Our friend Nick Rowe pays homage to Milton Friedman in one of his latest posts on what monetary policy cannot do. Indeed, Friedman’s speech to the AEA in 1967 should be required reading for any who wish to learn about monetary policy (it is indeed required reading for my Money and Banking students). The purpose of this homage is to absolve monetary policy of any wrong-doing in the current recession and preceding housing boom. Specifically, he argues:
It was in that paper that Friedman introduced the concept of the natural rate of unemployment. Prior to Friedman, most economists thought there was a downward-sloping Phillips curve, and that monetary policy could keep unemployment low provided we were willing to accept higher inflation. Friedman argued that this was true in the short run only, and that in the long run, when expected inflation equaled actual inflation, the Phillips curve was vertical. Monetary policy could target any rate of inflation, but the result would be the same long run equilibrium rate of unemployment.
Friedman needed a name for that long run equilibrium rate of unemployment, and he chose to call it the “natural rate of unemployment”. He chose that name to draw a parallel to Wicksell’s concept of the natural rate of interest.
There is nothing special about unemployment or interest rates in Friedman’s argument. Everything he says about them applies equally well to any real variable. Just as there is a natural rate of unemployment, and natural rate of interest, so there is a natural rate of output, employment, real wages, relative price of houses, or relative price of sardines. The underlying vision is one of the long-run super-neutrality of money.
Monetary policy has real effects in the short run, because it takes time for prices and expectations to adjust to a change in monetary policy. But if we compare alternative worlds where prices and expectations have adjusted to alternative monetary policies with different average money growth rates and average inflation rates, real variables should not be affected. They are pinned down at their natural rates by real, not monetary forces.
I am in agreement with Nick on nearly every point of this argument. Nevertheless, I am puzzled regarding his conclusion about relative prices. When discussing long-run superneutrality of money, he is referencing the idea that a change in money growth will only cause a change in the rate of inflation in the long run and thus have no effect on real variables. However, even accepting long run money neutrality, isn’t it possible (and in all likelihood probable) that relative prices have changed? In fact, this was Hayek’s main point in extending Wicksell’s idea of a natural rate of interest. In Prices and Production, for example, he writes:
. . . it seems obvious as soon as one once begins to think about it that almost any change in the amount of money, whether it does influence the price level or not, must always influence relative prices. And, as there can be no doubt that it is relative prices which determine the amount and the direction of production, almost any change in the amount of money must necessarily also influence production.
The appropriate question is thus whether superneutrality of money not only implies that the change in the rate of inflation is proportionate to the change in the rate of money growth, but also that individual price changes are equiproportional. Nick seems to believe that it is the case, whereas I find this conclusion wanting.
There is no greater illustration of our opposing views than the idea of inflation targeting. Nick argues that monetary policy did not play a role in the Canadian housing boom:
Did a change in monetary policy cause the house price bubble? In Canada, absolutely not. There was no change in monetary policy in Canada. As I argued in my previous post, The Bank of Canada hit its inflation target almost exactly on average over the period when Canadian house prices were rising. With actual CPI inflation at the 2% target, and expected CPI inflation presumably at the same 2% target, there was no sign of the unexpected inflation that is the signature of the short run effects of a change in monetary policy.
The inflation target was 2% and actual inflation was 2%. Nick views as suggesting that monetary policy could not possibly be to blame for rising house prices. I do not find this evidence convincing in the least. What an inflation target does is establish transparency and accountability for the central bank. If the central bank hits its target, all this tells us is that they have hit their goal. It does not tell us about the desirability of the outcome.
It is entirely possible (if not probable) that monetary policy has an influence on relative prices and the allocation of resources without the aggregate level (growth) of money having an effect on the aggregate level (growth) of output. Indeed, the fact that housing prices rose 85% in Canada while the inflation rate was 2% poses interesting questions. Does the price index targeted by the Bank of Canada underweight housing? Is housing measured properly in the price indices? Doesn’t the rise in housing prices suggest that relative prices have changed (in real terms)?
I think that this is the fundamental point surrounding inflation targeting. If one focuses exclusively on the overall rate of inflation and monetary policy affects relative prices, then monetary policy directed in this manner has the potential to create asset price bubbles, resource misallocations (or simply reallocations), and boom-bust scenarios — even if super-neutrality holds.