Contra General Theory Pro "The General Theory …"

Robert Waldmann

In the post below, I totally humiliate myself by arguing that they way to appreciate “The General Theory of Employment, Interest, and Money” is to totally reject the idea that economists should aim for general theories.

So which is it ? Do I reject “The General Theory …” or do I accept general theories ?

I think “The General Theory …” is a wonderful contribution to human understanding. However, I disagree with the title. Keynes doesn’t use the word “Theory” the way it is used by today’s economists (so what ?). More importantly he doesn’t use the word the way natural scientists and philosophers of science use it.

The word is now generally used to refer to a hypothesis which has survived many confrontations with the data — that is a statement which is falsifiable but not yet falsified. “The General Theory …” does not purport to present a model which can fit all macroeconomic data and which could be falsified by, say, post WWII US time series. Keynes specifically says that he can’t make predictions on a key variable, that he thinks it is a mistake to aim for a simple equation with that variable on the left hand side. The variable is nominal wages.

For most of the book, he divides all other variables by the nominal wage level. This book ends with a conclusory chapter 18 “The general Theory of Employment Re-Stated.” Then he goes on to discuss the things he left out — nominal quantities not quantities expressed as multiples of the wage level. When discussing this in Chapter 19 “Changes in Money Wages,” he notes that, in his incomplete model, a decline in nominal wages is just like an expansion of the money supply.

In 2 other chapters (including chapter 14 Appendix on the Rate of Interest in Marshall’s Principles …) he notes that, in a liquidity trap, neither increases in the money supply nor reductions in money wages cause increased real GNP. Now this is a special case, and it is very irritating for economics to depend on something silly like the real return to money hidden under a mattress, but here we are. A nice general model should not have “regimes” that is cases which are very different depending on the exact value of a variable (here is the safe short term nominal interest rate zero or not). Yuck. Even Keynes can barely stand the fact that the theory for a liquidity trap is not at all like the theory for normal times. But here we are in a liquidity trap. The problem is that two models with quite different properties are needed for trapped and untrapped liquidity. Keynes himself almost seems eager to hide the discussion of the liquidity trap in the most obscure possible part of his book.

Sorry for the digression. Back to Keynes on the determination of nominal wages. We get to chapter 20 “The employment function” which includes the best footnote in the economics literature “Those who (rightly) dislike algebra will lose little by omitting the first section of this chapter.” Keynes never wrote truer words (well maybe “nothing” would be better than “little.” The algebra consists of the definition of elasticity, defining elasticities, and deriving equations from those definitions. There are equations but there is no content. The only point is to set up a warning about what not to do — assume an elasticity is approximately constant because you have defined it. The second and valuable half of the chapter is devoted to arguing that the pointless equations are not useful at all, because the painfully defined elasticities are not constant.

One of those pointless equations is better known as the Phillips curve. Keynes specifically notes that it is vertical in the case of “flight from the currency” by which he clearly means hyperinflation. He argues here and elsewhere that it is unwise to hope for a simple relationship between anything else and the wage level. Basically he specifically and repeatedly warns against treating the Phillips curve as a causal law or a menu of options open to policy makers.

Thus the failure of the Phillips curve in the 70s is just one more triumph of Keynes. He warned that there was no simple theory of the wage level and that adding it to the General Theory (implicitly the IS-LM model) would lead to false predictions.

Then when things happened exactly as predicted by Keynes — or rather the thing which he said was unpredictable turned out to be unpredictable — the economics profession decided that Keynes’s general theory had been refuted by the data and it was necessary to throw away “The General Theory …” and start over.

Sargent, for example, looked at the case of hyperinflations and noted that that the claim in “The General Theory …” was exactly right and concluded that this showed that Keynes was wrong. Sargent is rather honest and I assume that he didn’t remember The General Theory of Employment Interest and Money or that he never got to chapter 20 (entirely possible I don’t think Sargent could stand to read the book).

Keynes managed to avoid being wrong by not presuming to explain everything always. He was modest enough to admit that some topics were not, or not yet, amenable to theory.

So you see the condemnation of general theories is consistent with praise of “The General Theory …” if you admit that Keynes’ rare virtue was his unusual humility.

Huh ?!? What ?!?!?! That’s crazy too, but I’ve already argued one oxymoron in this post. The humility of Keynes will have to be the topic of another post.