Laidler on Friedman’s puzzling views on the Phillips Curve
Last weekend I attended a conference on Karl Brunner. It was a treat meeting a number of famous monetarists, particularly Bennett McCallum, who I described last year as my favorite monetary economist. One of the conference readings was written by David Laidler, an outstanding monetary historian. Anyone interested in learning about the development of macro theory between 1900 and 1940 should read his two books on the subject.
I was struck by the following passage in the Laidler article, which describes an inconsistency in Friedman’s view of the Phillips Curve (or aggregate supply curve.) The passages in quotation marks were written by Friedman, the other portions are Laidler:
Once inflation got going in response to monetary expansion,
“Employees will start to reckon on rising prices of things they buy and to demand higher nominal wages for the future. ‘Market’ unemployment is below the ‘natural’ level. There is an excess demand for labor so real wages will tend to rise toward their initial level.” (1968, p. 104, italics added.)
This story of a market out of equilibrium is of course quite compatible with Friedman’s earlier observation that, after an unanticipated monetary shock, “To begin with, much or most of the rise in income will take the form of an increase in output and employment rather than prices.” (1968 p. 103), but less so with another remark that occurs in an intervening paragraph.
“Because selling prices of products typically respond to an unanticipated rise in nominal demand faster than prices of factors of production, real wages received have gone down — though real wages anticipated by employees went up . . . Indeed, the simultaneous fall ex post in real wages to employers and rise ex ante to employees is what enable employment to increase.” (1968 pp. 103-104, italics added)
Here the behaviour of a key real variable, employment, is presented as an equilibrium response to price changes that have already occurred, but have been perceived differently on different sides of the market. It is hard to see how the employment change in question could simultaneously have precededthose same price changes. What we have in this passage, then, is not an alternative description of a disequilibrium interpretation of the Phillips curve augmented by the idea of inflation expectations, but something rather closer to Lucas’s alternative aggregate supply curve interpretation, albeit without the idea of rational expectations.
This sounds like Irving Fisher’s view of the Phillips curve, which I have always preferred to the NAIRU approach. A few lines later, Laidler quotes another passage that confirms this is what Friedman had in mind.
“There was, however, a crucial difference between Fisher’s analysis and Phillips’, between the truth of 1926 and the error of 1958, which had to do with the direction of causation. Fisher took the rate of change of prices to be the independent variable that set the process going.” (1975 p. 12 italics in original)
I think both Friedman and I visualize the process in terms of nominal shocks having real effects. In contrast, many Keynesians see inflation as resulting from an overheated economy. So who’s right? I do see why Laidler is critical of Friedman; he also claims that demand shocks initially produce higher output, and only later do prices rise.
I see a three step process. A monetary shock (money supply or demand) causes flexible asset prices to change immediately (stocks, commodities, exchange rates, etc.) This causes output to rise, and consumer prices and wages respond with a lag.
Recently I seem to see advantages to NGDP almost everywhere I look. (Yes, I know; confirmation bias.) I think Friedman made a mistake focusing on inflation. His nominal aggregate should have been NGDP, not the price level. In that case the nominal shock is an unexpected change in NGDP, and the real effect is a change in employment. Inflation doesn’t have to play any causal role in the story. This avoids the inconsistency of claiming prices are slow to rise, yet rising prices cause employment to change. NGDP can rise almost immediately, even if many prices are sticky. And we can model wages as responding with a lag to NGDP, not to prices. Of course this is exactly what happens in the real world—otherwise we would have seen wages soaring in 2007-08, when inflation shot up, but NGDP growth remained subdued.
So unexpected changes in NGDP are the nominal shock and (because wages and prices are sticky) employment responds to these nominal shocks in the short run, and then moves back to the natural rate in the long run. A very Fisherian approach, which does not need to rely on awkward assumptions about inflation only occurring at full employment, and which avoids the logical inconsistencies involved in using the price level as the indicator of nominal shocks.
In the next post I’ll use these ideas to address an issue raised in a recent Krugman post; is it possible to produce moderate inflation expectations in a liquidity trap?
PS. The Laidler and Friedman quotations were from “Some Aspects of Monetarism Circa 1970: A View from 1994.” In Macroeconomics in Retrospect: The Selected Essays of David Laidler.
Tags: Irving Fisher, Milton Friedman
5. November 2010 at 17:21
Professor Sumner,
Looking through the lens of the hot potato process, where does inflation come from? I can’t figure it out. What causes prices to go up in response to demand, as opposed to income as it does at first?
Joe
5. November 2010 at 17:21
“A monetary shock (money supply or demand) causes flexible asset prices to change immediately (stocks, commodities, exchange rates, etc.) This causes output to rise, and consumer prices and wages respond with a lag.”
Where is a good starting point in the literature to find studies of this?
5. November 2010 at 17:46
Joe, Income goes up first, and also flexible prices. The mechanism is that the only way the public can get rid of cash is to buy other goods, services and assets. This raises the value of these assets, and encourages more production.
JTapp, I’m not sure, but I’d try reading some of the old monetarists like Friedman, Meltzer and Brunner.
Perhaps some of the interwar economists would also be able to shed light on these issues. The effects on asset prices are most noticeable for the interwar monetary shocks, which were quite large.
6. November 2010 at 00:27
Have you considered taking on the task of writing a macro-economics text book that takes nominal GDP seriously?
6. November 2010 at 02:34
Wish I had my old lecture notes here (they are in my office). I remember David Laidler making the exact same distinction between the two interpretations of the Phillips Curve in 1977. A Fisher/Lucas supply curve, vs a disequilibrium price adjustment equation. And how Milton Friedman flipped between the two.
Hume was in the second camp. So were Lipsey and Phelps. Modern New Keynesians are too. Both camps were present in the Phelps 68 volume, and they were not distinguished.
There’s a problem with using NGDP though. Theory says that in the long run all nominal variables should adjust together one-for-one. Theory says that real variables should adjust, but not necessarily one-for one. For example, the demand for nominal money should be unit elastic wrt P. But the real income elasticity of the demand for money is an empirical question; it may not be exactly one.
6. November 2010 at 05:18
Mikko, No. Don’t have time. If I do another book, it would probably be a set of essays on misconceptions in macro. Call it “Monetary Illusions.”
Nick. You said;
“There’s a problem with using NGDP though. Theory says that in the long run all nominal variables should adjust together one-for-one.”
I don’t see any problem here:
1. In the long run neither P nor NGDP are proporational to M, as both V and real money demand change for various reasons.
2. In the long run any given change in M will CAUSE both P and NGDP to rise by the same percentage. This is because in the long run M doesn’t affect RGDP, so any impact of M in the long run will be solely on P, (and the P portion of NGDP.)
3. In the short run M affects RGDP, but for that reason neither P or NGDP are proportional to M (as you say income elasticity is not exactly one.)
I can’t think of any stylistic fact that is true of the relationship between M and P, that is not also true of the relationship between M and NGDP.
Or have I missed something here.
6. November 2010 at 05:23
Scott, have you read Roger Garrison’s “Time and Money”? I would love to hear your thoughts on his efforts to explore Friedman, the philips curve and his “plucking” model of Bust-Boom.
6. November 2010 at 05:25
One other question: through what mechanisms do rising commodity and investment good prices cause output to rise? Is it the increased supply of loanable funds for business investment? Is it a wealth effect that accelerates consumption? A mix of both?
6. November 2010 at 05:44
Papola:
I don’t get it.
Higher commodity and investment goods (capital goods) prices, if caused by higher demand, results in greater profits from additional production. To capture those profits, the firms that produce those goods, produce more. That is more output.
Perhaps consumption and loanable funds rise, but surelyl that is secondary.
6. November 2010 at 06:33
Scott,
If you’re interested, there’s a scathing review of Laidler’s
“Fabricating the Keynesian Revolution: Studies of the Inter-war Literature”, by an economist named Michael Emmett Brady. This is the book you link to in your post.
http://www.amazon.com/Fabricating-Keynesian-Revolution-Inter-war-Literature/product-reviews/0521645964/ref=cm_cr_dp_all_summary?ie=UTF8&showViewpoints=1&sortBy=bySubmissionDateDescending
6. November 2010 at 06:45
Scott:
“Mikko, No. Don’t have time. If I do another book, it would probably be a set of essays on misconceptions in macro. Call it “Monetary Illusions.”
So since you don’t plan on writing a textbook is there anywhere a curious Danish grad student can see the curriculum for the courses in Monetary Economics you are teaching?
6. November 2010 at 08:53
Scott, would you mind commenting on what you think of the following?
The cash balance mechanism leads to higher prices. The marginal cost to increase production that is the labor component rises less than other prices, ie. wages are sticky. So companies increase production by having workers work more hours or by hiring more people.
6. November 2010 at 10:31
John, No, I hope to read it someday.
Imagine a rise in home prices–it’s an easy one to visualize. When home prices rise, producers have more incentive to produce homes. The same is true when wheat prices rise, or copper prices, or the prices of corporate assets, or the prices of shopping centers.
Bill, I agree.
Mike, I disagree with that review–it’s an outstanding book.
Jakob, I use Mishkin’s textbook. I don’t have my syllabus at home, but I spend a lot of time on 4 subjects:
1. How monetary policy affects the yield curve. Liquidity, income, price level and expected inflation effects.
2. How money affects prices. I do money supply and demand with 1/P on the vertical axis. I talk about different types of inflation, such as one time increases (supply shocks and fiscal stimulus) and persistent inflation (monetary policy). I discuss reasons for monetary inflations, such as monetizing the debt, or having a flawed macro model (such as 1966-81)
3. How money affects exchange rates (especially PPP and the interest parity theorem.) I also discuss the Balassa Samuelson effect–especially in regard to China.
4. How money affects the business cycle. I use the Fisher/Friedman approach to the Phillips curve, and also critique the NAIRU approach.
Wherever possible it’s partial equilibrium approach. I don’t use IS/LM.
I also discuss different types of monetary policy, such as the Taylor Rule. I usually don’t have time to do futures targeting, or NGDP targeting, as this is an undergraduate class.
How does it feel living in the happiest, most civic-minded and most market-friendly country on Earth? 🙂
TravisA, Yes, I think that’s correct, although I think there are other mechanisms too, such as sticky prices.