1. Skip to navigation
  2. Skip to content
  3. Skip to sidebar
Source link: http://archive.mises.org/8059/the-dangers-of-samuelsons-economic-method/

The Dangers of Samuelson’s Economic Method

April 29, 2008 by

A memorably painful part of my graduate education consisted of my attempts to read and understand Samuelson’s landmark book Foundations of Economic Analysis (1947), a treatise in mathematical economic theory, patterned after classical thermodynamics, that set the tone for much of what the cleverest mainstream economists would do for decades to come. The protocol became: build a mathematical model of abstract actors engaged in constrained maximization or minimization of an objective function; prove that the model has a stable equilibrium; show how the model’s equilibrium conditions change when its parameters are changed (the so-called method of comparative statics). FULL ARTICLE


Rafe April 29, 2008 at 8:48 am

The most eloquent testimony to the failure of Samuelson as an economist is the way the wrote about the Soviet experience in his popular textbook. Year after year, in successive editions of the book, up to the Fall of the Wall he wrote that the Soviet economy was doing well and catching up with the West.

Two Italian scholars wrote a historical account of the rise of mathematics in economics, these are some reading notes that might be interesting for young players. They were only planned for my own use so they are a bit ragged but they should convey the flavour…

Inquisitor April 29, 2008 at 9:06 am

Painlevé’s contributions to the book ESSAYS IN EUROPEAN ECONOMIC THOUGHT also allegedly constitute a good case against mathematics in economics.

Owen April 29, 2008 at 10:03 am

Did Samuelson every go to Russia? Did he have some special access to russian economic data that probably even the Russians didn’t have? So how was he supposed to know exactly how “well” the Russians were doing relative to the west?

Samuelson looked for a way of determining market prices, not for the purpose of knowing them exactly but of modelling their predicted behaviour.

Seems like Menger (and his followers) just throw his hands in the air and says “This is too hard” and then derides anyone for trying to find a way to explain them.

Equilibruim analysis if nothing else provides a useful conceptual way of understanding the interactions within micro and macro economic environments even though the models have proven particularly lousy in prediction in many cases.

Sometimes what is best theoretically is not useful for teaching and what has shaky theoretical foundations may just be the best way to understand an interaction at the elementary level before going on to more detailed analysis.

Inquisitor April 29, 2008 at 10:13 am

Right, and that is why Menger had an entire chapter in Principles of Economics, for the purpose of throwing up his hands in the air and avoiding explaining prices. Illuminating.

Mises also used the ERE as a means of understanding certain phenomena better… what he did not do was to use this as a model for reality or confuse it with it.

Person April 29, 2008 at 11:09 am

I think the real danger of Samuelson’s method is that it makes testable, falsifiable predictions. I think that real, serious economics should make claims that can’t be falsified by any observations, like the claims of math, because they’re true a priori.

Economics is just too complicated to predict, people should stick to apriorism.

Owen April 29, 2008 at 11:30 am

Menger did not attempt to build a model to predict prices. He threw his hands up and said “this is too hard” – his analysis is based on analogies as complicated as “swapping 40 bottles of wine for 80 sacks of rice”.

Samuelson also “got” that prices are determined by interactions but Samuelson refined models that allowed people to introduce multiple influences on price and predict the directional movement likely to occur.

Menger didn’t do this.

Jim Waddell April 29, 2008 at 11:48 am

PJ O’Rourke (from Eat the Rich) on Samuelson:

And here was another shock. Professor Samuelson, who wrote the early editions by himself, turns out to be almost as much of a goof as my friends and I were in the 1960s. “Marx was the most influential and perceptive critic of the market economy ever,” he says on page seven. Influential, yes. Marx nearly caused World War III. But perceptive? Samuelson continues: “Marx was wrong about many things … but that does not diminish his stature as an important economist.” Well, what would? If Marx was wrong about many things *and* screwed the baby-sitter?

Owen April 29, 2008 at 11:56 am

Is Diego Maradona any worse a Soccer player because he snorted cocaine after he retired?

Was Marx any less right about one thing because he was completely wrong about another?

If your answer is yes, then you need to go back and check your logic 101 textbook again.

What was done was done. Later events didn’t change that.

fundamentalist April 29, 2008 at 12:09 pm

Owen: “Was Marx any less right about one thing because he was completely wrong about another?”

I would be interested to learn just one thing that Marx was right about. So far, I haven’t found anything.

Danny April 29, 2008 at 12:12 pm


You clearly do not understand the Austrian attack on Samuelson. It is a philosophical error, trying to apply the scientific method to a phenomena so complex that trying to do so will lead one to conclusions that are much more often incorrect than correct, evidenced by the constant predictive error in the most praised of the mathematical/Keynesian economists, including Keynes himself.

I would recommend going back a few days and reading the article on Mises and a priori to gain a better understanding of why you’re criticism is off the mark.

Owen April 29, 2008 at 12:12 pm

I dunno either…I don’t subscribe to him. It seems Samuelson did though…

“…Marx was wrong about many things … but that does not diminish his stature as an important economist.”

fundamentalist April 29, 2008 at 12:17 pm

I don’t think Austrians would have objected to the use of math so much is the premises on which the equations were built were more realistic. It’s not so much the math that is the problem as the poor foundations. Equilibrium analysis is a fine teaching tool. Hayek used it and Mises used the similar Evenly Rotating Economy. But the real world is dynamic and never reaches any kind of equilibrium even for a very short time. Mainstream econ goes wrong because it can’t get away from equilibrium thinking and abandon its unrealistic assumptions.

billwald April 29, 2008 at 12:26 pm

“Samuelson maintains that “[e]very science is based squarely on induction — on observation of empirical facts. … [D]eduction has the modest linguistic role of translating certain empirical hypotheses into their ‘logical equivalents’” (p. 57).”

Seems reasonable to me. What has deduction to do with the real world? It is only true if every premise is true.

1. God has ordained a metallic money system.
2. Our money system is faith – based on electronic transfer.
3. Our money system is evil.

Isn’t this the bottom line of Mises Theology>

“One wonders what status Samuelson would accord the Action Axiom — the statement that human beings have ends and use means in conscious attempts to attain those ends.”

And the Sin Axiom? That we are all born in trespasses and sin and that there is no good in us? Does this not define the end result of the Action Axiom?

“Would he deny that it is apodictically certain? And would he also deny that it is empirically valid?”

Anyone deny that the doctrine of our sin nature is also empirically valid? Has there been any improvement in human morals in the last 4,000 years of human history?

Owen April 29, 2008 at 12:32 pm


By your logic most areas of scientific investigation in the world should also be abandoned because they seek to explain phenomena that are equally if not even more complex than economics.

Stranger April 29, 2008 at 1:00 pm

I hate those physicists who throw their hands up in the air and fail to explain how to travel faster than light. I want my physicist to offer models for faster than light travel. That ought to be real science.

Keith April 29, 2008 at 1:09 pm

Quote from Owen: “By your logic most areas of scientific investigation in the world should also be abandoned because they seek to explain phenomena that are equally if not even more complex than economics.”

I’m not sure how economics is less complex than almost anything else.

I would think its one thing to try to model a chaotic system of inanimate objects (e.g., the atmosphere) and quite another thing to try to model a system of animate objects with free will, that often act irrationally, and where new unforeseen variables can be added or subtracted at random.

At some point the later model becomes little more than guessing.

Owen April 29, 2008 at 1:13 pm

…as if discrediting 0.001% of a discipline that has so immesurably benefitted mankind makes any point at all.

Please. Try and discredit the other 99% of physics, chemistry, biology, psychology, psychiatry, geology, engineering, computer science…I could go on all day.

Please name me model(s), even mainstream ones within those disciplines that draw critisizm due to the complex nature of the phenomena they attempt to explain. Don’t worry, i will save you the trouble there are THOUSANDS of models that are disputed.

Hell, they still haven’t decided if light is a wave or if it is particles.

So should we get the eraser and rub them all out because what they attempt to explain is complex?

Owen April 29, 2008 at 1:18 pm


Have you ever tried to:

- predict geothermal or ecological phenomena on a global scale?

- explain what goes on inside the human brain?

- tested the behaviour of diseases within the human body or within a population?

- forecasted the weather?

- modelled electrons and atoms?

We are talking about phenomena in which the constituent parts cannot even yet be totally identified.

Economics is complicated but so are those areas and you yourself have no idea of how much so how can you say economics is more complicated?

James Redford April 29, 2008 at 1:28 pm

Prof. Paul Samuelson was a leading light in the field of masturbatonomics, which is defined as “the discipline of whatever as applied to nothing, and done so in a mathematically precise way.” Critics of the discipline are on solid ground to criticize one of its practitioners if they had found an error of mathematics; but to criticize the discipline on the grounds that it has nothing to do with reality, well, that would just be stating the obvious.

eric lansing April 29, 2008 at 1:41 pm

didn’t Rothbard get his undergrad in Math from Columbia? I think his dad was a mathematician who named him (middle name) after Isaac Newton. Point being, I’m pretty sure Rothbard never threw his hands up and said anything of the sort. He was just too wise to realise that ivory tower math is not useful in economics.

Speaking of ivory tower math, that is all that comes out of the economics departments at the Ivy league schools. google Bernanke, Blinder, Cecchetti etc if you want a good laugh.

Max Raskin April 29, 2008 at 2:32 pm

I wonder, how does Samuelson justify this claim:

“[N]o a priori empirical truths can exist in any field. If a thing has a priori irrefutable truth, it must be empty of empirical content” (p. 62).

The above statement is posited as an absolute truth. What is the status then of this supposed truth? It is stated in such a way that it is a priori…but according the the statement itself, such statements don’t say anything meaningful about the word. So either the statement is without any empirical content or not a priori; either case he does not prove his point. Hoppe often points out this flaw as a basic contradiction of mainstream methodology.

Francisco Torres April 29, 2008 at 3:39 pm

Economics is complicated but so are those areas and you yourself have no idea of how much so how can you say economics is more complicated?

Owen, people make decisions based on their preferences – you cannot predict them. So, trying to model a system based on purposeful decisions that you cannot account does not make sense. The problem with your thinking is that you assume a complex system like human activity or market activity can be compared to complex physical phenomenon where the parts behave according to physical rules and not out of whim.

fundamentalist April 29, 2008 at 3:45 pm

Owen: “Economics is complicated but so are those areas and you yourself have no idea of how much so how can you say economics is more complicated?”

When Mises used the term “complicated” he referred to something very specific. He did not mean that natural science is not complex in the common usage of the term. Today, we might use the word “deterministic” instead, which is closer to what Mises meant. In natural science, all things being equal, light, mass, energy, velocity, all work exactly the same way every time. They can be isolated and analyzed in a lab and many experiments performed with exactly the same results achieved every time if the variables are exactly the same.

On the other hand, in studying human behavior we are studying ourselves. Even with all things being equal, humans often won’t act the same way. We can’t use lab experiments and repeated random tests don’t return the same results. Where natural science experiments might return a fixed result, or results with a very narrow range, experiments with humans can only return a range of results with probabilities attached to each. That’s what Mises meant by complex.

Keith April 29, 2008 at 4:17 pm

Quote from Owen: “Economics is complicated but so are those areas and you yourself have no idea of how much so how can you say economics is more complicated?”

Humans are more than a collection of atoms. (I don’t mean anything metaphysical.) Your examples are models of inanimate objects, not animate acting humans.

Chrysotheras April 29, 2008 at 5:00 pm

My professor of “Political Economy” at the Aristotle University of Thessalonike (30 years ago!) was definitely a Samuelson admirer and proponent of his methodology. No wonder why his lectures were so difficult to understand …
As I read this post I couldn’t help a feeling of indignation from emerging at the thought of countless generations of misled graduates that were “educated” in the Samuelsonian ways, and who perhaps may never come to realise the rot of their pseudo-scientific foundations.
I consider myself as one of the few lucky “victims” that managed at a later stage to escape from this “omphaloscopic” mainstream economic thought. This I owe to the crystalline clarity of Austrian Economics.

Inquisitor April 29, 2008 at 5:27 pm

Owen, perhaps because it was Menger’s goal to explain, and not “predict”, whatever that means. Get your damn facts straight for once and stop infesting this blog with your stupidity and ignorance.

Billwald, do you deny that you act? But that is an action. So much, then, for your silly comments. Stick to the positivist theology, if you will. If you’re not going to be serious about epistemological claims, then don’t waste our time.

Owen April 29, 2008 at 5:50 pm

How did Mises profess to know the limits of the complexity of chemistry, biology and physics when the best practitioners today still cannot claim to do so?

Sounds like he was moving a little bit beyond his area of “expertise”.

How can one then impose a demarcation between the complexities of two areas of investigation, neither of which he knows the limits to the complexities of?

Sounds like a guess to me.

scineram April 29, 2008 at 6:07 pm

I can make any arbitrary choice anytime at will. No one can model that. You can stick your equations up to your ass.

Inquisitor April 29, 2008 at 6:08 pm

Of course, one could neglect that Mises was a polymath and heavily involved in the intellectual debates of his day, in addition to being well connected with various intellectuals in differing spheres, one of which being his brother, Richard von Mises. If anything, he was more well placed than most modern economists or even philosophers of science to make pronouncements on methodological matters.

Eric April 29, 2008 at 6:14 pm

Any math assertion which is shown to make but a single incorrect prediction is provably false. I spent my sophomore year proving theorems by using this technique. In fact, the old reductio proofs seem to be all my instructors ever used.

Therefore, if current mainstream economic models error sometimes in their predictions, then they are false. Just given the current predicaments of our economy, I think we can find plentiful examples of false economic predictions.

On the other hand, if a theory or model makes predictions about something that only exists in theory, then even if it is entirely self consistent, of what use is it if says nothing about the real world. Models are only useful if they can be proven to agree with future reality. Most economic models are limited to agreement with past reality.

I have a theory about an alternate universe where electrons have free will. In this universe, Ohm’s law does not always work. Some electrons drifting down a conductor sometimes complain and so refuse to be fools dancing on strings held by big shot protons – but usually only after talking it over with their neighbors. Now go prove me wrong!

Owen April 29, 2008 at 7:05 pm


So Mises, not being an expert in all other areas of science professed to know more about the complexities and limits of models and scientific understanding that were not to be devised 100 years after his death?

So how then could he draw the distinction between their validity as testable models and the invalidity of economics as testable models? Because they produce many false predictions?

If that is the case then we should throw out most of modern science because there are few theories that can claim to model exactly every known variable in the environment they model because many times they do not even know what these variables are. And yes, often times these models explain behaviour and interactions that are affected by humans and other animals.

Inquisitor April 29, 2008 at 7:18 pm

You just ignored every single pertinent point I made, based on your silly idea that one must be an expert in every single area of science in order to be able to comment on methodology. You hoist yourself by your own petard of course, because this likewise precludes you, by your own standards, from making any statements on methodology whatsoever – as does it preclude Samuelson, Friedman or any other economist, or philosopher of science, or scientist. Mises certainly knew enough to comment on economics, more than someone like you will ever know. I want to know, have you read any of Mises works on epistemology (beyond what the comments here have intimated)? Any of those of his successors? Or, like in every other topic you’ve commented, are you making up shit as you go along? Do you admit then to being a reductionist? Please though, continue making your ignorance plain for everyone to see. We could use a clown like you on this blog.

Owen April 29, 2008 at 7:33 pm

How can one claim something to be more complicated than another when one does not know the relative extent of the complications of either.

Mises claimed that economics was “too complicated” to model but natural sciences are not. How can one make such an assertion about the natural sciences without knowing the full extent of the complexity of it?

The processes inherent in biological and neuroscience dawned on him in his sleep? Please. The best neuroscientists and nanotech scientists do not even claim to know 10% of how the human brain or molecules function…so theses are “bunk” sciences according to Mises?

This “bunk” science have produced many concrete benefits to mankind in recent years. But their models attempt to explain complexity they are only beginning to understand.

How is this unknown complexity any different from the unknown complexity in economics? How can one claim to know the difference between two things if one isn’t even able to explain what they actually are?

Owen April 29, 2008 at 7:35 pm


Newsflash. None of your points were pertinent. They were laughable.

Danny April 29, 2008 at 8:03 pm


Mises does not call other sciences ‘bunk’ sciences. Sciences like biology, neuroscience, etc., will eventually yield correct models. Economics will never yield a correct model, not because it is necessarily more ‘complex’, but because it involves millions/billions of unpredictable actors. Every moment, each of these actors are absorbing information (hopefully) which will then influence how they act next. Some will act rationally, some willl act irrationally, but it will not be predictable. Austrians do not attack physical sciences that properly apply the scientific method where it is appropriate, but rather only when the scientific method is applied inappropriately, and its results are taken as scientific certainty.

D. Saul Weiner April 29, 2008 at 8:20 pm

“Mises claimed that economics was “too complicated” to model but natural sciences are not. How can one make such an assertion about the natural sciences without knowing the full extent of the complexity of it?”

For one thing, Mises observed that in all of the history of economics, nobody had ever discovered any constants, unlike in the natural sciences. That would seem to tell you something, if you were paying attention.

Petrochemist April 29, 2008 at 8:49 pm

Physical science also uses deduction from apriori axioms as well as postulates to frame hypotheses. This must be done before experiments can be planned. Without some apriori understanding experiments are impossible.

Physical science differs from economics and the other social sciences in the constancy of the parameters in the differential equations of our models. Economics has no such constant parameters, since human tastes and conditions are constantly changing. A good example of this inconstancy of parameters is a recent paper saying that the elasticity of oil demand has changed greatly since the seventies.

If another scientist said that the rate coefficient for the reaction of hydrogen and oxygen had changed over the last twenty years, all scientists would think him crazy, since the constancy of physical relations and constants is the prime apriori axiom of all science.

When I first encountered neo-classical economics I called it the science fiction school of economics. The pretence of being a positive science and the misuse of mathematics makes economics and other social sciences a laughing stock in the meetings of “real” scientists. Just evesdrop on the discussions in any academic physical science department to get an earful of ridicule.

Owen April 29, 2008 at 9:09 pm


That is a nice explanation, thanks. You seem to know what you are talking about.

Ron Brown April 29, 2008 at 9:25 pm

Thanks for your comment, I hadn’t thought about the contradiction at the time I was reading it.

To others: There’s no such thing as an irrational act, provided the person is capable of purposeful action. You may not like what someone else does, even if it’s criminal, but that doesn’t make it irrational.

Vedran April 29, 2008 at 10:41 pm


I think there are three main problems with mathematical economics.

1. Non repeatable experiments

2. Econometric studies which contradict each other constantly.

3. Sticking to math models that are obviously flawed . (professors are not truly put to the market test and believe in ideas long after any possibility of correctness has been dispelled)

The last I believe is one of the most important. Professors too often try to squeeze reality into a model rather than trying to model reality. Any excuses are deemed acceptable: ex. this model works if the elasticity of labor is 3, this one works if the relative risk aversion coefficient is 25, this one works if wages are completely flexible, this one works if people become more productive with higher wages, etc.

Instead of approaching problems by saying “let’s try this model, it failed. let’s try another”. The current approach is “let’s try this model and if it fails let’s spend our collective efforts trying to excuse the particular failure of the model for the next ten years.” Models in themselves are almost more respected than their applicability to reality. Certainly many are impressive but this shouldn’t be the goal of economists.

Owen April 29, 2008 at 11:22 pm

I must admit that I was fed an aweful number of models and letters and graphs at university micro and macro which meant little to me 5 minutes after I left the final exam.

Rafe April 30, 2008 at 1:13 am

Owen, it was not necessary to wait for the fall of the Soviet empire to realise the futility of socialist interventions in the economy, Samuelson could have learned from the New Deal that locked the US into depression to the end of the 1930s. But of course he was intoxicated by the trifecta of Keynesianism, the welfare state and mathematics.

Koen Deconinck April 30, 2008 at 6:45 am

Owen, I’d suggest you read Mises’ introductory chapters to Human Action (the first one-hundred pages or so), where he explains his reasons to reject positivism and where he elaborates his distinction between ‘theory’ and ‘history’ and explains the role of economics as an a priori logic – praxeology.

Also, Hans-Hermann Hoppe’s booklet “Economic Science and the Austrian Method” provides a good exposition and defense of the Misesian position.

Ofcourse, Mises’ “Theory and History” is devoted entirely to this topic. There’s more depth to it than just a disdain of maths and statistics.

Inquisitor April 30, 2008 at 7:24 am


Newsflash. None of your points were pertinent. They were laughable.”

Does anyone need MORE proof that you’re a troll? Confronted with a direct refutation and questioning of your self-refuting arguments, you proceed to saying my points were “laughable”. You sawed the very branch you were sitting on, so of course you’d want them to be “laughable”. Learn how to argue and come back when you’ve actually grasped the essence of epistemological/methodological debates. It has already been explained to you in what sense economics is “complicated” (that is to say, teleological) that the natural sciences are not (they are strictly deterministic.)

fundamentalist April 30, 2008 at 8:46 am

Petrochemist, Nice explanation! Thanks!

Inquisitor April 30, 2008 at 9:26 am

I have to wonder though, why it was necessary for Petrochemist to elaborate on this, when another article released just the other day gave the reasons for Mises’ position, and left little room for misunderstanding Mises.


The entire section entitled “Mises’ Methodological Dualism” expounds his rationale for his position, as Petrochemist graciously summarized here. But he shouldn’t have had to.

Paul Marks April 30, 2008 at 10:25 am

Owen “predicting prices” (in the sense you are using these words) is no more the work of an economist than predicting the future of individual human beings is the work of an astronomer – as opposed to the bunk “science” of astrology.

As for Karl Marx – his doctrines are a tissue of absurdities. Karl Marx added nothing true to the subject of economics.

To return to Samuelson:

Actually his claim that all science is virtually nothing but induction not only contradicts the Austrian school – it contradicts such thinkers as different from Ludwig Von Mises as Karl Popper.

Karl Popper was a great defender of the empirical method in the natural sciences. Which, perhaps, led to his friend F.A. Hayek having rather different opinions to Mises as regards the methods of work in such non natural sciences as the study of human agency in relation to scarce resources that is economics.

However, even Karl Popper made clear that it is deduction (human reasoning) that is important – even in the natural sciences.

Of course, as the article points out, Samulason also contradicts himself.

He bases his support for induction on his supposed respect for physical reality – but rejects Carl Menger real respect for reality, and instead embraces mathematical abstractions.

Perhaps because such mathematical abstractions offer Samuelson such fool’s gold as the magical ability to “predict prices” (in the sense that Owen uses these words).

David C April 30, 2008 at 10:50 am

If I could interject in the dialogue between Petrochemist and Owen, perhaps I can add this:

the problem with mathematrically rigorous economics a la Samuelson is that people, the elementary particles of economic science, are not like the elementary particles of the physical sciences, whereby each electron is identical to another and each proton is identical to all othe rprotons. Neoclassics, whose predecessors were originally inspired by the post-Newtonian rigour of the physical sciences, has built a fantastic ( and it must be said, quite beautiful and internally self-consistent) scientific model of the economic world, by the simple expedient of assuming away all of the parts of humans that make them human ( and for what its worth, all uncertainties and information acess differentials), leaving them ciphers with no discretion or choice, each constrained to behave in EXACTLY the same way in response to any particular circumstance or change decreed by the modeller.

Unfortunately, as rigorous and technical and mathematically pure as it is, it simply doesnt work. It is a model that describes a lot of – Im not sure what – but the one thing it doesn’t describe is economics, namely human interaction to mutual benefit!

Applying clever mathematics and ‘scientific’ rigour to anything involving lots of human brains making decisions yields nonsense. Just like the statistical average quoted to 4 decimal places, its precision is entirely bogus, and can’t even be tested empirically anyway, for the simple reason that any real-world data result cannot be used to invalidate or confirm the predictions of the model, because the very precisely defined assumptions demanded by the model itself cannot be replicated in any experiment involving human brains making choices!

By contrast, Austrian economics, as developed by Von Mises and his successors has only one central axiomatic assumption: At any point in time, a person faced with a set of circumstances, does exactly what seems like a good idea to him at the time ( OK so MIses wouldn’t have put it quite like that, but you get the point). That’s it. The entire Austrian edifice is built on that one ‘assumption’, which puts human beings and the choices they make at the centre of this very human science, rather than the Neoclassical approach of eliminating them!

Inquisitor April 30, 2008 at 11:32 am

That’s it’s the a priori element in it, yes. In practice, concept-formation also involves some non-logical assumptions, e.g. the disutility of labour. I’ve posted this paper up here several times, but it’s a great explanation of Austrian methodology from an Aristotelian vantage point:


Inquisitor April 30, 2008 at 11:35 am

that is*

Comments on this entry are closed.

Previous post:

Next post: