Monday, September 6, 2010


Leafing through Milton Friedman's Free to Choose. In Chapter 9 Friedman writes:

Financing government spending by increasing the quantity of money looks like magic, like getting something for nothing. To take a simple example, government builds a road, paying for the expenses incurred with newly printed Federal Reserve Notes. It looks as if everybody is better off. The workers who build the road get their pay and can buy food, clothing, and housing with it. Nobody has paid higher taxes. Yet there is now a road where there was none before. Who has paid for it?

The answer is that all holders of money have paid for the road. The extra money raises prices when it is used to induce the workers to build the road instead of engage in some other productive activity...

What about the possibility that the workers who built the road didn't do it instead of some other productive activity. What if the economy was slow and they were out of work, engaged in no productive activity at all? Friedman's logic of 'instead' is a lot like Fama and Cochrane's logic of would otherwise. It builds generalizations on top of faulty assumptions.

But what bothers me most about Friedman's statement is that he presents us with a static economy, an economy not growing. If the government builds a new road, that's infrastructure. One can imagine communities arising along that road. One can imagine growth. That is, after all, the whole reason for having infrastructure in the first place.

Milton Friedman is famous for his view that inflation is linked to "the quantity of money relative to output." But in order to make his road-building example work, Friedman ignores the normal expectation that the economy will grow. And he ignores the commonsense rule that says expansion of infrastructure encourages growth.

In order to make his road-building example work, Friedman must ignore the growth of output. He must assume a static economy. Because if output increases, an increase in the quantity of money is justified, according to Friedman's own rules and standards.

But it was years ago that I was troubled by Milton Friedman's assumption of a static economy. The reason the "static argument" arises now is that I find the same argument in the work of the father of Ricardian Equivalence. No, not Ricardo.

In Chapter 14 of Robert J. Barro's Macroeconomics: A Modern Approach, Chapter 14, we find a table of U.S. and U.K. debt, a couple graphs of debt, and lots of copyright warnings.

Beginning on page 349 there is a simple example of "Ricardian Equivalence." This example begins with several simplifying assumptions:

  • interest rates are stable
  • prices are stable
  • the quantity of money is stable
  • "real transfers" are zero
  • the government starts with no debt, and
  • government spending plans, whatever they are, do not change in this example.

Note that among the assumptions we do not explicitly find the following:

  • population is stable
  • technology is stable
  • productivity is stable
  • output is stable

Barro's simple example is a two-year run. In the example, government cuts taxes for one year, then in the second year raises them enough to pay the principal and interest.

But the example does not allow for economic growth. If the economy grows enough, there is no need to raise taxes in the second year. If we assume the economy doesn't grow, then taxes have to be raised or revenue won't cover expenses. But the normal, everyday assumption is that the economy will grow. Barrow has to go out of his way, avoiding the normal presumption of growth, in order to make the example work.

Of course, in the real world, the economy tends to grow more slowly than the deficit, so that taxes are liable to increase. And it is true that slow growth is a key part of the economic problem of our time. But our topic here is not slow growth. Our topic is the presumption of no growth, which is a trick too often used to make bad examples work.

Barro's example is invalid because it fails to consider the possibility of growth. Milton Friedman's example is invalid for exactly the same reason.

No comments: