings that allows us to understand and predict invisible-hand outcomes.

Yet, the complexity problem remains. Can we be certain, as a logical matter, that if certain conditions obtain, certain definite types of outcomes will result from the free market process? It is on this type of thinking that the more sophisticated Keynesians might base their arguments for benevolent and effective intervention in the face of economic crisis. And this is the subject of my next blog. 

Comments

[Daniel Kuehn]

re: "Can we be certain, as a logical matter, that if certain conditions obtain, certain definite types of outcomes will result from the free market process? It is on this type of thinking that the more sophisticated Keynesians might base their arguments for benevolent and effective intervention in the face of economic crisis."

Much of Keynesian theory lies at the intersection of this complex non-automata behavior (Keynesian uncertainty, liquidity preference, the psychological factors Keynes discussed) and the few constant "laws" we can be said to have - namely, a few important accounting identities and some empirical regularities (although those aren't strict "laws", but consistent enough to work with). There is a structure in which we non-automata act.

I'm curious what you think of advances in agent-based modeling. I'm told it's promising, but then again I'm told that by partisans. I think you're exactly right that we're not automata, but the question is if we build automata with enough of the characteristics we do have, could we get comparable emergent structures to help us understand the way the economy works? I think you can answer "yes" to that without thinking we're automata. We've been fruitfully simplifying and abstracting human behavior for a long time, after all.

[Peter Lewin]

On agent based modelling, i would have to plead ignorance. I understand the idea and I worry about being able to "build in" enough relevant complexity into agent responses, but I don't know enough to render a firm opinion either way.

[Raoul]

Hello,

”In Human Action and elsewhere, Mises emphasizes the difference between the natural and social sciences by pointing out that there are no constants in the latter. He means that there are no natural laws in social science equivalent to the law of gravity, or the constant speed of light. Human beings are not automata.”

Mises doesn’t exactly say that. He says that, presently, we don’t know if human actions are caused (and, as a consequence, if human beings are automata) and, anyway, even if they were, we ignore what are the causes of the human actions.

Thus, Mises writes:

”Concrete value judgments and definite human actions are not open to further analysis. We may fairly assume or believe that they are absolutely dependent upon and conditioned by their causes. But as long as we do not know how external facts--physical and physiological--produce in a human mind definite thoughts and volitions resulting in concrete acts, we have to face an insurmountable methodological dualism. In the present state of our knowledge the fundamental statements of positivism, monism and panphysicalism are mere metaphysical postulates devoid of any scientific foundation and both meaningless and useless for scientific research. Reason and experience show us two separate realms: the external world of physical, chemical, and physiological phenomena and the internal world of thought, feeling, valuation, and purposeful action. No bridge connects--as far as we can see today--these two spheres. Identical external events result sometimes in different human responses, and different external events produce sometimes the same human response. We do not know why.

In the face of this state of affairs we cannot help withholding judgment on the essential statements of monism and materialism. We may or may not believe that the natural sciences will succeed one day in explaining the production of definite ideas, judgments of value, and actions in the same way in which they explain the production of a chemical compound as the necessary and unavoidable outcome of a certain combination of elements. In the meantime we are bound to acquiesce in a methodological dualism. (HU, I, 3)."

”Biological, chemical, astronomical and cybernetic systems for example, are "complex phenomena" for which prediction of a quantitative kind is impossible.”

As I understand it, such a prediction regarding natural phenomena can actually be impossible for a human being, but it’s not theoretically impossible . It’s impossible only because both our intelligence and our knowledge of the data are limited. These qualifications are important in the Laplace’s sentence quoted by Hayek in The Counter Revolution of Science:

“A mind that in a given instance knew all the forces by which nature is animated and the position of all the bodies of which it is composed, if it were vast enough to include all these data within his analysis, could embrace in one single formula the movements of the largest bodies of the Universe and of the smallest atoms; nothing would be uncertain for him; the future and the past would be equally before his eyes."

[Ken S]

"If human beings were automata (given that human societies are characterized by huge numbers of multi-level interactions) we would be at a total loss to predict likely outcomes."

Wolfram divides his automata into 4 classes, class 3 automata are the ones that are chaotic and unpredictable, but class 4 are ones capable of 'universal' computation. The behavior of these can be predicted using the same methods a programmer might use to write and test correct software, once you discover the extra encoding needed to do this. Uncertainty is also compatible with a less philsophical usage of the term: http://en.wikipedia.org/wiki/Probabilistic_automaton

My position would be that the Austrians have been doing 'agent-based modeling' all along, just not always with the help of computers 'in silico'...

[Roger Koppl]

Peter: As you know, I like this sort of think; I think we can learn a lot from the mathematics of cellular automata and stuff like that. I keep citing and citing and citing

Velupillai, K., 2007. The impossibility of an effective theory of policy in a complex economy. In: Salzano, M., Colander, D. (Eds.), Complexity Hints for Economic Policy. Springer, Berlin.

Velupillai shows that “an effective theory of economic policy is impossible” for an economy that can be modeled as “a dynamical system capable of computation universality (pp. 273 and 280). He links this important result to F.A. Hayek’s “lifelong skepticism on the scope for policy in economies that emerge and form spontaneous orders” (p. 288).

Daniel: For my money, there are some real successes in the ACE literature. The artificial stock market of Arthur LeBaron and others. Gode and Sunder's study of zero-intelligence traders. Arthur's El Farol problem. Howitt, Alchian, Holland, Axtell, Schelling, and so on. Indeed, in some sense all of spontaneous-order theory is agent-based modeling, which brings in Adam Smith, Carl Menger, and F. A. Hayek. De Grauwe has a model of animal spirits that might count in the category. And so on.

Raoul: I don't think you said anything wrong about methodological dualism in Mises. But I don't think anything you said in any way corrects what Peter said. I do think you said something wrong about predicting automatons. LaPlace was mistaken; you could not predict the universe even assuming you somehow knew the position and motion of each particle. The best statement of that that I know of personally was given by David H. Wolpert

Wolpert, D. “Computational Capabilities of Physical Systems”, Physical Review E, Vol. 65, 016128, 12/20/01.

Ungated version here: http://ti.arc.nasa.gov/m/profile/dhw/papers/71.pdf

Wolfram notes that for universal Turing machines "meaningful prediction is impossible," which is the point Peter was picking up on.

[Troy Camplin]

The point Wolfram makes regarding the complex pattern-creating cellular automata is not that the pattern is incalculable, but that it would take longer to do the calculations necessary to predict the pattern of, say, line 500, than to just wait and see what happens when line 500 arrives. Wolfram is proposing a completely deterministic model, which means one could in fact do the math to determine the outcome of the process that creates the complex patterns. But since we find out what happens before we can do the calculations, such processes are in this sense incalculable (even to Laplace's demon). Unless one believes that the world is fully deterministic, this is somewhat different from the argument that human minds or real economies, both of which are in far-from-equilibrium states, are incalculable.

[Allan Walstad]

"Can we be certain, as a logical matter, that if certain conditions obtain, certain definite types of outcomes will result from the free market process? It is on this type of thinking that the more sophisticated Keynesians might base their arguments for benevolent and effective intervention in the face of economic crisis."

But, can we be certain that government intervention will lead to definite -- and desirable -- outcomes?

On "agent-based" modeling, meaning the mathematical Tinkertoys that we find in, for example, the El Farol Bar problem and its variants, if this is supposed to be success, then it is "success" in a very limited and potentially misleading way. In the El Farol Bar problem, everybody is trying to outguess everybody else, in order to show up on a night when the bar is not overcrowded -- based on the pattern of previous attendance numbers, which somehow everybody knows, whether or not they showed up. This game focuses the ingenuity of programmers on what mathematical heuristics -- algorithms -- their agents might use, how those heuristics might be updated, how they might be distributed throughout the population, etc. And sure enough, tweaking the parameters, one can get fluctuations in attendance that mimic the statistics of fluctuations in financial markets. So what? So perhaps financial markets are just big zero-sum games of chance? The problem is that one is being led to think within a highly constrained box. Nobody in the El Farol problem gets the idea to expand the facility or build another bar or franchise it out or call ahead, or a million other things that come to mind when one is thinking about human beings rather than mathematical stick figures. But it seems so obvious that a supervising authority could manage the zero-sum situation so everyone gets their equal and fair share. The real world has land, labor, capital, and entrepreneurship, as well as, unfortunately, central banks that jerk around with the money supply and government meddling that distorts incentives. I submit that when we construct verbal scenarios of human action, we have human beings in mind; when we construct mathematical models we have machines in mind, and it is far too easy to be led down the primrose path of social engineering.

[Roger Koppl]

Troy is right on what Wolfram said about level IV cellular automata. In the WP version of the Wolpert paper I linked to above he assumes a deterministic and computable context. He shows that you still can't always predict the system *ahead of time*. Wolpert's analysis in not the same as Wolfram's. Wolpert gets his result by assuming the system contains a copy of the computer the imaginary observer is using. the published version generalizes. I can't find link to the old Santa Fe WP, "An incompleteness theorem for calculating the future."

[Roger Koppl]

Allan: I don't think I understand the structure of your argument. The way it's written seems to suggest that Arthur's model is no good because is did not included the entrepreneurial market process. But I don't recall that he was drawing any general conclusions about markets vs. control in the paper. Maybe I just blanked out the bits I didn't care for? I think the point was to show how there is an ecology of expectations and models in social systems, and how we can model that. The more applied version would be the artificial stock market, whose policy implications, BTW, are not interventionist as far as I can tell. Oh, and I think the sort of scenario he describes really does crop up often enough as reflected in the quote from Yogi Berra: "That place is so crowded, no body goes there anymore."

[Seth]

Maybe we're not automata. We could be made up of it.

[Allan Walstad]

Roger, I don't know what you might have been looking for under the heading of "structure," but I am happy to elaborate a bit on my comment. The El Farol Bar model is a constant denizen of the so-called "econophysics" literature. In my reading and experience, the practitioners of econophysics tend to be determined to find the causes of instability within the market, not from outside shocks or interventions such as Fed monetary machinations. There is a strong strain of criticism of modern mainstream economics for its supposed devotion to the self-correcting nature of markets, even though, of course, much of the mainstream is always finding fault with markets and pushing interventions, especially monetary. One major difference between the two camps is that the econ mainstream faults markets for failing to match the idealized model, namely general equilibrium, whereas the econophysicists fault markets for resembling their models, "agent-based" models involving mathematically characterized stick figures with pathetically limited capabilities and options, the results of whose "actions" and "interactions," if taken seriously, patently lend credence to the idea of fruitful interventionism. It is in the context of these observations that I judge the El Farol Bar model and its variants to be not merely weak but misleading, as discussed also in my previous comment. I don't know what Arthur personally thinks or what his agenda might be. It's not about one person.

[Raoul]

Roger,

Indeed, in retrospect, I see I was not correcting what Peter Lewin was saying.

Thanks for the interesting paper.

[Roger Koppl]

Allan,

I don't think we should criticize Arthur or his El Farol model is some econophysicists have taken it in a bad direction. If they have, we should criticize them! FWIW, here is what I said in a book review about one important econophysicist:

McCauley’s contribution gives us an overview of some techniques of econophysics appropriate for forecasting financial markets and, perhaps, uncovering some facts that will require theoretical explanation using the concepts and techniques of social science. Unfortunately, it does not go deeper. An enthusiast for techniques borrowed from physics, McCauley seems to have no pene- tration into the real problems of either social science or statistical inference. He says, for instance, “Market equilibrium is equivalent to time translational invariance of the price distribution” (p. 63), thus making economics a branch of physics in spite of his own suggestive remark that “Every- thing in economics reflects ‘mind”’ (p. 65).

[Allan Walstad]

Roger, I'm happy with my original statement that if the El Farol Bar model is an example of success, then it is "success" in a very limited and potentially misleading way. It is the sort of thing that lends itself, as math modeling in econ tends to do, to a narrowing of focus away from human beings and toward mechanism, and to the social engineering impulse. I've got nothing against Arthur for playing around with this stuff, but if El Farol is on the short list of successes in agent-based computational economics, it does not bode well for ACE in my view.

You mention McCauley. I exchanged a few emails with him a couple years ago. I've pushed that conversation pretty far out of my mind, but do recall he kept going on and on about "excess demand." I regard him as a quack, an example of how physics is being used as window dressing for simplistic thinking or worse. An upcoming issue of American Journal of Physics will have my review of a recent econophys book.

[Anamaria Berea]

Yes, we can predict certain outcomes of free market in an agent-based model. The difference between cellular automata and agent based models is that in an ABM the agents have individual behavioral specifications, may or may not have cognitive/decision maps and may or may not change their behavior within the same simulation. The decisions, behaviors and rules are specified either theoretically or empirically or both. Cellular automata is about computational and mathematical complexity, agent based modeling is about emergent behavior from heterogeneous agent interaction. In cellular automata, the agent behavior is actually missing. The predictive power of a free market in an agent based model comes from specifying or relaxing the assumptions of the same model with an environment or parameters that are representing some policies or constraints to the free market. The difference in outcomes of the two simulations would likely give predictive power to either of the models, as long as the other one is well validated by real world phenomena.

[Peter Lewin]

Glad to see the post sparked some interest. I don't want to answer point by point - its not a seminar and I will to some of these issues in subsequent posts, but just add a couple of points.

What Mises really said and meant is actually less important here that the main question I pose - which is: are we better off if human's have "free" (undetermined) choice than if they programmed automata? I actually agree that a careful reading of Mises puts him closer to Hayek than usually believed, in that he allows the possibility that a physiological explanation of human action is at some level possible, but that this would be hopelessly complex (not his words). As a practical matter, this is irrelevant since (as with Hayek) such a model, even if we could comprehend it, would not give us the "understanding" of human nature that we get from using our knowledge of humans as intentional beings - a subjectivist approach. (See Mises, Theory and History and a footnote on this in my 1997 article in J. of Ec. Methodology).

Wolfram: Yes, many of the programs are simple and determinate but a significant minority are not; they exhibit unpredictability. The latter type of process would be intractable as a model for human action.

A theme running below the surface in many of these comments is the question of whether the subjectivism we invoke is "deep" subjectivism or just a practical matter (I guess ontology vs. epistemology). I am not sure I want to get into this.

[Barkley Rosser]

On the El Farol bar model, whatever its faults it is not a model of market instability. If anything it shows self-organizing stability.

[Troy Camplin]

In regards to agent based modeling, I cannot recommend enough Kenrick's work on Dynamical Evolutionary Psychology:

http://homepage.psy.utexas.edu/homepage/group/LiLab/KenrickLiButner2003.pdf

http://www2.psych.ubc.ca/~schaller/Kenricketal2002.pdf

[marris]

@Roger

Googling for "An incompleteness theorem for calculating the future." took me to http://www.santafe.edu/research/working-papers/abstract/cc5f73e4586cc5eba99458f4dc6ea884/, but clicking on the link brings up a PDF titled "Computation capabilities of physical systems" [http://www.santafe.edu/media/workingpapers/96-03-008.pdf].

Maybe the paper was renamed.

[Jerry O'Driscoll]

Mises certainly believed there are no constants in the social sciences. But he has seemingly conflicting positions on whether there are natural laws of social science.

In Human Action (p.761), he states that: "The reality of the laws of praxeology is revealed to man by the same signs that reveal the reality of natural law..." He also states, however, that: "As the idea of natural law is quite arbitrary, such discussions are not open to settlement" (p. 720).

I've tried to rationalize these seemingly conflicting statements. But there is no entirely satisfactory resolution of the conflict.

BTW, I am curious how the date of the comments on Peter's post are dated earlier than the post.

[Allan Walstad]

From "Is Economics the Next Physical Science?" by J. Doyne Farmer, Martin Shubik, and Eric Smith. _Physics Today_ September 2005, page 37:

"The El Farol game was simplified and abstracted by Damien Challet and Yi-Cheng Zhang as the minority game....Moreover, as in real markets, payoffs display clustered volatility and have a distribution with a power-law tail."

That's what I'm talking about, just in case Barkley Rosser was referring to my earlier comments.

[Roger Koppl]

Jerry:

I'm not sure I see where there's a contradiction between Mises's "same signs" comment and his "no constants" comment. The complete "same signs" comment says, "The reality of the laws of praxeology is revealed to man by the same signs that reveal the reality of natural law, namely, the fact that his power to attain chosen ends is restricted and conditioned." There are many possible analogies between physics and economics. Mises denies one and affirms another. He denies that economics has the kind of constants you find in physics. He affirms however, that we must adjust our actions to the "laws" of economics if those actions are to succeed, just as we much adjust our actions to the laws of physics if they are to succeed.

As I always say, the king of France knows that he cannot declare a moratorium on gravity and expect that he will float after pitching himself off the top of a tower. But the king of France (= a literal monarch or the democratic electorate) does not know that he cannot feed the hungry by declaring a new and lower price of bread.

[Roger Koppl]

Jerry, I forgot to mention that several of us have purchased these new tachyon drives for our computers. They are so fast that comments arrive earlier -- sometimes several days earlier -- than the posts they comment on. Amazon.com has the best prices for them, I'd say.

[Jerry O'Driscoll]

@Roger,

The contradiction is between passages affirming and denying the existence of natural laws. He could have denied the existence of constants, and either aiirmed or denied that economic laws are a kind of natural law.

[Roger Koppl]

@Jerry,

"Natural law" means two different things in the two different passages. In the "same signs" passage, Mises is referring to scientific laws. In the "arbitrary" passage Mises is referring to notions of justice rooted in a supposed "natural law" of what is just and unjust. I don't think there's a contradiction between those two passages.

[Josh S]

What do you mean by "certain conditions?" Western capitalism? Because you could argue that the Amish are a "free market," since they (a) recognize property and (b) have very little government control of their communities. I think all you can say about a free market is that people will realize their values and preferences better than they do with continual government intervention, which is just a small group of people imposing their preferences on everyone else.

As to what those preferences are going to be, you really can't say.

[Troy Camplin]

"Self-organization" and evolutionary selection mechanisms are both natural laws without being constants in the same way gravitation is a constant. Simple laws vs. complex rules, I'd say.

[George Machen]

It's not so much that that there are no *constants* in Austrian economics; rather, it's that there are no *conservation laws* in economics, as there are in physics. Therefore, economics does not have a Dynamics — it has no time equations. But it does have a Statics (the realm of what's being discussed in this thread), and there are of course plenty of (innocuous) constants in Statics. (Individual human time preferences & pecking order of valuations not among them.)

Israel M. Kirzner talked about this:
http://oll.libertyfund.org/?option=com_staticxt&staticfile=show.php%3Ftitle=304&chapter=5959&layout=html&Itemid=27#lf0723_endnote_nt359

...and a fuller treatment of this issue in question by F. S. C. Northrop is here:
http://www.jstor.org/discover/10.2307/1883027?uid=3739776&uid=2&uid=4&uid=3739256&sid=56175840633

Sorry, that's a JSTOR link for which all may not have free academic access. So there's this:
http://tinyurl.com/PDFofDubiousCopyrightStatus

Greg Ransom objected to my prior adducing that analysis here:
[Oh drat, Greg's comment link doesn't resolve directly, but his post, beginning with, "The whole 'testing' and 'postulates' picture of 'science' and 'knowledge' is the problem," is among these]:
http://www.coordinationproblem.org/2010/08/some-clarifications/comments/page/1/#comments

...but I'll be interested in his sentiments after completing Northrop's _The Logic of the Sciences and the Humanities_, and whether Greg then finds Northrop a better fit with Kuhn, Hanson, Popper & Hayek. I may be mistaken, but I think Northrop at least partially reconciles Greg's excellent points that he's been making here for years about the perils of formalism — with Austrian economics, Darwin, etc., especially once the significance of the "Copernican twist" in Austrian economics is recognized.

[Peter Lewin]

@Jerry. You are very observant. The comments appearing before the post are an indication of the ability to engage in time travel, thus violating "natural law." My bad. I am learning how to use the dashboard for this site and I inadvertently removed then reposted the post, at which point it reset the date of the post but not the comments.

Further to what Roger said on Mises and natural laws, in Theory and History Mises seems to me to say that he believes everything that happens in the universe may be the working out of certain natural laws, but that this is irrelevant for us social scientists. We need to assume that humans have "free will" if we are to understand them.

[Jerry O'Driscoll]

@Roger,

I rationalized the distinction much as you did. Mises accepted "positive" natural law (science), but not "normative" natural law (ethics).

But Mises' distinction reflects a defective view of natural law, at least in its classical development. Rothbard argues that forcefully in The Ethics of Liberty. Natural Law ethics is based on reason and is science.

@Peter,

Thank you for explaining time travel. I must try it sometime.

Again, I don't think Mises understood Natural Law theorizing. There is no conflict between Natural Law and free will. Natural Law ethics predicts the consequences of bad decisions. It doesn't preclude them.

[Troy Camplin]

Can one actually understand a dynamical process using statics?

[George Machen]

Troy, I don't think that the absence of time equations in economics that permit a deterministic prediction of the future state of a static system's particulars from the configuration of a past state means that we can't understand a dynamical process. If you read my Northrop link, he's just saying that Austrian economics is constrained to more general conclusions about dynamical processes than is possible in physics. Austrian economics is still no small accomplishment.

[Troy Camplin]

Of course Austrian economics is constrained to more general conclusions about dynamical processes than is possible in physics. Physical processes are far, far, far, far, far simpler processes, even at their most complex, than are economic processes. But statics misleads when it comes to understanding dynamical processes just as equilibrium misleads when it comes to understanding far-from-equilibrium states (which economies are in). Austrian economics is the great accomplishment it is because it does in fact take time into consideration. Using agent-based modeling of dynamical processes will only help Austrian economics -- particularly since the former have demonstrated to what degree the latter were already right on many things.

[Roger Koppl]

I'm starting to get confused by some of the methodological claims whirling around on this thread. "Equilibrium," "disequilibrium," "statics," "dynamics." I feel like I'm caught in a kaleidoscope, and it's making me dizzy.

What model or argument would you use to work out what would happen to the price of domestic tradeables if the authorities imposed a general tariff? How does your answer to that question depend on "dynamics," "disequilibrium," "agent-based modeling," "Austrian economics," "uncertainty," or "time"?

[Peter Lewin]

Regarding Mises on methodological dualism.

"What the sciences of human action must reject is not determinism but the positivstic and panphysicalistic distortion of determinism. They stress the fact that ideas determine human action and that at least in the present state of human science it is impossible to reduce the emergence and transformation of ideas to physical, chemical or biological factors. It is this impossibility that constitutes the autonomy of the sciences of human action (Theory and History.: 93)."

"Saying that judgments of value are ultimately given facts means that the human mind is unable to trace them back to those facts and happenings with which the natural sciences deal" (Theory and History: 69).

[Troy Camplin]

"What model or argument would you use to work out what would happen to the price of domestic tradeables if the authorities imposed a general tariff?"

When? Immediately? In the short run? In the long run? In the long run relative to the short run?

Some of the simple questions, like this one, do not necessarily need more complex modeling. Indeed, we are talking about a simple, single feedback mechanism, which gives fairly predictable results.

But suppose I was asking, "How much will the new Nike shoes cost?"

Here time matters. The cost will change over time due to a number of factors, including the fact that one has to take into consideration both diminishing marginal utility and how popular the shoes will be, meaning both negative and positive feedback are involved. And fads, being positive feedback, go through boom-bust cycles, meaning a fad will first drive up price, then drive it down. And that's just on the demand side.

More complex models would also address many of the arguments against economic concepts. For example, the simple models show that a minimum wage creates more unemployment. The economic logic is unassailable. Yet, if you look at the data, you do not see unemployment rising every time the minimum wage is increased -- even if you look two years down the road. I have had people use this argument against the economic argument against the minimum wage. With more complex models, we might be able to show under what conditions the effects of having or increasing a minimum wage are mediated or masked.

Much like an interdisciplinary scholar needs disciplinary scholars, we need the simple models to understand the complex models -- but we need the complex models. The simple models are necessary, but not sufficient.

[Roger Koppl]

Troy:

I'm not what economic theory has to do with predicting the price of Nike shoes. I thought that's why we had entrepreneurs. I asked about tariffs and tradeables. You speak of "simple models." Well, what "simple model" would you use to work out what a general would imply for the price of domestic tradeables? Do we have any models for that in economics?

[Roger Koppl]

what a general tariff would imply . . .

[Troy Camplin]

I was using the price of new Nike shoes as an example of how complex prices really are. One can include many more factors that influence prices. How can one have a proper theory of prices based only on the negative feedback that create the supply and demand curves?

But regarding tariffs, if we include the time element, a general tariff would cause an immediate increase in prices of those goods under the tariff. This would then result in more domestic players entering the market, which would eventually drive prices back down, though likely not to the levels without the tariff. Thus there would be more domestic manufacturers and more domestic jobs in that field. It is this latter that proponents of tariffs focus on, though the consumer is harmed both in the short term and in the long term, even if prices do come back down slightly. There are thus a variety of feedbacks driving variation over time. Opponents of tariffs focus on the short term spike in prices and consumers; proponents of tariffs focus on more domestic production of that good and workers. Recognizing and, indeed, acknowledging all these things are taking place I think helps one to make the argument against tariffs by being able to demonstrate that the costs outweigh the benefits.

I do find it a little odd that you, of all people, are giving me pushback on complexity and agent-based modeling,both of which imply the necessary use of time, Roger.

[Roger Koppl]

You need the right tool for the right job, Troy. My SDAE presidential address (http://www.gmu.edu/depts/rae/archives/VOL19_4_2006/1-Koppl.pdf) was a paean to complexity theory. But near the end of the talk I also said,

---------
there is a sense in which the Austrian school is deeply neoclassical. The Austrian school has never departed from supply and demand, marginalism, opportunity cost reasoning, and the basic theory of markets. The Austrian school has never departed from Caldwell’s “basic economic reasoning.” Austrians should continue to practice basic economic reasoning and to defend basic economic reasoning.
-----------

Bruce Caldwell's 2004 intellectual biography of Hayek is also a paean to complexity theory. But he too comes down hard on the importance of "basic economic reasoning," which I cited in the passage above. In a 2008 article in the Quarterly Journal of Austrian Economics, Peter Klein has emphasized "The Mundane Economics of the Austrian School." (Ungated version: http://mises.org/journals/scholar/klein4.pdf) Peter says, "Without a commitment to preserving and extending the hard core of Austrian price theory, the distinct place of the Austrian School will be lost." Now I suppose I have to admit that Peter and I pick up on rather different threads within the Austrian movement in America. But that makes it all the more meaningful, I think, that he too emphasizes the "mundane" core of Austrian theory.

I'm all for complexity economics. I have published on complexity and on computability. I have defended "Verstehen" and even attempted to link it to mathematical complexity theory. I have written on process analysis and subjectivism and said that they are characteristic of Austrian economics. All these things. And I don't take back a word. But we fall into error, I think, when we make unguarded statements suggesting that good economics is identical with complexity, or out-of-equilibrium reasoning, or radical subjectivism, or anything else that might exclude "basic economic reasoning" and "mundane economics." I don't know if your comments count as examples, Troy, but certainly others have fallen into that error and it is an error we should guard against. Concern over that error motivated my comments that you described as "pushback on complexity and agent-based modeling."

[Troy Camplin]

Nor should one depart from supply and demand, marginalism, opportunity cost reasoning, and the basic theory of market. These are all necessary to understanding how the economy works. And it is a great approximation to what is happening at the microeconomic level, where negative feedback dominates. I grant all that, and do not disagree with it. When I talk to people about economics, discussing why prices go up or down, I most definitely talk about how competition among consumers drives up price and how competition among suppliers drives down price.

But it was the belief that equilibrium prices were a real feature of the economy that made people think that economic planning was possible, that one could calculate what the price of a given good should be. What prevents such calculations from ever being possible, even with quantum computers? The fact that prices are in fact set by bipolar feedback and constant mutual adjustments.

If there are better models we can develop, we should, shouldn't we? But before a better model can be developed, one has to realize that, as good as one's current models may be, they are not as good as they could be. I don't think that such modeling in any way goes against what the Austrians have done -- quite the contrary, I think these models go a long way to proving their points. Complexity models can prove incalculability. Why wouldn't one want that?

[Szayel]

i actually don't see waste and cooutprirn as the big issues that have put us where we are today; instead i see a lot of bad policy.how'd we really get here?2 3/4 un-paid-for wars, a giant not-paid-for tax cut, and that medicare drug program, which is, again, unfunded.and then there's that pesky health care cost inflation problem.underneath that are special tax breaks for special groups, but the reality is that the (unfunded) tax break we give parents for having kids is a lot more expensive than the tax breaks we hear about for corporate jet owners, which apparently are good for about $4 billion.and that's why revenues matter here so much, because in the end, tackling waste and cooutprirn gets you billions but undoing the bush tax cuts gets you trillions, and bending the cost curve on medicare gets you even more money than that.

[Edne]

Wait until Part II. They say there is charity care, but you shloud expect to receive a hefty bill. Then they will demand that you send in all of your proof of income, assets and expenses before they will actually approve the charity care. In the meantime they will harass you with bills. I hope that you will follow up with articles about whether you actually receive the charity care they promised and how much fighting you have to go through to prove it.

[Barbara]

, I don't think it's true that Austrian economics doesn't offer much in the way of mreatnoy theory. Menger's On the Origins of Money and Mises's regression theorem in The Theory of Money and Credit are indispensable. Mises in particular is the link between Wicksellian loanable funds theory and modern understandings of mreatnoy equilibrium. Also, let's not forget Hayek's Denationalization of Money , which was the spark that ignited all the modern research on free banking. These aren't just exercises in history of thought; the ideas in these works are directly relevant to current issues.

[Camilo]

Kedves Ilona!Ez valf3ban egy e9rdekes e9s teljesen faj bekfteete9si terfclet. Magas hozamot csak kocke1zatos terme9kkel lehet ele9rni, de ezekne9l a robotokne1l a kocke1zat kezele9se teljesen faj elveket alapszik e9s megteremti a hosszfa te1vfa, nagyobb f6sszegű pe9nzkerese9s lehetőse9ge9t. Levelet is kfcldf6k a re9szletekkel, de javaslom, hogy tale1lkozzunk szeme9lyesen.