Scott Sumner  

Reasoning from a price change, example #213

Mao's Murderousness... Forced Labor in Venezuela...

Noah Smith has a new post on interest rates:

Traditionally, macroeconomists have believed that low interest rates encourage inflation. But first Japan, and now the U.S. and Europe have kept rates low for years now, and inflation has stayed stubbornly low. A radical group of macroeconomists, including Stephen Williamson of the Federal Reserve Bank of St. Louis and John Cochrane of the Hoover Institution, have introduced a new theory called Neo-Fisherism, which says that a long period of low interest rates actually holds prices down instead of pushing them up. Williamson and Cochrane have both repeatedly stressed that New Keynesian models -- the most mainstream type of macroeconomic theory -- can easily yield the Neo-Fisherian result instead of the traditional view.
The first sentence is flat out wrong. There is nothing in traditional macroeconomics that suggests low interest rates lead to high inflation. Paul Krugman would define traditional macro as the IS-LM model, so let's use that workhorse. If the LM curve shifts right, then interest rates fall and inflation rises. If the IS curve shifts left, then interest rates fall and inflation falls. So there is no necessary relationship between interest rates and inflation. If Japan has experienced low rates and low inflation, presumably the Japanese IS curve has been sluggish---no big surprise given their falling population. So contra Smith, there is no "puzzle" to be explained.

As far as Neo-Fisherism, we've known for a long time that low nominal interest rates tend to be correlated with low inflation. Milton Friedman said something similar back in 1997.

Friedman is saying that a tight money policy will cause low inflation, and that over time this will lead to low interest rates. So if you observe low interest rates, it's a sign that policy was tight in previous periods. Friedman understood all this, and hence there is no puzzle for the Neo-Fisherians to explain. If you want to claim that a sudden unexpected cut in interest rates will lead to lower inflation expectations, that's different. There are rare occasions where it is true (such as Switzerland in January 2015) but only when accompanied by other contractionary steps. Normally an unexpected rate cut leads to higher expected inflation (in the asset markets).

Smith goes on to discuss how new ideas in behavioral economics may be able to resolve some of these interest rate puzzles. I'm all for trying new approaches, including behavioral economics (indeed my other blog is named after a popular behavioral economic theory) but in this case I simply don't see any puzzles to explain. Rates are low because tight money has led to very slow NGDP growth, plus a few other non-monetary factors impacting global saving/investment.

HT: Marcus Nunes

Comments and Sharing

COMMENTS (9 to date)
marcus nunes writes:

With inflation expectations below target and expectations of low nominal and real growth going forward, there´s no reason for yields to rise! And so they fall. Notice that the start of the rate hike talk in mid-2014 clinches the “low nominal growth-low inflation-falling yield” scenario.

ThaomasH writes:

Smith is just making a sloppy reference to a central bank policy of lowering rates/increasing the base will be inflationary. It's what you rightly criticize BOJ for NOT doing enough of.

Dennis writes:

Typically, people put comically high numbers in titles like "example #213". Searching , there are 122 pages of results with 5 links each for "reason from a price change, meaning you have over 600 examples there alone. You also gave example #656 in 2013, so you are probably into quadruple digits by now.

Richard A. writes:

According to Neo-Fisherist logic, the increase in IOER late last year should have been stimulative. I think the slowdown in nominal GDP growth rate is proving them wrong.

It is the supply of and demand for money that determines nominal GDP. If the Fed were to contract the money supply to prop up the interest rate above the natural rate, it would produce a needless contraction in GNPn, a lower level of inflation, and a decrease in nominal interest rates as a result of a decrease in the inflation risk premium.

Rajat writes:

Speaking of behavioural biases, reasoning from a price change has to be the biggest one there is.

Market Fiscalist writes:

'There is nothing in traditional macroeconomics that suggests low interest rates lead to high inflation'

Hang on a minute. There is lots in traditional macro to suggest that if a CB consistently held interest rates below the natural rate it would lead to high inflation. Isn't that the central point in Wicksell that neo-fisherists are challenging ?

Scott Sumner writes:

Thaomas, You said:

"Smith is just making a sloppy reference to a central bank policy of lowering rates/increasing the base will be inflationary."

But then there is no puzzle to explain. Smith suggests that what's been going on recently in Japan is somehow at variance with macro theory, and it is not. The entire premise of his article is wrong.

Japan has seen its IS curve shift left. Yes, the BOJ responds to that by cutting its policy rate, but it's still an IS story.

Market Fiscalist, I am not at all sure that is the case. Do the NeoFisherians even have the Wicksellian rate in their models? You'll have to ask them.

In any case, the BOJ has obviously not been holding rates under the Wicksellian rate, so there is no puzzle to explain.

bill writes:

Another test of the Neo-Fisherist logic was conducted by the ECB about 5 years ago. They are just plain wrong.

Matt writes:

I'm definitely missing something here. Where is inflation accounted for in the IS-LM model? The horizontal axis represents RGDP not inflation.

Re: Wicksell, just to be clear, the argument is that we get inflation when the interest rate is held below the Wicksellian rate *wherever it might be*, not simply when interest rates are "low," correct? So, e.g. if the natural rate is 15% but central bank manipulation brings the market rate to 13%. In that case is 13% a "low" rate of interest? Likewise it very well may be the case today that the natural rate in the US is 50 bps. If so is the current FFR target "low?" The whole point is the discrepancy between the natural and market rates not the level of the market rate by itself, yes?

Comments for this entry have been closed
Return to top