Friday, April 12, 2019

The Washington Post: The Phillips Curve Is Not Working

A.W. Phillips
By Robert Wenzel

The Phillips Curve is failing in a big way for the second time in the last 50 years.

Robert J. Samuelson, the economics columnist for The Washington Post, explains the curve:
You may recall that Phillips refers to economist A.W. Phillips, whose 1958 landmark paper argued that there was a trade-off between unemployment and inflation: Lower unemployment generally produced higher inflation, and vice versa.
The logic seemed impeccable. Tight labor markets would raise wages, which would be passed along to consumers in higher prices. Phillips’s study of British wages and unemployment from 1861 to 1913 seemed to verify the relationship, which also seemed to hold for later decades. American economists found similar patterns in the United States...
The Phillips Curve was soon modified in one crucial respect. As economists Milton Friedman and Edmund Phelps pointed out, the effort to cut unemployment below a certain level wouldn’t just raise inflation. It would also cause accelerating inflation — inflation that kept getting worse. 
And he explains how this econometric-based theory has collapsed:
 One of today’s economic mysteries is: Why is inflation so low? The unemployment rate is a puny 3.8 percent. The recovery from the 2007-2009 Great Recession is nearly a decade old, just when tight labor markets and strong demand usually push up wages and prices. Yet inflation (measured by the consumer price index) has averaged only 1.5 percent annually since 2014.
One answer is that the Phillips Curve — the relationship between unemployment and inflation — is breaking down, says a study from the Peterson Institute for International Economics...
In recent years, the economy has defied the Phillips Curve. As unemployment has dropped, wages and inflation haven’t risen sharply.
This isn't the first time that the Curve has failed. In the 1980s, the Curve failed in the opposite direction. Inflation and unemployment were both skyrocketing.

Murray Rothbard wrote in 1984:
Every time someone calls for the government to abandon its inflationary policies, Establishment economists and politicians warn that the result can only be severe unemployment. We are trapped, therefore, into playing off inflation against high unemployment, and become persuaded that we must therefore accept some of both.

This doctrine is the fallback position for Keynesians. Originally, the Keynesians promised us that by manipulating and fine-tuning deficits and government spending, they could and would bring us permanent prosperity and full employment without inflation. Then, when inflation became chronic and ever-greater, they changed their tune to warn of the alleged tradeoff, so as to weaken any possible pressure upon the government to stop its inflationary creation of new money.

The tradeoff doctrine is based on the alleged "Phillips curve," a curve invented many years ago by the British economist A. W. Phillips. Phillips correlated wage rate increases with unemployment, and claimed that the two move inversely: the higher the increases in wage rates, the lower the unemployment. On its face, this is a peculiar doctrine, since it flies in the face of logical, commonsense theory. Theory tells us that the higher the wage rates, the greater the unemployment, and vice versa. If everyone went to their employer tomorrow and insisted on double or triple the wage rate, many of us would be promptly out of a job. Yet this bizarre finding was accepted as gospel by the Keynesian economic establishment.

By now, it should be clear that this statistical finding violates the facts as well as logical theory. For during the 1950s, inflation was only about one to two percent per year, and unemployment hovered around three or four percent, whereas nowadays unemployment ranges between eight and 11 percent, and inflation between five and 13 percent. In the last two or three decades, in short, both inflation and unemployment have increased sharply and severely. If anything, we have had a reverse Phillips curve. There has been anything but an inflation-unemployment tradeoff.

But ideologues seldom give way to the facts, even as they continually claim to "test" their theories by facts. To save the concept, they have simply concluded that the Phillips curve still remains as an inflation-unemployment tradeoff, except that the curve has unaccountably "shifted" to a new set of alleged tradeoffs. On this sort of mind-set, of course, no one could ever refute any theory.
So in the 1980s, the claim was that both inflation and unemployment were high because the curve had shifted.

Now, we are experiencing the exact opposite and we get the shift explanation once again.

Here's Samuelson once again:
In their paper and blog post, Christopher G. Collins and Joseph E. Gagnon of the Peterson Institute find that “the relationship between inflation and unemployment has shifted.” But they don’t conclude that the change is permanent or irreversible.
In other words, the Curve often doesn't work but mainstream economists never call it what it is, a failure, they just state it has shifted.

Quite a tool, its forecast often fails but the failure is merely a shift that wasn't forecasted.

It is best to recognize the wisdom of Ludwig von Mises on  econometric modelling:
As a method of economic analysis econometrics is a childish play with figures that does not contribute anything to the elucidation of the problems of economic reality.-The Ultimate Foundation of Economic Science
The specific experience with which economics and economic
statistics are concerned always refers to the past. It is history, and
as such does not provide knowledge about a regularity that will
manifest itself also in the future.-Epistemological Problems of Economics
The fact of the matter is that there is no direct correlation between price inflation and employment. A central bank can be aggressively printing money during a period of low unemployment but if productivity and/or the desire to hold cash balances is strong, there might be little to no inflation.

On the other hand, if a central bank is printing aggressively following a Fed manipulated boom but the printing isn't enough to keep the boom going, it is possible to have high price inflation and high unemployment at the same time.

Econometricians, including Milton Friedman, fail to understand the complex nature of the world and try to squeeze the world into their limited in scope models and always end up with egg on their face---and that's about a close to a correlation in the social sciences that is always going to work.

Robert Wenzel is Editor & Publisher of and Target Liberty. He also writes EPJ Daily Alert and is author of The Fed Flunks: My Speech at the New York Federal Reserve Bank and most recently Foundations of Private Property Society Theory: Anarchism for the Civilized Person Follow him on twitter:@wenzeleconomics and on LinkedIn. His youtube series is here: Robert Wenzel Talks Economics. More about Wenzel here.


  1. I remember watching a documentary about mathematics or economics - I can't recall, I'm sorry, it was several years ago, but I do remember the narrator talking about A. W. Phillips, a person who had a great curiosity and affinity for all things mechanical when he was a young lad living in New Zealand. He was a gifted tinkerer, so the narrator said. Phillips had a life one can say of great adventure and harrowing happenings, first as a traveler in the 30, then a prisoner of the Japanese. He wanted to study sociology while studying at the London School of Economics but switched his interest towards economics. He then developed, mustering his tinkerer nature, the so-called MONIAC machine, a kind of hydraulically-operated analog machine that purported to simulate an economy, you know, with various inputs becoming various outputs and such. Because, you know, people are like components of a big analog computer, or so we must gather. It is clear that people who are good at engineering don't necessarily become good economists and Phillips was indeed a bad economist. He was a good tinkerer, though.

    What fascinates me is that a correlation 'discovered' by Phillips is still considered useful enough by some economists to the point they're willing to forgo the logical fallacy behind the premise, that a correlation does not mean causation.

  2. 2+2 = 5. Hey, didja blink and miss that shift? Thought ya did!

  3. I think that there is a deeper fallacy in the Phillips Curve theory, namely, that there is no way to measure "price inflation." There are many well-documented issues with the CPI and PCE in terms of what they include and exclude, how they are calculated, etc. No one actually lives the CPI or PCE in their daily lives; everyone has a different bundle of consumption goods that they purchase and assets that they own. Moreover, there can be other factors at play which are price deflationary, and all we see is the net result, so actual price inflation is hidden.