As an economist who’s spent most of my career trying to understand how a free-market monetary system would work and how we ended up with central banking instead, I’ve naturally been keenly interested in the reactions of economics bloggers to the questions Jon Stewart asked of Alan Greenspan on Tuesday:

Why do we have a Fed? Is the free market – wouldn’t the market take care of interest rates and all that? Why do we have someone adjusting the rates if we’re a free-market society?

(My homemade transcript of the interview is here. The video clip is still available on Comedy Central for the time being.)

A colleague emailed me to ask whether I’d written Stewart’s questions. Actually, I suspect they were drawn from Brad DeLong’s review of Greenspan’s book in the L. A. Times. After describing how the Fed conducts monetary policy, DeLong asked:

Isn't this odd? Don't we have a market economy? Why should a central planner be setting interest rates?

Greg Mankiw’s remarks on the interview are particularly interesting. My thanks to those who cited my work with George Selgin in the comments section of Mankiw’s blog, but the fact that Mankiw was unaware of that work (or thought it not worth mentioning) is a sobering commentary on how little impact we’ve had, even on the free-market side of the mainstream, on the profession’s thinking about the question of why we have a central bank.

Below the fold I comment on Mankiw’s remarks in detail.

Mankiw writes:

Alan's answer is not satisfying, but I don't blame him: The economics profession does not have a good answer.

We economists have rigorous and fundamental theory to explain why we have environmental regulation (externalities) and to explain why we have antitrust laws (market power), but there is no consensus about what market failure calls for the existence of a central bank. The answer has something to do with the benefits of a system of fiat money. And it has something to do with the possibility of short-run monetary nonneutrality (due to sticky prices and/or imperfect information about prices). But the precise combination of elements that would yield a satisfying answer is still elusive.

Stewart stumbled upon a fundamental question of monetary economics. If anyone has a good answer, let me know, or publish it in the American Economic Review.

There are really two questions here:

(1) Is there a market failure in a free-market money and banking system that a central bank could in principle remedy? (I.e. do we need a central bank?)

(2) Why, historically, did we get central banks?

Mankiw seems to take it for granted that a “yes” answer to (1) must be the key to answering (2). But that would be a mistake. Logically, we may have gotten central banks for other reasons. Just because they exist, does not mean that they must be optimal.

The historical reason we have central banks have in fact nothing to do with any free-market failures. Parliament historically gave the Bank of England the privileges that made it a central bank for fiscal reasons: a quid pro quo for lending the government money. Congress passed the Federal Reserve Act to remedy the panics under the predecessor National Banking regulatory system, panics that were not free-market failures but the result of legal restrictions (on branching and note-issue) that inadvertently weakened the U. S. banking system. Consider Canada, by contrast: no such restrictions, no panics, no call for a central bank. When Canada established the McMillan Commission as a preliminary to creating a central bank for nationalistic reasons, in the 1930s, the bankers on the McMillan Commission actually opposed it.

My own long answer to (1), examining the market-failure arguments, can be found in chapters 5 and 6 of The Theory of Monetary Institutions. The short answer is “no”.

Mankiw speaks of “the benefits of a system of fiat money”. But those “benefits” are only potential, and in practice need not be positive. The resource costs of a silver or gold standard with free banking are in fact less than the deadweight costs of inflation have been under fiat money standards. (See: Theory of Monetary Institutions, chapter 2.)