Friday, August 16, 2013

The Key Discovery, the One Success


What caused the Great Inflation? According to Arthur Burns, the Fed Chairman who presided over much of that inflation, cost-push pressures caused it. In a post I return to again and again, Marcus Nunes summarizes that other Arthur's view:

According to Robert Hetzel, during the period of the Great Inflation, the prevailing view, and the one embraced by Arthur Burns, Fed chairman from 1970 to 1977, was that inflation was a real (cost-push), and therefore non-monetary, phenomenon.


Nick Rowe writes

I know what it's like to live in a demand-side world, because I used to live in one....

I left that world: when I discovered we lived in a monetary exchange economy (like a fish discovering it swims in water); when I saw the Phillips Curve shift in the 1970's; when I discovered Milton Friedman, and learned not only that units didn't matter but that the rate of change of units didn't eventually matter either.

In comments, Scott Sumner responds:

I remember those days. The key discovery was that changing the trend rate of inflation (within reason) had no impact on the average rate of unemployment.

I cropped a "But mostly..." sentence from Nick, to focus on Phillips and Friedman. That's the part Sumner calls the "key discovery". Obviously, it is pretty important to Nick Rowe, as well.


Let me re-summarize a little history from Stephen Williamson, some U.S. macroeconomic economic history starting with the 1960s:
Let's review that.
(i) Samuelson/Solow and others think that the Phillips curve is a structural relationship - a stable relationship between unemployment and inflation that represents a policy choice for the Fed.
(ii) Friedman (in words) says that this is not so. There is no long-run tradeoff between unemployment and inflation. It is possible to have high inflation and high unemployment.
(iii) Macroeconomic events play out in a way consistent with what Friedman stated. We have high inflation and high unemployment.
(iv) Lucas writes down a theory that makes rigorous what Friedman said...
(v) Paul Volcker takes Friedman seriously.

Nothing technical. There's never anything technical. There is only a sort of reverence, a kneeling before the great and wonderful Oz.

Since that Friedman thing, his 1968 speech to the AEA and its aftermath, since it is "the key discovery" and all, shouldn't there be a bit of technical presentation somewhere that lays it out like science? Don't you think? Not too technical, or I would fail to understand. But something more than a blind, religious embrace would be nice.

Religious, yeah. Nick Rowe's post, linked above, is a story of his conversion to the new faith. That's what bothers me. That is what I had in mind when I wrote to Marcus,

It seems people say he [Friedman] predicted the high-inflation-high-unemployment combination that showed the mis-use of the Phillips curve.

It seems the Friedman event brought “expectations” into economics. But I never see any but non-technical acceptance of what Friedman [supposedly] said.

Nothing technical. Perhaps this is what Williamson meant when he wrote "Friedman (in words)". Mark Thoma lays it out well:

Estimates of the Phillips curve that produced stable looking relationships were based upon data from time periods when inflation expectations were stable and unchanging. Friedman warned that if policymakers tried to exploit this relationship and inflation expectations changed, the Phillips curve would shift in a way that would give policymakers the inflation they were after, but the unemployment rate would be unchanged... When subsequent data appeared to validate Friedman's prediction, the New Classical, rational expectations, microfoundations view of the world began to gain credibility over the old Keynesian model...

The subsequent data appeared to validate Friedman's prediction.


At Angry Bear, back in May 2012, Robert asked a question:

Did Samuelson and Solow claim that the Phillips Curve was a structural relationship?

Did they claim that it showed a permanent tradeoff between inflation and unemployment.

James Forder says no.

This is an important post for me, Robert's post. Nick Rowe and Sumner and Williamson and Nunes and many others point to Milton Friedman's put-down of the Phillips curve. That put-down, it seems, was the birth of an alternative to Keynesian economics. That alternative was based on Samuelson and Solow being wrong about the trade-off.

But there are doubts about that. That's what Robert's post is about. And that's why his post is important. Robert writes:

Paul Krugman, John Quiggin and others (including me [Robert writes]) have argued that the one success of the critics of old Keynesian economics is the prediction that high inflation would become persistent and lead to stagflation.

That's the key point, for me: one success, parlayed. Friedman predicted stagflation, and when his prediction came true, Friedman became like a religious icon. But, see, I have a different explanation of stagflation than Friedman. So if Friedman is right, I must be wrong.

I don't think I'm wrong.

13 comments:

Nick Rowe said...

Arthur: Two points:

1. "That alternative was based on Samuelson and Solow being wrong about the trade-off.

But there are doubts about that. That's what Robert's post is about."

Suppose Robert is right about that, which means that Samuelson and Solow had already anticipated what Friedman would say, and said it earlier. And suppose they had said it clearly enough that we had understood them to be saying it. We would have had exactly the same conversion to the same set of beliefs, except we would have given credit to S and S, instead of to Friedman. Ideas matter more than people. People matter mostly because some people can articulate those ideas better than others. (Austrians say that Hayek was the one who said it much earlier, and they may be right, but nobody understood what Hayek was saying, and we didn't understand it till Friedman said it much better.)

2. The bit you cropped from me is rather important (well, maybe just to me). Let me copy it here:

"But mostly I left that world when I discovered how to do macroeconomics with imperfect competition and learned how to see the world both ways at once."

That meant I could see inflation both as cost-push and as monetary/demand-pull and reconcile those two apparently contradictory theories. That happened in 1985, and had nothing to do with Friedman (or S&S or Hayek), because I figured that one out for myself.

Cost-push shifts the Phillips Curve in the bad direction, but whether that results in higher inflation or only higher unemployment in the long run depends on how the central bank responds.

The Arthurian said...

Nick, thank you so much.

"We would have had exactly the same conversion to the same set of beliefs, except we would have given credit to S and S, instead of to Friedman."

Yes, I see that.

I always crop your important parts, perhaps because they are hardest for me to understand.

"That happened in 1985"

You know the date! Well, I would like to read that story.

I spent a lot of time with Free to Choose and in my younger days and I treasure the handwritten replies I got in response to my letters to Friedman. And I readily accept that prices tend to vary with spending, as does the Q of M ordinarily. But cost pressures are a separate matter.

Cost pressures are not dissipated by monetary restriction, the way demand-side pressures are.

More than a year ago now, in A last ditch defense of Inflation Targeting, Marcus Nunes wrote:

To Volker, the policy adopted by the FOMC “rests on a simple premise, documented by centuries of experience, that the inflation process is ultimately related to excessive growth in money and credit”.

This view, an overhaul of Fed doctrine, implicitly accepts that rising inflation is caused by “demand-pull” or excess aggregate demand or nominal spending.


That last sentence is key. It is clear to me that Volcker restricted the increase in prices without doing anything to defuse the underlying cost pressures.

"Cost-push shifts the Phillips Curve in the bad direction, but whether that results in higher inflation or only higher unemployment in the long run depends on how the central bank responds."

Sure. but if we can get the Phillips Curve shifting back toward the lower left corner of the graph, people won't so much care how the central bank responds.

Oh, and the underlying cost-push pressure is the cost of finance. "The interest of money is always a derivative revenue," Smith said, derived from wages or from profit (or from rent).

The Arthurian said...

... "Free to Choose" and "Money Mischief" in my younger days...

Nick Rowe said...

Arthur:

In 1985 I figured out how to build a macro model where firms had monopoly power, and see it both from the micro and macro perspective. And I then asked myself: if we start in equilibrium, and then monopoly power increases, so every firm wants to raise its price, what happens in long run equilibrium?

I learned that equilibrium output and employment would go down to a new level that was independent of how the central bank responded. Same vertical LRAS and LRPC as under perfect competition. (Actually, I learned that's not strictly accurate, because it is possible if you rig the model right to get multiple vertical LRAS curves or LRPC, but none of those rigs seemed very plausible to me.)

And that what happened to the price level and inflation depended totally on how the central bank responded to the increase in monopoly power. If, for example, the CB kept M constant, there would be a one-time rise in the price level but no ongoing rise in the inflation rate. But holding M constant is only one possible monetary policy.

In other words, I learned that it wasn't strictly correct to say "Cost pressures are not dissipated by monetary restriction, the way demand-side pressures are.". Or rather, it depends on what you mean by that.

Nick Rowe said...

Let me clarify what I just wrote when I said: "Same vertical LRAS and LRPC as under perfect competition."

It's still vertical, but it shifts to the left. And an increase in monopoly power shifts it even further left.

jim said...

Hi Art,
As Harry Dent likes to say, its all rooted in demographics.

Here is a chart related to this discussion.

http://research.stlouisfed.org/fred2/graph/?g=lvX

The rise and fall of the great inflation is correlated to the the change in debt. But when you factor in the change in labor force participation, the correlation becomes much closer.

If this chart is predictive we are headed towards deflation as both the expansion of debt and the baby boomers leaving the workforce kicks in.

-jim

Jazzbumpa said...

Art -

Did Friedman actually make a prediction, or just recognize a possibility? Is he getting too much credit?

We discussed the Phillips curve a couple of years back.

http://jazzbumpa.blogspot.com/2011/08/is-phillips-curve-valid.html

A lot of the scatter and low R^2 values for the curves in my graph are due to 1) transition periods where the data is clearly migrating from the realm of one line to another (obvious in the red set) and 2) my force fitting the data into arbitrary chronology packets rather than circumstance-defined packets. Your criticism in comments was cogent.

But is the fact that there is no LONG RUN trade off between inflation and unemployment particularly relevant? Econ policy happens in the here and now, and the short run effects look pretty valid.

BTW, I think expectations are a very important element in Keynesian econ.

Nick -
I don't think seeing inflation as both cost-push and as monetary/demand-pull are contradictory theories. I see it as a recognition that the real world is complex, circumstances change, and different causes can have similar effects.

So you have to understand the cause if you're going to construct the right policy response. Something the current crop of austerians is totally unable to do.

Jazzbumpa said...

jim -

That's an interesting correlation between CPI and CIVPART*TCMDO.

I'm having a little trouble wrapping my had around what the multiplication gives you.

Debt normalized to the percentage of people with the ability to pay - something along those lines?

I'll use myself as a counter example. I've been retired for almost 5 years, but pension and SS are sufficient to cover my living needs, including servicing debt.

But I'm not in labor participation.

Cheers!
JzB

jim said...

@Jazz

It is multiplying a percent by percent. That is not an out of the ordinary calculation. 10% of 10% is one percent for instance.

Both the change in labor force participation and the change in total debt seem to run on similar paths to inflation. When combined it gets very close

-jim

Jazzbumpa said...

Jim -

I get the math.

You can always multiply numbers.

But conceptually, what does the product mean?

(Annual % change of A)*(Annual % change of B)

What is the significance of it?

Cheers!
JzB

Nick Rowe said...

Jazzbumpa: "Nick -
I don't think seeing inflation as both cost-push and as monetary/demand-pull are contradictory theories. I see it as a recognition that the real world is complex, circumstances change, and different causes can have similar effects."

Let me restate your point like this:

The short run real effects of supply shocks will depend on the monetary policy being followed. It matters whether, for example, the central bank targets inflation or NGDP.

I would agree with that. So would nearly all macroeconomists.

jim said...

Hi jazz,

The way I look at it is all three time series are cyclical and related to the same periodicity so it isn't just multiplying any arbitrary numbers. All of them tend to go down in recessions and up in good times. Both the change in total debt and the change in workforce participation have a feedback relationship with inflation, but that relationship is obviously not exactly linear. It gets a lot closer to linear when debt and workforce participation are combined by multiplication.

It also seems to be a bit predictive. Shift the blue curve to the right a couple years and the blue and red look even closer.

Jazzbumpa said...

jim -

I'm not trying to be difficult, I'm trying to understand.

Granted the two numbers have similar cyclicality. So multiplying them magnifies that effect. And it ends up being similar to the cyclicality of inflation, with a lag.

But in the absence of some justifying narrative it might just be coincidental.

I need to think about this some more. If the relationship meaningful, then the meaning should be describable.

Otherwise, it's just a curiosity.

Cheers!
JzB