“…the debate over how much faith the Fed puts in the Phillips curve shows the broader dilemma of economic policy. For all the researchers over the decades and centuries who have tried to understand how the economy really works and to predict its course with precision, our ability to know where the economy is heading next year is no better than the ability of weather forecasters to predict whether it will rain three weeks from today…
…The United States economy is, after all, determined largely by the endlessly complicated interactions of 320 million people producing $17 trillion worth of stuff, which even relatively complex models can’t keep up with.”, Neil Irwin, “The 57-Year-Old Chart That Is Dividing the Fed”, The New York Times, October 25, 2015
“Isn’t it better to let 320 million people (a free market) decide the cost of money (interest rates) in an unknowable future, than a handful of government “economic experts” at the Fed, with archaic tools like the Phillips curve? And, to me and so many others, it’s unbelievable after all the asset bubbles and busts caused by the Fed and its policies (and companies and jobs and economic activity destroyed as a result of this volatility)….that the Fed, only considers “prices” to be a basket of prices for consumer goods and services, and they ignore asset prices like stocks, bonds, and real estate, where the bubbles and busts have occurred time and again, causing financial crises and economic volatility, not stability? It makes no sense.”, Mike Perry, former Chairman and CEO, IndyMac Bank
The 57-Year-Old Chart That Is Dividing the Fed
Next week, when Federal Reserve officials meet to decide whether to raise interest rates for the first time in nine years, one question will be front and center: How much faith should be placed in a line on a graph first drawn by a New Zealand economist nearly six decades ago, based on data on wages and employment in Britain dating to the 1860s?
That would be the Phillips curve, one of the most important concepts in macroeconomics. It shows how inflation changes when unemployment changes and vice versa. The intuition is simple: When joblessness is low, employers have to pay ever higher wages to attract workers, which feeds through into higher prices more broadly. And inflation is particularly prone to rise when the unemployment rate falls below the “natural rate” at which pretty much everybody who wants a job either has one or can find one quickly.
As the Fed’s chairwoman, Janet L. Yellen, put it in a 2007 speech, the Phillips curve “is a core component of every realistic macroeconomic model.”
Credit Tim Cook
Except it doesn’t work. Or at least, it hasn’t worked very well in the last few decades in the United States. And it has proved particularly problematic to try to use that historical relationship to predict where inflation is going.
That is why a longstanding academic debate is now at the core of the Fed’s policy debate. Ms. Yellen and many of her Fed colleagues have indicated that they think they should raise interest rates this year, in part because the Phillips curve suggests there will be excessive inflation if they don’t. The unemployment rate was 5.1 percent in September, just a smidgen above the 4.9 percent that Fed leaders believe is the appropriate jobless rate in the longer run.
In other words, if you believe in the traditional Phillips curve, inflation should be taking off any day now.
But this month, two Fed governors, Lael Brainard and Daniel K. Tarullo, argued against a rate move. Ms. Brainard said that the Phillips curve relationship was “at best, very weak at the moment.” Mr. Tarullosaid that it was “probably wise not to be counting so much on past correlations, things like the Phillips curve, which haven’t been working effectively for 10 years now.”
It’s only a slight exaggeration to say that the Fed’s rate decision this year will be based on whether its leaders really believe that the Phillips curve is useful in describing how the economy works in 2015.
The idea of the Phillips curve has been under attack almost since William Phillips, the aforementioned New Zealander, wrote his 1958 paper “The Relation Between Unemployment and the Rate of Change of Money Wage Rates in the United Kingdom, 1861-1957.”
The most crude versions of the Phillips curve have indeed, in recent decades at least, been nearly useless. Any attempt to estimate it requires a researcher to decide what measure of employment to use, what measure of inflation and what time lags to assume, among other choices. So there are nearly as many versions of the Phillips curve as there are researchers who study it.
If you simply look at the unemployment rate in the United States versus the Consumer Price Index, excluding volatile food and energy prices for every year since 1958, there is nearly no statistical relationship at all, just a jumble of dots. (A best-fit line actually points the wrong direction, correlating higher unemployment with higher inflation, albeit very weakly.)
If you take only subsets of that period, the relationship looks stronger. For example, research from the Federal Reserve Bank of Minneapolis shows a fairly clear (negative) correlation between unemployment and inflation from 1977 to 1990, but suggests that relationship basically disappeared in the 1990s and was barely evident in the first decade of the 2000s. But in some ways an ever shifting curve raises more questions than it answers.
About 400 job seekers posed in Tyrone, Pa., in 1971 for an ad aimed at attracting new industry to the small town after a local paper plant closed. Unemployment is a factor the Fed will consider next week when it meets to decide whether to raise interest rates. Credit United Press International
After all, if the relationship between unemployment and inflation is constantly shifting, for reasons that are impossible to predict, it isn’t a particularly useful tool for deciding what policy should be today.
That said, a generation or two of economists have been working to understand what those simplistic versions of the Phillips curve get wrong, and have created more complex versions that incorporate more variables and might prove useful at predictions.
In particular, after the combination of high inflation and high unemployment of the 1970s — a condition that would be impossible in a world in which the Phillips curve worked perfectly — scholars began adding inflation expectations to their models to help explain the paradox.
Robert J. Gordon, an economist at Northwestern University, has his own version that he argues explains inflation levels throughout recent decades. But it is hardly simple. Its prediction for inflation relies not just on joblessness but also on measures of productivity growth, shifts in food and energy prices and overall inflation over the six preceding years.
In other words, just knowing the unemployment rate may tell you very little about what inflation will be. But if you add a few more variables, you can do a better job at predictions.
“If you do it right, the Phillips curve relationship has been very stable over the last 40 years,” Mr. Gordon said. “It’s the old ocean liner analogy — it takes a long time to turn an ocean liner, and it takes a long time for the inflation rate to respond to high unemployment or low unemployment.”
An important implication of Mr. Gordon’s model is that much of the inflation rate in 2016 or even 2020 is decided by economic shifts that have already happened (recent productivity levels, for example) or by factors that policy makers have little control over (like what happens to oil prices).
His model predicts that if unemployment levels keep falling the way they have recently, inflation won’t reach the Fed’s 2 percent target until 2020. Mr. Gordon doubts that whether the central bank raises interest rates this week or in December or six months from now will make much difference to that outcome.
In a way, the debate over how much faith the Fed puts in the Phillips curve shows the broader dilemma of economic policy. For all the researchers over the decades and centuries who have tried to understand how the economy really works and to predict its course with precision, our ability to know where the economy is heading next year is no better than the ability of weather forecasters to predict whether it will rain three weeks from today. The United States economy is, after all, determined largely by the endlessly complicated interactions of 320 million people producing $17 trillion worth of stuff, which even relatively complex models can’t keep up with.
There’s a difference, though, between economic forecasting and weather forecasting. People don’t have any short-term influence over the weather, but central banks and other economic policy makers do have influence over the short-term course of the economy. So for all its faults, the Phillips curve — particularly its more sophisticated varieties — may be the best tool that Janet Yellen and company have to work with.
A version of this article appears in print on October 25, 2015, on page BU1 of the New York edition with the headline: A Guiding Principle That May Lead the Fed Astray.