Wednesday, June 10, 2009


Thinking Outside the Box on Monetary Issues

Deep Thought #1: Let's assume for the moment that there really is a cartel of international bankers who basically run the various governments and central banks. Wouldn't it be to their advantage to reproduce the Japanese lost decade in as many places as possible? You get a bad economy, so the public allows you to impose new regulations and spend tons of money to fix it, and yet you don't suffer the fallout from rapid price inflation. What's not to like?

Deep Thought #2: Robert Wenzel has really been spot-on for most of this crisis. At this point, I don't even bother linking every good post. Suffice it to say, Wenzel is just about the only blogger on the planet who (1) knows Austrian economics, (2) follows financial events like a PI, and (3) has a very cynical view of politicians and big business. (For example, check out his take on banks being able to repay TARP money.)

Anyway, Wenzel has been watching M2 for signs of when US price inflation will kick in, and I decided I had no particular reason for preferring the narrower M1 aggregate (which is what I normally plot on this blog, in addition to the monetary base). So I wanted to check whether M2 in fact has been a good predictor of price inflation. I may have blogged about this before, but if so it bears repeating:

If you didn't know what those lines above represented, wouldn't you say they were quite clearly things that moved in opposite directions? It's not perfect, but they are basically mirror images.

And yet the standard monetarist story is that--with a conveniently difficult-to-falsify "long and variable lag"--high/low M2 growth causes high/low price inflation.

Scott Sumner stole my thunder and mocked the "long and variable lag" notion already; I had a few good zingers myself at my talk at the Mises Circle in Fort Worth. But basically, I don't think it takes 6 months for the market to adjust to a bunch of new money dumped into it. If the Saudis announce that they are increasing production, oil prices respond almost immediately. So why does it take the market a year or two years to change the price of money (i.e. its purchasing power)?

I think the graph above shows that maybe in practice, given the market forming expectations and the Fed officials trying to operate in that environment, what actually happens is that the rising inflation rate causes the Fed to cut down M2 growth. On the other hand, in periods of tame inflation, the Fed can get away with inflating.

To clarify, I am not denying that an injection of money may lead to higher prices after a lag. But I am saying the mechanism is that the demand for money has increased at first, and then later returns to the old level. So for example, both M1 and M2 have risen at a very fast clip since the fall of 2008. But prices actually fell at first, and now are rising at a much more moderate pace. The answer is that people want to hold higher cash balances than they did before Paulson told them the world was ending. When people stop expecting further price declines, then they will want to reduce the size of their holdings, and you will see prices rise at a higher rate, even if (at that time) M1 and M2 aren't growing particularly quickly.

In one sense, I'm just quibbling about sloppy language, I suppose. But people say stuff like "it takes a while for new money to have an impact." No it doesn't. Price deflation would have been a lot more severe in 4q 2008 had Bernanke kept M1 and M2 level.

I don't know. My econometrics text has an example of distributed lag analysis in which it regressed the CPI on lags of M1 upt to 24 quarters, or six years. The change in the CPI began immediately, but was small. It increased with each lag until it reached its maximum effect at 18 quarters, or 4.5 years.

Another problem with following just a monetary aggregate is that it doesn't capture velocity changes which will have an impact on CPI but not the monetary measure.

Also, and this may be related to velocity, monetary pumping affects prices differently in a depression than in the later stages of a boom. Because of the pile up of idle resources during a depression, massive monetary pumping won't cause price inflation for quite a while. Finally, productivity increases reduce the effects of monetary pumping.

I think the link between money and price inflation is pretty loose. It seems to me that Hayek used to say the biggest mistake economists can make is to ignore the quantity theory of money. The second biggest is to take it too seriously.
Here is a study from the Cleveland Fed that shows a strong - and immediate - correlation between M2 and inflation from 1900 to 1960, but much less so - with more lag - since then. (My own analysis shows the correlation appears to lag an extra year every decade or so).

The question is why?

See Figures 11a-b and 12a-b
Post a Comment

Subscribe to Post Comments [Atom]

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to Posts [Atom]