Rothbard’s Mistake


Being interested in the history of the 1930s, I recently picked up a copy of America’s Great Depression by the influential libertarian Murray Rothbard (1926–1995). I choked on the introduction, where Rothbard lays out his theory about theory, which makes no sense to me.

“This book rests squarely on the Misesian interpretation of the business cycle,” he writes, referring to the theories of the older libertarian economist, Ludwig von Mises (1881–1973). “Note that I make no pretense of using the historical facts to ‘test’ the truth of the theory. On the contrary, I contend that economic theories cannot be ‘tested’ by historical or statistical fact. These historical facts are complex and cannot, like the controlled and isolable physical facts of the scientific laboratory, be used to test theory . . . The only test of a theory is the correctness of the premises and of the logical chain of reasoning.”

You have to keep in mind that the map sometimes lies, or maybe tells you a truth different from the one you need to know.

Philosophers make a distinction between statements that are valid and statements that are true. Validity is like math. It’s about logic. If P then Q. It’s theory. Truth is about what’s real, which is not the same thing. Logic is useful, but ultimately what we care about is what’s real.

I am reminded of the accounting classes I took many years ago. I gave up on accounting, but one thing has stuck in my mind: the professor described accounting as a map of the “territory” of a firm, and warned us not to confuse the map with the territory. The “map” might say the company is making money, but the truth might be that it runs out of cash before the owners are paid. (As a business journalist I wrote about some companies like that.) The map is useful; to steer the company you need the map. But you have to keep in mind that it sometimes lies, or maybe tells you a truth different from the one you need to know.

Back to Rothbard. He says that an economic theory is “a priori to all other historical facts.” It can be used to explain the historical record, but it cannot be tested. Here is his argument:

Suppose a theory asserts that a certain policy will cure a depression. The government, obedient to the theory, puts the policy into effect. The depression is not cured. The critics and advocates of the theory now leap to the fore with interpretations. The critics say that failure proves the theory incorrect. The advocates say that the government erred in not pursuing the theory boldly enough, and that what is needed is stronger measures in the same direction. Now the point is that empirically there is no possible way of deciding between them. Where is the empirical “test” to resolve the debate? How can the government rationally decide upon its next step? Clearly, the only possible way of resolving the issue is in the realm of pure theory — by examining the conflicting premises and chains of reasoning.

This strikes me as piffle. There are several ways of deciding between the two claimants. You can compare what happened at times when the policy was imposed with what happened at times when it wasn’t. You might compare the depression of the 1930s with the depressions of 1920–21 or 1893–97 or 1873–79, etc., and see that the one in the 1930s featured the slowest recovery in US history. That is evidence (not proof) that whatever policies were tried didn’t work too well. You can dig deeper. How did investors, entrepreneurs, company managers, workers, and other people in the 1930s respond to the National Recovery Administration? To mass unionization? To the retained-earnings tax? To the abandonment of gold? What did supporters and opponents predict the players would do, and what did they do?

Robert Higgs asks these kinds of questions in Depression, War and Cold War. You can reject what he does — none of his arguments amount to a drop-dead test such as you find in a chemistry lab — but they are ingenious. They are instructive. They make a case.

The social life of humans is more complicated than a test tube.

Rothbard argues, in essence, that such questions are too messy to answer. A theory cannot be “tested” in the way a question in chemistry can be “tested” by heating compounds in a test tube. He’s right in thinking that you can’t test that way with economic policies, but it doesn’t mean that “empirically there is no possible way of deciding between them.” You can look at what lawyers call “the preponderance of the evidence.” “Test” is a high-hurdle word, the wrong word. You can evaluate. You won’t get to 100% certainty, but it’s unlikely that you’ll be stuck at 50-50, either. You can decide, but you have to look at the territory as well as your map — and you may find yourself correcting your map to make it fit the territory better.

Essentially Rothbard denies this.

“Clearly,” he asserts, “the only possible way of resolving the issue [of choosing the best economic policy] is in the realm of pure theory — by examining the conflicting premises and chains of reasoning.” In other words, the only way to decide what to do “in the territory” is to pick the best-looking map without looking at the territory.

No, no, no! Because the social life of humans is more complicated than a test tube, and because cause and effect are mixed up and piled on each other, you have to check your “map” against the territory all the time. Because your theory is only an approximation. A simplification. It is not life.

Praxeology is not primary. Supply and demand curves are not reality.

To quote the philosopher Robert Heinlein: “What are the facts? Again and again — what are the facts?”

If you say, “I don’t care about what facts you have. What experiences, or what statistics, or anything. I have my theory, I’m sure it’s right, and I don’t need to ‘test’ it,” you become irrelevant. You become ignorable. You become the frog at the bottom of the well.

Share This


David M. Brown

Ramsey confuses the character of fundamental economic principles with the job of applying them to particular historical circumstances. Harrison confuses Austrian-school economic method with that of economists who want to copy the method of physics.

Rothbard did not argue that one derives economic principles without reference to reality, including what we observe of the nature of human beings and what human action consists of. But does Ramsey see no difference between the fundamental fact that human beings purposefully apply means in the pursuit of ends, on the one hand--and the requirements of judging exactly what interplay of human actions caused a particular drop in prices of commodity X on day Y in human history, on the other? Must we re-discover from one historical moment to the next all relevant facts of human nature, the nature and requirements of production, etc? Must we, every time we investigate a new historical event, learn and prove from scratch--and using only the historical facts we are seeking to understand--that a precipitous drop in supply of a good will, all other relevant economic factors remaining the same, lead to either no change in the price of that good or a rise in the price of that good? Have we no way of knowing that if the prices of oranges drop instead of rise in the weeks after a frost kills half the crop of oranges, that some countervailing change affecting supply of or demand for oranges in addition to the frost must have occurred, and that we can know this even before we find reporting on plausible candidates?

If it is granted that we do know something about human nature and the nature of economics before when we go to analyze specific historical facts, then Rothbard is vindicated in his contention that fundamental economic truths applicable to specific historical facts can be learned independently of knowledge of specific historical facts--but not independently of facts per se. Rothbard argues explicitly that observation of reality is required to arrive at the fundamental truths of economics. The interested reader may pursue his elaboration in Man, Economy, and State.

Jon Harrison

Both Rothbard and Ramsey are wrong, or at least partly wrong, for the simple reason that economics is a non-science founded on shaky, ever-shifting ground. "Facts" change over time in economics, and it's impossible to demonstrate that certain actions or policies are right for all societies at all times. The human factor overrides all proposed economic "laws" at some time and in some places.

Economists of course will never admit this, because they yearn for their discipline to be taken seriously as an exact science like physics or chemistry. They create formulae and equations because they crave the same respect that the physicists get on campus and in the news. The Nobel Committee's decision (opposed by the family) to create a prize for economics exposed the nonsense for the whole world to see. Both Samuelson and Friedman got the Nobel, proving that economics is the fuzziest "science" ever conjured up by the intellectual-academic class.

Scott Robinson

Dear Jon,

Well stated about the problem of calling economics a science. Economics is based on choice by individuals who compose the economy. These individuals may choose different responses to the exact same stimuli. Ramsey's point that it's more like a model that needs to be frequently evaluated is good. Just like his accounting reference, a profit and loss report just puts chances that placing your money into a certain endeavor will result in you earning a profit. It is not guaranteed.

Good Comment,

© Copyright 2020 Liberty Foundation. All rights reserved.

Opinions expressed in Liberty are those of the authors and not necessarily those of the Liberty Foundation.

All letters to the editor are assumed to be for publication unless otherwise indicated.