Samuelson on Whether Economics Is a Science

Paul Solman: More excerpts today from an interview I did with economist Paul Samuelson nearly a decade ago. Today, his thoughts on whether economics is a science and where math fits in the discipline. Click through to read the transcript.

SOLMAN: Is economics a science?

SAMUELSON: Economics is not an exact science, it’s a combination of an art and elements of science. And that’s almost the first and last lesson to be learned about economics: that in my judgment, we are not converging toward exactitude, but we’re improving our data bases and our ways of reasoning about them.

SOLMAN: I asked [philosopher] Nelson Goodman this question once, and Nelson Goodman said economics is as much of a science as physics. I said, well, how could that be? He said: Physics can explain how a leaf falls from a tree and everything that happens to it, but it can’t tell you where the leaf’s going to land. Economics is the same.

SAMUELSON: I think that it’s more important for an economist to be wise and sophisticated in scientific method than it is for a physicist because with controlled laboratory experiments possible, they practically guide you, you couldn’t go astray. Whereas in economics, by dogma and misunderstanding, you can go very sadly astray.

SOLMAN: So you need to know more, in some sense. But is Nelson Goodman right?

SAMUELSON: Yes, absolutely right. I use that all the time. The parts of physics that are exact are the parts of physics that are exact. The parts that are inexact are vastly greater. Sensible scientists don’t waste their time pushing against doors that endlessly will not give. They are opportunistic and go where they can, but there are pitfalls in that. It’s like the drunk who is looking for his keys under the lamplight in the street, but he wasn’t near the lamplight. He said, yes, but that’s the only place I can see anything, therefore I’m looking here.

SOLMAN: Is math overdone in economics, particularly at the introductory level? Because that’s where so many people seem to get stuck.

SAMUELSON: When I agreed to do an elementary textbook, primarily to serve our students, I chose not to make it a heavily mathematical thing, because the MIT engineers were very good at the routine math, but they wouldn’t see the forest for the trees. Everything would just become a little homework problem of how to mix two gasolines so as to get the best diesel fuel, and I wanted them to see the big principles involved. So there’s a mathematical substructure, but it’s like armor worn under clothing; it doesn’t show.

SOLMAN: But let’s say we were talking about high school students, or even junior high school students. The concept of opportunity cost, the concept of putting a price on everything that you do, because there’s some utility, the concept of diminishing returns — everything we’ve talked about, you could teach, presumably, to an 8th grade kid, and it would be tremendously useful to that kid. Does math, then, get in the way?

SAMUELSON: It can, and if one was going to study only one year of economics, I don’t think it should be a heavily mathematical one, unless you happen to be extremely adept at it. There is a big problem of the profession becoming too technical. What’s remarkable to me is that the best people on qualitative important social policy questions, the Art Okuns, the Jim Tobins, the Milton Friedmans, and so forth, all happen to be very good in the mathematical side. They’re very rare exceptions these days. Now, this is not relevant to your question, but if I were advising a careerist, going into economics today, I’d say get a lot of math in while you’re young because you’ll never get it when you’re not young, and you can only fight fire with fire.

SOLMAN: But if I’m trying to speak to a broader and broader base of the pyramid they’re scared by the math? [I was putting together a group of introductory economics videos to accompany McGraw-Hill textbooks like Samuelson’s — a DVD of them is available to buy on the web, even “used” ones. They’re all the same.]

SAMUELSON: I don’t want to throw any wet cloth of doubt on you, but you may end up with a group of listeners who have a pretty good understanding of each of your little topics, but can’t really put it together, in the whole picture.

SOLMAN: Do you need math for that?

SAMUELSON: No, but when you want to have closure of a set of arguments — I’m not talking about pedagogy now — being able to write it down mathematically is very convenient.

SOLMAN: Convenient, to be sure, but necessary?

SAMUELSON: Well, my belief is that nothing that can be expressed by mathematics cannot be expressed by careful use of literary words. All the mathematics itself originally grew out of axioms — assumptions stated in words — and [it reminds me of] the only speech that the great Willard Gibbs, at Yale, ever gave at a faculty meeting, when there was a debate between studying languages and mathematics. He said four words: “mathematics IS a language.” And that’s on the frontispiece of my Foundations of Economic Analysis, but I’ve written somewhere that he should have cut it down to three words: “Mathematics is language.” It’s just a very compact form of language.